Regulatory Impact Analysis of the Proposed
Revisions to the National Ambient Air
Quality Standards for Ground-Level Ozone
-------
[This page intentionally left blank]
-------
EPA-452/P-14-006
November 2014
Regulatory Impact Analysis of the Proposed Revisions
to the National Ambient Air Quality Standards for Ground-Level Ozone
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
in
-------
CONTACT INFORMATION
This document has been prepared by staff from the Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency. Questions related to this document should be
addressed to Robin Langdon, U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, C439-02, Research Triangle Park, North Carolina 27711 (email:
langdon.robin@epa.gov).
ACKNOWLEDGEMENTS
In addition to EPA staff from the Office of Air Quality Planning and Standards, personnel from
the Office of Atmospheric Programs, the Office of Transportation and Air Quality, the Office of
Policy Analysis and Review, and the Office of Policy's National Center for Environmental
Economics contributed data and analysis to this document.
IV
-------
TABLE OF CONTENTS
LIST OF TABLES xiii
LIST OF FIGURES xxiv
EXECUTIVE SUMMARY ES-1
Overview ES-1
ES. 1 Overview of Analytical Approach ES-3
ES. 1.1 Establishing the Baseline ES-5
ES. 1.2 Control Strategies and Emissions Reductions ES-6
ES. 1.2.1 Emissions Reductions from Known Controls in 2025 ES-7
ES.1.2.2 Emissions Reductions beyond Known Controls in 2025 ES-9
ES.1.2.3 Emissions Reductions beyond Known Controls for Post-2025 ES-9
ES.1.3 Human Health Benefits ES-10
ES.1.4 Welfare Co-Benefits of the Primary Standard ES-12
ES.2 Results of Benefit-Cost Analysis ES-13
ES.3 References ES-18
CHAPTER 1: INTRODUCTION AND BACKGROUND 1-1
Overview 1-1
1.1 Background 1-1
1.1.1 NAAQS 1-1
1.1.2 Ozone NAAQS 1-2
1.2 Role of this RIA in the Process of Setting the NAAQS 1-2
1.2.1 Legislative Roles 1-2
1.2.2 Role of Statutory and Executive Orders 1-3
1.2.3 The Need for National Ambient Air Quality Standards 1-4
1.2.4 Illustrative Nature of the Analysis 1-4
1.3 Overview and Design of the RIA 1-5
1.3.1 Existing and Revised Ozone National Ambient Air Quality Standards 1-5
1.3.2 Establishing Attainment with the Current Ozone National Ambient Air
Quality Standard 1-6
1.3.3 Establishing the Baseline for Evaluation of Alternative Standards 1-8
1.4 Health and Welfare Benefits Analysis Approach 1-9
1.4.1 Health Benefits 1-9
1.4.2 Welfare Co-Benefits 1-10
1.5 Cost Analysis Approach 1-10
1.6 Organization of this Regulatory Impact Analysis 1-10
1.7 References 1-11
CHAPTER 2: DEFINING THE OZONE AIR QUALITY PROBLEM 2-1
Overview 2-1
2.1 Emissions and Atmospheric Chemistry 2-1
-------
2.2 Spatial and Temporal Variations in Ambient Ozone Concentrations 2-3
2.3 Ozone Monitoring 2-5
2.3.1 Ozone Monitoring Network 2-5
2.3.2 Recent Ozone Monitoring Data and Trends 2-7
2.4 Background Ozone 2-10
2.4.1 Seasonal Mean Background Ozone in the U.S 2-12
2.4.2 Seasonal Mean Background Ozone in the U.S. as a Proportion of Total
Ozone 2-15
2.4.3 Daily Distributions of Background Ozone within the Seasonal Mean 2-16
2.5 References 2-19
CHAPTERS: AIR QUALITY MODELING AND ANALYSIS 3-1
Overview 3-1
3.1 Modeling Ozone Levels in the Future 3-2
3.1.1 Selection of Future Analytic Year 3-2
3.1.2 Air Quality Modeling Platform 3-3
3.1.3 Emissions Inventories 3-7
3.1.4 Emissions Sensitivity Simulations 3-11
3.2 Methods for Calculating Current and Future Year Ozone Design Values 3-16
3.2.1 Current Year Ozone Design Value Calculations 3-16
3.2.2 Future Year Ozone Design Value Projections 3-17
3.3 Determining Tons of Emissions Reductions to Meet Various NAAQS Levels 3-18
3.3.1 Determining Ozone Response from Each Emissions Sensitivity 3-18
3.3.2 Combining Response from Multiple Sensitivity Runs To Construct
Baseline And Alternative Standard Scenarios 3-22
3.3.3 Creation of the Baseline Scenario 3-24
3.3.4 Creation of the 70, 65, and 60 ppb Alternative Standard Level Scenarios ...3-25
3.3.5 Monitoring Sites Excluded from Quantitative Analysis 3-26
3.4 Creating Spatial Surfaces 3-28
3.4.1 BenMap Surfaces 3-28
3.4.2 W126 surfaces 3-35
3.5 References 3-40
APPENDIX 3: ADDITIONAL AIR QUALITY ANALYSIS AND RESULTS 3 A-1
3A.I 2011 Model Evaluation for Ozone 3A-1
3A.2 California Sub-Regions and Areas of Influence 3A-17
3A.3 VOC Impact Areas 3A-22
3 A.4 Numeric Examples of Calculation Methodology for Changes in Design
Values 3A-23
3 A. 5 Emissions Reductions Applied to Create Baseline and Alternative Standard
Level Scenarios 3 A-25
3 A.6 Design Values for All Monitors included in the Quantitative Analysis 3 A-27
3 A.7 Monitors Excluded from the Quantitative Analysis 3 A-51
3 A.7.1 Sites without Projections Due to Insufficient Days 3 A-51
3A.7.2 Winter Ozone 3A-53
VI
-------
3 A.7.3 Monitoring Sites in Rural/Remote Areas of the West and Southwest 3 A-54
3A.8 Calculation Methodology for W126 Metric 3A-63
3 A. 9 References 3A-64
CHAPTER 4: CONTROL STRATEGIES AND EMISSIONS REDUCTIONS 4-1
Overview 4-1
4.1 Control Strategy Analysis Steps 4-3
4.2 Baseline Control Strategy 4-5
4.3 Alternative Standard Analyses 4-12
4.3.1 Identifying Known Controls Needed to Meet the Alternative Standards 4-14
4.3.2 Known Control Measures Analyzed 4-19
4.3.3 Emissions Reductions beyond Known Controls Needed to Meet the
Alternative Standards 4-21
4.3.4 Summary of Emissions Reductions Needed to Meet the Alternative
Standards 4-22
4.4 Limitations and Uncertainties 4-23
4.5 References 4-25
APPENDIX 4: CONTROL STRATEGIES AND EMISSIGNS REDUCTIONS 4A-1
Overview 4 A-1
4 A. 1 Types of Control Measures 4 A-1
4A.2 Application of Control Measures in Geographic Areas 4A-1
4A. 3 NOX Control Measures for NonEGU Point Sources 4A-4
4A.4 VOC Control Measures for Non-EGU Point Sources 4A-12
4A. 5 NOX Control Measures for Nonpoint (Area) and Nonroad Sources 4A-12
4A.6 VOC Control Measures for Nonpoint (Area) Sources 4A-123
CHAPTER 5: HUMAN HEALTH BENEFITS ANALYSIS APPROACH AND RESULTS ..5-1
5.1 Synopsis 5-1
5.2 Overview 5-4
5.3 Updated Methodology Presented in this RIA 5-8
5.4 Human Health Benefits Analysis Methods 5-10
5.4.1 Health Impact Assessment 5-12
5.4.2 Economic Valuation of Health Impacts 5-14
5.4.3 Estimating Benefits for the 2025 and Post-2025 Scenarios 5-16
5.4.4 Benefit-per-ton Estimates for PM2.5 5-17
5.5 Characterizing Uncertainty 5-19
5.5.1 Monte Carlo Assessment 5-21
5.5.2 Sensitivity Analysis Addressing Both Incidence and Dollar Benefit
Valuation 5-22
5.5.3 Supplemental Analyses 5-24
5.5.4 Qualitative Assessment of Uncertainty and Other Analysis Limitations 5-25
5.6 Benefits Analysis Data Inputs 5-26
5.6.1 Demographic Data 5-26
vn
-------
5.6.2 Baseline Incidence and Prevalence Estimates 5-27
5.6.3 Effect Coefficients 5-30
5.6.3.1 Ozone Premature Mortality Effect Coefficients 5-38
5.6.3.2 Hospital Admissions and Emergency Department Visits 5-48
5.6.3.3 Acute Health Events and School/Work Loss Days 5-50
5.6.3.4 Unquantified Human Health Effects 5-52
5.6.4 Economic Valuation Estimates 5-55
5.6.4.1 Mortality Valuation 5-56
5.6.4.2 Hospital Admissions and Emergency Department Valuation 5-67
5.6.4.3 Nonfatal Myocardial Infarctions Valuation 5-68
5.6.4.4 Valuation of Acute Health Events 5-70
5.6.4.5 Growth in WTP Reflecting National Income Growth over Time 5-71
5.7 Benefits Results 5-755
5.7.1 Benefits of the Proposed and Alternative Annual Primary Ozone
Standards for the 2025 Scenario 5-766
5.7.2 Benefits of the Proposed and Alternative Annual Primary Ozone
Standards for the post-2025 Scenario 5-822
5.7.3 Uncertainty in Benefits Results (including Discussion of Sensitivity
Analyses and Supplemental Analyses) 5-877
5.7.3.1 Sensitivity Analyses 5-89
5.7.3.2 Supplemental Analyses 5-911
5.8 Discussion 5-922
5.9 References 5-955
APPENDIX 5A: COMPREHENSIVE CHARACTERIZATION OF UNCERTAINTY IN
OZONE BENEFITS ANALYSIS 5A-1
Overview 5 A-1
5 A. 1 Description of Classifications Applied in the Uncertainty Characterization ... 5 A-1
5A.1.1 Direction of Bias 5A-2
5 A. 1.2 Magnitude of Impact 5 A-2
5A. 1.3 Confidence in Analytic Approach 5A-3
5 A. 1.4 Uncertainty Quantification 5 A-4
5 A.2 Organization of the Qualitative Uncertainty Table 5 A-5
5 A. 3 References 5 A-16
Appendix 5B: ADDITIONAL SENSITIVITY ANALYSES RELATED TO THE OZONE
HEALTH BENEFITS ANALYSIS 5B-1
Overview 5B-1
5B. 1 Threshold Sensitivity Analysis for Premature Mortality Incidence and
Benefits from Long-term Exposure to Ozone 5B-2
5B.2 Alternative Concentration-Response Functions for PM2.s-Related Mortality. 5B-4
5B.3 Income Elasticity of Willingness-to-Pay 5B-7
5B.4 References 5B-9
Vlll
-------
APPENDIX 5C: SUPPLEMENTAL ANALYSES RELATED TO THE OZONE HEALTH
BENEFITS ANALYSIS 5C-1
Overview 5C-1
5C. 1 Age Group-Differentiated Aspects of Short-Term Ozone Exposure-Related
Mortality 5C-1
5C.2 Evaluation of Mortality Impacts Relative to the Baseline Pollutant
Concentrations (used in generating those mortality estimates) for both Short-Term
Ozone Exposure-Related Mortality and Long-Term PMi.s Exposure-Related
Mortality 5C-6
5C.3 Core Incidence and Dollar Benefits Estimates Reflecting Application of
Known Controls for the 2025 Scenario 5C-14
5C.4 References 5C-19
APPENDIX 5D: DISCUSSION OF EFFECT ESTIMATES REFLECTED IN THE
DEVELOPMENT OF DOLLAR-PER-TON VALUES USED IN MODELING PM2.5
COBENEFITS 5D-1
Overview 5D-1
5D.1 PM2.5 Premature Mortality Effect Coefficients 5D-5
5D.2 Hospital Admissions and Emergency Department Visits 5D-14
5D.3 Acute Health Events and School/Work Loss Days 5D-16
5D.4 Nonfatal Acute Myocardial Infarctions (AMI) (Heart Attacks) 5D-19
5D.5 References 5D-21
APPENDIX 5E: INPUTS TO PM2.5 COBENEFITS MODELING 5E-1
Overview 5E-1
CHAPTER 6: IMPACTS ON PUBLIC WELFARE OF ATTAINMENT STRATEGIES TO
MEET PRIMARY AND SECONDARY OZONE NAAQS 6-1
Overview 6-1
6.1 Welfare Benefits of Strategies to Attain Primary and Secondary Ozone Standards . 6-2
6.2 Welfare Benefits of Reducing Ozone 6-3
6.3 Additional Welfare Benefits of Strategies to Meet the Ozone NAAQS 6-6
6.4 Analysis of Commercial Agricultural and Forestry Related Benefits Using the
Forest and Agricultural Sector Optimization Model - Greenhouse Gas Version
(FASOMGHG) 6-8
6.4.1 Summary of the Analytical Approach 6-8
6.4.2 Summary of FASOMGHG Results 6-10
6.5 References 6-16
APPENDIX 6A: Methods and Data Used to Develop estimates of Ozone Effects on Crop
and Forest Productivity 6A-1
6 A. 1 Methodology 6 A-1
IX
-------
6A. 1.1 Calculation of Relative Yield Loss 6A-4
6A. 1.1.1 Relative Yield Loss for Crops 6A-5
6A. 1.1.2 Relative Yield Loss for Trees 6A-7
6A.1.1.3 Calculation of Relative Yield Gain 6A-7
6A.1.2 Conducting Model Scenarios in FASOMGHG 6A-8
6A.1.3 Data Inputs 6A-9
6A. 1.3.1 Ambient Ozone Concentration Data 6A-10
6A. 1.3.2 Changes in Crop and Forest Yields with Respect to 75 ppb
Scenario 6 A-16
6A.2 Model Results 6A-25
6A.2.1 Agricultural Sector 6A-25
6A.2.1.1 Production and Prices 6A-25
6A.2.1.2 Crop Acreage 6A-28
6A.2.2 Forestry Sector 6A-30
6A.2.2.1 Production and Prices 6A-30
6A.2.2.2 Forest Acres Harvested 6A-33
6A.2.2.3 Forest Inventory 6A-34
6A.2.3 Cross-Sectoral Policy Impacts 6A-35
6A2.3.1 Land Use 6A-36
6A2.3.2 Welfare 6A-37
6A.2.3.3 Greenhouse Mitigation Potential 6A-40
6A.3 Summary 6A-43
6A.4 References 6A-45
CHAPTER 7: ENGINEERING COST ANALYSIS and Economic Impacts 7-1
Overview 7-1
7.1 Estimating Engineering Compliance Costs 7-2
7.1.1 Methods and Data 7-2
7.1.2 Compliance Cost Estimates for Known Controls 7-4
7.2 The Challenge of Estimating Costs for Unknown Controls 7-10
7.2.1 Incomplete Characterization of NOx Marginal Abatement Cost Curves 7-11
7.2.2 Comparison of Baseline Emissions and Controls across Ozone NAAQS
RIAs from 1997 to 2014 7-13
7.2.3 Impact of Technological Innovation and Diffusion 7-18
7.2.4 Learning by Doing 7-20
7.2.5 Using Regional NOx Offset Prices to Estimate Costs of Unknown
Emissions Controls 7-22
7.2.6 Conclusion 7-25
7.3 Compliance Cost Estimates for Unknown Emissions Controls 7-26
7.3.1 Methods 7-26
7.3.2 Unknown Compliance Cost Estimates 7-30
7.4 Total Compliance Cost Estimates 7-31
7.5 Updated Methodology Presented in this RIA 7-33
7.6 Economic Impacts 7-36
7.6.1 Introduction 7-36
7.6.2 Summary of Market Impacts 7-37
-------
7.7 Uncertainties and Limitations 7-38
7.8 References 7-40
APPENDIX 7A: ENGINEERING COST ANALYSIS 7A-8
Overview 7A-8
7 A. 1 Cost of Known Controls in Alternative Standards Analyses 7A-8
7A.2 Alternative Estimates of Costs Associated with Emissions Reductions from
Unknown Controls 7A-8
CHAPTERS: COMPARISON OF COSTS AND BENEFITS 8-1
Overview 8-1
8.1 Results 8-1
8.2 Discussion of Results 8-8
8.2.1 Relative Contribution of PM Benefits to Total Benefits 8-9
8.2.2 Developing Future Control Strategies with Limited Data 8-9
8.3 Framing Uncertainty 8-11
8.4 Key Observations from the Analysis 8-14
8.5 References 8-16
CHAPTER 9: STATUTORY AND EXECUTIVE ORDER IMPACT ANALYSIS 9-1
Overview 9-1
9.1 National Technology Transfer and Advancement Act 9-1
9.2 Paperwork Reduction Act 9-1
9.3 Regulatory Flexibility Act 9-2
9.4 Unfunded Mandates Reform Act 9-3
9.5 Executive Order 12866: Regulatory Planning and Review 9-5
9.6 Executive Order 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations 9-5
9.7 Executive Order 13045: Protection of Children from Environmental Health &
Safety Risks 9-7
9.8 Executive Order 13132: Federalism 9-8
9.9 Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments 9-9
9.10 Executive Order 13211: Actions that Significantly Affect Energy Supply,
Distribution, or Use 9-10
APPENDIX 9A: SOCIO-DEMOGRAPHIC CHARACTERISTICS OF POPULATIONS IN
CORE-BASED STATISTICAL AREAS WITH OZONE MONITORS EXCEEDING
PROPOSED OZONE STANDARDS 9A-1
9 A. 1 Design of Analysis 9 A-1
9A1.1 Demographic Variables Included in Analysis 9A-3
9A.2 Considerations in Evaluating and Interpreting Results 9A-6
9A.3 Presentation of Results 9A-7
XI
-------
CHAPTER 10: QUALITATIVE DISCUSSION OF EMPLOYMENT IMPACTS OF AIR
QUALITY 10-1
Overview 10-1
10.1 Economic Theory and Employment 10-1
10.2 Current State of Knowledge Based on the Peer-Reviewed Literature 10-6
10.2.1 Regulated Sectors 10-6
10.2.2 Labor Supply Impacts 10-7
10.3 Employment Related to Installation and Maintenance of NOx Control Equipment 10-8
10.3.1 Employment Resulting from Addition of NOx Controls at EGUs 10-9
10.3.2 Assessment of Employment Impacts for Individual Industrial,
Commercial, and Institutional (ICI) Boilers and Cement Kilns 10-12
10.4 References 10-15
xn
-------
LIST OF TABLES
Table ES-1. Summary of Emission Reductions by Sector for Known Controls for 70 ppb
Proposed Alternative Standard Level for 2025, except California (1,000 tons/year)a..ES-8
Table ES-2. Summary of Emission Reductions by Sector for Known Controls for 65 ppb
Proposed Alternative Standard Level for 2025 - except California (1,000 tons/year)21. ES-8
Table ES-3. Summary of Emission Reductions by Sector for Known Controls for 60 ppb
Alternative Standard Level for 2025 - except California (1,000 tons/year)a ES-8
Table ES-4. Summary of Emissions Reductions by Alternative Standard for Unknown
Controls for 2025 - except California (1,000 tons/year)21 ES-9
Table ES-5. Summary of Emissions Reductions by Alternative Standard Level for
Unknown Controls for post-2025 - California (1,000 tons/year)a ES-10
Table ES-6. Total Annual Costs and Benefits" for U.S., except California in 2025
(billions of 2011$, 7% Discount Rate)b ES-14
Table ES-7. Summary of Total Number of Annual Ozone and PM-Related Premature
Mortalities and Premature Morbidity: 2025 National Benefits a ES-14
Table ES-8. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S., except California (billions of 2011$, 7% Discount Rate)a ES-15
Table ES-9. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
2025 Scenario (nationwide benefits of attaining each alternative standard everywhere
in the U.S. except California) -Full Attainment a ES-16
Table ES-10. Total Annual Costs and Benefits'1 of Control Strategies Applied in
California, post-2025 (billions of 2011$, 7% Discount Rate)b ES-17
Table ES-11. Summary of Total Number of Annual Ozone and PM-Related Premature
Mortalities and Premature Morbidity: Post-2025a ES-17
Table ES-12. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for post-2025 - California (billions of 2011 $, 7% Discount Rate)a ES-18
Table ES-13. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
post-2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) - Full Attainment21 3-18
Table 3-1. 2011 and 2025 Base Case NOX and VOC Emissions by Sector (thousand
tons) 11
Table 3-2. List of Emissions Sensitivity Cases that Were Modeled in CAMx to
Determine Ozone Response Factors 3-15
xni
-------
Table 3-3. Anthropogenic NOX and VOC Emissions from the 2025 Base and Explicit
Control Cases* 3-16
Table 3A-1. Daily Maximum 8-hour Ozone Performance Statistics > 60 ppb by Region,
by Network 3 A-6
Table 3A-2. Key Monitoring Sites Used for the Ozone Time Series Analysis 3 A-11
Table 3 A-3. Emissions Reductions Applied to Create the Baseline Scenario* 3 A-25
Table 3 A-4. Emissions Reductions Applied Beyond the Baseline Scenario to Create the
70 ppb Scenario 3 A-25
Table 3A-5. Emissions Reductions Applied Beyond the Baseline Scenario to Create the
65 ppb Scenario 3 A-26
Table 3 A-6. Emissions Reductions Applied Beyond the Baseline Scenario to Create the
60 ppb Scenario 3 A-26
Table 3 A-7. Design Values for California Region Monitors 3 A-27
Table 3 A-8. Design Values for Southwest Monitors 3 A-31
Table 3A-9. Design Values for Central Region Monitors 3A-35
Table 3 A-10. Design Values for Midwest Monitors 3 A-39
Table 3A-11. Design Values for Northeast Monitors 3A-45
Table 3 A-12. Monitors Without Projections due to Insufficient High Modeling Days to
Meet EPA Guidance for Projecting Design Values 3 A-52
Table 3A-13. Monitors Determined to Have Design Values Affected by Winter Ozone
Events 53
Table 3 A-14. Monitors with Limited Response to Regional NOx and National VOC
Emissions Reductions in the 2025 Baseline 3 A-56
Table 4-1. Controls Applied for the Alternative Standard Analyses Control Strategy 4-7
Table 4-2. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Attainment of the Current Standard for the 2025 Baseline - U.S., except
California (1,000 tons/year)21 4-11
Table 4-3. Summary of Emission Reductions (Known and Unknown Controls) Applied
to Demonstrate Attainment in California for the post-2025 Baseline (1,000
tons/year)a 4-11
xiv
-------
Table 4-4. Number of Counties with Exceedances and Number of Additional Counties
Where Reductions Were Applied for the 2025 Alternative Standards Analyses - U.S.,
except California 4-18
Table 4-5. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Nationwide Attainment with a 70 ppb Ozone Standard in 2025, except
California (1,000 tons/year)21 4-18
Table 4-6. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Nationwide Attainment with a 65 ppb Ozone Standard in 2025 - except
California (1,000 tons/year)21 4-19
Table 4-7. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Nationwide Attainment with a 60 ppb Ozone Standard in 2025 - except
California (1,000 tons/year)a 4-19
Table 4-8. Summary of Emissions Reductions by Alternative Standard for Unknown
Controls for 2025 - except California (1,000 tons/year)21 4-21
Table 4-9. Summary of Emissions Reductions by Alternative Level for Unknown
Controls for post-2025 - California (1,000 tons/year)21 4-21
Table 4-10. Summary of Known and Unknown Emissions Reductions by Alternative
Standard Levels in 2025, Except California (1,000 tons/year)21 4-22
Table 4-11. Summary of Known and Unknown Emissions Reductions by Alternative
Standard Levels for post-2025-California (1,000 tons/year)21 4-23
Table 4A-1. Geographic Areas for Application of NOx Controls in the Baseline and
Alternative Standard Analyses - U.S., except California21 4A-2
Table 4A-2. Geographic Areas for Application of VOC Controls in the Baseline and
Alternative Standard Analyses - U.S., except California21 4A-3
Table 4A-3. Geographic Areas for Application of NOX Controls in the Baseline and
Alternative Standard Analyses - California21 4A-3
Table 4A-4. Geographic Areas for Application of VOC Controls in the Baseline and
Alternative Standard Analyses - California21 4A-4
Table 4A-5. NOX Control Measures Applied in the Baseline Analysis 4A-5
Table 4A-6. VOC Control Measures Applied in the Baseline Analysis 4A-6
Table 4A-7. NOx Control Measures Applied in the 70 ppb Alternative Standard
Analysis 4A-6
xv
-------
Table 4A-8. VOC Control Measures Applied in the 70 ppb Alternative Standard
Analysis 4A-8
Table 4A-9. NOx Control Measures Applied in the 65 ppb Alternative Standard
Analysis 4A-8
Table 4A-10. VOC Control Measures Applied in the 65 ppb Alternative Standard
Analysis 4A-10
Table 4A-11. NOx Control Measures Applied in the 60 ppb Alternative Standard
Analysis 4A-10
Table 4A-12. VOC Control Measures Applied in the 60 ppb Alternative Standard
Analysis 4A-12
Table 5-1. Estimated Monetized Benefits of Attainment of the Alternative Ozone
Standards for the 2025 Scenario (nationwide benefits of attaining each alternative
standard everywhere in the U.S. except California) - Full Attainment (billions of
2011$)a 5-3
Table 5-2. Estimated Monetized Benefits of Attainment of the Alternative Ozone
Standards for the Post-2025 Scenario (nationwide benefits of attaining each
alternative standard just in California) -Full Attainment (billions of 2011$) a 5-3
Table 5-3. Human Health Effects of Pollutants Potentially Affected by Strategies to
Attain the Primary Ozone Standards (endpoints included in the core analysis are
identified with a red checkmark) 5-5
Table 5-4. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population 5-29
Table 5-5. Asthma Prevalence Rates 5-29
Table 5-6. Criteria Used When Selecting C-RFunctions 5-33
Table 5-7. Health Endpoints and Epidemiological Studies Used to Quantify Ozone-
Related Health Impacts in the Core Analysis a 5-35
Table 5-8. Health Endpoints and Epidemiological Studies Used to Quantify PM2.5-
Related Health Impacts in the Core Analysis a 5-36
Table 5-9. Health Endpoints and Epidemiological Studies Used to Quantify Ozone-
Related Health Impacts in the Sensitivity Analysis a 5-37
Table 5-10. Unit Values for Economic Valuation of Health Endpoints (2011$) a 5-57
Table 5-11. Influence of Applied VSL Attributes on the Size of the Economic Benefits
of Reductions in the Risk of Premature Death (U.S. EPA, 2006a) 5-62
xvi
-------
Table 5-12. Unit Values for Hospital Admissions a 5-67
Table 5-13. Alternative Direct Medical Cost of Illness Estimates for Nonfatal Heart
Attacks21 5-69
Table 5-14. Estimated Costs Over a 5-Year Period of a Nonfatal Myocardial Infarction
(in2011$)a 5-69
Table 5-15. Elasticity Values Used to Account for Projected Real Income Growth a 5-74
Table 5-16. Adjustment Factors Used to Account for Projected Real Income Growth a 5-75
Table 5-17. Population-Weighted Air Quality Change for the Proposed and Alternative
Annual Primary Ozone Standards Relative to Analytical Baseline for 2025a 5-77
Table 5-18. Emission Reductions in Illustrative Emission Reduction Strategies for the
Proposed and Alternative Annual Primary Ozone Standards, by Pollutant and Region
Relative to Analytical Baseline -Full Attainment (tons)a 5-77
Table 5-19. Estimated Number of Avoided Ozone-Only Health Impacts for the Proposed
and Alternative Annual Ozone Standards (Incremental to the Analytical Baseline) for
the 2025 Scenario (nationwide benefits of attaining each alternative standard
everywhere in the U.S. except California) a'b 5-78
Table 5-20. Total Monetized Ozone-Only Benefits for the Proposed and Alternative
Annual Ozone Standards (Incremental to the Analytical Baseline) for the 2025
Scenario (nationwide benefits of attaining each alternative standard everywhere in
the U.S. except California) (millions of 2011) "^ 5-79
Table 5-21. Estimated Number of Avoided PM2.s-Related Health Impacts for the
Proposed and Alternative Annual Ozone Standards (Incremental to the Analytical
Baseline) for the 2025 Scenario (nationwide benefits of attaining each alternative
standard everywhere in the U.S. except California) a 5-80
Table 5-22. Monetized PM2.s-Related Health Co-Benefits for the Proposed and
Alternative Annual Ozone Standards (Incremental to Analytical Baseline) for the
2025 Scenario (nationwide benefits of attaining each alternative standard everywhere
in the U.S. except California) (Millions of 2011 )a'b'c 5-80
Table 5-23. Estimate of Monetized Ozone and PM2.5 Benefits for Proposed and
Alternative Annual Ozone Standards Incremental to the Analytical Baseline for the
2025 Scenario (nationwide benefits of attaining each alternative standard everywhere
in the U.S. except California)-Full Attainment (billions of 2011$) a 5-81
Table 5-24. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
2025 Scenario (nationwide benefits of attaining each alternative standard everywhere
in the U.S. except California) -Full Attainment a 5-82
xvn
-------
Table 5-25. Population-Weighted Air Quality Change for the Proposed and Alternative
Annual Primary Ozone Standards Relative to Analytical Baseline for post-2025 a 5-82
Table 5-26. Estimated Number of Avoided Ozone-Only Health Impacts for the Proposed
and Alternative Annual Ozone Standards (Incremental to the Analytical Baseline) for
the Post-2025 Scenario (nationwide benefits of attaining each alternative standard
just in California) a'b 5-83
Table 5-27. Total Monetized Ozone-Only Benefits for the Proposed and Alternative
Annual Ozone Standards (Incremental to the Analytical Baseline) for the post-2025
Scenario (nationwide benefits of attaining each alternative standard just in
California) (millions of 2011) a'b 5-84
Table 5-28. Estimated Number of Avoided PIVh.s-Related Health Impacts for the
Proposed and Alternative Annual Ozone Standards (Incremental to the Analytical
Baseline) for the post-2025 Scenario (nationwide benefits of attaining each
alternative standard just in California) a 5-85
Table 5-29. Monetized PM2.s-Related Health Co-Benefits for the Proposed and
Alternative Annual Ozone Standards (Incremental to Analytical Baseline) for the
post-2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) (Millions of 20ll)a'b 5-86
Table 5-30. Estimate of Monetized Ozone and PM2.5 Benefits for Proposed and Alternative
Annual Ozone Standards Incremental to the Analytical Baseline for the post-2025
Scenario (nationwide benefits of attaining each alternative standard just in
California)-Full Attainment (billions of 2011$) a 5-86
Table 5-31. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
post-2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) - Full Attainment a 5-87
Table 5A-1. Summary of Qualitative Uncertainty for Key Modeling Elements in Ozone
Benefits 5 A-6
Table 5B-1. Long-term Ozone Mortality Incidence at Various Assumed Thresholds a 5B-4
Table 5B-2. Application of Alternative (Expert Elicitation-Based Effect Estimates) to the
Modeling of PM2.5 Co-benefit Estimates for PM2.5 (avoided incidence) 5B-7
Table 5B-3. Ranges of Elasticity Values Used to Account for Projected Real Income
Growth a 5B-8
Table 5B-4. Ranges of Adjustment Factors Used to Account for Projected Real Income
Growth to 2024 a 5B-8
Table 5B-5. Sensitivity of Monetized Ozone Benefits to Alternative Income Elasticities
in 2025 (Millions of 2011$) a 5B-8
xvm
-------
Table 5C-1. Potential Reduction in Premature Mortality by Age Range from Attaining
Alternate Ozone Standards (2025 scenario) a'b 5C-4
Table 5C-2. Potential Years of Life Gained by Age Range from Attaining Alternate
Ozone Standards (2025 Scenario) a'b 5C-4
Table 5C-3. Estimated Percent Reduction in All-Cause Mortality Attributed to the
Proposed Primary Ozone Standards (2025 Scenario) a 5C-5
Table 5C-4. Population Exposure in the Baseline Sector Modeling (used to generate the
benefit-per-ton estimates) Above and Below Various Concentrations Benchmarks in
the Underlying Epidemiology Studies a 5C-12
Table 5C-5. Fraction of Total Core Benefits Associated with Partial Attainment
(application of known controls) (2025 Scenario) 5C-14
Table 5C-6. Estimated Number of Avoided Ozone-Only Health Impacts for the
Alternative Annual Primary Ozone Standards (Incremental to the Analytical
Baseline) for the Partial Attainment of the 2025 Scenario (known controls) a>b 5C-15
Table 5C-7. Total Monetized Ozone-Only Benefits for the Alternative Annual Primary
Ozone Standards (Incremental to the Analytical Baseline) for the Partial Attainment
of the 2025 Scenario (using known controls) ^ 5C-16
Table 5C-8. Estimated Number of Avoided PM2.s-Related Health Impacts for the
Alternative Annual Primary Ozone Standards (Incremental to the Analytical
Baseline) for the Partial Attainment of the 2025 Scenario (using known controls) a>b5C-17
Table 5C-9. Monetized PM2.s-Related Health Co-Benefits for the Alternative Annual
Primary Ozone Standards (Incremental to Analytical Baseline) for the Partial
Attainment of the 2025 Scenario (using known controls) a'b'c 5C-18
Table 5C-10. Combined Estimate of Monetized Ozone and PM2.5 Benefits for the
Alternative Annual Primary Ozone Standards for the Partial Attainment of the 2025
Scenario (using known controls) (billions of 2011$) a>b 5C-18
Table 5D-1. Health Endpoints and Epidemiological Studies Used to Quantify PM2.5-
related Health Impacts in the Core Analysis a 5D-4
Table 5D-2. Summary of Effect Estimates from Associated with Change in Long-Term
Exposure to PM2.5 in Recent Cohort Studies in North America 5D-10
Table 6-1. Welfare Effects of NOx and VOC Emissions 6-3
Table 6-2. Change in Consumer and Producer Surplus in the Forestry Sector from
Attaining Alternative Ozone Standard Levels Compared to Attaining the Current
Ozone Standard (Million 2011$) 6-11
xix
-------
Table 6-3. Percent Change in Consumer and Producer Surplus in the Forestry Sector
from Attaining Alternative Ozone Standard Levels Compared to Attaining the
Current Ozone Standard 6-11
Table 6-4. Change in Consumer and Producer Surplus in the Agricultural Sector from
Attaining Alternative Ozone Standard Levels Compared to Attaining the Current
Ozone Standard (Million 2011$) 6-11
Table 6-5. Percent Change in Consumer and Producer Surplus in the Agricultural
Sector from Attaining Alternative Ozone Standard Levels Compared to Attaining the
Current Ozone Standard 6-12
Table 6-6. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry from Attaining Alternative Ozone Standard Levels Compared to Attaining
the Current Ozone Standard, 2010-2040, Million 2011$ (3% Discount Rate) 6-13
Table 6-7. Annualized Percent Changes in Consumer and Producer Surplus in
Agriculture and Forestry from Attaining Alternative Ozone Standard Levels
Compared to Attaining the Current Ozone Standard, 2010-2040, (3% Discount Rate) 6-13
Table 6-8. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry from Attaining Alternative Ozone Standard Levels Compared to Attaining
the Current Ozone Standard, 2010-2040, Million 2011$ (7% Discount Rate) 6-13
Table 6-9. Annualized Percent Changes in Consumer and Producer Surplus in
Agriculture and Forestry from Attaining Alternative Ozone Standard Levels
Compared to Attaining the Current Ozone Standard, 2010-2040, (7% Discount Rate) 6-14
Table 6-10. Increase in Carbon Sequestration from Attaining Alternative Ozone Standard
Levels Compared to Attaining the Current Ozone Standard, MMtCO2e 6-15
Table 6-11. Percent Change in Carbon Sequestration from Attaining Alternative Ozone
Standard Levels Compared to Attaining the Current Ozone Standard 6-16
Table 6 A-1. Comparison of Total Cropland and Forestland NLCD Area (sq m) 6 A-2
Table 6A-2. Parameter Values Used for Crops and Tree Species 6A-5
Table 6A-3. Mapping of Ozone Impacts on Crops to FASOMGHG Crops 6A-6
Table 6A-4. Mapping of Ozone Impacts on Forests to FASOMGHG Forest Types 6A-7
Table 6A-5. Forestland W126 Ozone Values under Alternative Scenarios 6A-10
Table 6A-6. Cropland W126 Ozone Values under Modeled Scenarios 6A-11
Table 6A-7. Agricultural Production Fisher Indices (Current conditions =100) 6A-25
Table 6A-7a. Agricultural Price Fisher Indices (Current Conditions = 100) 6A-25
xx
-------
Table 6A-8. Major Crop Acreage, Million Acres 6A-28
Table 6A-9. Forest Products Production, Million Cubic Feet 6A-30
Table 6A-10. Forest Product Prices, U. S. Dollars per Cubic Foot 6A-31
Table 6A-11. Forest Product Prices and Percentage Change, U.S. Dollars per Cubic Foot 6A-32
Table 6A-12. Forest Acres Harvested, Thousand Acres 6A-33
Table 6A-13. Existing and New Forest Inventory, Million Cubic Feet 6A-34
Table 6A-14. Land Use by Major Category, Thousand Acres 6A-35
Table 6A-15. Consumer and Producer Surplus in Agriculture, Million 2010 U.S. Dollars.6A-37
Table 6A-16. Consumer and Producer Surplus in Forestry, Million 2010 U.S. Dollars 6A-38
Table 6A-17. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry, 2010-2044, Million 2010 U.S. Dollars (3% Discount Rate) 6A-39
Table 6A-18. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry, 2010-2044, Million 2010 U.S. Dollars (7% Discount Rate) 6A-39
Table 6A-19. Carbon Storage, MMtCO2e 6A-41
Table 6A-20. Forestry Carbon Sequestration, MMtCO2e 6A-42
Table 7-1. Summary of Known Annualized Control Costs by Sector for 70 ppb for
2025-U.S., except California (millions of 2011$) 7-8
Table 7-2. Summary of Known Annualized Control Costs by Sector for 65 ppb for
2025 U.S., except California (millions of 2011$) 7-9
Table 7-3. Summary of Known Annualized Control Costs by Sector for 60 ppb for
2025 U.S., except California (millions of 2011$) 7-9
Table 7-4. Emissions and Cost Information for Major NOx Rules Issued Between 1997
and 2008 7-14
Table 7-5. Comparison of Key Assumptions Used in Developing Estimates of NOx
Emissions Controls and Costs across Past and Current Ozone NAAQS Regulatory
Impact Analyses a 7-16
Table 7-6. Average NOX Offset Prices for Four Areas (2011$)a 7-24
Table 7-7. Annualized NOx Offset Prices for Four Areas (2011 $)a 7-25
xxi
-------
Table 7-8. Extrapolated Control Costs in 2025 by Alternative Standard for 2025 -- U.S.,
except California (millions of 2011 $) 7-31
Table 7-9. Extrapolated Control Costs in 2025 by Alternative Standard for Post-2025 -
California (millions of 2011$) 7-31
Table 7-10. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S., except California (millions of 2011$, 7% Discount Rate)a 7-33
Table 7-11. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for post-2025 - California (millions of 2011$, 7% Discount Rate)a 7-33
Table 7A-1. Costs for Known NOx Controls in the 70 ppb Analysis (millions, 2011 $)a.... 7A-1
Table 7A-2. Costs for Known VOC Controls in the 70 ppb Analysis (millions, 2011$) a...7A-2
Table 7A-3. Costs for Known NOX Controls in the 65 ppb Analysis (millions, 2011 $) a.... 7A-3
Table 7A-4. Costs for Known VOC Controls in the 65 ppb Analysis (millions, 2011$) a...7A-4
Table 7A-5. Costs for Known NOx Controls in the 60 ppb Analysis (millions, 2011 $) a... 7A-5
Table 7A-6. Costs for Known VOC Controls in the 60 ppb Analysis (millions, 2011$) a...7A-6
Table 7A-7. Extrapolated Control Costs in 2025 by Alternative Standard for 2025 U.S.,
except California, using Alternative Average Cost Assumptions (millions of 2011$) .7A-7
Table 7A-8. Extrapolated Control Costs in 2025 by Alternative Standard for Post-2025
California, using Alternative Average Cost Assumptions (millions of 2011$) 7A-7
Table 7A-9. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S. using Alternative Cost Assumption for Extrapolated Costs,
except California (millions of 20 ll$)a 7A-7
Table 7A-10. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for Post-2025 California - U.S. using Alternative Cost Assumption for
Extrapolated Costs (millions of 20 ll$)a 7A-8
Table 8-1. Total Costs, Total Monetized Benefits, and Net Benefits in 2025 for U.S.,
except California (billions of 20 ll$)a 8-4
Table 8-2. Total Costs, Total Monetized Benefits, and Net Benefits of Control
Strategies Applied in California, Post-2025 (billions of 2011 $)a 8-4
Table 8-3. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S., except California (billions of 2011$, 7% Discount Rate)a 8-5
Table 8-4. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for post-2025 - California (billions of 2011 $, 7% Discount Rate)a 8-5
xxn
-------
Table 8-5. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
2025 Scenario (nationwide benefits of attaining each alternative standard everywhere
in the U.S. except California) - Full Attainment a 8-5
Table 8-6. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
post-2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) - Full Attainment a 8-6
Table 8-7. Human Health Effects of Pollutants Potentially Affected by Strategies to
Attain the Primary Ozone Standards 8-6
Table 8-8. Relevant Factors and Their Potential Implications for Attainment 8-12
Table 9-1. Summary of Population Totals and Demographic Categories for Areas of Interest
and National Perspective 9-6
Table 9-2. Summary of Population Totals and Demographic Categories for Areas of Interest
and National Perspective 9-6
Table 10-1. Summary of Direct Labor Impacts for SCR Installation at EGUs 10-11
Table 10-2. Key Assumptions in Labor Analysis for EGUs 10-12
Table 10-3. Summary of Direct Labor Impacts for Individual ICI Boilers 10-14
Table 10-4. Estimated Direct Labor Impacts for Individual SNCR Applied to a Mid-
Sized Cement Kiln (125-208 tons clinker/hr) 10-15
xxm
-------
LIST OF FIGURES
Figure ES-1. Analytical Flowchart for Primary Standards Analyses ES-5
Figure ES-2. Projected Ozone Design Values in the 2025 Baseline Scenario ES-7
Figure ES-3. Projected Ozone Design Values in the post-2025 Baseline Scenario ES-10
Figure 2-1. Map of U.S. Ambient Os Monitoring Sites Reporting Data to EPA During
the 2009-2013 Period 2-6
Figure 2-2. Trend in U.S. Annual 4th Highest Daily Maximum 8-hour Ozone
Concentrations in ppb, 2000 to 2013. Solid center line represents the median value
across monitoring sites, dashed lines represent 25th and 75th percentile values, and
top/bottom lines represent 10th and 90th percentile values 2-8
Figure 2-3. Map of 8-hour Ozone Design Values in ppb, Averaged Across the 2009-
2011,2010-2012, and 2011-2013 Periods 2-8
Figure 2-4. Trend in U.S. Annual W126 Concentrations in ppm-hrs, 2000 to 2013. Solid
center line represents the median value across monitoring sites, dashed lines
represent 25th and 75th percentile values, and top/bottom lines represent 10th and
90th percentile values 2-10
Figure 2-5. Map of 3-year Average W126 Values in ppm-hrs, Averaged Across the
2009-2011,2010-2012, and 2011-2013 Periods 2-10
Figure 2-6. Map of 2007 CMAQ-estimated Seasonal Mean of 8-hour Daily Maximum
Ozone from Natural Background (ppb) based on Zero-Out Modeling 2-13
Figure 2-7. Map of 2007 CMAQ-estimated Seasonal Mean of 8-hour Daily Maximum
Ozone from North American Background (ppb) based on Zero-out Modeling 2-14
Figure 2-8. Map of 2007 CMAQ-estimated Seasonal Mean of 8-hour Daily Maximum
Ozone from United States Background (ppb) based on Zero-Out Modeling 2-15
Figure 2-9. Map of Site-Specific Ratios of U.S. Background to Total Seasonal Mean
Ozone based on 2007 CMAQ Zero-Out Modeling 2-17
Figure 2-10. Map of Site-Specific Ratios of Apportionment-Based U.S. Background to
Seasonal Mean Ozone based on 2007 CAMx Source Apportionment Modeling 2-18
Figure 2-11. Distributions of Absolute Estimates of Apportionment-Based U.S.
Background (all site-days), Binned by Modeled MDA8 from the 2007 Source
Apportionment Simulation 2-18
xxiv
-------
Figure 2-12. Distributions of the Relative Proportion of Apportionment-Based U.S.
Background to Total Ozone (all site-days), Binned by Modeled MDA8 from the
2007 Source Apportionment Simulation 2-19
Figure 3 -1. Map of the C AMx Modeling Domain Used for Ozone NAAQ S RIA 3-4
Figure 3-2. Map of Counties for Which Explicit Emissions Controls Were Identified and
Modeled in CAMx (shaded in orange) and Counties that Contained One or More
Monitor Projected above 70 ppb in the 2025 Base Case Modeling (shaded in blue). ...3-13
Figure 3-3. Five U.S. Regions Used to Create Across-the-Board Emissions Reduction
and Combination Cases 3-15
Figure 3-4. Map of VOC Impact Areas Applied in the Evaluation of a 60 ppb
Alternative Standard Level 3-22
Figure 3-5. Projected post-2025 Baseline Scenario May-September Mean of 8-hr Daily
Maximum Ozone (ppb) 3-32
Figure 3-6. Change in May-September Mean of 8-hr daily Maximum Ozone (ppb)
between the post-2025 Baseline Scenario and the post-2025 70 ppb Scenario 3-33
Figure 3-7. Change in May-September Mean of 8-hr daily Maximum Ozone (ppb)
between the post-2025 Baseline Scenario and the post-2025 65 ppb Scenario 3-34
Figure 3-8. Change in May-September Mean of 8-hr Daily Maximum Ozone (ppb)
between the post-2025 Baseline Scenario and the post-2025 60 ppb Scenario 3-35
Figure 3-9. Projected post-2025 Baseline Scenario W126 Values (ppm-hrs) 3-37
Figure 3-10. Projected post-2025 70 ppb Scenario W126 Values (ppm-hrs) 3-38
Figure 3-11. Projected post-2025 65 ppb Scenario W126 Values (ppm-hrs) 3-39
Figure 3-12. Projected post-2025 60 ppb Scenario W126 Values (ppm-hrs) 3-40
Figure 3A-1. Distribution of observed and predicted MDA8 ozone by month for the
period May through September for the Northeast subregion, (a) AQS network and
(b) CASTNet network, [symbol = median; top/bottom of box = 75th/25th
percentiles; top/bottom line = max/min values] 3 A-6
Figure 3A-2. Distribution of observed and predicted MDA8 ozone by month for the
period May through September for the Southeast subregion, (a) AQS network and
(b) CASTNet network 3A-7
Figure 3A-3. Distribution of observed and predicted MDA8 ozone by month for the
period May through September for the Midwest subregion, (a) AQS network and (b)
CASTNet network 3A-7
XXV
-------
Figure 3A-4. Distribution of observed and predicted MDA8 ozone by month for the
period May through September for the Central states, (a) AQS network and (b)
CASTNet network 3A-8
Figure 3A-5. Distribution of observed and predicted MDA8 ozone by month for the
period May through September for the West, (a) AQS network and (b) CASTNet
network 3 A-8
Figure 3 A-6. Mean Bias (ppb) of MDA8 ozone greater than 60 ppb over the period May-
September 2011 at AQS and CASTNet monitoring sites in 12-km U.S. modeling
domain 3 A-9
Figure 3 A-7. Normalized Mean Bias (%) of MDA8 ozone greater than 60 ppb over the
period May-September 2011 at AQS and CASTNet monitoring sites in 12-km U.S.
modeling domain 3 A-9
Figure 3 A-8. Mean Error (ppb) of MDA8 ozone greater than 60 ppb over the period May-
September 2011 at AQS and CASTNet monitoring sites in 12-km U.S. modeling
domain 3 A-10
Figure 3 A-9. Normalized Mean Error (%) of MDA8 ozone greater than 60 ppb over the
period May-September 2011 at AQS and CASTNet monitoring sites in 12-km U.S.
modeling domain 3 A-10
Figure 3A-10a. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 60195001 in Fresno Co., California 3A-11
Figure 3A-10b. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 60710005 in San Bernardino Co., California 3A-12
Figure 3A-10c. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 80350004 in Douglas Co., Colorado 3A-12
Figure 3A-10d. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 484392003 in Tarrant Co., Texas 3A-13
Figure 3A-10e. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 480391004 in Brazoria Co., Texas 3A-13
Figure 3A-10f Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 551170006 in Sheboygan Co., Wisconsin 3A-14
Figure 3A-10g. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 261630019 in Wayne Co., Michigan 3A-14
Figure 3A-10L Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 211110067 in Jefferson Co., Kentucky 3A-15
xxvi
-------
Figure 3A-10L Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 420031005 in Allegheny Co., Pennsylvania 3A-15
Figure 3A-10J . Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 240210037 in Frederick Co., Maryland 3 A-16
Figure 3 A-10k. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 240251001 in Harford Co., Maryland 3 A-16
Figure 3 A-101. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 360810124 in Queens, New York 3A-17
Figure 3 A-10m. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 361030002 in Suffolk County, New York 3A-17
Figure 3 A-l 1. a) Depiction of governing wind patterns and topography, b) California
Air Basins and sub-regions used for this analysis. Northern sub-region is outlined in
pink and southern sub-region is outlined in blue 3A-19
Figure 3 A-12. Impact of 50% anthropogenic California NOx cuts (ppb) on 8-hr daily
average ozone concentrations on three days in 2011 3A-20
Figure 3A-13. Downwind California receptor regions for Northern California (green)
and Southern California (purple) 3 A-21
Figure 3A-14. Impact of 50% US anthropogenic VOC cuts (ppb) on 8-hr daily average
ozone concentrations on three days in 2011 3A-22
Figure 3 A-l 5. Change in July average of 8-hr daily maximum ozone concentration (ppb)
due to 50% cut in US anthropogenic VOC emissions 3 A-23
Figure 3A-16. Projected change in 2025 ozone design values with an additional 75%
regional NOx control (Southwest region; stars represent sites identified in Table 3 A-
14) 3A-55
Figure 3A-17. Location of sites identified in Table 3A-14 3A-58
Figure 3A-18. Projected change in 2025 ozone design values with an additional 90%
California NOx control (Southwest region; stars represent sites identified in Table
3A-14) 3A-60
Figure 3 A-19. Daily 8-hr maximum ozone values at ozone monitor in Weminuche
Wilderness area in La Plata County Colorado from 2009-2013. Horizontal line
provided at 65 ppb 3 A-62
Figure 4-1. Counties Projected to Exceed the Baseline Level of the Current Ozone
Standard (75 ppb) in 2025 Base Case 4-9
XXVll
-------
Figure 4-2. Counties Where Emissions Reductions Were Applied to Demonstrate
Attainment of the Current Standard for the Baseline Analysis 4-10
Figure 4-3. Projected Ozone Design Values in the 2025 Baseline Scenario 4-13
Figure 4-4. Projected Ozone Design Values in the post-2025 Baseline Scenario 4-14
Figure 4-5. Counties Where Emissions Reductions Were Applied to Demonstrate
Attainment with a 70 ppb Ozone Standard in the 2025 Analysis 4-16
Figure 4-6. Counties Where Emissions Reductions Were Applied to Demonstrate
Attainment with 65 and 60 ppb Ozone Standards in the 2025 Analyses 4-17
Figure 5-1. Illustration of BenMAP-CE Approach 5-14
Figure 5-2. Data Inputs and Outputs for the BenMAP-CE Program 5-16
Figure 5-3. Procedure for Generating Benefits Estimates for the 2025 and Post-2025
Scenarios 5-17
Figure 5C-1. Premature Ozone-related Deaths Avoided for the Alternative Standards
(2025 scenario) According to the Baseline Ozone Concentrations 5C-7
Figure 5C-2. Cumulative Probability Plot of Premature Ozone-related Deaths Avoided for
the Alternative Standards (2025 scenario) According to the Baseline Ozone
Concentrations 5C-8
Figure 5C-3. Percentage of Adult Population (age 30+) by Annual Mean PM2.5 Exposure
in the Baseline Sector Modeling (used to generate the benefit-per-ton estimates)* ... 5C-12
Figure 5C-4. Cumulative Distribution of Adult Population (age 30+) by Annual Mean
PM2.5 Exposure in the Baseline Sector Modeling (used to generate the benefit-per-ton
estimates)* 5C-13
Figure 6A-1. Cropland Area (sq m) by County according to NLCD 2006 Data 6A-2
Figure 6A-2. Cropland Area (sq m) by County according to NLCD 2011 Data 6A-3
Figure 6A-3. Forest Area (sq m) by County according to NLCD 2006 Data 6A-3
Figure 6A-4. Forest Area (sq m) by County according to NLCD 2011 Data 6A-4
Figure 6A-5. FASOMGHG Modeling Flowchart 6A-9
Figure 6A-6. Ozone Reductions with Respect to 75 ppb under Alternative Scenarios 6A-14
Figure 6A-7. Percentage Changes in Corn RYGs with Respect to the 75 ppb Scenario ....6A-16
Figure 6A-8. Percentage Changes in Cotton RYGs with Respect to the 75 ppb Scenario..6A-17
xxvm
-------
Figure 6A-9. Percentage Changes in Potato RYGs with Respect to the 75 ppb Scenario ..6A-18
Figure 6A-10. Percentage Changes in Sorghum RYGs with Respect to the 75 ppb
Scenario 6A-19
Figure 6A-11. Percentage Changes in Soybean RYGs with Respect to the 75 ppb
Scenario 6 A-20
Figure 6A-12. Percentage Changes in Winter Wheat RYGs with Respect to the 75 ppb
Scenario 6 A-21
Figure 6A-13. Percentage Changes in Softwood RYGs with Respect to the 75 ppb
Scenario 6 A-22
Figure 6A-14. Percentage Changes in Hardwood RYGs with Respect to the 75 ppb
Scenario 6 A-23
Figure 6A-15. Carbon Storage in Forestry Sector, MMtCChe 6A-41
Figure 7-1. Marginal Costs for Known NOX Controls for All Source Sectors (EGU, non-
EGU Point, Nonpoint, and Nonroad) 7-5
Figure 7-2. Observed but incomplete MACC (solid line) based on known controls
identified by current tools and complete MACC (dashed line) where gaps indicate
abatement not identified by current tools 7-12
Figure 10-1. Size Distribution of 145 Existing Coal-Fired EGU Units without SCR NOx
Controls 10-11
XXIX
-------
EXECUTIVE SUMMARY
Overview
In setting primary national ambient air quality standards (NAAQS), the EPA's
responsibility under the law is to establish standards that protect public health. The Clean Air Act
(the Act) requires the EPA, for each criteria pollutant, to set a standard that protects public health
with "an adequate margin of safety." As interpreted by the Agency and the courts, the Act
requires the EPA to base this decision on health considerations only; economic factors cannot be
considered. The prohibition against considering cost in the setting of the primary air quality
standards does not mean that costs, benefits or other economic considerations are unimportant.
The Agency believes that consideration of costs and benefits is an essential decision-making tool
for the efficient implementation of these standards. The impacts of costs, benefits, and efficiency
are considered by the States when they make decisions regarding what timelines, strategies, and
policies are appropriate for their circumstances.
The EPA is proposing to revise the level of the ozone NAAQS to within a range of 65
ppb to 70 ppb and is soliciting comment on alternative standard levels below 65 ppb, as low as
60 ppb. The EPA is also proposing to revise the level of the secondary standard to within the
range of 65 ppb to 70 ppb to provide increased protection against vegetation-related effects on
public welfare.1 The EPA performed an illustrative analysis of the potential costs, human health
benefits, and welfare co-benefits of nationally attaining primary alternative ozone standard levels
and did not estimate any incremental costs and benefits associated with attaining a revised
secondary standard. Per Executive Order 12866 and the guidelines of OMB Circular A-4, this
Regulatory Impact Analysis (RIA) presents the analyses of the following alternative standard
levels — 70 ppb, 65 ppb, and 60 ppb. The cost and benefit estimates below are calculated
incremental to a 2025 baseline that incorporates air quality improvements achieved through the
projected implementation of existing regulations and full attainment of the existing ozone
1 As an initial matter, the EPA is proposing that ambient ozone concentrations in terms of a three-year average
W126 index value within the range from 13 parts per million-hours (ppm-hours) to 17 ppm-hours would provide the
requisite protection against known or anticipated adverse effects to the public welfare, which data analyses indicate
would provide air quality in terms of three-year average W126 index values of a range at or below 13 ppm-hours to
17 ppm-hours. Data analyses also indicate that actions taken to attain a standard in the range of 65 ppb to 70 ppb
would also improve air quality as measured by the W126 metric.
ES-1
-------
NAAQS (75 ppb). The 2025 baseline reflects, among other existing regulations, the Mercury
and Air Toxics Standard, the Clean Air Interstate Rule, the Tier 3 Motor Vehicle Emission and
Fuel Standards, and adjustments for the Clean Power Plan, all of which will help many areas
move toward attainment of the existing ozone standard (see Chapter 3, Section 3.1.3 for
additional information).
In this RIA we present the primary costs and benefits estimates for 2025. We assume
that potential nonattainment areas everywhere in the U.S., excluding California, will be
designated such that they are required to reach attainment by 2025, and we developed our
projected baselines for emissions, air quality, and populations for 2025.
The EPA will likely finalize designations for a revised ozone NAAQS in late
2017. Depending on the precise timing of the effective date of those designations, nonattainment
areas classified as Marginal will likely have to attain in either late 2020 or early
2021. Nonattainment areas classified as Moderate will likely have to attain in either late 2023 or
early 2024. If a Moderate nonattainment area qualifies for two 1-year extensions, the area may
have as late as 2026 to attain. Lastly, Serious nonattainment areas will likely have to attain in
late 2026 or early 2027. We selected 2025 as the primary year of analysis because most areas of
the U.S. will likely be required to meet a revised ozone standard by 2025 and because it provided
a good representation of the remaining air quality concerns that Moderate nonattainment areas
would face; states with areas classified as Moderate and higher are required to develop
attainment demonstration plans for those nonattainment areas.
In estimating the incremental costs and benefits of potential alternative standards, we
recognize that there are several areas that are not required to meet the existing ozone standard by
2025. The Clean Air Act allows areas with more significant air quality problems to take
additional time to reach the existing standard. Several areas in California are not required to
meet the existing standard by 2025 and may not be required to meet a revised standard until
sometime between 2032 and 2037.2 We were not able to project emissions and air quality
2 The EPA will likely finalize designations for a revised ozone NAAQS in late 2017. Depending on the precise
timing of the effective date of those designations, nonattainment areas classified as Severe 15 will likely have to
ES-2
-------
beyond 2025 for California, however, we adjusted baseline air quality to reflect mobile source
emissions reductions for California that would occur between 2025 and 2030; these emissions
reductions were the result of mobile source regulations expected to be fully implemented by
2030. While there is uncertainty about the precise timing of emissions reductions and related
costs for California, we assume costs occur through the end of 2037 and beginning of 2038. In
addition, we model benefits for California using projected population demographics for 2038.
Because of the different timing for incurring costs and accruing benefits and for ease of
discussion throughout the analyses, we refer to the different time periods for potential attainment
as 2025 and post-2025 to reflect that (1) we did not project emissions and air quality for any year
other than 2025; (2) for California, emissions controls and associated costs are assumed to occur
through the end of 2037 and beginning of 2038; and (3) for California benefits are modeled using
population demographics in 2038. It is not straightforward to discount the post-2025 results for
California to compare with or add to the 2025 results for the rest of the U.S. While we estimate
benefits using 2038 information, we do not have good information on precisely when the costs of
controls will be incurred. Because of these differences in timing related to California attaining a
revised standard, the separate costs and benefits estimates for post-2025 should not be added to
the primary estimates for 2025.
ES.l Overview of Analytical Approach
This RIA consists of multiple analyses including an assessment of the nature and sources
of ambient ozone (Chapter 2 - Defining the Air Quality Problem); estimates of current and
future emissions of relevant precursors that contribute to the problem; air quality analyses of
baseline and alternative control strategies (Chapter 3 - Air Quality Modeling and Analysis);
development of illustrative control strategies to attain the primary alternative standard levels
(Chapter 4 - Control Strategies and Emissions Reductions); estimates of the incremental benefits
of attaining the primary alternative standard levels (Chapter 5 - Human Health Benefits); a
qualitative discussion of the welfare co-benefits of attaining the primary alternative standard
levels (Chapter 6 - Welfare Co-Benefits of the Primary Standard); estimates of the incremental
attain sometime between late 2032 and early 2033 and nonattainment areas classified as Extreme will likely have to
attain by December 31, 2037.
ES-2
-------
costs of attaining the primary alternative standard levels (Chapter 7 - Engineering Cost Analysis
and Economic Impacts); a comparison and discussion of the benefits and costs (Chapter 8 -
Comparison of Costs and Benefits); an analysis of the impacts of the relevant statutory and
executive orders (Chapter 9 - Statutory and Executive Order Impact Analysis); and a discussion
of the theoretical framework used to analyze regulation-induced employment impacts, as well as
information on employment related to installation of NOx controls on coal and gas-fired electric
generating units, industrial boilers, and cement kilns (Chapter 10 - Qualitative Discussion of
Employment Impacts of Air Quality).
Because States are ultimately responsible for implementing strategies to meet revised
standards, this RIA provides insights and analysis of a limited number of illustrative control
strategies that states might adopt to meet a revised standard. The goal of this RIA is to provide
estimates of the costs and benefits of the illustrative attainment strategies to the meet each
alternative standard level. The flowchart below (Figure ES-1) outlines the analytical steps taken
to illustrate attainment with the potential alternative standard levels, and the following
discussion, by primary flowchart section, describes the steps taken.
ES-4
-------
Project Base Year (Emissions & Air
Quality)
Establish 2023 Baseline Air
Qualm-
Establish
Baseline
Estimate Air Qualify for
Attainment of Alternative
NAAQS Levels
Costs Associated with Known
Controls Analysis for Partial
Attainment
Benefits Analysis for Partial
Attainment
Control Strategies,
Emissions
Reductions from
Known Controls,
Benefits
Extrapolated Costs For Full
Attainment
Benefits Analysis for Full
Attainment
Emissions
Reductions Beyond
Known Controls
Estimate Total Costs
Estimate Total Benefits
Figure ES-1. Analytical Flowchart for Primary Standards Analyses
ES. 1.1 Establishing the Baseline
The future year base case reflects emissions projected from 2011 to 2025 and
incorporates current state and federal programs, including the Tier 3 Motor Vehicle Emission
and Fuel Standards (U.S. EPA, 2014a) (see Chapter 3, Section 3.1.3 for a discussion of the rules
included in the base case). The base case does not include control programs specifically for the
purpose of attaining the existing ozone standard (75 ppb). The baseline builds on the future year
base case and reflects the additional emissions reductions needed to reach attainment of the
current ozone standard (75 ppb), as well as adjustments for the Clean Power Plan (U.S. EPA,
2014b).
We performed a national scale air quality modeling analysis to estimate ozone
concentrations for the future base case year of 2025. To accomplish this, we modeled multiple
ES-5
-------
emissions cases for 2025, including the 2025 base case and twelve 2025 emissions sensitivity
simulations. The twelve emissions sensitivity simulations were used to develop ozone sensitivity
factors (ppb/ton) from the modeled response of ozone to changes in NOx and VOC emissions
from various sources and locations. These ozone sensitivity factors were then used to determine
the amount of emissions reductions needed to reach the 2025 baseline and evaluate potential
alternative standard levels of 70, 65, and 60 ppb incremental to the baseline. We used the
estimated emissions reductions needed to reach each of these standard levels to analyze the costs
and benefits of alternative standard levels.
ES. 1.2 Control Strategies and Emissions Reductions
The EPA analyzed illustrative control strategies that areas across the U.S. might employ
to attain alternative revised primary ozone standard levels of 70, 65, and 60 ppb. The EPA
analyzed the impact that additional emissions control technologies and measures, across
numerous sectors, would have on predicted ambient ozone concentrations incremental to the
baseline. These control measures, also referred to as known controls, are based on information
available at the time of this analysis and include primarily end-of-pipe control technologies. In
addition, to attain some of the alternative primary standard levels analyzed, some areas needed
additional emissions reductions beyond the known controls, and we refer to these as unknown
controls (see Chapter 7, Section 7.2 for additional information).
Using average ozone response factors, we estimated the portion of the emissions
reductions required to meet the baseline, including any additional emissions reductions beyond
known controls. We then estimated the emissions reductions incremental to the baseline that
were needed to meet the alternative standard levels of 70, 65, and 60 ppb. Costs of controls
incremental to (i.e., over and above) the baseline emissions reductions are attributed to the costs
of meeting the alternative standard levels. These emissions reductions can come from both
specific known controls, as well as unknown controls. The baseline shows that by 2025, while
ozone air quality would be significantly better than today under current requirements, depending
on the alternative standard level analyzed, several areas in the Eastern, Central, and Western U.S.
would need to develop and adopt additional controls to attain alternative standard levels (see
Chapter 4, Section 4.3).
ES-6
-------
ES. 1.2.1 Emissions Reductions from Known Controls in 2025
Figure ES-2 shows the counties projected to exceed the alternative standard levels
analyzed for 2025 for areas other than California. For the 70 ppb alternative standard level,
emissions reductions were required for monitors in the Central and Northeast regions. For the 65
and 60 ppb alternative standard levels, emissions reductions were applied in all regions with
projected baseline design values (DVs) above these levels.3 For the 60 ppb alternative standard
level, additional VOC emissions reductions were identified in Chicago because some sites in that
area experienced NOX disbenefits, meaning that the regional NOX emissions reductions resulted
in ozone increases from below 60 ppb to above 60 ppb. Tables ES-1 through ES-3 show the
emissions reductions from known controls for the alternative standard levels analyzed.
Legend
I 9 counties are projected tc exceed 70 ppb
59 additions! counties are pr ejected tc be betow 70 but exceed ef ppb
173 additional counties are projected tc be beiow 65 but exceed 60 ppb
0 200 400
I i i i I
800 Kilometers
i I
Figure ES-2. Projected Ozone Design Values in the 2025 Baseline Scenario
3 A design value is a statistic that describes the air quality status of a given area relative to the level of the NAAQS.
Design values are typically used to classify nonattainment areas, assess progress toward meeting the NAAQS, and
develop control strategies.
ES-7
-------
Table ES-1. Summary of Emission Reductions by Sector for Known Controls for 70 ppb
Proposed Alternative Standard Level for 2025, except California (1,000
tons/year)"
Geographic Area
East
West
Emissions Sector
ECU
Non-EGU Point
Nonpoint
Nonroad
Total
ECU
Non-EGU Point
Nonpoint
Nonroad
Total
NOx
25
210
260
5
490
-
-
-
-
-
voc
-
0.98
54
-
55
-
-
-
-
-
"Estimates are rounded to two significant figures.
Table ES-2. Summary of Emission Reductions by Sector for Known Controls for 65 ppb
Proposed Alternative Standard Level for 2025 - except California (1,000
tons/year)"
Geographic Area Emissions Sector NOx VOC
EGU 170
Non-EGU Point 410 3.6
East Nonpoint 420 95
Nonroad 12
Total 1,000 99
EGU 36
Non-EGU Point 38 0.47
West Nonpoint 37 6.6
Nonroad L3 -
Total 110 7
"Estimates are rounded to two significant figures.
Table ES-3. Summary of Emission Reductions by Sector for Known Controls for 60 ppb
Alternative Standard Level for 2025 - except California (1,000 tons/year)"
Geographic Area Emissions Sector
EGU
Non-EGU Point
East Nonpoint
Nonroad
Total
EGU
Non-EGU Point
West Nonpoint
Nonroad
Total
NOx
170
410
420
12
1,000
62
48
39
1.3
150
VOC
-
4.2
99
-
100
-
0.47
6.6
-
7
"Estimates are rounded to two significant figures.
ES-8
-------
ES. 1.2.2 Emissions Reductions beyond Known Controls in 2025
There were several areas where known controls did not achieve enough emissions
reductions to attain the proposed alternative standard levels of 70 and 65 as well as the more
stringent alternative standard level of 60 ppb. To complete the analysis, the EPA then estimated
the additional emissions reductions beyond known controls needed to reach attainment (i.e.,
unknown controls). Table ES-4 shows the emissions reductions needed from unknown controls
in 2025 for the U.S., except California, for the alternative standard levels analyzed.
Table ES-4. Summary of Emissions Reductions by Alternative Standard for Unknown
Controls for 2025 - except California (1,000 tons/year)"
Proposed Alternative Standard
70 ppb
65 ppb
Region
Levels
East
West
East
West
NOx
150
750
voc
-
-
Alternative Standard Level
60 ppb
East
West
1,900
350
41
ES. 1.2.3 Emissions Reductions beyond Known Controls for Post-2025
Figure ES-3 shows the counties projected to exceed the alternative standard levels
analyzed for the post-2025 analysis for California. For the California post-2025 alternative
standard level analyses, all known controls were applied in the baseline, so incremental
emissions reductions are from unknown controls. Table ES-5 shows the emissions reductions
needed from unknown controls for post-2025 for California for the alternative standard levels
analyzed.
ES-9
-------
Legend
counties are projected t exceed 7D ppb
additonsl comities are projected to be below 70 but exceed 0S ppb
Figure ES-3. Projected Ozone Design Values in the post-2025 Baseline Scenario
Table ES-5. Summary of Emissions Reductions by Alternative Standard Level for
Unknown Controls for post-2025 - California (1,000 tons/year)"
Region
NOx
VOC
Proposed Alternative Standard Levels
70 ppb
CA
53
65 ppb
CA
110
Alternative Standard Level
60 ppb
CA
140
a Estimates are rounded to two significant figures.
ES. 1.3 Human Health Benefits
To estimate benefits, we follow a "damage-function" approach in calculating total
benefits of the modeled changes in environmental quality. This approach estimates changes in
individual health endpoints (specific effects that can be associated with changes in air quality)
and assigns values to those changes assuming independence of the values for those individual
endpoints. Total benefits are calculated as the sum of the values for all non-overlapping health
endpoints. The "damage-function" approach is the standard method for assessing costs and
benefits of environmental quality programs and has been used in several recent published
analyses (Levy et al, 2009; Fann et al, 2012a; Tagaris et al, 2009).
ES-10
-------
To assess economic values in a damage-function framework, the changes in
environmental quality must be translated into effects on people or on the things that people
value. In some cases, the changes in environmental quality can be directly valued, as is the case
for changes in visibility. In other cases, such as for changes in ozone and PM, an impact analysis
must first be conducted to convert air quality changes into effects that can be assigned dollar
values. For the purposes of this RIA, the health impacts analysis is limited to those health effects
that are directly linked to ambient levels of air pollution and specifically to those linked to ozone
and PM2.5.
Benefits estimates for ozone were generated using the damage function approach outlined
above wherein potential changes in ambient ozone levels (associated with future attainment of
alternative standard levels) were explicitly modeled and then translated into reductions in the
incidence of specific health endpoints. In generating ozone benefits estimates for the two
attainment timeframes considered in the RIA (2025 and post-2025), we used three distinct
benefits simulations including one completed for 2025 and two completed for 2038. The way in
which these three benefits simulations were used to generate estimates for the two timeframes is
detailed in Chapter 5, Section 5.4.3.
In contrast to ozone, we used a benefit-per-ton approach in modeling PIVh.s co-benefits.
With this approach, we use the results of previous benefits analysis simulations focusing on
PM2.5 to derive benefits-per-ton estimates for NOx.4 We then combine these dollar-per-ton
estimates with projected reductions in NOx associated with meeting a given alternative standard
level to project cobenefits associated with PIVh.s. We acknowledge increased uncertainty
associated with the dollar-per-ton approach for PIVh.s, relative to explicitly modeling benefits
using gridded PIVb.s surfaces specific to the baseline and alternative standard levels (see
Appendix 5 A, Table 5A-1 for additional discussion).
In addition to ozone and PIVb.s benefits, implementing emissions controls to reach some of
the alternative ozone standard levels would reduce other ambient pollutants. However, because
the methods used in this analysis to simulate attainment do not account for changes in ambient
4 In addition to dollar-per-ton estimates for NOx, we also used incidence-per-ton values (also for NOx) for specific
health endpoints to generate incidence reduction estimates associated with the dollar benefits.
ES-11
-------
concentrations of other pollutants, we were not able to quantify the co-benefits of reduced
exposure to these pollutants. In addition, due to data and methodology limitations, we were
unable to estimate some anticipated health benefits associated with exposure to ozone and PIVh.s.
ES. 1.4 Welfare Co-Benefits of the Primary Standard
Section 109 of the Clean Air Act defines welfare effects to include any non-health effects,
including direct economic damages in the form of lost productivity of crops and trees, indirect
damages through alteration of ecosystem functions, indirect economic damages through the loss
in value of recreational experiences or the existence value of important resources, and direct
damages to property, either through impacts on material structures or by soiling of surfaces (42
U.S.C. 7409). Ozone can affect ecological systems, leading to changes in the ecological
community and influencing the diversity, health, and vigor of individual species (U.S. EPA,
2013). Ozone causes discernible injury to a wide array of vegetation (U.S. EPA, 2013). In terms
of forest productivity and ecosystem diversity, ozone may be the pollutant with the greatest
potential for region-scale forest impacts (U.S. EPA, 2013). Studies have demonstrated repeatedly
that ozone concentrations observed in polluted areas can have substantial impacts on plant
function (De Steiguer et al,. 1990; Pye, 1988).
In this RIA, we are able to quantify only a small portion of the welfare impacts associated
with reductions in ozone concentrations to meet alternative ozone standards. Using a model of
commercial agriculture and forest markets, we are able to analyze the effects on consumers and
producers of forest and agricultural products of changes in the W126 index resulting from
meeting alternative standards within the proposed range of 70 to 65 ppb, as well as a lower
standard level of 60 ppb. We also assess the effects of those changes in commercial agricultural
and forest yields on carbon sequestration and storage. This analysis provides limited quantitative
information on the welfare co-benefits of meeting these alternative standards, focused only on
one subset of ecosystem services. Commercial and non-commercial forests provide a number of
additional services, including medicinal uses, non-commercial food and fiber production, arts
and crafts uses, habitat, recreational uses, and cultural uses for Native American tribes. A more
complete discussion of these additional ecosystem services is provided in the final Welfare Risk
and Exposure Assessment for Ozone (U.S. EPA, 2014c).
ES-12
-------
ES.2 Results of Benefit-Cost Analysis
Below in Table ES-6, we present the primary costs and benefits estimates for 2025 for all
areas except California. In addition, Tables 5-1 and 5-23 in Chapter 5 provide a breakdown of
ozone-only and PIVh.s-only benefits, as well as total benefits at 3 percent. We anticipate that
benefits and costs will likely begin occurring earlier, as states begin implementing control
measures to show progress towards attainment. In these tables, ranges within the total benefits
rows reflect variability in the studies upon which the estimates associated with premature
mortality were derived. PIVb.s co-benefits account for approximately two-thirds to three-quarters
of the estimated benefits, depending on the standard analyzed and on the choice of ozone and
PM mortality functions used. In addition for 2025, Table ES-7 presents the numbers of
premature deaths avoided for the alternative standard levels analyzed, as well as the other health
effects avoided. Table ES-8 provides information on the costs by geographic region for the U.S.,
except California in 2025, and Table ES-9 provides a regional breakdown of benefits for 2025.
In the RIA we provide estimates of costs of emissions reductions to attain the proposed
standards in three regions — California, the rest of the western U.S., and the eastern U.S. In
addition, we provide estimates of the benefits that accrue to each of these three regions resulting
from (i) control strategies applied within the region, (ii) reductions in transport of ozone
associated with emissions reductions in other regions, and (iii) the control strategies for which
the regional cost estimates are generated. These benefits are not directly comparable to the costs
of control strategies in a region because the benefits include benefits not associated with those
control strategies.
The net benefits of emissions reductions strategies in a specific region would be the
benefits of the emissions reductions occurring both within and outside of the region minus the
costs of the emissions reductions. Because the air quality modeling is done the national level, we
do not estimate separately the nationwide benefits associated with the emissions reductions
occurring in any specific region.5 As a result, we are only able to provide net benefits estimates
at the national level. The difference between the costs for a specific region and the benefits
5 For California, we provide separate estimates of the costs and nationwide estimates of benefits, so it is appropriate
to calculate net benefits. As such, we provide net benefits for the post-2025 California analysis.
ES-13
-------
accruing to that region is not an estimate of net benefits of the emissions reductions in that
region.
Table ES-6. Total Annual Costs and Benefits3 for U.S., except California in 2025 (billions
of 2011$, 7% Discount Rate)b
Total Costs (7%)
Total Health Benefits (7%)c
Net Benefits (7%)
Proposed Alternative Standard Levels
70 ppb 65 ppb
$3.9 $15
$6.4 to $13.0 $19 to $38
$2.5 to $9.1 $4 to $23
Alternative Standard
Level
60 ppb
$39
$34 to $70
($5) to $31
a Benefits are nationwide benefits of attainment everywhere except California.
b EPA believes that providing comparisons of social costs and social benefits at 3 and 7 percent is appropriate.
Estimating multiple years of costs and benefits is not possible for this RIA due to data and resource limitations. As
a result, we provide a snapshot of costs and benefits in 2025, using the best available information to approximate
social costs and social benefits recognizing uncertainties and limitations in those estimates.
0 The benefits range reflects the LOW and UPPER core estimates of short-term ozone and long-term PM mortality.
EPA believes that providing comparisons of social costs and social benefits at 3 and 7
percent is appropriate. Ideally, streams of social costs and social benefits over time would be
estimated and the net present values of each would be compared to determine net benefits of the
illustrative attainment strategies. The three different uses of discounting in the RIA - (i)
construction of annualized engineering costs, (ii) adjusting the value of mortality risk for lags in
mortality risk decreases, and (iii) adjusting the cost of illness for non-fatal heart attacks to adjust
for lags in follow up costs - are all appropriate. Our estimates of net benefits are the
approximations of the net value (in 2025) of benefits attributable to emissions reductions needed
to attain just for the year 2025.
Table ES-7. Summary of Total Number of Annual Ozone and PM-Related Premature
Mortalities and Premature Morbidity: 2025 National Benefits a
Proposed Alternative Standard Levels
(95th percentile confidence intervals)13
Alternative Standard
Level
(95th percentile
confidence intervals)
Short-term exposure-related
premature deaths avoided (all ages)
(Ozone - 2 studies)
70 ppb
200 to 340
(97 to 300)
(180 to 490)
65 ppb
630 to 1,000
(3 10 to 940)
(560 to 1,500)
60 ppb
1,100 to 1,900
(560 to 1,700)
(1,000 to 2,800)
Long-term exposure-related
premature deaths avoided (age
30+) (PM - 2 studies)
O3: 680
(230 to 1,100)
PM2 5: 510 to 1,100°
O3: 2,100
(710 to 3,500)
PM25: 1,400 to 3,300C
O3: 3,900
(1,300 to 6,400)
zs: 2,600 to 6,000°
Other health effects avoided"1
Non-fatal heart attacks (age 18-99) (5
studies) PM
64 to 600
180 to 1,700
330 to 3,100
ES-14
-------
Alternative Standard
Proposed Alternative Standard Levels Level
(95th percentile confidence intervals)13 (95th percentile
confidence intervals)
Respiratory hospital admissions (age
0-99)O3> PM
Cardiovascular hospital admissions
(age 18-99) PM
Asthma emergency department visits
(age 0-99) 03>PM
Acute bronchitis (age 8-12) PM
Asthma exacerbation (age 6-18) °3' PM
Lost work days (age 18-65)PM
Minor restricted activity days (age
18.65)03,PM
Upper & lower respiratory symptoms
(children 7-14) PM
School loss days (age 5-17) °3
70ppb
510
180
1,400
790
320,000
65,000
1,300,000
24,000
330,000
65 ppb
1,500
530
4,300
2,300
960,000
180,000
4,000,000
70,000
1,000,000
60 ppb
2,900
950
8,000
4,100
1,800,000
340,000
7,300,000
130,000
1,900,000
a Nationwide benefits of attainment everywhere except California.
b We present a confidence interval in parentheses for each study on short-term or long-term ozone-related mortality.
0 These estimates were generated using benefit-per-ton estimates and confidence intervals are not available. In
general, the 95th percentile confidence interval for the health impact function alone ranges from j^30 percent for
mortality incidence based on Krewski et al. (2009) and + 46 percent based on Lepeule et al. (2012).
d See Table 5-19 in Chapter 5 for detailed information on confidence intervals related to ozone-related morbidity
incidence estimates. The PM2 5 morbidity incidence estimates were generated using benefit-per-ton estimates and
confidence intervals are not available.
Table ES-8. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S., except California (billions of 2011$, 7% Discount Rate)"
„ , . . Total Control Costs
.,, T , Geographic Area ,T, ,
Alternative Level (Known and
Extrapolated)
East3.9
Total $3.9
East 15
65ppb West 0.40
Total $15
East 33
6°Ppb
Total $39
' All values are rounded to two significant figures. Extrapolated costs are based on the average-cost methodology.
ES-15
-------
Table ES-9. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
2025 Scenario (nationwide benefits of attaining each alternative standard
everywhere in the U.S. except California) - Full Attainment a
. Proposed and Alterative Standards
egl0° 70ppb 65ppb 60ppb
Eastb 99% 96% 92%
California 0% 0% 0%
Rest of West 1% 4% 7%
a Because we use benefit-per-ton estimates to calculate the PM2 5 co-benefits, a regional breakdown for the co-
benefits is not available. Therefore, this table only reflects the ozone benefits.
b Includes Texas and those states to the north and east. Several recent rules such as Tier 3 will have substantially
reduced ozone concentrations by 2025 in the East, thus few additional controls would be needed to reach 70 ppb.
To understand possible additional costs and benefits of fully attaining in California in a
post-2025 timeframe, we provide separate results for California in Table ES-10. In addition,
Tables 5-2 and 5-30 in Chapter 5 provide a breakdown of ozone-only and PIVh.s-only benefits, as
well as total benefits at 3 percent. Relative to the primary cost and benefits estimates, the
California cost estimates are between 5 and 20 percent and the benefits estimates are between 8
and 15 percent of the national estimates. Because of the differences in the timing of achieving
needed emissions reductions, incurring costs, and accruing benefits for California, the separate
costs and benefits estimates for post-2025 should not be added to the primary estimates for 2025.
For the post-2025 timeframe, Table ES-11 presents the numbers of premature deaths avoided for
the alternative standard levels analyzed, as well as the other health effects avoided. Table ES-12
provides information on the costs for California for post-2025, and Table ES-13 provides a
regional breakdown of benefits for post-2025.
The EPA presents separate costs and benefits results for California because forcing
attainment in an earlier year than would be required under the Clean Air Act would likely lead to
an overstatement of costs because California might benefit from some existing federal or state
programs that would be implemented between 2025 and the ultimate attainment years; because
additional new technologies may become available between 2025 and the attainment years; and
because the cost of existing technologies might fall over time. As such, we use the best available
data to estimate costs and benefits for California in a post-2025 timeframe, but because of data
limitations and additional uncertainty associated with not projecting emissions and air quality
beyond 2025, we recognize that the estimates of costs and benefits for California in a post-2025
ES-16
-------
timeframe are likely to be relatively more uncertain than the national attainment estimates for
2025.
Table ES-10. Total Annual Costs and Benefits" of Control Strategies Applied in California,
post-2025 (billions of 2011$, 7% Discount Rate)b
Proposed Alternative Standard Levels
Total Costs (7%)
Total Health Benefits (7%)c
Net Benefits (7%)
70ppb
$0.80
$1.1 to $2
$0.3 to $1.2
65 ppb
$1.6
$2.2 to $4.1
$0.60 to $2.5
Alternative Standard
Level
60 ppb
$2.2
$3.2 to $5.9
$1 to $3.7
a Benefits are nationwide benefits of attainment in California.
b EPA believes that providing comparisons of social costs and social benefits at 3 and 7 percent is appropriate.
Estimating multiple years of costs and benefits is not possible for this RIA due to data and resource limitations. As
a result, we provide a snapshot of costs and benefits in 2025, using the best available information to approximate
social costs and social benefits recognizing uncertainties and limitations in those estimates.
0 The benefits range reflects the LOW and UPPER core estimates of short-term ozone and long-term PM mortality.
Table ES-11. Summary of Total Number of Annual Ozone and PM-Related Premature
Mortalities and Premature Morbidity: Post-2025a
Proposed Alternative Standard Levels
(95th percentile confidence intervals)13
Short-term exposure-related
premature deaths avoided (all ages)
(Ozone - 2 studies)
Long-term exposure-related
premature deaths avoided (age
30+) (PM - 2 studies)
70 ppb
65 to 110
(31 to 97)
(57 to 160)
O3: 260
(88 to 430)
PM25: 45 to 100C
65 ppb
140 to 230
(68 to 210)
(120 to 340)
O3: 560
(190 to 930)
PM25: 89 to 200C
Alternative Standard
Level
(95th percentile
confidence intervals)
60 ppb
210 to 350
(100 to 320)
(190 to 5 10)
O3: 840
(290 to 1,400)
PM25: 120 to 280C
Other health effects avoided"1
Non-fatal heart attacks (age 18-99) (5
studies) PM
Respiratory hospital admissions (age
0-99)°3' PM
Cardiovascular hospital admissions
(age 18-99) PM
Asthma emergency department visits
(age 0-99) 03>PM
Acute bronchitis (age 8-12) PM
Asthma exacerbation (age 6-18) 03> PM
Lost work days (age 18-65)PM
Minor restricted activity days (age
18-65)03>PM
Upper & lower respiratory symptoms
(children 7-14) PM
School loss days (age 5-17) °3
6 to 54
130
16
340
67
99,000
5,500
320,000
2,100
110,000
11 to 110
290
32
740
130
210,000
11,000
690,000
4,100
230,000
16 to 140
430
45
1,100
180
320,000
15,000
1,000,000
5,600
350,000
a Nationwide benefits of attainment in California.
b We present a confidence interval in parentheses for each study on short-term or long-term ozone-related mortality.
ES-17
-------
0 These estimates were generated using benefit-per-ton estimates and confidence intervals are not available. In
general, the 95th percentile confidence interval for the health impact function alone ranges from + 30 percent for
mortality incidence based on Krewski et al. (2009) and + 46 percent based on Lepeule et al. (2012).
d See Table 5-26 in Chapter 5 for detailed information on confidence intervals related to ozone-related morbidity
incidence estimates. The PM2.5 morbidity incidence estimates were generated using benefit-per-ton estimates and
confidence intervals are not available.
Table ES-12. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for post-2025 - California (billions of 2011$, 7% Discount Rate)"
Total Control Costs
Alternative Level Geographic Area (Known and
Extrapolated)
70ppb California $0.80
65 ppb California $1.6
60 ppb California $2.2
a All values are rounded to two significant figures. Extrapolated costs are based on the average-cost methodology.
Table ES-13. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
post-2025 Scenario (nationwide benefits of attaining each alternative standard
just in California) - Full Attainment"
Proposed and Alterative Standards
Region
70 ppb 65 ppb 60 ppb
East 0% 0% 0%
California 93% 94% 94%
Rest of West 6% 6% 6%
a Because we use benefit-per-ton estimates to calculate the PM2 5 co-benefits, a regional breakdown for the co-
benefits is not available. Therefore, this table only reflects the ozone benefits.
Despite uncertainties inherent in any complex, quantitative analysis, the overall underlying
analytical methods used in this RIA have been peer-reviewed. For a detailed discussion on
uncertainty associated with developing illustrative control strategies to attain the alternative
standard levels, see Chapter 4, Section 4.4. For a description of the key assumptions and
uncertainties related to the modeling of ozone benefits, see Chapter 5, Section 5.7.3, and for an
additional qualitative discussion of sources of uncertainty associated with both the modeling of
ozone-related benefits and PIVh.s-related co-benefits, see Appendix 5A. For a discussion of the
limitations and uncertainties in the engineering cost analyses, see Chapter 7, Section 7.7. For a
discussion about generally framing uncertainty, see Chapter 8, Section 8.3.
ES.3 References
De Steiguer, I, Pye, I, Love, C. 1990. Air Pollution Damage to U.S. Forests. Journal of Forestry, 88(8), 17-22.
ES-18
-------
Fann N, Lamson A, Wesson K, Risley D, Anenberg SC, Hubbell BJ. 2012a. "Estimating the National Public Health
Burden Associated with Exposure to Ambient PM2s and ozone. Risk Analysis," Risk Analysis 32(1): 81-95.
Levy JI, Baxter LK, Schwartz J. 2009. "Uncertainty and variability in health-related damages from coal-fired power
plants in the United States." Risk Analysis 29(7) 1000-1014.
Pye, J.M. 1988. Impact of ozone on the growth and yield of trees: A review. Journal of Environmental Quality, 17,
347-360.
Tagaris E, Liao KJ, Delucia AJ, et al. 2009. "Potential impact of climate change on air-pollution related human
health effects." Environmental Science & Technolology 43: 4979-4988.
U.S. Environmental Protection Agency (U.S. EPA). 2013. Integrated Science Assessment of Ozone and Related
Photochemical Oxidants (Final Report). EPA-600/R-10/076F. February. Available on the Internet at
http://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=247492.
U.S. Environmental Protection Agency (U.S. EPA). 2014a. Control of Air Pollution from Motor Vehicles: Tier 3
Motor Vehicle Emission and Fuel Standards. Office of Transportation and Air Quality. Available at
http://www.epa.gov/otaq/tier3.htm.
U.S. Environmental Protection Agency (U.S. EPA). 2014b. Proposed Carbon Pollution Guidelines for Existing
Power Plants and Emission Standards for Modified and Reconstructed Power Plants. Available at
http://www2. epa. gov/sites/production/files/2014 -06/documents/20140602ria-clean-power-plan.pdf
U.S. Environmental Protection Agency. 2014c. Policy Assessment for the Review of the Ozone NAAQS, Final
Report. August. Available on the Internet at
http://www.epa.gov/ttn/naaqs/standards/ozone/data/20140829pa.pdf.
ES-19
-------
CHAPTER 1: INTRODUCTION AND BACKGROUND
Overview
The EPA Administrator is proposing to revise the level of the ozone National Ambient
Air Quality Standards (NAAQS) to within a range of 65 to 70 ppb and is soliciting comment on
alternative standard levels below 65 ppb, as low as 60 ppb. This chapter summarizes the purpose
and background of this Regulatory Impact Analysis (RIA). In the RIA we estimate the human
health and welfare benefits and costs of alternative standards of 65 ppb and 70 ppb, which
represent the lower and upper bounds of the range of proposed levels, as well as a more stringent
alternative level of 60 ppb. According to the Clean Air Act ("the Act"), the Environmental
Protection Agency (EPA) must use health-based criteria in setting the NAAQS and cannot
consider estimates of compliance cost. The EPA is producing this RIA both to provide the public
a sense of the benefits and costs of meeting a revised ozone NAAQS and to meet the
requirements of Executive Orders 12866 and 13563.
1.1 Background
7.7.7 NAAQS
Two sections of the Act govern the establishment and revision of NAAQS. Section 108
(42 U.S.C. 7408) directs the Administrator to identify pollutants that "may reasonably be
anticipated to endanger public health or welfare" and 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 all identifiable effects on public health or welfare which may be
expected from the presence of [a] pollutant in the ambient air." Ozone is one of six pollutants for
which the EPA has developed air quality criteria.
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 an ambient air quality standard "the attainment and
maintenance of which in the judgment of the Administrator, based on [the] criteria and allowing
an adequate margin of safety, [is] requisite to protect the public health." 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 [the] criteria, is requisite to protect the
1-1
-------
public welfare from any known or anticipated adverse effects associated with the presence of
[the] pollutant in the ambient air." 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, 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."
Section 109(d) of the Act directs the Administrator to review existing criteria and
standards at 5-year intervals. When warranted by such review, the Administrator is to retain or
revise the NAAQS. After promulgation or revision of the NAAQS, the standards are
implemented by the states.
7.7.2 Ozone NAAQS
The EPA initiated the current ozone NAAQS review in September 2008. Between 2008
and 2014, the EPA prepared draft and final versions of the Integrated Science Assessment, the
Health and Welfare Risk and Exposure Assessments, and the Policy Assessment. Multiple drafts
of these documents were available for public review and comment, and as required by the Clean
Air Act, were peer-reviewed by the Clean Air Scientific Advisory Committee (CASAC), the
Administrator's independent advisory committee established by the CAA. The final documents
reflect the EPA staffs consideration of the comments and recommendations made by CASAC
and the public on draft versions of these documents.
1.2 Role of this RIA in the Process of Setting the NAAQS
7.2.1 Legislative Roles
The EPA Administrator is proposing to revise the level of the ozone National Ambient
Air Quality Standards (NAAQS) to within a range of 65 to 70 ppb and is soliciting comment on
alternative standard levels below 65 ppb, as low as 60 ppb. The EPA Administrator is also
proposing to revise the level of the current secondary standard to within the range of 65 ppb to
70 ppb. As such, the RIA analyzes a range of potential alternative primary standard levels. In
setting primary ambient air quality standards, the EPA's responsibility under the law is to
establish standards that protect public health, regardless of the costs of implementing those
standards. The Act requires the EPA, for each criteria pollutant, to set standards that protect
1-2
-------
public health with "an adequate margin of safety." As interpreted by the Agency and the courts,
the Act requires the EPA to create standards based on health considerations only.
The prohibition against the consideration of cost in the setting of the primary air quality
standards, however, does not mean that costs or other economic considerations are unimportant
or should be ignored. The Agency believes that consideration of costs and benefits is essential to
making efficient, cost-effective decisions for implementing these standards. The impact of cost
and efficiency is considered by states during this process, as they decide what timelines,
strategies, and policies make the most sense. This RIA is intended to inform the public about the
potential costs and benefits that may result when new standards are implemented, but it is not
relevant to establishing the standards themselves.
7.2.2 Role of Statutory and Executive Orders
This RIA is separate from the NAAQS decision-making process, but several statutes and
executive orders still apply to any public documentation. The analysis required by these statutes
and executive orders is presented in Chapter 9.
The EPA presents this RIA pursuant to Executive Orders 12866 and 13563 and the
guidelines of Office of Management and Budget (OMB) Circular A-4 (U.S. OMB, 2003). In
accordance with these guidelines, the RIA analyzes the benefits and costs associated with
emissions controls to attain the upper and lower bounds of the proposed 8-hour ozone standard
of 65 parts per billion (ppb) to 70 ppb in ambient air, incremental to a baseline of attaining the
existing standard (8-hour ozone standard of 75 ppb). OMB Circular A-4 requires analysis of one
potential alternative standard level more stringent than the proposed range and one less stringent
than the proposed range. In this RIA, we analyze a more stringent alternative standard level of
60 ppb. The existing standard of 75 ppb represents the less stringent alternative standard and the
costs and benefits of this standard were presented in the 2008 ozone NAAQS RIA (U.S. EPA,
2008a). The available scientific evidence and quantitative exposure and risk information
indicate that reducing ambient ozone concentrations will reduce the occurrence of harmful health
effects. As discussed in the Notice, this evidence and information provide strong support for
considering alternative standard levels from 65 to 70 ppb, but do not identify a bright line within
this range that indicates exactly where to set a standard. Similarly, the available scientific
1-2
-------
information does not provide a basis for identifying any specific standard level between 70 and
75 ppb for analysis in the RIA.
The control strategies presented in this RIA are illustrative and represent one set of control
strategies states might choose to implement in order to meet the final standards. As a result,
benefit and cost estimates provided in the RIA are not additive to benefits and costs from other
regulations, and, further, the costs and benefits identified in this RIA will not be realized until
specific controls are mandated by State Implementation Plans (SIPs) or other federal regulations.
1.2.3 The Need for National Ambient Air Quality Standards
OMB Circular A-4 indicates that one of the reasons a regulation such as the NAAQS may
be issued is to address existing "externalities." A market failure or externality occurs when one
party's actions impose uncompensated costs on another party. Environmental problems are a
classic case of an externality. Setting and implementing primary and secondary air quality
standards is one way the government can address an externality and thereby increase air quality
and improve overall public health and welfare.
1.2.4 Illustrative Nature of the Analysis
This NAAQS RIA is an illustrative analysis that provides useful insights into a limited
number of emissions control scenarios that states might implement to achieve revised NAAQS.
Because states are ultimately responsible for implementing strategies to meet any revised
standard, the control scenarios in this RIA are necessarily hypothetical in nature. Important
uncertainties and limitations are documented in the relevant portions of the analysis.
The illustrative goals of this RIA are somewhat different from other EPA analyses of
national rules, or the implementation plans states develop, and the distinctions are worth brief
mention. This RIA does not assess the regulatory impact of an EPA-prescribed national rule, nor
does it attempt to model the specific actions that any state would take to implement a revised
standard. This analysis attempts to estimate the costs and human and welfare benefits of cost-
effective implementation strategies that might be undertaken to achieve national attainment of
new standards. These hypothetical strategies represent a scenario where states use one set of
cost-effective controls to attain a revised NAAQS. Because states—not the EPA—will
1-4
-------
implement any revised NAAQS, they will ultimately determine appropriate emissions control
scenarios. SIPs would likely vary from the EPA's estimates due to differences in the data and
assumptions that states use to develop these plans. The illustrative attainment scenarios
presented in this RIA were constructed with the understanding that there are inherent
uncertainties in projecting emissions and controls.
1.3 Overview and Design of the RIA
The RIA evaluates the costs and benefits of hypothetical national control strategies to
attain three alternative ozone standard levels of 60, 65 and 70 ppb.
1.3.1 Existing and Revised Ozone National Ambient Air Quality Standards
The EPA is proposing to retain the indicator, averaging time and form of the existing
primary ozone standard and is proposing to revise the level of that standard to within the range of
65 ppb to 70 ppb. The EPA is proposing this revision to increase public health protection,
including for "at-risk" populations such as children, older adults, and people with asthma or
other lung diseases, against an array of ozone-related adverse health effects. For short-term
ozone exposures, these effects include decreased lung function, increased respiratory symptoms
and pulmonary inflammation, effects that result in serious indicators of respiratory morbidity,
such as emergency department visits and hospital admissions, and all-cause (total non-
accidental) mortality. For long-term ozone exposures, these health effects include a variety of
respiratory morbidity effects and respiratory mortality. In recognition that levels as low as 60
ppb could potentially be supported, but would place very little weight on the uncertainties in the
health effects evidence and exposure/risk information, the EPA is also soliciting comment on
alternative standard levels below 65 ppb, as low as 60 ppb. In addition, the EPA is taking
comment on the option of retaining the current 8-hour primary ozone standard of 75 ppb.
The EPA is proposing to revise the level of the secondary standard to within the range of
65 ppb to 70 ppb to provide increased protection against vegetation-related effects on public
welfare. As an initial matter, the EPA is proposing that ambient ozone concentrations in terms of
a three-year average W126 index value within the range from 13 parts per million-hours (ppm-
hours) to 17 ppm-hours would provide the requisite protection against known or anticipated
adverse effects to the public welfare, which data analyses indicate would provide air quality in
1-5
-------
terms of three-year average W126 index values of a range at or below 13 ppm-hours to 17 ppm-
hours. Data analyses also indicate that actions taken to attain a standard in the range of 65 ppb to
70 ppb would also improve air quality as measured by the W126 metric. The quantitative
analysis assesses the welfare benefits of strategies to attain the proposed secondary standard
levels of 65 to 70 ppb.
1.3.2 Establishing Attainment with the Current Ozone National Ambient Air Quality Standard
The RIA is intended to evaluate the costs and benefits of reaching attainment with
alternative ozone standard levels. To develop and evaluate control strategies for attaining a more
stringent primary standard, it is important to first estimate ozone levels in the future after
attaining the current NAAQS (75 ppb) and taking into account projections of future air quality
reflecting on-the-books Federal regulations, enforcement actions, state regulations, and
population and economic growth. This allows us to then estimate the incremental costs and
benefits of attaining alternative primary standard levels.
Attaining 75 ppb reflects emissions reductions already achieved as a result of national
regulations, emissions reductions expected prior to 2025 from recently promulgated national
regulations (i.e., reductions that were not realized before promulgation of the previous standard,
but are expected prior to attainment of the existing ozone standard), and reductions from
additional controls that the EPA estimates need to be included to attain the existing standard (75
ppb). Emissions reductions achieved as a result of state and local agency regulations and
voluntary programs are reflected to the extent that they are represented in emissions inventory
information submitted to the EPA by state and local agencies. We took two steps to develop the
baseline reflecting attainment of 75 ppb. First, national ozone concentrations were projected
based on population and economic growth and the application of emissions controls resulting
from national rules promulgated prior to this analysis, as well as state programs and enforcement
actions. Second, we apply an illustrative control strategy to estimate emissions reductions for the
current standard of 75 ppb, also referred to as the baseline.
Below is a list of some of the national rules reflected in the baseline. For a more complete
list, please see the Technical Support Document: Preparation of Emissions Inventories for the
Version 6.1, 2011 Emissions Modeling Platform (US EPA, 2014a).
1-6
-------
• Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility
Generating Units (U.S. EPA, 2014b)
• Tier 3 Motor Vehicle Emission and Fuel Standards (U.S. EPA, 2014c)
• Mercury and Air Toxics Standards (U.S. EPA, 2011)
• Reciprocating Internal Combustion Engines (RICE) NESHAPs (U. S. EPA, 2010)
• Hospital/Medical/Infectious Waste Incinerators: New Source Performance Standards and
Emission Guidelines: Final Rule Amendments (U.S. EPA, 2009)
• C3 Oceangoing Vessels (U.S. EPA, 2010)
• Emissions Standards for Locomotives and Marine Compression-Ignition Engines (U.S.
EPA, 2008b)
• Control of Emissions for Nonroad Spark Ignition Engines and Equipment (U.S. EPA,
2008c)
• Regional Haze Regulations and Guidelines for Best Available Retrofit Technology
Determinations (U.S. EPA, 2005b)
• NOX Emission Standard for New Commercial Aircraft Engines (U.S. EPA, 2005)
• Clean Air Nonroad Diesel Rule (U.S. EPA, 2004)
• Heavy Duty Diesel Rule (U.S. EPA, 2000)
• Light-Duty Vehicle Tier 2 Rule (U.S. EPA, 1999)
The baseline for this analysis does not assume emissions controls that might be
implemented to meet the other NAAQS for PM2.5, NCh, or SO2. We did not conduct this analysis
incremental to controls applied as part of previous NAAQS analyses because the data and
modeling on which these previous analyses were based are now considered outdated and are not
compatible with the current ozone NAAQS analysis.6 In addition, all control strategies analyzed
in NAAQS RIAs are hypothetical. This analysis presents one scenario that states may employ
but does not prescribe how attainment must be achieved.
6 There were no additional NOx controls applied in the PM2 5 NAAQS RIA, and therefore there would be little to no
impact on the controls selected as part of this analysis. In addition, the only geographic areas that exceed the
alternative ozone standard levels analyzed in this RIA and in the 2012 PM25 NAAQS RIA are in California. The
attainment dates for a new PM2 5 NAAQS would likely precede attainment dates for a revised ozone NAAQS.
While the 2012 PM25 NAAQS RIA concluded that controls on directly emitted PM25 were the most cost-
effective on a $/ug basis, states may choose to adopt different control options. These options could include NOx
controls. It is difficult to determine the impact on costs and benefits for this RIA because it is highly dependent
upon the control measures that would be chosen and the costs of these measures.
1-7
-------
1.3.3 Establishing the Baseline for Evaluation of Alternative Standards
The RIA evaluates, to the extent possible, the costs and benefits of attaining the proposed
and alternative ozone standards incremental to attaining the existing ozone standard and
implementing existing and expected regulations. We assume that potential nonattainment areas
everywhere in the U.S., excluding California, will be designated such that they are required to
reach attainment by 2025, and we developed our projected baselines for emissions, air quality,
and populations for 2025.
The EPA will likely finalize designations for a revised ozone NAAQS in late
2017. Depending on the precise timing of the effective date of those designations, nonattainment
areas classified as Marginal will likely have to attain in either late 2020 or early
2021. Nonattainment areas classified as Moderate will likely have to attain in either late 2023 or
early 2024. If a Moderate nonattainment area qualifies for two 1-year extensions, the area may
have as late as early 2026 to attain. Lastly, Serious nonattainment areas will likely have to attain
in late 2026 or early 2027. We selected 2025 as the primary year of analysis because it provided
a good representation of the remaining air quality concerns that moderate nonattainment areas
would face and because most areas of the U.S. will likely be required to meet a revised ozone
standard by 2025. States with areas classified as Moderate and higher are required to develop
attainment demonstration plans for those nonattainment areas. In this RIA we present the
primary costs and benefits estimates for 2025.
In estimating the incremental costs and benefits of potential alternative standards, we
recognize that there are several areas that are not required to meet the existing ozone standard by
2025. The Clean Air Act allows areas with more significant air quality problems to take
additional time to reach the existing standard. Several areas in California are not required to
meet the existing standard by 2025 and may not be required to meet a revised standard until
sometime between 2032 and 2037.7 We were not able to project emissions and air quality
7 The EPA will likely finalize designations for a revised ozone NAAQS in late 2017. Depending on the precise
timing of the effective date of those designations, nonattainment areas classified as Severe 15 will likely have to
attain sometime between late 2032 and early 2033 and nonattainment areas classified as Extreme will likely have
to attain sometime between late 2037 and early 2038cember 31,.
-------
beyond 2025 for California, however, we adjusted baseline air quality to reflect mobile source
emissions reductions for California that would occur between 2025 and 2030; these emissions
reductions were the result of mobile source regulations expected to be fully implemented by
2030. While there is uncertainty about the precise timing of emissions reductions and related
costs for California, we assume costs occur through the end of 2037 and beginning of 2038. In
addition, we model benefits for California using projected population demographics for 2038.
Because of the different timing for incurring costs and accruing benefits and for ease of
discussion throughout the analyses, we refer to the different time periods for potential attainment
as 2025 and post-2025 to reflect that (1) we did not project emissions and air quality for any year
other than 2025; (2) for California, emissions controls and associated costs are assumed to occur
through the end of 2037 and beginning of 2038; and (3) for California benefits are modeled using
population demographics in 2038. It is not straightforward to discount the post-2025 results for
California to compare with or add to the 2025 results for the rest of the U.S. While we estimate
benefits using 2038 information, we do not have good information on precisely when the costs of
controls will be incurred. Because of these differences in timing related to California attaining a
revised standard, the separate costs and benefits estimates for post-2025 should not be added to
the primary estimates for 2025.
1.4 Health and Welfare Benefits Analysis Approach
1.4.1 Health Benefits
The EPA estimated human health (e.g., mortality and morbidity effects) under both
partial and full attainment of the three alternative ozone standards. We considered an array of
health impacts attributable to changes in ozone and PM 2.5 exposure and estimated these benefits
using the BenMAP tool (US EPA, 2014), which has been used in many recent RIAs (e.g., U.S.
EPA, 2006, 201 la, 201 Ib), and The Benefits and Costs of the Clean Air Act 1990 to 2020 (U.S.
EPA, 201 Ic). The EPA has incorporated an array of policy and technical updates to the benefits
analysis approach applied in this RIA, including incorporation of the most recent epidemiology
studies evaluating mortality and morbidity associated with ozone and PIVh.s exposure, and an
expanded uncertainty assessment. Each of these updates is fully described in the health benefits
chapter (Chapter 5). In addition, unquantified health benefits are also discussed in Chapter 5.
1-9
-------
1.4.2 Welfare Co-Benefits
Even though the primary standards are designed to protect against adverse effects to
human health, the emissions reductions would have welfare co-benefits in addition to the direct
human health benefits. The term welfare co-benefits covers both environmental and societal
benefits of reducing pollution. Welfare co-benefits of the primary ozone standard include
reduced vegetation effects resulting from ozone exposure, reduced ecological effects from
particulate matter deposition and from nitrogen emissions, reduced climate effects, and changes
in visibility. Both welfare co-benefits are discussed further in Chapter 6.
1.5 Cost Analysis Approach
The EPA estimated total costs under partial and full attainment of the three alternative
ozone standards. These cost estimates reflect only engineering costs, which generally include the
costs of purchasing, installing, and operating the referenced control technologies. The
technologies and control strategies selected for analysis are illustrative of one way in which
nonattainment areas could meet a revised standard. There are numerous ways to construct and
evaluate potential control programs that would bring areas into attainment with alternative
standards, and the EPA anticipates that state and local governments will consider programs that
are best suited for local conditions.
The partial-attainment cost analysis reflects the engineering costs associated with
applying end-of-pipe controls, or known controls. Costs for full attainment include estimates for
the costs associated with the additional emissions reductions that are needed beyond known
controls, referred to as unknown controls. The EPA recognizes that the portion of the cost
estimates from unknown controls reflects substantial uncertainty about which sectors and which
technologies might become available for cost-effective application in the future.
1.6 Organization of this Regulatory Impact Analysis
This RIA includes the following ten chapters:
• Chapter 1: Introduction and Background. This chapter introduces the purpose of the
RIA.
• Chapter 2: Defining the Ozone Air Quality Problem. This chapter characterizes the
nature, scope, and magnitude of the current-year ozone problem.
1-10
-------
Chapter 3: Air Quality Modeling and Analysis. The data, tools, and methodology used for
the air quality modeling are described in this chapter, as well as the post-processing
techniques used to produce a number of air quality metrics for input into the analysis of
costs and benefits.
Chapter 4: Control Strategies. This chapter presents the hypothetical control strategies,
the geographic areas where controls were applied, and the results of the modeling that
predicted ozone concentrations in 2025 after applying the control strategies.
Chapter 5: Human Health Benefits Analysis. This chapter quantifies the health-related
benefits of the ozone-related air quality improvements associated with several alternative
standards.
Chapter 6: Welfare Co-Benefits of the Primary Standard. This chapter quantifies and
monetizes selected other welfare effects, including vegetation effects from ozone
exposure, ecological effects from nitrogen and sulfur emissions, changes in visibility,
materials damage, ecological effects from PM deposition, ecological effects from
mercury deposition, and climate effects.
Chapter 7: Engineering Cost Analysis. This chapter summarizes the data sources and
methodology used to estimate the engineering costs of partial and full attainment of
several alternative standards.
Chapter 8: Comparison of Benefits and Costs. This chapter compares estimates of the
total benefits with total costs and summarizes the net benefits of several alternative
standards.
Chapter 9: Statutory and Executive Order Impact Analyses. This chapter summarizes the
Statutory and Executive Order impact analyses.
Chapter 10: Qualitative Discussion of Employment Impacts of Air Quality. This chapter
provides a discussion of employment impacts of reducing emissions of ozone precursors.
1.7 References
U.S. Environmental Protection Agency (U.S. EPA). 1999. Control of Air Pollution from New Motor Vehicles: Tier
2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control Requirements. Office of Transportation and Air
Quality. Available at http://www.epa.gov/tier2/frm/fr-t2reg.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2000. Heavy-Duty Engine and Vehicle Standards and Highway
Diesel Fuel Sulfur Control Requirements. Office of Transportation and Air Quality. Available at
http://www.epa.gov/otaq/highway-diesel/index.htm.
U.S. Environmental Protection Agency (U.S. EPA). 2004. Control of Emissions of Air Pollution from Nonroad
Diesel Engines and Fuel. Office of Transportation and Air Quality. Available at
http://www.regulations.gov/search/Regs/contentStreamer?objectId=09000064800be203&disposition=attachment&c
ontentType=pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2005. New Emission Standards for New Commercial Aircraft
Engines. Office of Transportation and Air Quality. Available at
http://www.epa.gov/oms/regs/nonroad/aviation/420f05015.htm.
1-11
-------
U.S. Environmental Protection Agency (U.S. EPA). 2005b. Regional Haze Regulations and Guidelines for Best
Available Retrofit Technology Determinations. Office of Air Quality Planning and Standards. Available at
http://www.epa.gov/fedrgstr/EPA-AIR/2005/July/Day-06/al2526.pdfand
http://www.epa.gov/visibility/fs_2005_6_15.html.
U.S. Environmental Protection Agency (U.S. EPA). 2005. Clean Air Interstate Rule. Office of Atmospheric
Programs, Washington, D.C. Available at http://edocket.access.gpo.gov/2005/pdf/05-5723.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2006. Air Quality Criteria for Ozone and Related
Photochemical Oxidants (Final). EPA/600/R-05/004aF-cF. Office of Research and Development, Research Triangle
Park, NC. Available at http://cfpub.epa.gov/ncea/CFM/recordisplay.cfm?deid= 149923.
U.S. Environmental Protection Agency (U.S. EPA). 2007. Guidance on the Use of Models and other Analyses for
Demonstrating Attainment of Air Quality Goals for Ozone, PM2 5, and Regional Haze. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/ttn/scram/guidance/guide/final-03-pm-rh-guidance.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Final Ozone NAAQS Regulatory Impact Analysis, US
EPA, Office of Air Quality Planning and Standards, Research Triangle Park, NC, EPA-452/R-08-003. Available at
http://www.epa.gov/ttnecasl/regdata/RIAs/452_R_08_003.pdf.
U.S. Environmental Protection Agency (U.S. EPA).2008b. Emissions Standards for Locomotives and Marine
Compression-Ignition Engines. Office of Transportation and Air Quality. Available at
http://www.epa.gov/otaq/regs/nonroad/420f08004.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2008c. Control of Emissions for Nonroad Spark Ignition
Engines and Equipment. Office of Transportation and Air Quality. Available at
http://www.epa.gov/otaq/regs/nomoad/marinesi-equipld^ndfrm.pdf
U.S. Environmental Protection Agency (U.S. EPA). 2009. Standards of Performance for New Stationary Sources
and Emission Guidelines for Existing Sources: Hospital/Medical/Infectious Waste Incinerators. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/ttn/atw/129/hmiwi/fr06oc09.pdf.
U.S. Environmental Protection Agency (U.S. EPA), 2009f. Regulatory Impact Analysis: Control of Emissions of Air
Pollution from Category 3 Marine Diesel Engines. Office of Transportation and Air Quality, Assessment and
Standards Division, Ann Arbor, MI. Report No. EPA-420-R-09-019. Available at:
http://www.epa.gov/otaq/regs/nonroad/marine/ci/420r09019.pdf
U.S. Environmental Protection Agency (U.S. EPA). 2010. Control of Emissions from New Marine Compression-
Ignition Engines at or Above 30 Liters per Cylinder. Office of Transportation and Air Quality. Available at
http://www.regulations.gov/search/Regs/contentStreamer?objectId=0900006480ae43a6&disposition=attachment&c
ontentType=pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2010. National Emission Standards for Hazardous Air
Pollutants for Reciprocating Internal Combustion Engines. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. Available at http://www.epa.gov/ttn/atw/icengines/fr/fr20aulO.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011. Cross-State Air Pollution Rule. Office of Atmospheric
Programs, Washington, D.C. Available at http://www.gpo.gov/fdsys/pkg/FR-2011-08-08/pdf/2011-17600.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011. Regulatory Impact Analysis for the Final Mercury and
Air Toxics Standards. Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-11-
011. Available at http://www.epa.gov/ttn/ecas/regdata/RIAs/matsriafinal.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011, The Benefits and Costs of the Clean Air Act from 1990 to
2020. Office of Air and Radiation, Washington, D.C. Available at:
http://www.epa.gov/cleanairactbenefits/febl l/fullreport_rev_a.pdf.
1-12
-------
U.S. Environmental Protection Agency (U.S. EPA). 2011. Regulatory Impact Analysis for the Federal
Implementation Plans to Reduce Interstate Transport of Fine Paniculate Matter and Ozone in 27 States; Correction
of SIP Approvals for 22 States. Available at http://www.epa.gov/airtransport/pdfs/FinalRIA.pdf.
U.S. Environmental Protection Agency, 2014. Policy Assessment for the Review of the Ozone National Ambient
Air Quality Standards, US EPA, OAQPS, 2014, RTF, NC, EPA-452/R-14-006.
U.S. Environmental Protection Agency, 2014a. Preparation of Emissions Inventories for the Version 6.1, 2011
Emissions Modeling Platform. Available at: http://www.epa.gov/ttn/chief/emch/index.htnuW2011.
U.S. Environmental Protection Agency, 2014b. Regulatory Impact Analysis for the Proposed Carbon Pollution
Guidelines for Existing Power Plants and Emission Standards for Modified and Reconstructed Power Plants, US
EPA, OAQPS, 2014, RTF, NC, EPA-452/R-14-002.
U.S. Environmental Protection Agency, 2014c. Control of Air Pollution from Motor Vehicles: Tier 3 Motor Vehicle
Emissions and Fuel Standards Rule, US EPA, OTAQ, 2014, Washington, DC, EPA-420-R-14-005.
U.S. Office of Management and Budget. Circular A-4, September 17, 2003, available at
http://www.whitehouse.gOv/omb/circulars/a004/-.pdf.
1-13
-------
CHAPTER 2: DEFINING THE OZONE AIR QUALITY PROBLEM
Overview
This section provides overviews of ozone precursor emissions and atmospheric chemistry
(section 2.1); ambient ozone concentrations (section 2.2); ambient ozone monitoring in the U.S.
(section 2.3); and available evidence and information related to background ozone (section 2.4).
2.1 Emissions and Atmospheric Chemistry
Ozone is formed through photochemical reactions of precursor gases and is not directly
emitted from specific sources. In the stratosphere, ozone occurs naturally and provides protection
against harmful solar ultraviolet radiation. In the troposphere, near ground level, ozone forms
through atmospheric reactions involving two main classes of precursor pollutants: volatile
organic compounds (VOCs) and nitrogen oxides (NOx). Carbon monoxide (CO) and methane
(CIHLt) are also important for ozone formation over longer time periods (US EPA, 2013, section
3.2.2).
Emissions of ozone precursor compounds can be divided into anthropogenic and natural
source categories, with natural sources further divided into biogenic emissions (from vegetation,
microbes, and animals) and abiotic emissions (from biomass burning, lightning, and geogenic
sources). Anthropogenic sources, including mobile sources and power plants, account for the
majority of NOx and CO emissions. Anthropogenic sources are also important for VOC
emissions, though in some locations and at certain times of the year (e.g., southeastern states
during summer) the majority of VOC emissions comes from vegetation (US EPA, 2013, section
3.2.1).
Rather than varying directly with emissions of its precursors, ozone changes in a nonlinear
fashion with the concentrations of its precursors. NOx emissions lead to both the formation and
destruction of ozone, depending on the local quantities of NOx, VOC, free radicals, and sunlight.
In areas dominated by fresh emissions of NOx, radicals are removed, which lowers the ozone
formation rate. In addition, the scavenging of ozone by reaction with NO is called "titration" and
is often found in downtown metropolitan areas, especially near busy streets and roads, as well as
in power plant plumes. This short-lived titration results in localized areas in which ozone
concentrations are suppressed compared to surrounding areas, but which contain NO2 that
2-1
-------
contributes to subsequent ozone formation further downwind. The NOx titration effect is most
pronounced in urban core areas that have a high volume of mobile source NOx emissions from
vehicles. In areas with relatively low NOx concentrations, such as those found in remote
continental areas and rural and suburban areas downwind of urban centers, ozone production
typically responds linearly to NOx concentrations (e.g., ozone decreases with decreasing NOx
emissions). Consequently, ozone response to reductions in NOx emissions is complex and may
include ozone decreases at some times and locations and increases of ozone at other times and
locations. As a general rule, as NOx emissions reductions occur, you can expect lower ozone
values to increase while the higher ozone values would be expected to decrease. NOx reductions
are expected to result in a compressed ozone distribution, relative to current conditions (EPA,
2014a).
The formation of ozone from precursor emissions is also affected by meteorological
parameters such as the intensity of sunlight and atmospheric mixing. Major episodes of high
ground-level ozone concentrations in the eastern United States are often associated with slow-
moving high pressure systems. High pressure systems during the warmer seasons are associated
with the sinking of air, resulting in warm, generally cloudless skies, with light winds. The
sinking of air results in the development of stable conditions near the surface that inhibit or
reduce the vertical mixing of ozone precursors. The combination of inhibited vertical mixing and
light winds minimizes the dispersal of pollutants, allowing their concentrations to build up. In
addition, in some parts of the United States (e.g., in Los Angeles), mountain barriers limit mixing
and result in a higher frequency and duration of days with elevated ozone concentrations.
Photochemical activity involving precursors is enhanced during warmer seasons because of the
greater availability of sunlight and higher temperatures (US EPA, 2013, section 3.2).
Elevated wintertime ozone concentrations have recently been measured in mountain
valleys in the Western U.S. (Schnell et al, 2009; Rappengluck et al, 2014; Helmig et al, 2014).
Hourly ozone concentrations during these winter events have been observed to reach 160 ppb.
This phenomenon is believed to result from the combination of several factors: 1) strong
wintertime inversions or "cold pools", which trap air in a shallow layer close to the ground, 2)
substantial emissions of NOx and VOC from nearby oil and gas operations, 3) high albedo of
deep snow, which leads to enhanced UV intensity and photochemical activity, and 4) possible
2-2
-------
uncharacterized sources of radicals. These wintertime ozone events have currently only been
observed in a limited number of locations in Wyoming, Utah, and Colorado. Events can last for
multiple days and can occur several times a year, but do not occur every winter in these
locations.
Ozone concentrations in a region are affected both by local formation and by transport of
ozone and its precursors from upwind areas. Ozone transport occurs on many spatial scales
including local transport between cities, regional transport over large regions of the U.S. and
international/long-range transport. In addition, ozone can be transferred into the troposphere
from the stratosphere, which is rich in ozone, through stratosphere-troposphere exchange (STE).
These intrusions usually occur behind cold fronts, bringing stratospheric air with them and
typically affect ozone concentrations in higher elevation areas (e.g. > 1500 m) more than areas at
lower elevations (U.S. EPA, 2013, section 3.4.1.1). The role of long-range transport of ozone and
other elements of ozone background are discussed in more detail in Section 2.4.
2.2 Spatial and Temporal Variations in Ambient Ozone Concentrations
Because ozone is a secondary pollutant formed in the atmosphere from precursor
emissions, concentrations are generally more regionally homogeneous than concentrations of
primary pollutants emitted directly from stationary and mobile sources (US EPA, 2013, section
3.6.2.1). However, variation in local emissions characteristics, meteorological conditions, and
topography can result in daily and seasonal temporal variability in ambient ozone concentrations,
as well as local and national-scale spatial variability.
Temporal variation in ambient ozone concentrations results largely from daily and seasonal
patterns in sunlight, precursor emissions, atmospheric stability, wind direction, and temperature
(US EPA, 2013, section 3.7.5). On average, ambient ozone concentrations follow well-
recognized daily and seasonal patterns, particularly in urban areas. Specifically, daily maximum
1-hour ozone concentrations in urban areas tend to occur in mid-afternoon, with more
pronounced peaks in the warm months of the ozone season than in the colder months (US EPA,
2013, Figures 3-54, 3-156 to 3-157). Rural sites also follow this general pattern, though it is less
pronounced in colder months (US EPA, 2013, Figure 3-55). With regard to day-to-day
variability, median maximum daily 8-hour average (MDA8) ozone concentrations in U.S. cities
-------
from 2007 through 2009 were approximately 47 ppb, with typical ranges between 35 to 60 ppb
and the highest MDA8 concentrations above 100 ppb in several U.S. cities (as noted further
below).
In addition to temporal variability, there is considerable spatial variability in ambient
ozone concentrations within cities and across different cities in the United States. With regard to
spatial variability within a city, local emissions characteristics, geography, and topography can
have important impacts. For example, as noted above, fresh NO emissions from motor vehicles
titrate ozone present in the urban background air, resulting in an ozone gradient around roadways
with ozone concentrations increasing as distance from the road increases (US EPA, 2013, section
3.6.2.1). Measured ozone concentrations are relatively uniform and well-correlated within some
cities (e.g., Atlanta) while they are more variable in others (e.g., Los Angeles) (US EPA, 2013,
section 3.6.2.1 and Figures 3-28 to 3-36).
Ozone concentrations also vary considerably across cities. Several cities had very high
measured ozone concentrations in 2007 through 2009 when the maximum recorded MDA8 was
137 ppb in Los Angeles, and was near or above 120 ppb in Atlanta, Baltimore, Dallas, New York
City, Philadelphia, and St. Louis (US EPA, 2013, Table 3-10). These same cities also had high
98th percentile ozone concentrations, with Los Angeles recording the highest 98th percentile
concentration (91 ppb) and many eastern and southern cities reporting 98th percentile
concentrations near or above 75 ppb. In contrast, somewhat lower 98th percentile ozone
concentrations were recorded in cities in the western United States outside of California (US
EPA, 2013, Table 3-10).
Rural sites can be affected by transport of ozone or ozone precursors from upwind urban
areas and by local anthropogenic sources such as motor vehicles, power generation, biomass
combustion, or oil and gas operations (US EPA, 2013, section 3.6.2.2). In addition, ozone tends
to persist longer in rural than in urban areas due to lower rates of chemical scavenging in non-
urban environments. At higher elevations, increased ozone concentrations can also result from
stratospheric intrusions (US EPA, 2013, sections 3.4, 3.6.2.2). As a result, ozone concentrations
measured in some rural sites can be higher than those measured in nearby urban areas (US EPA,
2013, section 3.6.2.2).
2-4
-------
2.3 Ozone Monitoring
2.3.1 Ozone Monitoring Network
To monitor compliance with the National Ambient Air Quality Standards (NAAQS), state
and local environmental agencies operate ozone monitoring sites at various locations, depending
on the population of the area and typical peak ozone concentrations.8 All of the state and local
monitoring stations that report data to the EPA Air Quality System (AQS) use ultraviolet (UV)
Federal Equivalent Methods (FEMs). In 2013, there were over 1,300 state, local, and tribal ozone
monitors reporting concentrations to EPA. The "State and Local Monitoring Stations" (SLAMS)
minimum monitoring requirements to meet the ozone design criteria are specified in 40 CFR Part
58, Appendix D. The requirements are both population and design value based.9 The minimum
number of ozone monitors required in a Metropolitan Statistical Area (MSA) ranges from zero
for areas with a population of at least 50,000 and under 350,000 with no recent history of an
ozone design value greater than 85 percent of the NAAQS, to four for areas with a population
greater than 10 million and an ozone design value greater than 85 percent of the NAAQS. At
least one site for each MSA, or Combined Statistical Area (CSA), must be sited to record the
maximum concentration for that particular metropolitan area. Since highest ozone concentrations
tend to be associated with particular seasons for various locations, EPA requires ozone
monitoring during specific ozone monitoring seasons, which vary by state.10
Figure 2-1 shows the locations of the U.S. ambient ozone monitoring sites reporting data to
EPA at any time during the 2009-2013 period. The gray dots that make up over 80% of the
ozone monitoring network are SLAMS monitors, which are operated by state and local
governments to meet regulatory requirements and provide air quality information to public health
agencies. Thus, the SLAMS monitoring sites are largely focused on urban and suburban areas.
The blue dots highlight two important subsets of monitoring sites within the SLAMS network:
8 The minimum ozone monitoring network requirements for urban areas are listed in Table D-2 of Appendix D to 40
CFR Part 58.
9 A design value is a statistic that describes the air quality status of a given area relative to the level of the NAAQS.
Design values are typically used to classify nonattainment areas, assess progress towards meeting the NAAQS,
and develop control strategies. See http://epa.gov/airtrends/values.html (U, 2010, 677582) for guidance on how
these values are defined.
10 The required ozone monitoring seasons for each state are listed in Table D-3 of Appendix D to 40 CFR Part 58.
Revised monitoring seasons are being proposed along with the proposed revision of the ozone NAAQS level.
2-5
-------
the "National Core" (NCore) multi-pollutant monitoring network and the "Photochemical
Assessment Monitoring Stations" (PAMS) network.
While the existing U.S. ozone monitoring network has a largely urban focus, to address
ecosystem impacts of ozone, such as biomass loss and foliar injury, it is equally important to
focus on ozone monitoring in rural areas. The green dots in Figure 2-1 represent the Clean Air
Status and Trends Network (CASTNET) monitors, which are located in rural areas. There were
about 80 CASTNET sites operating in 2013, with sites in the eastern U.S. being operated by
EPA and sites in the western U.S. being operated by the National Park Service (NPS).11 In total,
there were about 120 rural ozone monitoring sites operating in the U.S. in 2013.
SLAMS
CASTNET
NCORE/PAMS • SPMS/OTHER
Figure 2-1. Map of U.S. Ambient Os Monitoring Sites Reporting Data to EPA During the
2009-2013 Period
11 Additionally, the black dots represent "Special Purpose Monitoring Stations" (SPMS), which include about 20
rural monitors as part of the "Portable Os Monitoring System" (POMS) network operated by the NPS.
2-6
-------
2.3.2 Recent Ozone Monitoring Data and Trends
To determine whether or not the ozone NAAQS has been met at an ambient monitoring
site, a statistic commonly referred to as a "design value" must be calculated based on three
consecutive years of data collected from that site. The form of the existing ozone NAAQS design
value (DV) statistic is the 3-year average of the annual 4th highest daily maximum 8-hour ozone
concentration in parts per billion (ppb), with decimal digits truncated. The existing primary and
secondary ozone NAAQS are met at an ambient monitoring site when the DV is less than or
equal to 75 ppb.12 In counties or other geographic areas with multiple monitoring sites, the area-
wide DV is defined as the DV at the highest individual monitoring site, and the area is said to
have met the NAAQS only if all monitoring sites in the area are meeting the NAAQS.
Figure 2-2 shows the trend in the annual 4th highest daily maximum 8-hour ozone
concentrations in ppb based on 910 "trends" sites with complete data records over the 2000 to
2013 period. The center line in this figure represents the median value across the trends sites,
while the dashed lines represent the 25th and 75th percentiles, and the bottom and top lines
represent the 10th and 90th percentiles. Figure 2-3 shows a map of the ozone DVs (in ppb)
averaged across the 2009-2011, 2010-2012, and 2011-2013 periods at all monitoring sites in the
contiguous U.S.13 The trend figure shows that the annual 4th highest daily maximum values
decreased for the vast majority of monitoring sites in the U.S. between 2000 and 2013. The
decreasing trend is especially sharp from 2002 to 2004, when EPA implemented the "NOx SIP
Call", a program designed to reduce summertime emissions of NOx in the eastern U.S., but has
continued to decrease since then, in part due to ongoing reductions in mobile source NOx
emissions. Within the overall downward trend, there are periodic short-term increases. These
variations from the overall trend are the result of inter-annual variability in meteorological
conditions.
12 For more details on the data handling procedures used to calculate design values for the existing ozone NAAQS,
see 40 CFR Part 50, Appendix P.
13
All monitoring sites in Alaska, Hawaii, and Puerto Rico had DVs below 60 ppb.
2-7
-------
Trend in Annual 4th Highest Daily Maximum 8-hour O3 Concentrations
100 -
90 -
80 -
o
70 -
60 -
50
National Trend Based on 910 Monitoring Sites
o
o
o o
CM CM
••- CM
o o o o
o o o o
CM CM CM CM
oo o> o
o o --
o o o
CM CM CM
-- CM
PJ
O
Year
Figure 2-2. Trend in U.S. Annual 4th Highest Daily Maximum 8-hour Ozone
Concentrations in ppb, 2000 to 2013. Solid center line represents the median
value across monitoring sites, dashed lines represent 25th and 75th percentile
values, and top/bottom lines represent 10th and 90th percentile values.
• 44-60 ppb (99 sites) o 66-70 ppb (334 sites) • 76-105 ppb (265 sites)
O 61 - 65 ppb (193 sites) © 71-75 ppb (334 sites)
Figure 2-3. Map of 8-hour Ozone Design Values in ppb, Averaged Across the 2009-2011,
2010-2012, and 2011-2013 Periods
2-8
-------
In addition to the DV described above, another ozone metric of interest is the W126 index
value, which has been found to correlate with ozone-related damage to plants and ecosystems
(EPA, 2014c). The W126 metric is a seasonal aggregate of daytime (8:00 AM to 8:00 PM)
hourly ozone concentrations designed to measure the cumulative effects of ozone exposure on
plant and tree species, with units in parts per million-hours (ppm-hrs). The W126 metric uses a
logistic weighting function to place less emphasis on exposure to low hourly ozone
concentrations and more emphasis on exposure to high hourly ozone concentrations (Lefohn et
al, 1988).
Figure 2-4 shows the trend in annual W126 concentrations in ppm-hrs based on 900
"trends" sites with complete data records over the 2000 to 2013 period. The center line in this
figure represents the median value across the trends sites, while the dashed lines represent the
25th and 75th percentiles, and the bottom and top lines represent the 10th and 90th percentiles.
Figure 2-5 shows a map of the 3-year average annual W126 concentrations in ppm-hrs averaged
across the 2009-2011, 2010-2012, and 2011-2013 periods at all monitoring sites in the
contiguous U.S. The general patterns seen in these figures are similar to those seen in the DV
metric for the existing standard.
35
Trend in Annual W126 Concentrations
5 -
National Trend Based on 900 Monitoring Sites
Year
I
o
2-9
-------
Figure 2-4. Trend in U.S. Annual W126 Concentrations in ppm-hrs, 2000 to 2013. Solid
center line represents the median value across monitoring sites, dashed lines
represent 25th and 75th percentile values, and top/bottom lines represent 10th
and 90th percentile values.
• 0 - 3 ppm-hr (106 sites) o 8 -11 ppm-hr (381 sites) • 16-55 ppm-hr (147 sites)
© 4 - 7 ppm-hr (326 sites) © 12-15 ppm-hr (241 sites)
Figure 2-5. Map of 3-year Average W126 Values in ppm-hrs, Averaged Across the 2009-
2011, 2010-2012, and 2011-2013 Periods
2.4 Background Ozone
One of the aspects of ozone that is unusual relative to the other pollutants with NAAQS is
that, periodically, in some locations, an appreciable fraction of the observed ozone results from
sources or processes other than local and domestic regional anthropogenic emissions of ozone
precursors (Fiore et a/., 2002). Any ozone formed by processes other than the chemical
conversion of local or regional ozone precursor emissions is generically referred to as
"background" ozone. Background ozone can originate from natural sources of ozone and ozone
precursors, as well as from manmade international emissions of ozone precursors. Natural
sources of ozone precursor emissions such as wildfires, lightning, and vegetation can lead to
ozone formation by chemical reactions with other natural sources. Another important component
2-10
-------
of background is ozone that is naturally formed in the stratosphere through interactions of
ultraviolet light with molecular oxygen. Stratospheric ozone can mix down to the surface at high
concentrations in discrete events called intrusions, especially at higher-altitude locations. The
manmade portion of the background includes any ozone formed due to anthropogenic sources of
ozone precursors emitted far away from the local area (e.g., international emissions). Finally,
both biogenic and international anthropogenic emissions of methane, which can be chemically
converted to ozone over relatively long time scales, can also contribute to global background
ozone levels. Away from the surface, ozone can have an atmospheric lifetime on the order of
weeks. As a result, background ozone can be transported long distances in the upper troposphere
and, when meteorological conditions are favorable, be available to mix down to the surface and
add to the ozone loading from non-background sources.
The definition of background ozone can vary depending upon context, but it generally
refers to ozone that is formed by sources or processes that cannot be influenced by actions within
the jurisdiction of concern. In the Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards (US EPA, 2014c), EPA identified three specific definitions of
background ozone: natural background (NB), North American background (NAB), and United
States background (USB). Natural background is the narrowest definition of background, and it
is defined as the ozone that would exist in the absence of any manmade ozone precursor
emissions. The other two definitions of background are based on a presumption that the U.S. has
little influence over anthropogenic emissions outside either our continental or domestic borders.
North American background is defined as that ozone that would exist in the absence of any
manmade ozone precursor emissions from North America. U.S. background is defined as that
ozone that would exist in the absence of any manmade emissions inside the United States.
Modeling studies have estimated what background levels would be in the absence of
certain sets of emissions by simply assessing the remaining ozone in a simulation in which
certain emissions were removed (Zhang et al. (2011), Emery et al. (2012), US EPA (2014c)).
This basic approach is often referred to as "zero-out" modeling or "emissions perturbation"
modeling. While the zero-out approach has traditionally been used to estimate natural
background, North American background, and U.S. background, the methodology has an
2-11
-------
acknowledged limitation. It cannot answer the question of how much of the existing observed
ozone results from background sources or processes.
A separate modeling technique can be used to estimate the contribution of background
ozone and other contributing source terms to total ozone within a model. This approach, referred
to as "source apportionment" modeling, has been described and evaluated in the peer-reviewed
literature (Dunker et al, 2002; Kemball-Cook et al., 2009). Source apportionment modeling has
frequently been used in other regulatory settings to estimate the "contribution" to ozone of
certain sets of emissions (EPA 2005, EPA 2011). The source apportionment technique provides a
means of estimating the contributions of each user-identified source category to ozone formation
in a single model simulation. This is achieved by using multiple tracer species to track the fate of
ozone precursor emissions (VOC and NOx) and the ozone formation resulting from these
emissions. The methodology is designed so that all ozone and precursor concentrations are
tracked and apportioned to the selected source categories at all times without perturbing the
inherent chemistry. The primary limitation of the source apportionment modeling is that its
estimations of background ozone are explicitly linked to the emissions scenarios modeled and
would change with different emissions scenarios.
2.4.1 Seasonal Mean Background Ozone in the U.S.
The ISA (US EPA 2013, section 3.4) previously established that background ozone
concentrations vary spatially and temporally and that simulated mean background concentrations
are highest at high-elevation sites within the western U.S. Background levels typically are
greatest over the U.S. in the spring and early summer. EPA modeling presented in the Policy
Assessment for the Review of the Ozone National Ambient Air Quality Standards (US EPA,
2014c) focused on the months from April to October for 2007 (note that the 2007 modeling is
separate from the 2011 modeling described in Chapter 3 and used as the basis of cost and benefit
numbers in this RIA). Emissions and model set-up for the 2007 analysis are described in more
detail in the Policy Assessment. Briefly, the emissions for 2007 were derived from the 2008
National Emissions Inventory but included 2007 year-specific emissions where available.
Wildfire emissions were based on a multi-year climatological average as this analysis was meant
to capture seasonal mean background and typical ranges rather than explicitly quantify
background ozone on specific days. Figure 2-6 displays the spatial patterns of seasonal mean
2-12
-------
natural background ozone as estimated by a 2007 zero-out scenario using the Community
Multiscale Air Quality (CMAQ) model. Seasonal means are computed over those seven months.
This figure shows the average daily maximum 8-hour ozone concentration that would exist in the
absence of any anthropogenic ozone precursor emissions at monitor locations. As shown,
seasonal meanNB levels range from approximately 15-35 ppb (i.e., +/- 1 standard deviation)
with the highest values at higher-elevation sites in the western U.S. The median value over these
locations is 24.2 ppb, and more than 50 percent of the locations have natural background levels
of 20-25 ppb. The highest modeled estimate of seasonal average, natural background, 8-hr daily
maximum ozone is 34.3 ppb at the high-elevation CASTNET site (Gothic) in Gunnison County,
CO. Natural background ozone levels are higher at these high-elevation locations primarily
because of natural stratospheric ozone impacts and international transport impacts that increase
with altitude (where ozone lifetimes are longer).
Ozone (ppb)
< 20 (104)
20-25 (740)
25-30 (331)
30-35(119)
35-40 (0)
O 40-45 (0)
O 45-50(0)
50-55 (0)
55-60 (0)
> 60 (0)
Figure 2-6. Map of 2007 CMAQ-estimated Seasonal Mean of 8-hour Daily Maximum
Ozone from Natural Background (ppb) based on Zero-Out Modeling
Figures 2-7 and 2-8 show the same information for the NAB and USB scenarios. In these
model runs, all anthropogenic ozone precursor emissions were removed from the U.S., Canada,
and Mexico portions of the modeling domain (NAB scenario) and then only from the U.S. (USB
scenario). The figures show that there is not a large difference between the NAB and USB
scenarios. Seasonal mean NAB and USB ozone levels range from 25-50 ppb, with the most
frequent values estimated in the 30-35 ppb range. The median seasonal mean background levels
2-13
-------
are 31.5 and 32.7 ppb (NAB and USB, respectively). Again, the highest levels of seasonal mean
background ozone are predicted over the intermountain western U.S. Locations with NAB and
USB concentrations greater than 40 ppb are confined to Colorado, Nevada, Utah, Wyoming,
northern Arizona, eastern California, and parts of New Mexico. The 2007 EPA modeling
suggests that seasonal mean USB concentrations are on average 1 -3 ppb higher than NAB
background. These results were similar to those reported by Wang et al. (2009). From a seasonal
mean perspective, background ozone levels are below the NAAQS thresholds.
Ozone (ppb)
< 20 (0)
20-25 (0)
25-30 (396)
30-35 (628)
35-40 (147)
O 40-45(121)
O 45 - 50 (2)
© 50-55 (0)
55-60 (0)
(0)
Figure 2-7. Map of 2007 CMAQ-estimated Seasonal Mean of 8-hour Daily Maximum
Ozone from North American Background (ppb) based on Zero-out Modeling
2-14
-------
Ozone (ppb)
• < 20 (0)
20 - 25 (0)
25-30(127)
30 - 35 (842)
35-40 (188)
O 40-45(132)
O 45-50 (5)
Q 50-55 (0)
55 - 60 (0)
Figure 2-8. Map of 2007 CMAQ-estimated Seasonal Mean of 8-hour Daily Maximum
Ozone from United States Background (ppb) based on Zero-Out Modeling
2.4.2 Seasonal Mean Background Ozone in the U. S. as a Proportion of Total Ozone
Another informative way to assess the importance of background ozone as part of
seasonal mean ozone levels across the U.S. is to consider the ratios of NB, NAB, and USB to
total modeled ozone at each monitoring location. Considering the proportional impact of
background ozone allows for an initial assessment of the relative importance of background and
non-background sources. Because ozone chemistry is non-linear, one should not assume that
individual perturbations (e.g., zero-out runs) are additive in all locations. Figures 2-9 and 2-10
show the ratio of U.S. background to total ozone using the metric of the seasonal mean 8-hr daily
maximum ozone concentrations as estimated by both the zero-out and source apportionment
modeling methodologies. Recall that the terms NB, NAB, and USB are explicitly linked to the
zero-out modeling approach. For comparison, in Figure 2-10 we are extending the definition of
USB to also include the source apportionment model estimates of the ozone that are attributable
to sources other than U.S. anthropogenic emissions. To preserve the original definition of USB,
this second term will be hereafter referred to as "apportionment-based USB". As noted earlier,
the advantage of the source apportionment modeling is that all of the modeled ozone is attributed
to various source terms without perturbing the inherent chemistry. Thus, this approach is not
affected by the confounding occurrences of background ozone values exceeding the base ozone
values as can happen in the zero-out modeling (i.e., background proportions > 100%).
2-15
-------
Consequently, one would expect the fractional background levels to be lower in the source
apportionment methodology as a result of removing this artifact.
When averaged over all sites, ozone from sources other than U.S. anthropogenic
emissions is estimated to comprise 66 (zero-out) and 59 (source apportionment) percent of the
total seasonal ozone mean. The spatial patterns of USB and apportionment-based USB are
similar across the two modeling exercises. Background ozone is a relatively larger percentage
(e.g., 70-80%) of the total seasonal mean ozone in locations within the intermountain western
U.S. and along the U.S. border. In locations where ozone levels are generally higher, like
California and the eastern U.S., the seasonal mean background fractions are relatively smaller
(e.g., 40-60%). The additional 2007 modeling confirms that background ozone, while generally
not approaching levels of the ozone standard, can comprise a considerable fraction of total
seasonal mean ozone across the U.S (EPA, 2014c).
2.4.3 Daily Distributions of Background Ozone within the Seasonal Mean
As a first-order understanding, it is valuable to be able to characterize seasonal mean
levels of background ozone. However, it is well established that background levels can vary
substantially from day-to-day within the seasonal mean. From an implementation perspective,
the values of background ozone on possible exceedance days are a more meaningful
consideration. The Policy Assessment for the Review of the Ozone National Ambient Air
Quality Standards (US EPA, 2014c) concluded that "anthropogenic sources within the U.S. are
largely responsible for 4th highest 8-hour daily maximum 63 concentrations" based on modeling
using a 2007 base year and the two distinct modeling methodologies described above. Figure 2-
11 and 2-12 show the distribution of daily MDA8 apportionment-based USB levels (absolute
magnitudes and relative fractions, respectively) from the CAMx simulation. The 2007 modeling
shows that the days with highest ozone levels have similar distributions (i.e., means, inter-
quartile ranges) of background ozone levels as days with lower values, down to approximately
40 ppb. As a result, the proportion of total ozone that has background origins is smaller on high
ozone days (e.g., days > 60 ppb) than on the more common lower ozone days that tend to drive
seasonal means. Figure 2-11 also indicates that there are cases in which the model predicts much
larger background proportions, as shown by the upper outliers in the figure. These infrequent
episodes usually occur in relation to a specific event, and occur more often in specific
2-16
-------
geographical locations, such as at high elevations or wildfire prone areas during the local dry
season.
It should be noted here that EPA has policies for treatment of air quality monitoring data
affected by these types of events. EPA's exceptional events policy allows exclusion of certain air
quality monitoring data from regulatory determinations if a State adequately demonstrates that an
exceptional event has caused the exceedance or violation of a NAAQS. In addition, Section
179B of the Clean Air Act (CAA) also provides for treatment of air quality data from
international transport when an exceedance or violation of a NAAQS would not have occurred
but for the emissions emanating from outside of the United States. Finally, CAA section 182(h)
authorizes the EPA Administrator to determine that an area designated nonattainment can be
treated as a "rural transport area". In accordance with the statute, a nonattainment area may
qualify for this distinction if it meets the following criteria: 1) the area does not contain
emissions sources that make a significant contribution to monitored ozone concentrations in the
area, or in other areas; and 2) the area does not include and is not adjacent to an MSA. More
information regarding how background ozone is addressed in Clean Air Act implementation is
provided in Section VII.F of the notice of proposed rulemaking.
< 40% 10)
40 -50% (2)
50 -60% (393)
60-70% (551)
Q 70-80% (263)
; 80% <84|
Sources: U3G3, EGRI. TANA; AND Sou
Figure 2-9. Map of Site-Specific Ratios of U.S. Background to Total Seasonal Mean
Ozone based on 2007 CMAQ Zero-Out Modeling
2-17
-------
< 40% (0)
40 - 50% (343)
50 - 60% (433)
60 - 70% (237)
70-80% (178)
> 80% (52)
M e x i r
Figure 2-10. Map of Site-Specific Ratios of Apportionment-Based U.S. Background to
Seasonal Mean Ozone based on 2007 CAMx Source Apportionment Modeling
IB 60-
E
5
E
i i
<25 25-30 30-35 35-40 40-45 45-50
93-95 35-133 - 133
Bins of Base Model MDA8 Ozone (ppb)
Figure 2-11. Distributions of Absolute Estimates of Apportionment-Based U.S.
Background (all site-days), Binned by Modeled MDA8 from the 2007 Source
Apportionment Simulation
2-18
-------
* 1.00-
m
9
o
E "
o
o
c
o
30.50-
03
1
O 0.25-
<25 25-30 30-35 35^0 40^t5 45-50 50-55 55*0 60-65 65-70 70-75 75-EO BO-B5 85-90 90-95 95-100 > 100
Bins of Base Model MDA8 Ozone (ppb)
Figure 2-12. Distributions of the Relative Proportion of Apportionment-Based U.S.
Background to Total Ozone (all site-days), Binned by Modeled MDA8 from the
2007 Source Apportionment Simulation
2.5 References
Camalier, L.; Cox, W.; Dolwick, P. (2007). The Effects of Meteorology on Ozone in Urban Areas and their use in
Assessing Ozone Trends. Atmos Environ 41: 7127-7137.
Dunker, A.M.; Yarwood, G; Ortmann, J.P.; Wilson, G.M. (2002). Comparison of source apportionment and source
sensitivity of ozone in a three-dimensional air quality model. Environmental Science & Technology 36: 2953-
2964.
Emery, C; Jung, J; Downey, N; Johnson, J; Jimenez, M; Yarwood, G; Morris, R. (2012). Regional and global
modeling estimates of policy relevant background ozone over the United States. Atmos Environ 47: 206-217.
http://dx.doi.org/10.1016/j.atmosenv.2011.11.012.
Helmig, D., Thompson, C.R., Evans, J., Boylan, P., Hueber, J., Park, J.H. (2014). Highly elevated atmospheric
levels of volatile organic compounds in the Uintah Basin, Utah, Environmental Science & Technology, 48, 4707-
4715.
Henderson, B.H., Possiel, N., Akhtar, F., Simon, H.A. (2012). Regional and Seasonal Analysis of North American
Background Ozone Estimates from Two Studies. Available on the Internet at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_td.html
Kemball-Cook, S.; Parrish, D.; Ryerson, T; Nopmongcol, U.; Johnson, J.; Tai, E.; Yarwood, G. (2009).
Contributions of regional transport and local sources to ozone exceedances in Houston and Dallas: Comparison of
results from a photochemical grid model to aircraft and surface measurements. Journal of Geophysical Research-
Atmospheres, 114: DOOF02. DOI: 10.1029/2008JDO10248.
2-19
-------
Lapina K., Henze D.K., Milford J.B., Huang M, Lin M, Fiore A.M., Carmichael G., Pfister G.G., Bowman K.
(2014). Assessment of source contributions to seasonal vegetative exposure to ozone in the U.S. Journal of
Geophysical Research-Atmospheres, 119:DOI: 10.1002/2013JD020905.
Lefohn, A. S.; Laurence, J. A.; Kohut, R. J. (1988). A comparison of indices that describe the relationship between
exposure to ozone and reduction in the yield of agricultural crops. Atmos. Environ. 22: 1229-1240.
Rappengluck, B., Ackermann, L., Alvarez, S., Golovko, J., Buhr, M., Field, R.A., Soltis, J., Montague, D.C., Hauze,
B., Adamson, S., Risch, D., Wilkerson, G., Bush, D., Stoeckenius, T., Keslar, C. (2014). Strong wintertime ozone
events in the Upper Green River basin, Wyoming, Atmospheric Chemistry and Physics, 14, 4909-4934.
Schnell, R.C., Oltmans, S.J., Neely, R.R., Endres, M.S., Molenar, J.V., White, A.B. (2009) Rapid photochemical
production of ozone at high concentrations in a rural site during winter, Nature Geoscience, 2, 120-122.
U.S. Environmental Protection Agency (2005). Technical Support Document for the Final Clean Air Interstate Rule
Air Quality Modeling. Office of Air Quality Planning and Standards, Research Triangle Park, NC, 285pp.
http://www.epa.gov/cair/technical.html.
U.S. Environmental Protection Agency (2011). Air Quality Modeling Final Rule Technical Support Document.
Office of Air Quality Planning and Standards, Research Triangle Park, NC, 363pp.
http://www.epa.gov/airtransport/CSAPR/techinfo.html.
U.S. Environmental Protection Agency. (2013). Integrated Science Assessment for Ozone and Related
Photochemical Oxidants, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA/600/R-
10/076.
U.S. Environmental Protection Agency (2014a): Health Risk and Exposure Assessment for Ozone, final report. US
EPA, OAQPS, 2014, RTF, NC, EPA-452/R-14-004a
U.S. Environmental Protection Agency. (2014b). Welfare Risk and Exposure Assessment for Ozone, final, U.S.
Environmental Protection Agency, Research Triangle Park, NC. EPA-452/P-14-005a.
U.S. Environmental Protection Agency. (2014c). Policy Assessment for the Review of the Ozone National Ambient
Air Quality Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA 452/R-14-006.
Wang, HQ; Jacob, DJ; Le Sager, P; Streets, DG; Park, PJ; Gilliland, AB; van Donkelaar, A. (2009). Surface ozone
background in the United States: Canadian and Mexican pollution influences. Atmos Environ 43: 1310-1319.
http://dx.doi.org/10.1016/j.atmosenv.2008.11.036.
Zhang, L; Jacob, DJ; Downey, NV; Wood, DA; Blewitt, D; Carouge, CC; Van donkelaar, A; Jones, DBA; Murray,
LT; Wang, Y. (2011). Improved estimate of the policy-relevant background ozone in the United States using the
GEOS-Chem global model with 1/2° x 2/3° horizontal resolution over North America. Atmos Environ 4.
2-20
-------
CHAPTER 3: AIR QUALITY MODELING AND ANALYSIS
Overview
This regulatory impacts analysis (RIA) evaluates the costs as well as the health and
environmental impacts associated with complying with alterative National Ambient Air Quality
Standards (NAAQS) for ozone. For this purpose, we use air quality modeling to project ozone
concentrations into the future. This chapter describes the data, tools and methodology used for
the analysis, as well as the post-processing techniques used to produce a number of ozone
metrics necessary for this analysis.
Throughout this chapter, the base year modeling refers to model simulations conducted
for 2011 while the 2025 base case simulation refers to a photochemical model run conducted
with emissions projected to the year 2025 assuming all current on-the-books federal regulations
will apply14. A series of 2025 emissions sensitivity cases are created to determine ozone
response to emissions changes incremental to the 2025 base case. Finally, a set of four scenarios
are developed based on the 2025 base case and emissions sensitivity cases: the baseline scenario
(a scenario which applies additional controls to the 2025 base case that would be required to
meet the current standard of 75 ppb), and 3 alternative standard scenarios which represent
incremental emissions reductions beyond the baseline to meet potential standard levels of 70, 65,
and 60 ppb.
Section 3.1 describes the air quality modeling simulations, section 3.2 describes how
current and future ozone design values are calculated, section 3.3 describes the methodology for
determining necessary emissions reductions for meeting various alternative NAAQS levels, and
section 3.4 describes the creation of spatial surfaces that act as inputs to health and welfare
benefits calculations.
14 The 2012 PM NAAQS is not included in the 2025 base case because the scenarios modeled in PM NAAQS RIA
did not reflect any NOx emissions reductions (US EPA, 2012)
5-1
-------
3.1 Modeling Ozone Levels in the Future
A national scale air quality modeling analysis was performed to estimate ozone
concentrations for the future year of 2025. Ozone sensitivity factors were developed using the
modeled response of ozone to changes in NOX and VOC emissions from various sources and
locations. The sensitivity factors were used to calculate ratios of changes in ozone to changes in
emissions in order to determine the amount of emissions reductions needed to reach the baseline
and evaluate potential alternative standard levels of 70, 65, and 60 ppb incremental to the
baseline. The resulting emissions reductions were then used to estimate how health- and welfare-
related ozone concentration metrics would change under each scenario. The metrics were used as
inputs to the calculation of expected costs and benefits associated with the precursor emissions
and ozone concentration changes resulting from just attaining the alternative ozone standards.
As described in section 3.2, air quality modeling was used in a relative sense to project
future concentrations of ozone. As part of this approach, ozone predictions from the 2011 base
year simulation are coupled with predictions from the 2025 modeling to calculate the relative
change (between 2011 and 2025) in concentrations. These relative response factors (RRFs) were
applied to the corresponding measured design values15 (DVs) to predict future DVs. Multiple
emissions cases were modeled for 2025 including a 2025 base case and twelve 2025 emissions
sensitivity simulations. Details on the 2011-based air quality modeling platform, the 2025 base
case and emissions sensitivity simulations, along with the methods and results for attaining these
NAAQS levels are provided below.
3.1.1 Selection of Future Analytic Year
The RIA evaluates, to the extent possible, the costs and benefits of attaining the proposed
alternative ozone standards, incremental to attaining the existing 75 ppb ozone standard and
implementing existing and expected regulations. We selected 2025 as the primary year of
analysis because most areas of the U.S. will likely be required to meet a revised ozone standard
by 2025. We assumed that potential nonattainment areas everywhere in the U.S., excluding
15 The design value is the metric that is compared to the standard level to determine whether a monitor is violating
the NAAQS. The ozone design value is described in more detail in section 3.2.
5-2
-------
California, will be designated such that they are required to attain by 2025, and we developed our
projected baselines for emissions, ozone, and populations for 2025.
In estimating the incremental costs and benefits of potential alternative standards, we
recognize that there are several areas that are not required to meet the existing ozone standard of
75 ppb set in 2008 by the year 2025. The Clean Air Act allows areas with more significant air
quality problems to take additional time to reach the existing standard. Several areas in
California are not required to meet the existing standard by 2025 and may not be required to
meet a revised standard until sometime between 2032 and 2037.
We projected emissions for and modeled a single future base case year (2025), however for
California we adjusted the future baseline ozone concentrations to reflect the effects of mobile
source emissions reductions that will occur in California between 2025 and 2030 as described in
Section 3.3. While there is uncertainty about the precise timing of emissions reductions and
related costs for California, we assume costs occur through the end of 2037 and beginning of
2038. In addition, as described in Chapter 5, we model benefits for California using projected
population demographics for 2038.
Because of the different timing for incurring costs and accruing benefits in California and
for ease of discussion throughout the analyses, we refer to the different time periods for potential
attainment as 2025 and post-2025, to reflect that (1) we did not project emissions and air quality
for any year other than 2025; (2) costs in California are assumed to be incurred starting in 2032
and later; and (3) benefits from attainment of alternative standards in California are modeled
using population demographics in 2038.
3.1.2 Air Quality Modeling Platform
The 2011-based air quality modeling platform was used to provide emissions,
meteorology and other inputs to the 2011 and 2025 air quality model simulations. This platform
was chosen because it represents the most recent, complete set of base year emissions
information currently available for national-scale modeling.
We use the Comprehensive Air Quality Model with Extensions (CAMx version 6.1) for
photochemical model simulations performed for the RIA. CAMx is a three-dimensional grid-
-------
based Eulerian air quality model designed to estimate the formation and fate of oxidant
precursors, primary and secondary particulate matter concentrations, and deposition over
regional and urban spatial scales (e.g., over the contiguous U.S.) (Environ, 2014). Consideration
of the different processes (e.g., transport and deposition) that affect primary (directly emitted)
and secondary (formed by atmospheric processes) pollutants at the regional scale in different
locations is fundamental to understanding and assessing the effects of emissions control
measures that affect air quality concentrations. Because it accounts for spatial and temporal
variations as well as differences in the reactivity of emissions, CAMx is useful for evaluating the
impacts of the control strategies on ozone concentrations. CAMx is applied with the carbon-bond
6 revision 2 (CB6r2) gas-phase chemistry mechanism (Ruiz and Yarwood, 2013).
12US2 domain , .
x.y origin: -24120001^1, (;162Q0l)Otn /
col: 396 row:246 J\ X. V
Figure 3-1. Map of the CAMx Modeling Domain Used for Ozone NAAQS RIA
5-4
-------
Figure 3-1 shows the geographic extent of the modeling domain that was used for air
quality modeling in this analysis. The domain covers the 48 contiguous states along with the
southern portions of Canada and the northern portions of Mexico. This modeling domain
contains 25 vertical layers with a top at about 17,600 meters, or 50 millibars (mb), and horizontal
resolution of 12 km x 12 km. The model simulations produce hourly air quality concentrations
for each 12 km grid cell across the modeling domain.
CAMx requires a variety of input files that contain information pertaining to the
modeling domain and simulation period. These include gridded, hourly emissions estimates and
meteorological data, and initial and boundary conditions. Separate emissions inventories were
prepared for the 2011 base year, the 2025 base case, and the 2025 emissions sensitivity
simulations. All other inputs (i.e. meteorological fields, initial conditions, and boundary
conditions) were specified for the 2011 base year model application and remained unchanged for
each future-year modeling simulation. The assumption of constant meteorology and boundary
conditions was applied for two reasons: 1) this allows us to isolate the impacts of U.S. emissions
changes, and 2) there is considerable uncertainty in the direction and magnitude in any changes
in these parameters. EPA recognizes that changes in climate and international emissions may
impact these model inputs. Specifically, climate change may lead to temperature increases,
higher stagnation frequency, and increased wildfire activity, all of which could lead to higher
ozone concentrations. In the western U.S. over the last 15 years, increasing wildfires have
already been observed (Dennison et al, 2014). Potential future elevated ozone concentrations
could, in turn, necessitate more stringent emissions reductions. However, there are significant
uncertainties regarding the precise location and timing of climate change impacts on ambient air
quality. Generally, climate projections are most robust for periods at least several decades in the
future because the forcing mechanisms that drive near-term natural variability in climate patterns
(e.g., El Nino, North American Oscillation) have substantially larger signals over short time
spans than the driving forces related to long-term climate change. Boundary conditions, which
are impacted by international emissions and may also influence future ozone concentrations, are
held constant in this analysis based on a similar rationale regarding the significant uncertainty in
estimating future levels.
5-5
-------
CAMx requires detailed emissions inventories containing temporally allocated (i.e.,
hourly) emissions for each grid-cell in the modeling domain for a large number of chemical
species that act as primary pollutants and precursors to secondary pollutants. The annual
emission inventories, described in Section 3.1.3, were preprocessed into CAMx-ready inputs
using the Sparse Matrix Operator Kernel Emissions (SMOKE) modeling system (Houyoux et al,
2000).
Meteorological inputs reflecting 2011 conditions across the contiguous U.S. were derived
from Version 3.4 of the Weather Research Forecasting Model (WRF) (Skamarock, 2008). These
inputs included hourly-varying horizontal wind components (i.e., speed and direction),
temperature, moisture, vertical diffusion rates, and rainfall rates for each grid cell in each vertical
layer. Details of the annual 2011 meteorological model simulation and evaluation are provided in
a separate technical support document (US EPA, 2014a).
The lateral boundary and initial species concentrations are provided by a three-
dimensional global atmospheric chemistry model, GEOS-Chem (Yantosca, 2004) standard
version 8-03-02 with 8-02-01 chemistry. The global GEOS-Chem model simulates atmospheric
chemical and physical processes driven by assimilated meteorological observations from the
NASA's Goddard Earth Observing System (GEOS-5; additional information available at:
http://gmao.gsfc.nasa.gov/GEOS/ and http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-
5). This model was run for 2011 with a grid resolution of 2.0 degrees x 2.5 degrees (latitude-
longitude). The predictions were used to provide one-way dynamic boundary conditions at one-
hour intervals and an initial concentration field for the CAMx simulations. A model evaluation
was conducted to validate the appropriateness of this version and model configuration of GEOS-
Chem for predicting selected measurements relevant to their use as boundary conditions for
CAMx. This evaluation included using satellite retrievals paired with GEOS-Chem grid cell
concentrations (Henderson, 2014). More information is available about the GEOS-Chem model
and other applications using this tool at: http://www-as.harvard.edu/chemistry/trop/geos.
An operational model performance evaluation for ozone was performed to estimate the
ability of the CAMx modeling system to replicate 2011 measured concentrations. This
evaluation focused on statistical assessments of model predictions versus observations paired in
5-6
-------
time and space depending on the sampling period of measured data. Details on the evaluation
methodology and the calculation of performance statistics are provided in Appendix 3 A. Overall,
the model performance statistics for ozone from the CAMx 2011 simulation are within or close
to the ranges found in other recent peer-reviewed applications (Simon et al, 2012). These model
performance results give us confidence that our application of CAMx using this 2011 modeling
platform provides a scientifically credible approach for assessing ozone concentrations for the
purposes of the RIA.
3.1.3 Emissions Inventories
The 2011 base year and 2025 base case emissions inventories are described in the
Technical Support Document: Preparation of Emissions Inventories for the Version 6.1, 2011
Emissions Modeling Platform (US EPA, 2014b). Section 4 of the technical support document
(TSD) summarizes the control and growth assumptions by source type that were used to create
the U.S. 2025 base case emissions inventory, and includes a table of such assumptions for each
major source sector. Below we summarize the characteristics of the 2025 base case emissions
for each major source category.
The 2025 electric generating unit (EGU) projected inventory represents demand growth,
fuel resource availability, generating technology cost and performance, and other economic
factors affecting power sector behavior. The EGU emissions were developed using the Integrated
Planning Model (IPM) version 5.13
(http://www.epa.gov/powersectormodeling/BaseCasev513.html). IPM is a multiregional,
dynamic, deterministic linear programming model of the U.S. electric power sector. IPM reflects
the expected 2025 emissions accounting for the effects of environmental rules and regulations,
consent decrees and settlements, plant closures, units built, control devices installed, and forecast
unit construction through the calendar year 2025. In this analysis, the projected EGU emissions
include impacts from the Final Mercury and Air Toxics Standard (MATS) announced on
December 21, 2011 and the Clean Air Interstate Rule (CAIR) issued March 10, 200516.
16 A sensitivity case described in Section 3.1.4 also included a representation of EPA's proposed carbon pollution
guidelines under section 11 l(d) of the Clean Air Act (CAA)
5-7
-------
Projections for most stationary emission sources other than EGUs (i.e., non-EGUs) were
developed by using the EPA Control Strategy Tool (CoST) to create future year inventories.
CoST is described at http://www.epa.gov/ttnecasl/cost.htm. The 2025 base case non-EGU
stationary source emissions inventory includes all enforceable national rules and programs
including the Reciprocating Internal Combustion Engines (RICE) and cement manufacturing
National Emissions Standards for Hazardous Air Pollutants (NESHAPs) and Boiler Maximum
Achievable Control Technology (MACT) reconsideration reductions. Projection factors and
percent reductions for non-EGU point sources reflect comments previously received by EPA,
along with emissions reductions due to national and local rules, control programs, plant closures,
consent decrees and settlements. Projection approaches for corn ethanol and biodiesel plants,
refineries and upstream impacts represent the Energy Independence and Security Act (EISA)
renewable fuel standards mandate in the Renewable Fuel Standards Program (RFS2). Airport-
specific terminal area forecast (TAP) data were used for aircraft to account for projected changes
in landing/takeoff activity.
Regional projection factors for point and nonpoint oil and gas emissions were developed
by product type using Annual Energy Outlook (AEO) 2013 projections to year 2025
(http://www.eia.gov/forecasts/aeo/). Stationary engine criteria air pollutant (CAP) co-benefit
reductions (i.e., from the RICE NESHAP) and New Source Performance Standards (NSPS) VOC
controls are reflected for oil and gas sources.
Projection factors for livestock are based on expected changes in animal population from
2005 Department of Agriculture data, updated according to EPA experts in July 2012; fertilizer
application NH3 emissions projections include upstream impacts representing EISA. Area
fugitive dust projection factors for categories related to livestock estimates are based on expected
changes in animal population and upstream impacts from EISA. Residential Wood Combustion
(RWC) projection factors reflect assumed growth of wood burning appliances based on sales
data, equipment replacement rates and change outs. These changes include growth in lower-
emitting stoves and a reduction in higher emitting stoves. Projection factors for the remaining
nonpoint sources such as stationary source fuel combustion, industrial processes, solvent
utilization, and waste disposal, implement comments received on the projection of these sources
as a result of recent rulemakings and outreach to states on emission inventories, and they also
-------
include emission reductions due to control programs. Portable fuel container (PFC) projection
factors reflect the impact of the final Mobile Source Air Toxics (MSAT2) rule. Upstream
impacts from EISA, including post-2011 cellulosic ethanol plants are also reflected.
For onroad, nonroad, and commercial marine vessel mobile sources, all national
measures for which data were available at the time of modeling have been included. The Tier 3
standards finalized in March, 2014 (see http://www.epa.gov/otaq/tier3.htm) are represented in
the onroad and nonroad emissions. The 2011 and 2025 onroad mobile source emissions were
developed using emissions factors derived from the Tier 3 FRM version of the MOtor Vehicle
Emission Simulator (MOVES; http://www.epa.gov/otaq/models/moves/). The emissions factors
for year 2025 were developed using the same meteorology and procedures used to produce the
2011 emission factors. The onroad mobile source emissions were computed by using SMOKE
to combine the county-, vehicle type-, and temperature-specific emission factors with vehicle
miles traveled and vehicle population activity data, while taking into account hourly gridded
temperature data.
The MOVES-based 2025 onroad emissions account for changes in activity data and the
impact of on-the-books national rules including: the Tier 3 Vehicle Emission and Fuel Standards
Program, the Light-Duty Vehicle Tier 2 Rule, the Heavy Duty Diesel Rule, the Mobile Source
Air Toxics Rule, the Renewable Fuel Standard (RFS2), the Light Duty Green House
Gas/Corporate Average Fuel Efficiency (CAFE) standards for 2012-2016, the Heavy-Duty
Vehicle Greenhouse Gas Rule, the 2017 and the Later Model Year Light-Duty Vehicle
Greenhouse Gas Emissions and Corporate Average Fuel Economy Standards; Final Rule (LD
GHG). The MOVES-based 2025 emissions also include state rules related to the adoption of
LEV standards, inspection and maintenance programs, Stage II refueling controls, and local fuel
restrictions. For California, the base case emissions included most of this state's on-the-books
regulations, such as those for idling of heavy-duty vehicles, chip reflash, public fleets, track
trucks, drayage trucks, and heavy duty trucks and buses. The California emissions do not reflect
the impacts of the GHG/Smartway regulation, nor do they reflect state GHG regulations for the
projection of other emission sectors because that information was not included in the provided
inventories.
5-9
-------
The nonroad mobile 2025 emissions, including railroads and commercial marine vessel
emissions also include all national control programs. These control programs include the
Locomotive-Marine Engine rule, the Nonroad Spark Ignition rule and the Class 3 commercial
marine vessel "ECA-IMO" program. For California, the 2025 emissions for these categories
reflect the state's Off-Road Construction Rule for "In-Use Diesel", cargo handling equipment
rules in place as of 2011 (see http://www.arb.ca.gov/ports/cargo/cargo.htm), and state rules
through 2011 related to Transportation Refrigeration Units, the Spark-Ignition Marine Engine
and Boat Regulations adopted on July 24, 2008 for pleasure craft, and the 2007 and 2010
regulations to reduce emissions from commercial harbor craft. For ocean-going vessels, the
emissions data reflect the 2005 voluntary Vessel Speed Reduction (VSR) within 20 nautical
miles, the 2007 and 2008 auxiliary engine rules, the 40 nautical mile VSR program, the 2009
Low Sulfur Fuel regulation, the 2009-2018 cold ironing regulation, the use of 1% sulfur fuel in
the Emissions Control Area (ECA) zone, the 2012-2015 Tier 2 NOX controls, the 2016 0.1%
sulfur fuel regulation in ECA zone, and the 2016 International Marine Organization (IMO) Tier
3 NOX controls. Control and growth-related assumptions for 2025 came from the Emissions
Modeling Platform and are described in more detail in EPA (2014b). Non-U.S. and U.S.
category 3 commercial marine emissions were projected to 2025 using consistent methods that
incorporated controls based on ECA and IMO global NOX and SO2 controls.
All modeled 2011 and 2025 emissions cases use the 2006 Canada emissions data. Note
that 2006 is the latest year for which Canada had provided data at the time the modeling was
performed, and no accompanying future-year projected base case inventories were provided in a
form suitable for this analysis. For Mexico, 2012 and 2018 projections of the 1999 Mexico
National Emissions Inventory were used as described in the Development of Mexico National
Emissions Inventory Projections for 2008, 2012, and 2030 (ERG, 2009) and the associated
technical memorandum titled Mexico 2018 Emissions Projections for Point, Area, On-Road
Motor Vehicle and Nonroad Mobile Sources (ERG, 2009). Mexico emissions were held at 2018
levels because no 2025 projected emissions were available. Offshore oil platform emissions for
the United States represent the year 2008 because 2011 emissions were not available as of the
time of the modeling. Biogenic and fire emissions were held constant for all emissions cases and
were based on 2011-specific data. Table 3-1 shows the modeled 2011 and 2025 NOx and VOC
5-10
-------
emissions by sector. Additional details on the emissions by state are given in the Emissions
Modeling TSD.
Table 3-1. 2011 and 2025 Base Case NOX and VOC Emissions by Sector (thousand tons)
Sector
EGU-point
NonEGU-point
Point oil and gas
Wild and Prescribed Fires
Nonpoint oil and gas
Residential wood
combustion
Other nonpoint
Nonroad
Onroad
C3 Commercial marine
vessel (CMV)
Locomotive and C1/C2
CMV
Biogenics
TOTAL
2011NOx
1,948
1,768
17
347
653
36
832
1,630
5,592
125
1,046
1,018
15,012
2025 NOx
1,508
1,803
22
347
874
42
856
796
1,492
105
666
1,018
9,530
2011 VOC
33
872
88
5,175
2,273
447
3,793
2,025
2,738
5
48
40,696
58,192
2025 VOC
42
881
107
5,175
2,551
489
3,605
1,188
1,060
8
24
40,696
55,826
3.1.4 Emissions Sensitivity Simulations
A total of 12 emissions sensitivity runs were conducted to determine ozone response to
emissions reductions of NOX and VOC in different locations (Table 3-2). We determined that
this was an efficient and flexible approach that allowed us to evaluate impacts from multiple
source regions and levels of emission reductions simultaneously. All emissions sensitivity
simulations were incremental to the 2025 base case emissions described in section 3.1.3. There
were three types of emissions cases that were modeled in these sensitivity runs:
1) Explicit emissions control cases
2) Across-the-board reductions in anthropogenic emissions for different pollutants and
locations
3) Combination runs that included both explicit emissions controls and across the board
reductions.
Explicit Emissions Controls: Four explicit emissions control sensitivity cases were created.
First, we modeled a case that represented one possible implementation of the EPA's proposed
carbon pollution guidelines under section 11 l(d) of the Clean Air Act (CAA) (i.e., option 1 state;
5-11
-------
hereafter referred to as the 11 l(d) sensitivity). Emissions for this simulation are described in the
regulatory impact analysis for that proposed rule (EPA, 2014c). Second, we modeled three
additional emissions cases that included NOX emissions controls applied to specific sources
centered around the three regions of the country projected to have nonattainment monitors above
70 ppb in the 2025 base case: California, Texas, and the Northeastern U.S. Figure 3-2 shows the
three areas for which the explicit emissions controls were identified and modeled. CoST was
used to determine potential controls in these areas. NOX controls were identified for all nonpoint,
non-EGU point, and nonroad sources that emitted more than 50 tons of NOX per year and which
had available known controls that could be applied for less than $15,000/ton (see chapter 7 for
additional discussion). These emissions cases are referred to as "explicit control cases" because
they represent the impact of specific controls rather than sensitivities to all emissions within a
region. The extent of the area was determined by creating 200 km buffers around all monitors
projected above 70 ppb in 2025. All counties that fell completely within the buffer were targeted
for controls. In Texas and California these buffers were restricted to state boundaries. In the
Northeast, buffers were restricted to states/counties that are currently under the jurisdiction of the
Ozone Transport Commission (OTC). More details on the specific emissions controls identified
for these emissions cases are provided in Chapter 4.
5-12
-------
Legend
| Counties projected to exceed 70 ppb
Counties within 200 Kms of projected exceedance areas
0 200 400 800 Kilometers
I i i i I i i i I
Figure 3-2. Map of Counties for Which Explicit Emissions Controls Were Identified and
Modeled in CAMx (shaded in orange) and Counties that Contained One or
More Monitor Projected above 70 ppb in the 2025 Base Case Modeling (shaded
in blue).17
Across-the-board Emissions Reductions: Areas of the U.S. projected to contain monitors
with ozone design values greater than 60 ppb were split into 5 regions for the purpose of
determining ozone response to emissions reductions (Figure 3-3). Three emissions sensitivity
cases with across-the-board cuts in emissions from the 2025 base case were created and
modeled:
1) 50% cut in all anthropogenic NOX in the Southwest region,
2) 50% cut in all anthropogenic NOX in the Midwest region, and
17 Note that no buffer was created for the Sheboygan, WI area because emissions reductions from the proposed
carbon pollution guidelines under section 111 (d) of the CAA are projected to be sufficient to bring that location
down to 70 ppb in 2025.
$-13
-------
3) 50% cut in all U.S. anthropogenic VOC emissions across the 48 contiguous states.
Combination Emissions Sensitivities: Five additional emissions sensitivity cases were
created and modeled that combined the explicit emissions controls with across-the-board
reductions. For all combination emissions sensitivity cases, the area over which emissions
reductions were applied in the explicit emissions control runs was a subset of the full area for
which across-the-board NOX reductions were applied. These runs included two cases for
California: explicit emissions controls in California + additional 50% cut in all California
anthropogenic NOX emissions and explicit emissions controls in California + additional 90% cut
in all California anthropogenic NOX emissions. Two more emission sensitivities were
investigated for the Northeast region: explicit emissions controls in the Northeast + additional
50% cut in all Northeast region anthropogenic NOX emissions and explicit emissions controls in
the Northeast + additional 90% cut in all Northeast region anthropogenic NOX emissions. We
identified California and the Northeast as the two regions most likely to need NOX reductions
beyond 50% to reach one or more of the alternative standard levels considered based on a
previous EPA analysis (EPA, 2014d). Therefore both a 50% and a 90% NOX cut were performed
for each of these regions to better capture nonlinearities in ozone response to large NOx
emissions changes. Finally, a single emissions sensitivity was created for the Central region:
explicit emissions controls in Texas + additional 50% cut in all Central region anthropogenic
NOX emissions. A summary of Anthropogenic NOX and VOC emissions that were considered for
controls in this analysis is given in Table 3-3 by region from the 2025 base case and explicit
emissions control cases. In other words, the emissions summarized in Table 3 -3 only include
sectors for which emissions reductions were considered to meet various levels of the ozone
standard. Conversely, Table 3-1 summarizes all U.S. emissions that were included in the
modeling simulation including sources which contribute to background ozone.
5-14
-------
Figure 3-3. Five U.S. Regions Used to Create Across-the-Board Emissions Reduction and
Combination Cases
Table 3-2. List of Emissions Sensitivity Cases that Were Modeled in CAMx to Determine
Ozone Response Factors
Emissions
Sensitivity
Case
Region Pollutant
Emissions Change
1 National
2 National
3 California
4 California
5 California
6 Southwest
7 Texas
8 Central
9 Midwest
10 Northeast
1 1 Northeast
12 Northeast
All
voc
NOx
NOx
NOx
NOx
NOx
NOx
NOx
NOx
NOx
NOx
lll(d) option 1 state
50% VOC cut
CA explicit emissions control case
CA explicit emissions control case + 50% NOx cut
CA explicit emissions control case + 90% NOx cut
50% NOx cut
TX explicit emissions control case
TX explicit emissions control case +
50% NOx cut (central)
50% NOx cut
Northeast explicit emissions control case
Northeast explicit emissions control case +
50% NOx cut
Northeast explicit emissions control case +
90% NOx cut
5-15
-------
Table 3-3. Anthropogenic NOX and VOC Emissions from the 2025 Base and Explicit
Control Cases*
Region
Northeast
Midwest
Central
Southwest
California
Other states
Total contiguous US
NOx emissions in
2025 base case
(thousand tons)
1,185
1,771
2,176
713
446
1,840
8,130
NOx emissions in 2025
explicit control cases
(thousand tons)
1,074
2,073
—
416
—
VOC emissions in 2025
base case
(thousand tons)
1,345
1,803
3,066
1,016
478
2,264
9,971
*Note that unlike Table 3-1, these numbers do not include tribal, biogenic or fire emissions.
3.2 Methods for Calculating Current and Future Year Ozone Design Values
3.2.7 Current Year Ozone Design Value Calculations
As described in chapter 2, hourly ozone concentrations are used to calculate a statistic
referred to as a "design value" (DV) which is then compared to the standard level to determine
whether a monitor is above or below the NAAQS level in question. For ozone, the DV is
calculated as the 3-year average of the annual 4th highest daily maximum 8-hour ozone
concentration in parts per billion (ppb), with decimal digits truncated. For the purpose of this
analysis, the data handling and data completeness criteria used are those being proposed for the
new NAAQS in the proposed appendix U to 40 CFR Part 50 - Interpretation of the Primary and
Secondary National Ambient Air Quality Standards for ozone. A standard level of 60 ppb was
used when determining data completeness criteria as this results in the most inclusive set of
monitoring site DVs for analysis. For the purpose of this analysis, ozone DVs came from data
reported in EPA's air quality system (AQS) for the years 2009-2013. The current-year DVs
were calculated as the average of 3 consecutive DVs (2009-2011, 2010-2012, and 2011-2013)
which creates a 5-year weighted average DV. The 5-year weighted average DV is used as the
base from which to project a future year DV as is recommended by the EPA in its SIP modeling
guidance (US EPA, 2014e) because it stabilizes year-to-year meteorologically driven variability
in ozone DVs given that the future year meteorology is unknown. For cases in which there are
fewer than five years of valid monitoring data at a site, the current year DV was calculated only
when there was at least three years of consecutive valid data (i.e., at least one complete DV). If a
monitor had less than three consecutive years of data, then no current year D V was calculated for
that site and the monitor was not used in this analysis.
5-16
-------
3.2.2 Future Year Ozone Design Value Projections
Future year ozone design values were calculated at monitor locations using the Model
Attainment Test Software (MATS) program (Abt Associates, 2014). MATS calculates the 5-
year weighted average DV based on observed data and projects future year values using the
relative response predicted by the model as described below. Equation (3-1) describes the
recommended model attainment test in its simplest form, as applied for monitoring site /':
(DVF)j = (RRF)i X (DVB)i Equation 3-1
DVFj is the estimated design value for the future year in which attainment is required at
monitoring site /'; RRF; is the relative response factor at monitoring site /'; and DVB; is the base
design value monitored at site /'. The relative response factor for each monitoring site (RRF\ is
the fractional change of the DV in the vicinity of the monitor that is simulated on high ozone
days due to emissions changes between the base and future years. The recently released draft
version of EPA's ozone and PM2.5 photochemical modeling guidance (US EPA, 2014e) includes
updates to the recommended ozone attainment test used to calculate future year design values for
attainment demonstrations. The guidance recommends calculating RRFs based on the highest 10
modeled ozone days in the ozone season near each monitor location. Given the similar goal of
this analysis relative to an attainment demonstration, we are using the recommended modeling
guidance attainment test approach for the analyses. Specifically, the RRF was only calculated
based on the 10 highest days in the base year modeling at the monitor location when the base 8-
hr daily maximum ozone values were greater than or equal to 60 ppb for that day. In cases for
which the base model simulation did not have 10 days with ozone values greater than or equal to
60 ppb at a site, we used all days where ozone >= 60 ppb, as long as there are at least 5 days that
meet that criteria. At monitor locations with less than 5 days with ozone >= 60 ppb, no RRF or
DVF was calculated for the site and the monitor in question was not included in this analysis.
In determining the ozone RRF we considered model response in grid cells immediately
surrounding the monitoring site along with the grid cell in which the monitor is located, as is
currently recommended by the EPA in its SIP modeling guidance (US EPA, 2014e). The RRF
was based on a 3 x 3 array of 12 km grid cells centered on the location of the grid cell containing
5-17
-------
the monitor. The grid cell with the highest base ozone value in the 3x3 array was used for both
the base and future components of the RRF calculation.
3.3 Determining Tons of Emissions Reductions to Meet Various NAAQS Levels
The following section describes how projected ozone DVs from the 2025 base case and
12 emissions sensitivity cases were used to determine the expected emissions reductions needed
to attain the current and potential alternative ozone NAAQS. The scenario for which all U.S.
ozone monitors are projected to meet the current ozone NAAQS of 75 ppb is referred to as the
"2025 baseline" scenario. The costs and benefits for meeting 70, 65, and 60 ppb standards will
be determined incrementally from this baseline. Note that the 2025 baseline is different from the
2025 base case, which is the emissions scenario described in section 3.1.3 and represents the
ozone concentrations that are projected to occur in 2025 if there were no distinct reductions
made for the purpose of meeting the current or alternative ozone NAAQS.
3.3.1 Determining Ozone Response from Each Emissions Sensitivity
Section 3.2.2 describes, in general terms, how the 2025 projections for ozone DVs were
computed. This procedure was followed for the 2025 base case modeling and for each of the 12
emissions sensitivity cases. Using the projected DVs and corresponding emissions changes, a
unique ppb per ton response factor was calculated for each ozone monitor and for each emissions
sensitivity case based on equation 3-2:
DVij_DV2025basej
Rt; = —-— Equation 3-2
AŁ"j
In equation 3-2, Ry represents the response at monitor j to emissions changes in emissions
sensitivity case i, DVy represents the DV at monitor j in emissions sensitivity i, DV2025base,j
represents the DV at monitor j in the 2025 base case and AE; represents the difference in NOX or
VOC emissions (tons) between the 2025 base case and emissions sensitivity case i. In cases for
which emissions reductions in sensitivity i, were incremental to emissions reductions in another
case (k), the following equation was used:
Equation 3-3
5-18
-------
in which AEik represents the difference in NOX or VOC emissions (tons) between the emissions
case k and emission case i. Thus at each monitoring site in regions with multiple emissions
sensitivity cases, we determined a set of incremental DV responses per ton of emissions
reductions. The modeled impacts from the individual cases were then combined in a linear
manner to estimate the net impacts from multiple cases. For example, in the Northeast, we would
use the following equation to determine the DVs that would result from a 75% reduction in
Northeast emissions beyond the explicit emissions control case:
+ (,^NE_explicitcontrol,j X ^•N
+ (RNE90Noxj X O AŁ9owox) Equation 3-4
In equation 3-4, hENE_expUcitcontroi represents the difference in NOX emissions between
the 2025 base case and the 2025 Northeast explicit emissions control case, &E50NOx represents
the difference in NOX emissions between the 2025 Northeast explicit emissions control case and
the combined Northeast explicit control case with 50% Northeast NOx cuts and &E90NOX
represents the difference in NOX emissions between the combined Northeast explicit control case
with 50% Northeast NOx cuts and the combined Northeast explicit control case with 90%
25
Northeast NOX cuts. The — multiplier represents the ratio of required to modeled emissions (i.e.,
the difference between the 50% and 90% NOX cut emissions sensitivities represent emissions
equivalent to 40% of the Northeast explicit emissions control case, while in the example above,
we only require an addition 25% emission reduction beyond the 50% NOX cut simulation).
In two cases, we determined it was appropriate to compute response factors for smaller
geographic areas than were modeled in the emissions sensitivity simulations described in section
3.1.4. One of the cases pertains to splitting the responses between different air basins in
California and the other case involves the geographic scale of ozone impacts associated with
emissions reductions of VOC. Both of these case are described below.
In California, 2025 base case ozone DVs were substantially higher in the South Coast Air
Basin located in the southern portion of the state than in the San Joaquin Valley and in areas
further north. Additionally, the Transverse Mountain Ranges in Southern California generally
isolate the air masses in the South Coast Air Basin from those in the San Joaquin Valley.
5-19
-------
Consequently, it is unrealistic to force emissions reductions in locations in Northern California to
bring Southern California ozone DVs into attainment with the current or alternative levels of the
NAAQS. Therefore, when applying the results of the 50% and 90% California NOX emissions
reduction sensitivities, we made a simplifying assumption that the ozone responses predicted in
the San Joaquin Valley and areas of California further north are solely due to emissions changes
in those areas. We made a similar assumption about the response of ozone to emissions changes
for the southern portion of California. Using this approach we created distinct response factors
based on changes in ozone DVs and emissions from each of the two California sub-regions. In
general, this assumption seems reasonable based on the topography and wind patterns in these
areas and the mountain ranges that separate Los Angeles from the San Joaquin Valley. This
approach may lead to either some underestimation or overestimation of the air quality impacts of
emissions reductions in the Central and Northern California locations depending on the extent to
which emissions reductions in Southern California actually impact ozone in the San Joaquin
Valley or vice versa. A more complete description of how these areas were delineated and the
rational is provided in Appendix 3 A.
We followed a conceptually similar approach for geographically allocating the response of
ozone DVs to the 50% reduction in US anthropogenic VOC emissions. Past work has shown that
impacts of anthropogenic VOC emissions on ozone DVs in the U.S. tend to be localized (Jin et
al, 2008; Nopmongcol et al, 2014) and so consistent with past analyses (US EPA, 2008) we
have made the assumption that VOC reductions do not impact ozone at distances more than
100km from the emissions source. Consequently, we created a series of VOC impact regions for
urban areas in which ozone is responsive to VOC emissions reductions. These VOC impact
regions were only created for the urban areas with the highest projected 2025 base case ozone
DVs in each region: New York City, Pittsburgh, and Baltimore in the Northeast; Detroit,
Chicago, and Louisville in the Midwest; Houston and Dallas in the Central region; Denver in the
Southwest; and Northern and Southern California. VOC impact regions were delineated by
creating a 100km buffer around counties containing violating monitors. Since the counties with
5-20
-------
violating monitors differed at each standard level, a separate set of VOC impact regions was
developed for standard levels of 70, 65, and 60 ppb.18
In addition, VOC impact regions were constrained by state boundaries except in cases
where a current nonattainment area straddled multiple states (for instance New Jersey and
Connecticut counties that are included in the New York City nonattainment area were also
included in the New York City VOC impact region). The in-state constraint was also waived for
the Chicago area since it is well established that emissions from Chicago and Milwaukee are
often advected over Lake Michigan where they photochemically react and then impact locations
in Wisconsin, Illinois, Indiana, and Michigan that border the lake (Dye et al, 1995). In cases
where a county fell within two adjacent overlapping buffer areas (i.e. Easton County which lies
in both the greater Chicago and Detroit buffers) the county was assigned to the VOC impact area
that is most likely to be upwind based on prevailing wind patterns (i.e. Detroit). Finally, for
California, the VOC impact areas were delineated identically to the Northern and Southern NOX
sub-regions described above but were restricted to counties included in the explicit NOX control
cases. For each monitoring site within a VOC impact area, an ozone DV response factor (Ry)
was calculated using the VOC emissions reductions that occurred within that area based on the
U.S. 50% VOC sensitivity simulation. Figure 3-4 shows the VOC impact areas that were
developed for the 60 ppb standard. Maps for the 65 and 70 ppb VOC impact areas look very
similar to the map in Figure 3-4, but in some cases they include fewer counties around the
outside of the region.
18
The 70 ppb VOC impact areas were also used to construct the baseline (75 ppb) scenario.
5-21
-------
Legend
New York, New Jersey. Long Island NY-NJ-CT
| Baltimore- MD
| Pittsburgh. Beaver Valley. PA
| Louisville KY
Chicago-Lake Michigan, Wl-IL-IN-MI
Detroit. Ml
| Houston-Galveston-Brazoria.TX
Dallas-Ft.Worth. TX
| Denver-Boulder-Greeley-FtCollins-Loveland. CO
Los Angeles. CA
| San Joaquin. CA
Figure 3-4. Map of VOC Impact Areas Applied in the Evaluation of a 60 ppb Alternative
Standard Level
3.3.2 Combining Response from Multiple Sensitivity Runs To Construct Baseline And
Alternative Standard Scenarios
Ozone DVs were calculated for the baseline scenario as well as the proposed range of 65-
70ppb and a more stringent alternative standard of 60ppb by applying response factors described
in section 3.3.1 and the emissions reductions from multiple modeled sensitivity scenarios using
Equation 3-5:.
= DV
202SJ
X
(R2J X
(R3
3J
X
Equation 3-5
$-22
-------
For the baseline as well as the three alternative standards analyzed, we determine the least
amount of emissions reductions (tons) needed to bring the ozone DVs at all monitors down to the
particular NAAQS level. Note that the emissions reductions were applied on a region-specific
basis for most monitors. That is, given the construct of the analytic approach, we did not account
for the co-benefits of inter-regional transport except for monitors that are located near the border
to two regions. For instance, to determine the requisite tons of NOX and VOC reductions
necessary to bring monitors in the Central region down to 65 ppb, only emissions in the central
region were considered (i.e., emissions from the Texas explicit emissions control case, the 50%
Central region NOX reductions sensitivity case, and the VOC emissions from the Houston and
Dallas VOC impact areas). This constraint was applied with the assumption that most states
would not take into account emissions reductions from upwind states when designing their State
Implementation Plans but also acknowledging that there has been a history of some states
cooperating with neighboring states in the same region to determine multi-state pollution
reduction plans.19 This approach avoids the complexity of determining the order in which
emissions reductions should be considered across the multiple regions modeled for this analysis.
The result of this constraint is that the amount of emissions reductions estimated for attainment
in this analysis may be larger than necessary because downwind states may benefit from upwind
reductions that are not accounted for when determining the regional emissions reduction needed
to attain. There were two cases where emissions reductions from two regions were applied to
given monitors in a border area: monitors in the Illinois suburbs of St. Louis (Midwest Region)
that are clearly affected by emissions in neighboring Missouri (Central Region), and monitors in
Pittsburgh, PA and in Buffalo, NY (Northeast Region) that are substantially impacted by
emissions in Ohio (Midwest Region).
There are several assumptions inherent in this methodology. First, when applying
responses from the across-the-board emissions reduction sensitivities we do not have any
information about how ozone DVs respond differently to emissions from different locations
within the region. Therefore, for the most part, we assume that every ton of NOX or VOC
19 For instance the Ozone Transport Commission (OTC) is an organization made up of states in the eastern U.S.
which is responsible for "developing and implementing regional solutions to the ground-level ozone problem in the
Northeast and Mid-Atlantic regions" (www.otcair.org)
5-23
-------
reduced within the region (or VOC impact area) results in the same ozone response regardless of
where the emissions reductions are identified. In locations for which we have both explicit
emissions control case reductions and regional across-the-board reductions, it is possible to make
some more distinctions as described below, but we are still not able to fully account for variable
response to emissions from different locations within the region. Where possible, we try to
locate emissions reductions closer to the highest DV monitors with the understanding that
emissions reductions are likely to have lower impact when they occur further from the monitor
location. A second assumption is that NOX and VOC responses are additive. In the case of
monitors impacted by emissions from multiple regions, we also assume that the responses from
multiple regions are additive. Again, we do not have any more refined information that would
allow us to account for nonlinear interactions of these emissions. Third, we assume that ozone
response within each of these sensitivity simulations is linear (i.e., the first ton of NOX reduced
results in the same ozone response as the last ton of NOX reduced). In cases for which we have
multiple levels of emissions reductions (i.e., California, the Central region, and the Northeast) we
assume linearity within each simulation but are able to capture discrete shifts in ozone response
for each sensitivity simulation (i.e., one response for explicit emissions control case reductions,
another response level up to 50% NOX reductions beyond the explicit emissions control case
emissions, and a third level of response between 50% and 90% NOX reductions beyond the
explicit emissions control case). For the Central states there are only two discrete response
levels, while in California and the Northeast there are three. Finally, for regions without 90%
NOx cut emissions sensitivity scenarios (Southwest, Central, and Midwest), response to NOX
reductions greater than 50% must be extrapolated beyond the modeled emissions reductions.
3.3.3 Creation of the Baseline Scenario
Computing the response of DVs to emissions reductions from each emissions sensitivity
simulation allowed us to determine what emissions reductions would be needed in each region to
create the baseline scenario (i.e. to reach 75 ppb at every monitor location). We determine how
those emissions reductions could be achieved by applying controls in the following order: (1)
emissions changes from the 11 l(d) sensitivity, (2) known controls of NOX emissions from
nonpoint, non-EGU point, and nonroad sources greater than 50 tons per year (explicit control
cases), (3) mobile source emissions changes between 2025 and 2030 (California only), (4)
5-24
-------
known controls of VOC emissions, and (5) additional NOX controls. All reductions identified
from these sources were above and beyond reductions from on-the-books regulations that were
included in the 2025 base case modeling. The emissions changes from the 11 l(d) sensitivity
were applied throughout the entire U.S. in creating the baseline scenario. Other emissions
changes were only applied in the areas projected to have DVs greater than 75 ppb in the 2025
base case scenario: California and Texas. In California, all five types of emissions reductions
were needed to meet the current 75 ppb standard while in Texas only the 11 l(d) emissions
changes and a portion of the explicit modeled controls were needed. The 2025 to 2030 mobile
source changes were applied in California because, as discussed at the beginning of this chapter,
many locations in California will likely have attainment dates substantially further out than 2025.
Although emissions projections for those years were not generally available, the state of
California did provide emissions projections for mobile sources in the year 2030. Emissions of
both VOC and NOX were available for onroad, nonroad, locomotive, and C1/C2 commercial
marine vessel sectors by county. There were both increases and decreases between 2025 and
2030 depending on the county and sector, but overall these mobile source changes resulted in
VOC emissions that were 1% less than those modeled in the California explicit emissions control
case and NOX emissions that were 4% less than those modeled in the California explicit
emissions control case in the Northern California sub-region and 3% less than those modeled in
the Southern California sub-region. The NOX and VOC mobile source emissions changes were
applied to create the baseline scenario in California using the response ratios developed from the
50% California NOx cut and the 50% U.S. VOC cut sensitivity simulations. Summaries of the
emissions reductions are presented by region in Appendix 3 A. In addition, resulting ozone DVs
at all evaluated monitors are provided in Appendix 3 A.
3.3.4 Creation of the 70, 65, and 60 ppb Alternative Standard Level Scenarios
To create the scenarios for the three alternative standard levels (i.e. 70, 65, and 60 ppb), we
started with the baseline and then identified additional controls for each region from the five
categories listed in section 3.3.3. Not all types of emissions reductions were required in each
region for each scenario. For regions that contained a NOX explicit emissions control case buffer,
only the known controls within the buffer were applied before the known VOC controls. In
those regions, after explicit emissions control case reductions and the VOC known controls were
5-25
-------
applied, then additional NOX controls were considered. In those regions, an additional constraint
was also applied that forced the tons of additional NOX reductions applied within and outside the
explicit emissions control case area to be applied proportionally to the starting NOX emissions
within and outside the explicit emissions control case buffer areas (e.g., if 40% of the starting
emissions in the explicit control scenario simulation were located within the buffer, then 40% of
the emissions reductions also had to come from within the buffer). This constraint was applied
because the monitors with the highest DVs were located within the buffers and the response
factors were based on an average ppb/ton across the region. Thus, the constraint ensured some
measure of spatial equivalence between the location of the modeled emissions reductions and
those applied to construct the scenario. In some cases, this constraint also resulted in including
unknown controls to create the scenario even when known controls were still available within the
region but outside of the explicit control case buffer. In regions without an explicit emissions
control case buffer area, all known controls of NOX emissions from nonpoint, non-EGU point,
and nonroad sources greater than 50 tons per year were applied throughout the entire region
before any VOC emissions reductions were applied. A numeric example of the calculation
methodology is provided in Appendix 3 A. Summaries of the emissions reductions are presented
by region in Appendix 3 A and by source category in Chapter 4. In addition, ozone DVs at all
evaluated monitors are provided for each scenario in Appendix 3 A.
3.3.5 Monitoring Sites Excluded from Quantitative Analysis
There were 1219 ozone monitors with complete ozone data for at least one DV period
covering the years 2009-2013. Of those sites, we quantitatively analyzed 1150 (94%) in this
analysis. In determining the necessary tons of emissions reductions for each of the four
scenarios, there were three types of sites that were not treated quantitatively, i.e. emissions
reductions necessary to reach the alternative standard levels at these sites were not quantified.
First, tons of emissions reductions were not determined for 36 sites that did not have a valid
projected 2025 base case DV due to less than 5 modeled days above 60 ppb in the 2011 CAMx
simulation as required to project a DV in the EPA SIP modeling guidance (US EPA, 2014e). It
is unlikely that these sites would have any substantial impact on resulting costs and benefits,
since the reason that projections could not be made is that they have no more than 4 modeled
5-26
-------
days above 60 ppb, in which case they would likely already be meeting all standard levels
evaluated in this analysis using the current year data. These sites are listed in appendix 3 A.
Second, 7 sites for which the DVs were influenced by wintertime ozone episodes were
not included because the modeling tools are not currently sufficient to properly characterize
ozone formation during wintertime ozone episodes. It is not appropriate to apply the model-
based response (RRF) developed based on summertime conditions to a wintertime ozone event,
which is driven by different types of chemistry and meteorology. Since there was no technically
feasible method for projecting DVs at these sites, these sites were not included in determining
required reductions in NOX and VOCs to meet current or alternative standard levels. Wintertime
ozone events tend to be very localized phenomena driven by local emissions from oil and gas
operations (Schnell et al, 2009; Rappengluck et al, 2014; Helmig et al, 2014). Consequently,
the emissions reductions needed to lower wintertime ozone levels would likely be different from
those targeted for summertime ozone events. It follows that there could be additional emissions
reductions required to lower ozone at these locations and thus potential additional costs and
benefits that are not quantified in this analysis. Appendix 3 A includes a list of sites influenced
by wintertime ozone and the methodology used to identify those sites.
Finally, while the majority of the sites had projected ozone exceedances primarily caused
by local and regional emissions, there were a set of 26 relatively remote, rural sites in the
Western U.S. with projected 2025 base case DVs between 62 and 69 ppb20 that showed limited
response to the regional NOX emission and national VOC emission sensitivities in our modeling.
Air agencies responsible for these locations may choose to pursue one or more of the Clean Air
Act provisions that offer varying degrees of regulatory relief. Regulatory relief may include:
• Relief from designation as a nonattainment area (through exclusion of data affected by
exceptional events)
• Relief from the more stringent requirements of higher nonattainment area classifications
(through treatment as a rural transport area; through exclusion of data affected by
exceptional events; or through international transport provisions)
20 Except in California where sites had projected 2025 base case DVs up to 75 ppb. The California sites all had
estimated ozone DVs below 70 ppb in the post-2025 baseline scenario.
5-27
-------
• Relief from adopting more than reasonable controls to demonstrate attainment (through
international transport provisions)
In addition, some of these sites could potentially benefit from the CAA's interstate transport
provisions found in sections 110(a)(2)(D) and 126. Appendix 3 A provides additional detail on
the treatment of these sites.
3.4 Creating Spatial Surfaces
The emissions reductions for attainment of the alternative NAAQS levels were used to
create spatial fields of ozone concentrations (i.e., spatial surfaces) for input to the calculation of
the benefits associated with attainment of each NAAQS level, incremental to the baseline. The
spatial surfaces used to calculate health-related benefits with the BenMap tool (Chapter 5) are
described below in section 3.4.1. Spatial surfaces used to calculate welfare-related benefits with
the FASOMGHG model (Chapter 6) are described below in section 3.4.2.
3.4.1 BenMap Surfaces
Two ozone metrics are used to evaluate health benefits associated with meeting different
ozone standard levels. These metrics and the studies that they are derived from are described in
more detail in Chapter 5. Briefly, ozone surfaces for the baseline and each alternative NAAQS
level were created for the following metrics: May-Sep seasonal mean of 8-hr daily maximum
ozone and Apr-Sep seasonal mean of 1-hr daily maximum ozone. For each metric, surfaces were
created for a total of 11 scenarios. These scenarios include:
• 2025 baseline
• post-2025 baseline
• 2025 70 ppb partial attainment
• 2025 70 ppb full attainment
• post-2025 70 ppb full attainment
• 2025 65 ppb partial attainment
• 2025 65 ppb full attainment
• post-2025 65 ppb full attainment
• 2025 60 ppb partial attainment
• 2025 60 ppb full attainment
5-28
-------
• post-2025 60 ppb full attainment
The surfaces created for the 2025 scenarios represent all continental U.S. monitors outside
of California attaining the standard being evaluated while the surfaces for the post-2025
scenarios represent all continental U.S. monitors including California meeting the standard being
evaluated. The effects due only to California meeting the standard are isolated in Chapter 5
through a series of BenMap simulations using these surfaces and varying assumptions about
population demographics. In addition, for the 2025 scenarios we include "partial" and "full"
attainment in which the partial attainment scenarios only include emissions reductions identified
from known control measures while the full attainment scenarios include emissions reductions
necessary to attain the standard from both known and unknown controls.
The ozone surfaces were created using the following steps. Each step is described in more
detail below:
• Step 1: Aggregate gridded hourly modeled concentrations into relevant seasonal
ozone metrics
o Inputs: Hourly gridded model concentrations for 2011, 2025 base case, and
12 2025 emissions sensitivity simulations detailed in Section 3.1.4
o Outputs: Seasonal ozone metrics for 2011, 2025 base case, and 12 2025
emissions sensitivity simulations
• Step 2: Calculate response factors for each seasonal ozone metric from each
emissions sensitivity simulation
o Inputs: Seasonal ozone metrics for 2011, 2025 base case, and 12 2025
emissions sensitivity simulations; Amount of emissions reductions (tons)
modeled in each emissions case
o Outputs: Gridded ppb/ton response factor for each seasonal ozone metric
from each emissions sensitivity simulation
• Step 3: Create gridded field for each attainment scenario and each seasonal ozone
metric
o Inputs: Gridded ppb/ton response factor for each seasonal ozone metric
from each emissions sensitivity simulation; Amount of emissions reductions
from each region described in Appendix 3 A.
o Outputs: Gridded seasonal ozone metrics for each attainment scenario
• Step 4: Create 2011 enhanced Voronoi Neighbor Averaging (eVNA) fused surface
of modeled and observed values for each seasonal ozone metric
5-29
-------
o Inputs: 2010-2012 observed ozone values (seasonal ozone metrics at each
monitor location); 2011 modeled ozone (seasonal ozone metrics at each grid
cell)
o Outputs: 2011 fused modeled/monitored surfaces for each seasonal ozone
metric
Step 5: Create eVNA fused modeled/monitored surface for each attainment
scenario and each seasonal ozone metric
o Inputs: 2011 fused model/obs surfaces for each seasonal ozone metric;
modeled seasonal ozone metrics (gridded fields) for 2011 and each
attainment scenario
o Outputs: Fused modeled/monitored surface for each attainment scenario and
each seasonal ozone metric
Step 1:
Gridded hourly ozone modeled concentrations were aggregated to the relevant metric for
the 201 1, 2025 base case, and each of the 12 emissions sensitivity simulations. This step
resulted in 15 ozone fields for each of the two metrics.
Step 2:
A gridded ppb/ton response factor was determined for each metric and for each emissions
sensitivity simulation.
Step 3:
Based on the emissions reductions provided in appendix 3 A, the response factors were
multiplied by the relevant tons of emissions reductions for each sensitivity and then summed to
create a gridded field representing the scenario in question (Equation 3-6)21
03^2025,771 + R X
Equation 3-6
21 An extra 3,500 tons of VOC reductions available in Northern California outside of the N California sub-region
was mistakenly applied in creating all surfaces. This lead to absolute changes in gridded ozone concentrations of
less than 0.01 ppb. Since this error was carried through all surfaces, the incremental changes in ozone between
surfaces were not impacted.
-------
In equation 3-6, ozonexy,s,m represents the ozone concentrations at grid cell x,y, for
scenario s, and using metric, m. Similarly ozonexy,2025,m represents the modeled ozone from the
2025 base simulation at grid cell x,y aggregated to metric m. Rxy,i,m represents the response
factor (ppb/ton) in grid cell x,y using metric m, for the sensitivity simulation #1. Finally AEi,s
represents the amount of emissions reductions from sources modeled in sensitivity #1 determined
necessary for scenario s. Partial attainment surfaces at each standard level were created by first
starting with the full attainment surface and then subtracting off impacts from emissions
reductions that were identified from unknown controls. For the 70 ppb scenario in which all
unknown controls were located in the explicit emissions control case buffer areas, the ppb/ton
response ratios from the explicit emissions control case sensitivity simulations were applied to
back out impacts from unknown controls. For the 65 and 60 ppb scenarios, unknown controls
were located both within and outside explicit emissions control case buffer areas. We did not
have any response ratios that represent emissions from outside the explicit emissions control case
buffer areas alone. Therefore, for the 65 and 60 ppb scenarios, we applied the response ratios
from the regional 50% NOX reduction sensitivity simulations. This leads to additional
uncertainty in the 65 and 60 pp partial attainment surfaces since the relative proportions of
unknown emissions reductions within and outside the buffer areas were different from the
relative proportions of emissions reductions within and outside of the buffer areas that were
applied in the 50% regional NOX reduction simulations.
Step 4:
The MATS tool was used to create a fused gridded 2011 field using both ambient and
modeled data using the eVNA technique (Abt, 2014). This method essentially takes an
interpolated field of observed data and adjusts it up or down based on the modeled spatial
gradients. For this purpose, the 2010-2012 ambient data was interpolated and fused with the
2011 model data. One "fused" eVNA surface was created for each of the two seasonal ozone
metrics.
Step 5:
-------
The 22 model-based surfaces (i.e., 11 scenarios and 2 metrics) were used as inputs in the
MATS tool along with the gridded 2011 eVNA surfaces. For each metric and each scenario a
gridded RRF field was created by dividing the gridded ozone field for scenario s by the gridded
2011 model field. This RRF field was then multiplied by the 2011 eVNA field to create a
gridded eVNA field for each scenario.
Results of this process for the May-September 8-hr daily maximum ozone metric are
shown in Figures 3-5, 3-6, 3-7, 3-8. These figures show the post-2025 baseline and the changes
in ozone between the post-2025 baseline and each of the post-2025 scenarios for lower standard
levels: 70, 65, and 60 ppb. The post-2025 baseline represents the case where all continental US
monitors meet the current 75 ppb standard and similarly the post-2025 alternative standard
scenarios represent the case where all continental US monitors meet the standard level being
evaluated.
0
10 20 30 40 50
60
Figure 3-5. Projected post-2025 Baseline Scenario May-September Mean of 8-hr Daily
Maximum Ozone (ppb)
-------
-10 -8
Figure 3-6. Change in May-September Mean of 8-hr daily Maximum Ozone (ppb)
between the post-2025 Baseline Scenario and the post-2025 70 ppb Scenario
-------
in
LTJ
en
CD
in
Figure 3-7. Change in May-September Mean of 8-hr daily Maximum Ozone (ppb)
between the post-2025 Baseline Scenario and the post-2025 65 ppb Scenario
-------
-10 -8
-2
Figure 3-8. Change in May-September Mean of 8-hr Daily Maximum Ozone (ppb)
between the post-2025 Baseline Scenario and the post-2025 60 ppb Scenario
3.4.2 W126 surfaces
This section describes the creation of ozone surfaces aggregated using the W126 metric.
The general methodology for calculating the W126 metric is provided in appendix 3A. Ozone
surfaces aggregated to the W126 metric were created for eight scenarios:
• 2025 baseline
• post-2025 baseline
• 2025 70 ppb full attainment
• post-2025 70 ppb
-------
• 2025 65 ppb full attainment
• post-2025 65 ppb
• 2025 60 ppb full attainment
• post-2025 60 ppb
Several steps were followed to create these surfaces. First, as was done with the projected
ozone DVs and the ozone surfaces for health benefits, ppb/ton response factors was determined
for each sensitivity simulation. In this case, the response factors were created based on hourly
ozone data in the 2025 base case and 12 emissions sensitivity simulations. Therefore, six months
of gridded hourly response factors were created for each emission sensitivity simulation. Then,
based on the emissions reductions described in Appendix 3 A, the hourly response factors were
multiplied by the relevant tons of emissions reductions from each emissions sensitivity and then
summed to create a gridded field representing the scenario in question (Equation 3 -6). For the
W126 calculations, the metric in equation 3-6 is hourly ozone. At the end of this step, there were
eight sets of hourly gridded ozone fields, one for each scenario. These gridded hourly ozone
fields, along with the 2011 modeled hourly ozone field, were then aggregated into the W126
metric, which is described in more detail in Chapter 2. The MATS tool was used to project
W126 values at each monitor location. The set-up was similar to the approach for projecting
DVs. In essence the 2011 and eight W126 scenarios gridded fields were used to create an RRF
for each monitor location using the 3x3 matrix of grid cells surrounding the monitor location as
described in section 3.2.2. These model-based RRFs were multiplied by the three year average
(2010-2012) of the measured W126 at each monitor. At the end of this step, there were a set of
W126 values at all monitor locations for each scenario. Finally, a gridded field of W126 values
was created for each scenario by spatially interpreting the projected monitor values using an
inverse distance weighted Voronoi Neighbor Averaging (VNA) technique (Gold, 1997; Chen et
al, 2004). This is similar to how W126 gridded fields have been created for previous EPA
analyses (EPA, 2014f). Figure 3-9 shows the W126 gridded field for the post-2025 baseline
scenarios. Figures 3-10, 3-11, and 3-12 show the W126 gridded field of the post-2025 scenarios
for 70, 65, and 60.
-------
in
3
3
9
10
CM
0
Figure 3-9. Projected post-2025 Baseline Scenario W126 Values (ppm-hrs)
-------
in
3
3
9
10
CM
0
Figure 3-10. Projected post-2025 70 ppb Scenario W126 Values (ppm-hrs)
-------
0
20
Figure 3-11. Projected post-2025 65 ppb Scenario W126 Values (ppm-hrs)
-------
0
20
Figure 3-12. Projected post-2025 60 ppb Scenario W126 Values (ppm-hrs)
3.5 References
Abt Associates, 2014. User's Guide: Modeled Attainment Test Software.
http://www.epa.gov/scram001/modelingapps_mats.htm
Chen, I, Zhao, R., Li, Z. (2004). Voronoi-based k-order neighbor relations for spatial analysis. ISPRS J
Photogrammetry Remote Sensing, 59(1-2), 60-72.
Dennison, P.E., Brewer, S.C., Arnold, J.D., Moritz, M. A. (2014) Large wildfire trends in the western United States,
1984-2011. Geophysical Research Letters, 41, 2928-2933.
Dye T.S., Roberts P.T., Korc, M.E. (1995) Observations of transport processes for ozone and ozone precursors
during the 1991 Lake Michigan ozone study, J. Appl. Met., 34 (8), 1877-1889
ENVIRON, 2014. User's Guide Comprehensive Air Quality Model with Extensions version 6.1, www.camx.com.
ENVIRON International Corporation, Novato, CA.
ERG, 2009. Development of'Mexico National Emissions Inventory Projections for 2008, 2012, and 2030. Available
at: http://www.azdeq.gov/environ/air/plan/download/dmneip.pdf.
3-i
-------
Gold, C. (1997). Voronoi methods in GIS. In: Algorithmic Foundation of Geographic Information Systems (va
Kereveld M, Nievergelt, I, Roos, T., Widmayer, P., eds). Lecture Notes in Computer Science, Vol 1340. Berlin:
Springer-Verlag, 21-35.
Helmig, D., Thompson, C.R., Evans, I, Boylan, P., Hueber, I, Park, J.H. (2014). Highly elevated atmospheric
levels of volatile organic compounds in the Uintah Basin, Utah, Environmental Science & Technology, 48, 4707-
4715.
Henderson, B.H., Akhtar, F., Pye, H.O.T., Napelenok, S.L., Hutzell, W.T. (2014). A database and tool for boundary
conditions for regional air quality modeling: description and evaluations, Geoscientific Model Development, 7,
339-360.
Houyoux, M.R., Vukovich, J.M., Coats, C.J., Wheeler, N.J.M., Kasibhatla, P.S. (2000), Emissions inventory
development and processing for the Seasonal Model for Regional Air Quality (SMRAQ) project, Journal of
Geophysical Research-Atmospheres, 105(D7), 9079-9090.
Jin, L., Tonse, S., Cohan, D. S., Mao, X. L., Harley, R. A., Brown, N. J. (2008) Sensitivity analysis of ozone
formation and transport for a central California air pollution episode. Environ. Sci. Technol., 42, 3683-3689.
Nopmongcol, U., Emery, C., Sakulyanontvittaya, T., Jung, J., Knipping, E., Yarwood, G. (2014) A modeling
analysis of alternative primary and secondary US ozone standards in urban and rural areas, Atmospheric
Environment, Available online 28 September 2014, ISSN 1352-2310,
http://dx.doi.0rg/10.1016/j.atmosenv.2014.09.062.
Rappengluck, B., Ackermann, L., Alvarez, S., Golovko, J., Buhr, M., Field, R.A., Soltis, J., Montague, D.C., Hauze,
B., Adamson, S., Risch, D., Wilkerson, G., Bush, D., Stoeckenius, T., Keslar, C. (2014). Strong wintertime ozone
events in the Upper Green River basin, Wyoming, Atmospheric Chemistry and Physics, 14, 4909-4934.
Ruiz, L.H., Yarwood, G., 2013. Interactions between Organic Aerosol and NOy: Influence on Oxidant Production.
http://aqrp.ceer.utexas.edu/projectinfoFY12_13/12-012/12-012%20Final%20Report.pdf.
Schnell, R.C., Oltmans, S.J., Neely, R.R., Endres, M.S., Molenar, J.V., White, A.B. (2009) Rapid photochemical
production of ozone at high concentrations in a rural site during winter, Nature Geoscience, 2, 120-122.
Simon, H., Baker, K.R., Phillips, S. (2012) Compilation and interpretation of photochemical model performance
statistics published between 2006 and 2012, Atmospheric Environment, 61, 124-139.
U.S. Environmental Protection Agency (2008) Final ozone NAAQS regulatory impact analysis, US EPA, OAQPS,
RTF, NC, EPA-452/R-08-003
U.S. Environmental Protection Agency (2012) Regulatory Impact Analysis for the Final Revision to the National
Ambient Air Quality Standards for Paniculate Matter, US EPA, OAQPS, 2012, RTF, NC, EPA-452/R-12-005.
http://www.epa.gov/ttnecasl/regdata/RIAs/fmalria.pdf
U.S. Environmental Protection Agency (2014a) Meteorological Model Performance for Annual 2011 Simulation
WRF v3.4, US EPA, OAQPS, 2014, RTF, NC http://www.epa.gov/scram001/
U.S. Environmental Protection Agency (2014b) Preparation of Emissions Inventories for the Version 6.1, 2011
Emissions Modeling Platform (http://www.epa.gov/ttn/chief/emch)
U.S. Environmental Protection Agency (2014c) Regulatory Impact Analysis for the Proposed Carbon Pollution
Guidelines for Existing Power Plants and Emission Standards for Modified and Reconstructed Power Plants, US
EPA, OAQPS, 2014, RTF, NC, EPA-542/R-14-002
5-41
-------
U.S. Environmental Protection Agency (2014d) Health Risk and Exposure Assessment for Ozone, final report. US
EPA, OAQPS, 2014, RTF, NC, EPA-452/R-14-004a
U.S. Environmental Protection Agency (2014e) Draft Modeling Guidance for demonstrating attainment of air
quality goals for ozone, PM2.5, and regional haze. September 2014, U.S. Environmental Protection Agency,
Research Triangle Park, NC, 27711.
U.S. Environmental Protection Agency (2014f) Welfare Risk and Exposure Assessment for Ozone, final report. US
EPA, OAQPS, 2014, RTF, NC, EPA-452/R-14-005a.
5-42
-------
APPENDIX 3: ADDITIONAL AIR QUALITY ANALYSIS AND RESULTS
3A.1 2011 Model Evaluation for Ozone
An operational model evaluation was conducted for the 2011 base year CAMx annual
model simulation performed for the 12-km U.S. modeling domain.22 The purpose of this
evaluation was to examine the ability of the Ozone NAAQS RIA air quality modeling platform
to replicate the magnitude and spatial and temporal variability of measured (i.e., observed) ozone
concentrations within the modeling domain. The model evaluation for ozone was based upon
comparisons of model predicted 8-hour daily maximum concentrations to the corresponding
observed data at monitoring sites in the EPA Air Quality System (AQS) and the Clean Air Status
and Trends Network (CASTNet). Included in the evaluation are statistical measures of model
performance based upon model-predicted versus observed concentrations that were paired in
space and time on an hourly basis.
Model performance statistics were calculated for several spatial scales and temporal
periods. Statistics were calculated for individual monitoring sites and for each of five regions of
the 12-km U.S. modeling domain. The regions include the Northeast, Midwest, Southeast, and
Central and Western states which are defined based upon the states contained within the
Regional Planning Organizations (RPOs)23. For maximum daily average 8-hour (MDA8) ozone,
the statistics for each site and region were calculated for the May through September ozone
season.24 In addition to the performance statistics, we prepared several graphical presentations of
22 See Chapter 3, section 3.1.2 of the RIA document for a description of the 12-km U.S. modeling domain.
23 The subregions are defined by States where: Midwest is IL, IN, MI, OH, and WI; Northeast is CT, DE,
MA, MD, ME, NH, NJ, NY, PA, RI, and VT; Southeast is AL, FL, GA, KY, MS, NC, SC, TN, VA, and
WV; Central is AR, IA, KS, LA, MN, MO, ME, OK, and TX; West is AK, CA, OR, WA, AZ, MM, CO, UT, WY,
SD, ND, MT, ID, and NV.
24 In calculating the ozone season statistics we limited the data to those observed and predicted pairs with
observations that exceeded 60 ppb in order to focus on concentrations at the upper portion of the distribution of
values.
5A-1
-------
model performance for MDA8 ozone which is the key pollutant for the Ozone NAAQS Rule.
These graphical presentations include:
(1) regional maps which show the mean bias and error as well as normalized mean bias and
error calculated for MDA8 > 60 ppb for May through September at individual monitoring sites,
(2) bar and whisker plots which show the distribution of the predicted and observed data
by month (May through September) and by region, and
(3) time series plots (May through September) of observed and predicted concentrations
for 13 representative high ozone sites in the urban areas with the highest projected ozone levels
in each region from the 2025 base case CAMx simulation.
The Atmospheric Model Evaluation Tool (AMET) was used to calculate the model
performance statistics used in this document (Gilliam et al, 2005). For this analysis and
summary of the 2011 model evaluation for ozone, we have selected the mean bias, mean error,
normalized mean bias, and normalized mean error to characterize model performance which are
consistent with the recommendations in Simon et al. (2012) and the draft SIP modeling guidance
(US EPA 2014). As noted above, we calculated the performance statistics by the May through
September ozone season.
Mean bias (MB) is used as average of the difference (predicted - observed) divided by the
total number of replicates (ri). Mean bias is given in units of ppb and is defined as:
MB = -Łi(P — 0) , where P = predicted and O = observed concentrations.
Mean error (ME) calculates the absolute value of the difference (predicted - observed)
divided by the total number of replicates (n). Mean error is given in units of ppb and is defined
as:
ME = -2? IP-01
n L ' '
Normalized mean bias (NMB) is used as a normalization to facilitate a range of
concentration magnitudes. This statistic averages the difference (predicted - observed) over the
5A-2
-------
sum of observed values. NMB is a useful model performance indicator because it avoids over
inflating the observed range of values, especially at low concentrations. Normalized mean bias is
given in units of % and is defined as:
Normalized mean error (NME) is also similar to NMB, where the performance statistic is
used as a normalization of the mean error. NME calculates the absolute value of the difference
(predicted - observed) over the sum of observed values. Normalized mean error is given in units
of % and is defined as:
In general, the model performance statistics indicate that the 8 -hour daily maximum ozone
concentrations predicted by the 201 1 CAMx modeling platform closely reflect the corresponding
8-hour observed ozone concentrations in space and time in each region of the 12 -km U.S.
modeling domain. The acceptability of model performance was judged by considering the 201 1
CAMx performance results in light of the range of performance found in recent regional ozone
model applications (NRC, 2002; Phillips et al, 2007; Simon et al, 2011; US EPA, 2005; US
EPA, 2009; US EPA, 201 1) These other modeling studies represent a wide range of modeling
analyses which cover various models, model configurations, domains, years and/or episodes,
chemical mechanisms, and aerosol modules. Overall, the ozone model performance results for
the 201 1 CAMx simulations performed for the Ozone NAAQS are within the range found in
other recent applications. The model performance results, as described in this document,
demonstrate that the predictions from the Ozone NAAQS modeling platform closely replicate the
corresponding observed concentrations in terms of the magnitude, temporal fluctuations, and
spatial differences for 8-hour daily maximum ozone.
Consistent with EPA's guidance for attainment demonstration modeling, we have applied
the model predictions performed as part of the Ozone NAAQS in a relative manner for
projecting future concentrations of ozone. The National Research Council (NRC, 2002) states
that using air quality modeling in a relative manner "may help reduce the bias introduced by
5A-3
-------
modeling errors and, therefore, may be more accurate than using model results directly (absolute
values) to estimate future pollutant levels". Thus, the results of this evaluation together with the
manner in which we are applying model predictions gives us confidence that our air quality
model applications using the CAMx 2011 modeling platform provides a scientifically credible
approach for assessing ozone for the Ozone NAAQS Rule.
The 8-hour ozone model performance bias and error statistics by network for the ozone
season (May-September average) for each region are provided in Table 3A-1. The statistics
shown were calculated using data pairs on days with observed 8-hour ozone of > 60 ppb. The
distributions of observed and predicted 8-hour ozone by month in the 5-month ozone season for
each region are shown in Figures 3 A-l through 3A-5. Spatial plots of the mean bias and error as
well as the normalized mean bias and error for individual monitors are shown in Figures 3 A-6
and 3 A-9. The statistics shown in these two figures were calculated over the ozone season using
data pairs on days with observed 8-hour ozone of > 60 ppb. Time series plots of observed and
predicted 8-hour ozone during the ozone season at the 13 representative high ozone monitoring
sites are provided in Figure 3 A-lOa-m. These sites are listed in Table 3 A-2.
As indicated by the statistics in Table 3A-l, bias and error for 8-hour daily maximum
ozone are relatively low in each region. Generally, MB for 8-hour ozone > 60 ppb during the
ozone season is less than 5 ppb except in the Western region and at rural (CASTNET) sites in the
central region for which ozone is somewhat under-predicted. The monthly distribution of 8-hour
daily maximum ozone during the ozone season generally corresponds well with that of the
observed concentrations, as indicated by the graphics in Figures 3A-l through 3A-5. The
predicted concentrations tend to be close to the observed 25th percentile, median and 75th
percentile values for each region, although there is a small persistent overestimation bias for
these metrics. The CAMx model also has a tendency to under-predict the highest observational
concentrations at both the AQS and CASTNet network sites.
Figures 3 A-6 through 3 A-9 show the spatial variability in bias and error at monitor
locations. Mean bias, as seen from Figure 3A-6, is less than 6 ppb at most of the sites across the
modeling domain. Figure 3 A-7 indicates that the normalized mean bias for days with observed 8-
hour daily maximum ozone greater than or equal to 60 ppb is within ±10 percent at the vast
5A-4
-------
majority of monitoring sites across the modeling domain. There are regional differences in model
performance, where the model tends to over-predict from the Southeast into the Mid-Atlantic
States and generally under predict in the Central and Western U.S. Model performance in the
Midwest states shows both under and over predictions.
Model error, as seen from Figure 3A-8, is 10 ppb or less at most of the sites across the
modeling domain. Figure 3 A-9 indicates that the normalized mean error for days with observed
8-hour daily maximum ozone greater than or equal to 60 ppb is within 10 percent at the vast
majority of monitoring sites across the modeling domain. Somewhat greater error is evident at
sites in several areas most notably along portions of the Northeast Corridor and in portions of
Florida, North Dakota, Illinois, Ohio, North Carolina, and the western most part of the modeling
domain.
In addition to the above analysis of overall model performance, we also examine how well
the modeling platform replicates day to day fluctuations in observed 8-hour daily maximum
concentrations at 13 high ozone monitoring sites. For this site specific analysis we present the
time series of observed and predicted 8-hour daily maximum concentrations by site over the
ozone season, May through September. These monitors were chosen as representative high
ozone sites in urban areas with the highest projected ozone levels in the 2025 base case
simulation. The results, as shown in Figures A-lOa through m, indicate that the modeling
platform replicates the day-to-day variability in ozone during this time period. For example,
several of the sites not only have minimal bias but also accurately capture both the seasonal and
day-to-day variability in the observations: Alleghany County, PA; Frederick County, MD;
Wayne County, MI; Jefferson County, KY. Many additional sites generally track well and
capture day-to-day variability but underestimate some of the peak ozone days: Tarrant County,
TX; Brazoria County, TX; Harford County, MD; Queens County, NY; Suffolk County, NY;
Sheboygan County, WI; Douglas County, CO. Finally, the daily modeled ozone at the two
California sites evaluated correlates well with observations but has a persistent low bias. Looking
across all 13 sites indicates that the modeling platform is able to capture the site to site
differences in the short-term variability of ozone concentrations.
5A-5
-------
Table 3A-1. Daily Maximum 8-hour Ozone Performance Statistics > 60 ppb by Region,
by Network
Network
AQS
CASTNet
Subregion
Northeast
Mid-West
Central
South
West
Northeast
Mid-West
Central
South
West
No. of
Obs
3,746
4,240
6,087
6,736
13,568
264
240
216
443
905
MB
0.6
-0.7
-4.4
2.2
-6.6
1.1
-4.2
-8.2
-0.7
-11.0
ME
7.3
7.8
8.2
7.1
9.2
5.9
6.6
8.7
5.7
11.5
NMB
(%)
0.9
-1.0
-6.4
3.3
-9.6
1.7
-6.3
-12.4
-1.1
-16.0
NME
(%)
10.7
11.5
11.9
10.6
13.4
8.7
9.8
13.1
8.8
16.7
20llef_v6_llg_camx6losoM2US2O3_ehrmaxforAQS_DallyforMay-Sep 20llef_v6_llg_camx6losoM2US2O3_8hrmaxforCASTNET_DallyforMay-Sep
150 - D AOS_Daily 150 -
1
AQS_Daily
D---- CAMx
RPO = MANE-VU
a
I
x
«
I
201L05 2011_06 2011_07 2011_08 2011JJ9
Months
0 CASTNET_Daily
D- - - - CAMx
RPO = MANE-VU
2011_05 2011_06 2011_07 201L08 2011_09
Months
Figure 3A-1. Distribution of observed and predicted MDA8 ozone by month for the period
May through September for the Northeast subregion, (a) AQS network and (b)
CASTNet network, [symbol = median; top/bottom of box = 75th/25th
percentiles; top/bottom line = max/min values]
5A-6
-------
201 let_v6_11g camx610soM2US2 O3_8hrmax for AQS_Dally for May-Sep 2011ef V6 11g camx610sol 12US2 O3_8hrmax for CASTNET^Dally for May-Sep
I
|
I
AQS_Daily
CAMx
RPO = VISTAS
a
2011_05 201U6 2011_07 2011_08 201U9
Months
CASTNET_Daily
CAMx
RPO = VISTAS
2011_05 2011J6 201U7 2011_08 2011_
Months
Figure 3A-2. Distribution of observed and predicted MDA8 ozone by month for the period
May through September for the Southeast subregion, (a) AQS network and (b)
CASTNet network
201 let v6_11g_camx610soM2US2 O3_8hrmax for AQS Dally for May-Sep 2011ef_v6 11g camx610sol 12US2 O3_8hrmax for CASTNET_Dally for May-Sep
150 - a AOS_Daily 150 -
1
D AQS_Daily
D---- CAMx
RPO = MWRPO
a
2011_05 2011JJ6 2011_07 2011_08 2011_09
Months
• CASTNET_Daily
D- - - - CAMx
RPO = MANE-VU
201L05 2011_06 2011_07 2011_08 2011_09
Months
[Figure 3A-3. Distribution of observed and predicted MDA8 ozone by month for the period
May through September for the Midwest subregion, (a) AQS network and (b)
CASTNet network
J/V-;
-------
201 lef_v6_11g_camx610soM2US2 O3_8hrmax for AQS_Dally for May-Sep 201lef_v6_1lg_camx610soM2US2 O3_8hrmax for CASTNET_Dally for May-Sep
130 H H AQS_Daily 150 -\ n CASTNET_Daily
I
D AQS_Daily
O---- CAMx
RPO=CENRAP
a
2011_05 2011JJ6 2011_07 2011_08 2011_09
Months
O - - - CAMx
RPO = CENRAP
2011_05 2011_06 2011_07 201L08 2011_09
Months
Figure 3A-4. Distribution of observed and predicted MDA8 ozone by month for the period
May through September for the Central states, (a) AQS network and (b)
CASTNet network
2011ef_v6_11g camx610sol 12US2 O3 Bhrmax for AQS_Dally for May-Sep 2011ef_v6_1lg camx610sol_12US2 O3_8hrmax for CASTNET Dally for May-Sep
150 H B AQS_Daily 150 -j D CASTNET_Daily
D AQS_Daily
O---- CAMx
RPO = WRAP
a
2011_05 2011_06 2011_07 2011_08 2011_09
Months
a---- CAMx
RPO = WRAP
2011_05 2011J
2011_07 2011_08 2011_09
Months
Figure 3A-5. Distribution of observed and predicted MDA8 ozone by month for the period
May through September for the West, (a) AQS network and (b) CASTNet
network
J/V-C
-------
O3_8hrmax MB (ppb) for run2011ef_v6_11g_camx610soM2US2 for 20110501 to 20110930
units = ppb
coverage limit - 75%
>20
15
10
5
0
-5
-10
-15
<-20
TRIANGLE=CASTNET_Daily;CIRCLE=AQS_Daily;
Figure 3A-6. Mean Bias (ppb) of MDA8 ozone greater than 60 ppb over the period May-
September 2011 at AQS and CASTNet monitoring sites in 12-km U.S. modeling
domain
03
8hrmax NMB (%) tor run 2011et,v6 11g,camx6tOsoM2US2 (Of May-Sep [O3 Shrmax ob>=60ppb]
nf.i - *.
jfy~\
owarag* Urrn - ?S%
>100
90
180
70
60
50
40
30
\20
110
10
1-10
-20
1-30
-40
1-50
-€0
-70
-«0
1-90
:-100
ClRCtE-AQS Dally; TRIANGLE-CASTNET Daily:
Figure 3A-7. Normalized Mean Bias (%) of MDA8 ozone greater than 60 ppb over the
period May-September 2011 at AQS and CASTNet monitoring sites in 12-km
U.S. modeling domain
3A-9
-------
O3 8hrmax ME (ppb) tor run2011et_v6.11g_camx6108Ol_12US2 tor May-Sep [03_8hrma» ob>=60ppb]
JT~\
... ..^,. - • .
>20
II
16
14
12
10
8
6
4
CIRCLE-AQS Daily; TRIANGLE-CASTNET Daily:
Figure 3A-8. Mean Error (ppb) of MDA8 ozone greater than 60 ppb over the period May-
September 2011 at AQS and CASTNet monitoring sites in 12-km U.S. modeling
domain
O3 Bhrmax NME (%) tor run 2011et v6 11g camx610sol 12US2 tor May-Sep [O3 Shnnax^ob>-60ppb]
>50
145
40
35
30
J 25
CIRCLE-AQS Daily; TRIANGLE-CASTNET Daily.
Figure 3A-9. Normalized Mean Error (%) of MDA8 ozone greater than 60 ppb over the
period May-September 2011 at AQS and CASTNet monitoring sites in 12-km
U.S. modeling domain
5 A-10
-------
Table 3A-2. Key Monitoring Sites Used for the Ozone Time Series Analysis
County
State
Monitoring
Site ID
Fresno
California
60195001
San Bernardino California
60710005
Tarrant
Brazoria
Allegheny
Frederick
Harford
Queens
Suffolk
Sheboygan
Wayne
Jefferson
Douglas
Texas
484392003
Texas
480391004
Pennsylvania 420031005
Maryland
Maryland
New York
New York
Wisconsin
Michigan
Kentucky
Colorado
240210037
240251001
360810124
361030002
551170006
261630019
211110067
80350004
2011ef_v6 11g camx610sol 12US2 O3 Shrmax for AQS Daily Site: 060195001 in CA
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
AQS_Daily
2011ef v6_11g_camx610sol_12US2
# of Sites: 1
Site: 060195001
MayOI May14 May 27 Jun 08 Jun 20 Jul 02 Jul 14 Jul 25 Aug 06 Aug18 Aug 30 Sep 12 Sep 25
Date
Figure 3A-10a. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 60195001 in Fresno Co., California
5 A-11
-------
2011ef_v6J1g_camx610sol_12US2 OSJhrmax for AQS_Daily Site: 060710005 in CA
140
130
120
-110
Ł100
AQS_Daily
2011ef_v6_11g_camx610sol_12US2
# of Sites: 1
Site: 060710005
/\
Figure 3A-10b. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 60710005 in San Bernardino Co.,
California
2011ef_v6_11g_camx610sol 12US2 O3 Shrmax for AQS Daily Site: 080350004 in CO
140 -
130 -
120 -
- 110 -
§: 100 -
~ 90 -
| 80 -
70 -
60 -
sg
O 50 -
40 -
30 -
20 -
CO
AQS_Daily
2011ef_v6_11g_camx610sol_12US2
# of Sites: 1
Site: 080350004
May 01 May 14 May 27 Jun 09 Jun 22 Jul 04 Jul 15 Jul 26 Aug 07 Aug 20 Sep 02 Sep15 Sep 28
Date
Figure 3A-10c. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 80350004 in Douglas Co., Colorado
JA-12
-------
2011ef_v6_11g_camx610sol 12US2 O3_8hrmax for AQS_Daily Site: 484392003 in TX
AQS_Daily
— 2011ef_v6_11g_camx610sol_12US2
# of Sites: 1
Site: 484392003
140 -
130 -
120 -
- 110-
8: 100 -
— 90 -
I 80 -
Ł 70 -
«' 6°-
O 50 -
40 -
30 -
20 -
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure 3A-10d. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 484392003 in Tarrant Co., Texas
2011ef_v6_11g_camx610sol_12US2 O3_8hrmax for AQS_Daily Site: 480391004 in TX
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
AQS_Daily
— 2011ef_v6_11g_camx610sol_12US2
# of Sites: 1
Site: 480391004
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure 3A-10e. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 480391004 in Brazoria Co., Texas
5 A-13
-------
2011ef_v6_11g_camx610sol_12US2 O3_8hrmax for AQS_Daily Site: 551170006 in Wl
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
AQS_Daily
— 2011ef_v6_11g_camx610sol_12US2
# of Sites: 1
Site: 551170006
May 01 May 14 May 27 Jun 09 Jun21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure 3A-10f. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 551170006 in Sheboygan Co., Wisconsin
2011ef_v6 11g camx610sol 12US2 O3 Shrmax for AQS Daily Site: 261630019 in Ml
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
AQS_Daily
- 2011ef_v6_11g_camx610soM2US2
# of Sites: 1
Site: 261630019
IIIMII Mill HIM Illlllll I
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure 3A-10g. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 261630019 in Wayne Co., Michigan
JA-14
-------
2011ef_v6 11 g camx61 Osol 12US2 O3 Shrmax for AQS Daily Site: 211110067 in KY
140
130
120
~ 110
§; 100
~ 90
| 80
^ 70
»' 6°
O 50
40
30
20
- AQS_Daily
- 2011ef_v6_11g_camx610sol_12US2
# of Sites: 1
Site: 211110067
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure 3A-10H. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 211110067 in Jefferson Co., Kentucky
2011ef_v6_11g_camx610sol_12US2 O3_8hrmax for AQS_Daily Site: 420031005 in PA
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
AQS_Daily
- 2011ef_v6_11g_camx610soM2US2
# of Sites: 1
Site: 420031005
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure SA-lOi. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 420031005 in Allegheny Co., Pennsylvania
5 A-15
-------
2011ef v6 11g camx610sol 12US2 O3 Shrmax for AQS Daily Site: 240210037 in MD
140
130
120
~ 110
a 100
~ 90
| 80
^ 70
«' 6°
O 50
40
30
20
AQS_Daily
2011 ef_v6_11 g_camx61 Osol_12US2
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure 3A-10J. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 240210037 in Frederick Co., Maryland
2011ef_v6_11g_camx610sol_12US2 O3_8hrmax for AQS_Daily Site: 240251001 in MD
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
AQS_Daily
- 2011ef_v6_11g_camx610soM2US2
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 15 Jul 26 Aug 07 Aug 20 Sep 02 Sep 15 Sep 28
Figure 3A-10k. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 240251001 in Harford Co., Maryland
5 A-16
-------
2011ef_v6_11g_camx610sol_12US2O3_8hrmax for AQS_Daily Site: 360810124 in NY
140
130
120
- 110
& 100
~ 90
| 80
Ł 70
"' 60
O 50
40
30
20
AQS_Daily
- 2011ef_v6_11g_camx610sol_12US2
# of Sites: 1
Site: 36081 0124
May 01 May 14 May 27 Jun 09 Jun21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure 3A-101. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 360810124 in Queens, New York
2011ef_v6_11g_camx610sol_12US2O3_8hrmax for AQS_Daily Site: 361030002 in NY
AQS_Daily
2011ef_v6_11g_camx610soM2US2
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
O 50 -
40 -
30 -
20 -
INN IIM i ii IIM i ii inll hini inn ii inn Minn mi ill mi ii i mi nil ill nil ill ii ii nil i ii nil ill nil inini inn ii inn ii inn ill in 11 Mini in
May 02 May 15 May 28 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 18 Aug 30 Sep 12 Sep 25
Date
Figure 3A-10m. Time series of observed (black) and predicted (red) MDA8 ozone for
May through September 2011 at site 361030002 in Suffolk County, New York.
3A.2 California Sub-Regions and Areas of Influence
As discussed in chapter 3 of the ozone RIA, we performed air quality modeling to gauge
the sensitivity of ozone to 50% and 90% cuts in NOx emissions statewide in California. When
applying the results of these model simulations to estimate emissions reductions to attain the
current and alternative NAAQS, we made a simplifying assumption that the model-predicted
ozone response at locations in the San Joaquin Valley and areas of California further north are
5 A-17
-------
solely due to emissions changes in these areas. That is, we associated the predicted ozone
changes in northern California to emissions changes in this portion of the state even though the
air quality model simulation included reductions statewide. We made a similar assumption about
the response of ozone to emissions changes for the southern portion of California. The northern
and southern source regions were identified for the purpose of determining which emissions
reductions would be considered to impact design values at specific monitors. In calculating the
impact of emissions changes on design values only monitors within each source regions were
assumed to be impacted from emissions from within that source region. For VOC emissions
reductions from available known controls, the source regions were defined to include only those
portions of the sub-regions that also fell within the NOx buffer area for which known NOx
controls were applied (see figure 3-4 from chapter 3 of the RIA). Therefore, when determining
the tons of emissions reductions needed to meet various standard levels, the impacts of emissions
reductions within each source region were applied to just those monitors located within that
regions. In addition, we determined the areas outside of California that are expected to be most
affected by emissions reductions from each of the two California sub-regions (i.e., downwind
impact areas) (Figure 3A-14). When creating the BenMap and FASOMGHG surfaces described
in section 3.4 of the main chapter, we determined the impact on ozone in these downwind areas
due to emissions reductions from within each of the two California sub-regions.
Several considerations were considered when delineating these two California sub-regions
for the various steps in this analysis. The spatial extent of the two California NOx source regions
were based on the geographic boundaries of California air basins as defined by the California Air
Resources Board (CARB). CARB designates Air Basins for the "purpose of managing air
resources" in areas with "generally . . . similar meteorology and geographic conditions
throughout" (http://www.arb.ca.gov/ei/maps/statemap/abmap.htm). The various Californian Air
Basins were then combined into two larger sub-regions for this analysis. The geographic
groupings were based on general air flow patterns which are governed by mountain topography
and onshore/offshore wind flows. The Air Basins, sub-regions, and predominant California wind
patterns are shown in Figure 3A-11. One county, Kern County, was split between Air Basins
that were assigned to different sub-regions: San Joaquin Valley Air Basin (Northern sub-region)
and Mojave Desert Air Basin (Southern sub-region). Since emissions inventories are categorized
5 A-18
-------
by county, we assigned all Kern County emissions to the Northern sub-region because
Bakersfield, the most populated area of Kern County, is located in the Northern sub-region.
a)
Upslope/downslopei
f low mixes-air.f ror» [•
the valley into the-^
foothills
Onshore/Offshore
flow transports LA
air to coastal areas
Mountains form
a natural barrier
to air flow
between San
Joaquin Valley
and Southern
California
b)
California Air Basins
NORTH COAST
LAKE COUNTY
SAN FRANCISCO
BAY
V NORTHEAST PLATEAU
[SACRAMENTO VALLEV
[MOUNTAIN COUNTIES
-LAKE TAHOE
SAN JOAQUIN VALLEY
GREAT BASIN
VALLEYS
NORTH CENTRAL
COAST
MOJAVE
UESERT
SOUTH CENTRAL
COAST
SOUTH COAST
SAN Dlf&'i
COUNTY
Figure 3A-ll.a) Depiction of governing wind patterns and topography, b) California Air
Basins and sub-regions used for this analysis. Northern sub-region is outlined in
pink and southern sub-region is outlined in blue.
The downwind receptor regions outside of California were determined by examining the
spatial patterns of ozone impacted by the the 50% cut California in NOx. Ozone changes due to
the state-wide emissions reductions appear to follow fairly distinct widespread plumes. From
this analysis it was determined that impacts of emissions reductions from the Northern sub-
region were generally limited to Oregon, Washington, and a few Nevada counties near Carson
City. Conversely, emissions reductions from the Southern sub-region appear to have widespread
impacts across much of the Southwestern and Central US. One caveat is that the geographic
extent of these downwind regions are based on general transport patterns, as determined by
examining model outputs from a single year of meteorology (i.e., 2011) and are not intended to
fully represent the downwind transport of ozone from California emissions on all days and at all
times. However, this approach was necessary in order to match the Northern and Southern
California sub-region emissions reductions to ozone impacts in downwind areas outside this
state. Figure 3A-12 shows examples of the impact on 8-hr daily maximum ozone from 50%
California NOx cuts on a few representative days. Figure 3 A-13 shows the downwind receptor
regions that were applied in this analysis. Note that VOC impacts were only applied within the
5 A-19
-------
source regions consistent with how VOC reductions were treated for other areas of the country.
Figure 3A-14 shows that ozone impacts from a 50% cut in US anthropogenic VOC emissions
were localized within the two California sub-regions and do not appear to impact downwind
states.
May 7, 2011
May 24, 2011
July 14, 2011
Figure 3A-12. Impact of 50% anthropogenic California NOx cuts (ppb) on 8-hr daily
average ozone concentrations on three days in 2011
5A-20
-------
Figure 3A-13. Downwind California receptor regions for Northern California (green) and
Southern California (purple)
JA-21
-------
June 21, 2011
JulyS, 2011
Figure 3A-14. Impact of 50% US anthropogenic VOC cuts (ppb) on 8-hr daily average
ozone concentrations on three days in 2011
3A.3 VOC Impact Areas
As described in chapter 3, we defined VOC impact regions for the following urban areas:
New York City, Pittsburgh, Baltimore, Detroit, Chicago, Louisville, Houston, Dallas, Denver,
Northern California and Southern California. Not only did these areas have the highest design
values in each region, but ozone in these areas was also sensitive to VOC emissions reductions in
our modeling. Figure 3A-15 shows the impact of 50% US anthropogenic VOC cuts on July
monthly average 8-hour daily maximum ozone concentrations across the US. Ozone in each of
5A-22
-------
the areas listed above is shown to have at least 0.2 ppb response to VOC emissions cuts.
Ozone Change from US 50% VO
July avg of 8-hr daily max
Figure 3A-15. Change in July average of 8-hr daily maximum ozone concentration (ppb)
due to 50% cut in US anthropogenic VOC emissions
3A.4 Numeric Examples of Calculation Methodology for Changes in Design Values
In this section we use the data for two monitoring sites to demonstrate how changes in
design values were calculated, as described in section 3.3. For each monitor, numerical
examples are given for calculating the emissions reductions necessary to attain the current 75
ppb NAAQS (i.e., the baseline scenario) as well as the 65 ppb scenario, which is incremental to
the baseline. Note that design values are truncated when they are compared to a standard level,
so a calculated design value of 75.9 is truncated to 75 ppb and, therefore, meets the current 75
ppb standard. Similarly, a design value of 65.9 would meet an alternative standard level of 65.
For each monitor, we start with the base case design value, then account for ozone changes
simulated in the 11 l(d) sensitivity simulation and then apply equation 3-5 from chapter 3.
= DV
202SJ
X
(R2J
X
(R3
3J
X
Eq3-5
Example 1. Fresno California monitor 60195001 (baseline):
5A-23
-------
^^60195001,
NOx
NOx+VOC /
83.5 + ^07 + -9.9 x IP"5 x 15,000
,iiid \^60\<)Zoo\,CAcontrl,NCA ^ECAcontrl,NCA'
NOx
+ -1.3 X IP"4 X 8100 + 32,000
R60-L9500-L,CAcontrol+50NOx,NCA
VOC
+ -9.9 x 10-6 x 3200 + 22,000 ) = 75.9 ppb
\K60-L9500-L,VOC_50,NCA
Example 2. Fresno California monitor 60195001 (65 ppb scenario):
NOx up to 50% of CA modeled control sensitivity
DVj:65 = 75.9 + ( -1.3 x IP"4 x 61,000 ]
DVj,baseline \K60l9500l,CAcontrol+50NOx,NCA AE /
NOx beyond 50% of CA modeled control sensitivity
+ -2.0 x IP-* x 12,000 =65.6 ppb
\K60l9500l,CAcontrol50-90NOx,NCA
Example 3. Dallas monitor 484392003 (baseline):
NOx
NOx+VOC 7
77A + +( -1-6 x 10-s x 4400
01/484392003,2025 &DV '484392003, Hid \K484392003,TXcontrol
ppb
Example 4. Dallas monitor 484392003 (65 ppb scenario):
5A-24
-------
DV*
484392003,65
NOx
75.4 + f -1.6 x 10-s x 56,000
^484392003,fcaseiine \R484392003,TXcontroi
NOx
-1.1 x IP"5 x 770,000
^484392003, TXcontrol+SOcentralNOx AE
VOC
+ ( -9.2 x 10-6 x 17,000 ] = 65.9 ppb
\K484392003,VOC_50,Dallas AE /
3A.5 Emissions Reductions Applied to Create Baseline and Alternative Standard Level
Scenarios
The following tables present emissions reductions applied in each region to create the
baseline and alternative standard level scenarios. These emissions reductions were determined
using the methodology described in section 3.3.2 of the main RIA and demonstrated in section
3 A. 4 of this appendix. Sector-specific controls used for these reductions are discussed in more
detail in chapter 4 of the RIA. These emissions reductions were used to create the ozone
surfaces described in section 3.4 of the RIA.
Table 3A-3. Emissions Reductions Applied to Create the Baseline Scenario*
Emissions reductions (thousand tons) applied from
2025-2030
California mobile
source changes
NOx reductions
from one of the
explicit control
cases
VOC reductions
identified from
maxcontrol CoST
run
Additional NOx
reductions
Northeast
Midwest
Central
Southwest
California
N/A
N/A
N/A
N/A
14 (NOx)
6 (VOC)
N/A
N/A
45 (TX explicit
emissions control
case)
N/A
29 (CA explicit
emissions control
case)
N/A
N/A
N/A
N/A
51 (inNandSCA
buffer region)
N/A
N/A
N/A
N/A
32 (N California);
130 (S California)
*These emission are in addition to changes modeled in the simulation representing option
1 (state) of the proposed carbon pollution guidelines under section 11 l(d) of the CAA.
Table 3A-4. Emissions Reductions Applied Beyond the Baseline Scenario to Create the 70
ppb Scenario
Emissions reductions (thousand tons) applied from
5A-25
-------
Northeast
Midwest
Central
Southwest
California
NOx reductions from
one of the explicit
control cases*
110
N/A
58 (TX explicit emissions
control case)
N/A
Exhausted in baseline
scenario
VOC reductions
identified from
maxcontrol CoST run
31 (NY area);
6 (Baltimore area)
N/A
1 8 (Houston area)
N/A
Exhausted in baseline
scenario
Additional NOx reductions
130 (98 within NE buffer; 3 1
outside the NE buffer)
N/A
350 (95 within TX buffer;
260 outside of TX buffer)
N/A
38 (N California);
15 (S California)
Table 3A-5. Emissions Reductions Applied Beyond the Baseline Scenario to Create the 65
ppb Scenario
Emissions reductions (thousand tons) applied from
NOx reductions from one of
the explicit control cases*
VOC reductions
identified from
maxcontrol CoST
Additional NOx reductions
run
Northeast
Midwest
Central
Southwest
California
110
250
58 (TX explicit emissions
control case)
77
Exhausted in baseline scenario
36 (NY area)
28 (Chicago area)
18 (Houston area);
17 (Dallas area)
7 (Denver area)
Exhausted in baseline
scenario
400 (300 within NE buffer;
97outside the NE buffer)
180
770 (210 within TX buffer;
560 outside of TX buffer)
36
73 (N California);
32 (S California)
*In regions without a modeled explicit control case (Southwest and Midwest) this represents
equivalent emissions reductions that would have been identified in an explicit control case that
covered the entire region
Table 3A-6. Emissions Reductions Applied Beyond the Baseline Scenario to Create the 60
ppb Scenario
Emissions reductions (thousand tons) applied from
Northeast
Midwest
Central
Southwest
California
NOx reductions from one
of the explicit control
cases*
110
250
58 (TX explicit emissions
control case)
77
Exhausted in baseline
scenario
VOC reductions
identified from
maxcontrol CoST run
36 (NY area);
6 (Baltimore area)
32 (Chicago area);
4 1 (additional Chicago
area VOC control)
18 (Houston area);
18 (Dallas area)
7 (Denver area)
Exhausted in baseline
scenario
Additional NOx reductions
610 (470 within NE buffer;
150 outside the NE buffer)
620
1,200 (340 within TX buffer;
9 10 outside of TX buffer)
240
97 (N California);
48 (S California)
5A-26
-------
*In regions without a modeled explicit control case (Southwest and Midwest) this represents
equivalent emissions reductions that would have been identified in an explicit control case that
covered the entire region
3A.6 Design Values for All Monitors included in the Quantitative Analysis
Table 3A-7. Design Values for California Region Monitors
Site ID
60010007
60010009
60010011
60012001
60050002
60070007
60070008
60090001
60111002
60130002
60131002
60131004
60170010
60170012
60170020
60190007
60190011
60190242
60192009
60194001
60195001
60210003
60250005
60254003
60254004
60290007
60290008
60290011
60290014
60290232
60295002
60296001
60311004
60333001
Lat
37.69
37.74
37.81
37.65
38.34
39.71
39.76
38.20
39.20
37.94
38.01
37.96
38.73
38.81
38.89
36.71
36.79
36.84
36.63
36.60
36.82
39.53
32.68
33.03
33.21
35.35
35.05
35.05
35.36
35.44
35.24
35.50
36.31
39.03
long
-121.78
-122.17
-122.28
-122.03
-120.76
-121.62
-121.84
-120.68
-122.02
-122.03
-121.64
-122.36
-120.82
-120.03
-121.00
-119.74
-119.77
-119.87
-120.38
-119.50
-119.72
-122.19
-115.48
-115.62
-115.55
-118.85
-119.40
-118.15
-119.04
-119.02
-118.79
-119.27
-119.64
-122.92
State
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
County
Alameda
Alameda
Alameda
Alameda
Amador
Butte
Butte
Calaveras
Colusa
Contra Costa
Contra Costa
Contra Costa
El Dorado
El Dorado
El Dorado
Fresno
Fresno
Fresno
Fresno
Fresno
Fresno
Glenn
Imperial
Imperial
Imperial
Kern
Kern
Kern
Kern
Kern
Kern
Kern
Kings
Lake
Base Case
66
47
46
54
60
63
53
63
53
65
65
51
67
61
69
82
81
81
64
76
83
56
69
64
63
81
72
71
77
77
74
73
74
48
Baseline
61
44
43
51
54
58
49
57
49
60
59
47
60
59
62
74
73
74
59
69
75
52
62
54
53
74
67
58
71
71
68
67
67
45
70
57
42
41
48
50
54
46
53
46
57
56
45
55
57
57
69
68
70
55
65
70
49
61
53
52
70
63
57
66
66
64
63
63
43
65
53
40
39
45
46
50
43
49
43
53
51
42
50
56
52
64
63
65
52
60
65
46
60
51
50
65
59
55
61
61
59
59
59
41
60
48
36
35
41
43
47
40
46
41
48
47
38
46
55
47
59
58
60
49
56
60
44
59
50
48
60
55
53
56
57
55
55
55
38
5A-27
-------
Site ID
60370002
60370016
60370113
60371002
60371103
60371201
60371302
60371602
60371701
60372005
60374002
60376012
60379033
60390004
60392010
60410001
60430006
60470003
60530002
60530008
60531003
60550003
60570005
60570007
60590007
60591003
60592022
60595001
60610003
60610004
60610006
60650004
60650008
60650009
60650012
60650016
60651016
60652002
60655001
60656001
Lat
34.14
34.14
34.05
34.18
34.07
34.20
33.90
34.01
34.07
34.13
33.82
34.38
34.67
36.87
36.95
37.97
37.55
37.28
36.50
36.21
36.70
38.31
39.23
39.32
33.83
33.67
33.63
33.93
38.94
39.10
38.75
34.01
33.74
33.45
33.92
33.58
33.95
33.71
33.85
33.79
long
-117.92
-117.85
-118.46
-118.32
-118.23
-118.53
-118.21
-118.07
-117.75
-118.13
-118.19
-118.53
-118.13
-120.01
-120.03
-122.52
-119.84
-120.43
-121.73
-121.13
-121.64
-122.30
-121.06
-120.85
-117.94
-117.93
-117.68
-117.95
-121.10
-120.95
-121.27
-117.52
-115.82
-117.09
-116.86
-117.08
-116.83
-116.22
-116.54
-117.23
State
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
County
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Madera
Madera
Marin
Mariposa
Merced
Monterey
Monterey
Monterey
Napa
Nevada
Nevada
Orange
Orange
Orange
Orange
Placer
Placer
Placer
Riverside
Riverside
Riverside
Riverside
Riverside
Riverside
Riverside
Riverside
Riverside
Base Case
75
88
62
73
63
83
57
66
80
74
55
88
79
70
73
47
65
71
50
51
46
53
63
61
60
58
60
66
69
61
70
77
56
60
85
64
87
74
81
78
Baseline
55
65
49
53
48
62
52
55
62
55
50
64
60
64
67
44
61
65
40
41
36
49
58
56
49
48
45
54
62
55
63
60
47
46
64
48
65
59
63
58
70
52
61
46
49
45
59
51
52
58
51
49
60
57
60
63
41
58
61
39
40
35
46
54
52
47
46
42
51
57
51
58
57
46
43
60
45
61
57
60
54
65
48
56
43
46
41
55
49
48
54
47
47
56
54
56
59
39
55
57
37
39
33
43
50
48
44
44
40
48
52
47
53
53
44
41
56
43
58
55
57
51
60
43
51
39
42
37
51
47
44
50
43
45
51
51
52
55
36
53
53
36
37
32
40
47
45
42
41
37
45
47
44
48
50
43
39
52
40
53
52
54
47
5A-28
-------
Site ID
60658001
60658005
60659001
60659003
60670002
60670006
60670010
60670011
60670012
60670014
60675003
60690002
60690003
60710001
60710005
60710012
60710306
60711004
60711234
60712002
60714001
60714003
60719002
60719004
60730001
60730003
60730006
60731001
60731002
60731006
60731008
60731010
60731016
60731201
60732007
60771002
60773005
60790005
60792006
60793001
Lat
34.00
34.00
33.68
33.61
38.71
38.61
38.56
38.30
38.68
38.65
38.49
36.84
36.49
34.90
34.24
34.43
34.51
34.10
35.76
34.10
34.42
34.06
34.07
34.11
32.63
32.79
32.84
32.95
33.13
32.84
33.22
32.70
32.85
33.36
32.55
37.95
37.68
35.63
35.26
35.37
long
-117
-117
.42
.49
-117.33
-114
-121
-121
-121
-121
-121
-121
-121
-121
-121
-117
-117
-117
.60
.38
.37
.49
.42
.16
.51
.21
.36
.16
.02
.27
.56
-117.33
-117
-117
-117
.63
.40
.49
-117.29
-117
-116
-117
-117
-116
-117
-117
-117
-116
-117
-117
-117
-117
-116
-121
-121
-120
-120
-120
.15
.39
.27
.06
.94
.13
.26
.08
.77
.40
.15
.12
.09
.94
.27
.44
.69
.67
.84
State
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
County
Riverside
Riverside
Riverside
Riverside
Sacramento
Sacramento
Sacramento
Sacramento
Sacramento
Sacramento
Sacramento
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
San
Benito
Benito
Bernardino
Bernardino
Bernardino
Bernardino
Bernardino
Bernardino
Bernardino
Bernardino
Bernardino
Bernardino
Bernardino
Diego
Diego
Diego
Diego
Diego
Diego
Diego
Diego
Diego
Diego
Diego
Joaquin
Joaquin
Luis Obispo
Luis Obispo
Luis Obispo
Base Case Baseline 70
89
85
74
60
65
66
61
62
77
60
72
54
61
69
99
85
77
91
65
96
88
97
82
91
60
61
62
57
58
69
56
55
59
58
54
59
70
55
47
46
68
65
55
54
59
60
55
56
69
54
65
42
48
57
75
65
60
71
60
74
68
73
66
69
54
49
50
48
44
55
44
49
47
45
48
54
65
46
38
39
64
61
52
53
54
55
51
52
64
50
60
40
46
55
70
62
57
66
59
69
64
68
63
64
53
47
49
46
42
52
43
48
46
43
47
50
61
44
37
38
65
60
57
49
52
49
50
47
48
58
46
54
38
45
53
65
58
54
62
59
64
60
64
61
60
52
45
47
45
40
50
41
47
44
41
46
46
56
43
36
38
60
55
53
45
51
45
45
42
44
52
42
49
36
43
51
60
54
50
56
58
59
55
59
58
55
51
43
45
44
38
47
39
46
42
39
45
42
52
42
35
37
5A-29
-------
Site ID
60794002
60798001
60798005
60798006
60811001
60830008
60830011
60831008
60831013
60831014
60831018
60831021
60831025
60832004
60832011
60833001
60834003
60850002
60850005
60851001
60852006
60852009
60870007
60890004
60890007
60890009
60893003
60950004
60950005
60953003
60970003
60990005
60990006
61010003
61010004
61030004
61030005
61070009
61072002
61072010
Lat
35.03
35.49
35.64
35.35
37.48
34.46
34.43
34.95
34.73
34.54
34.53
34.40
34.49
34.64
34.45
34.61
34.60
37.00
37.35
37.23
37.08
37.32
36.98
40.55
40.45
40.69
40.54
38.10
38.23
38.36
38.44
37.64
37.49
39.14
39.21
40.26
40.18
36.49
36.33
36.03
long
-120.50
-120.67
-120.23
-120.04
-122.20
-120.03
-119.69
-120.44
-120.43
-119.79
-120.20
-119.46
-120.05
-120.46
-119.83
-120.08
-120.63
-121.57
-121.89
-121.98
-121.60
-122.07
-121.99
-122.38
-122.30
-122.40
-121.57
-122.24
-122.08
-121.95
-122.71
-120.99
-120.84
-121.62
-121.82
-122.09
-122.24
-118.83
-119.29
-119.06
State
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
County
San Luis Obispo
San Luis Obispo
San Luis Obispo
San Luis Obispo
San Mateo
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Clara
Santa Clara
Santa Clara
Santa Clara
Santa Clara
Santa Cruz
Shasta
Shasta
Shasta
Shasta
Solano
Solano
Solano
Sonoma
Stanislaus
Stanislaus
Sutter
Sutter
Tehama
Tehama
Tulare
Tulare
Tulare
Base Case
50
53
67
65
53
51
49
43
54
58
49
58
60
47
49
52
54
58
56
59
62
56
47
52
58
60
58
53
58
58
39
66
76
55
64
65
63
79
71
74
Baseline
41
44
55
53
50
44
42
35
45
49
43
49
52
40
42
44
47
53
52
54
57
52
36
47
53
54
55
49
53
53
36
60
69
50
59
60
58
72
64
68
70
40
43
53
51
49
43
41
34
44
48
43
48
50
39
41
43
46
50
49
51
53
49
34
44
49
51
53
46
50
50
34
56
64
47
55
56
55
68
60
64
65
39
42
52
49
46
42
40
33
42
47
42
47
49
38
41
41
45
46
45
47
49
45
32
41
46
47
51
43
46
46
32
52
59
43
52
53
51
64
56
60
60
37
40
50
47
42
41
39
32
41
45
41
46
48
37
40
40
44
42
41
43
45
41
30
38
43
44
50
39
43
43
31
48
55
41
49
50
49
60
52
56
5A-30
-------
Site ID
61090005
61110007
61110009
61111004
61112002
61113001
61130004
61131003
Lat
37.98
34.21
34.40
34.45
34.28
34.25
38.53
38.66
long
-120.38
-118.87
-118.81
-119.23
-118.68
-119.14
-121.77
-121.73
State
California
California
California
California
California
California
California
California
County
Tuolumne
Ventura
Ventura
Ventura
Ventura
Ventura
Yolo
Yolo
*The design value from the monitor(s) with the highest projected
shown in bold blue text
Table 3A-8. Design Values for Southwest Monitors
Site ID
40051008
40070010
40130019
40131004
40131010
40132001
40132005
40133002
40133003
40134003
40134004
40134005
40134008
40134010
40134011
40137003
40137020
40137021
40137022
40137024
40139508
40139702
40139704
40139706
40139997
40170119
40190021
lat
35.21
33.65
33.48
33.56
33.45
33.57
33.71
33.46
33.48
33.40
33.30
33.41
33.82
33.64
33.37
33.29
33.49
33.51
33.47
33.51
33.98
33.55
33.61
33.72
33.50
34.82
32.17
long
-111.65
-111.11
-112.14
-112.07
-111.73
-112.19
-111.86
-112.05
-111.92
-112.08
-111.88
-111.93
-112.02
-112.34
-112.62
-112.16
-111.86
-111.76
-111.81
-111.84
-111.80
-111.61
-111.73
-111.67
-112.10
-109.89
-110.74
State
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
County
Coconino
Gila
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Maricopa
Navajo
Pima
Base
62
64
66
67
72
55
57
60
ozone
Base
63
62
65
66
58
62
62
62
63
64
61
59
62
58
56
60
62
63
60
61
59
62
62
61
64
61
62
Case Baseline
58
49
51
56
55
44
52
54
in each scenario
Case Baseline
63
62
65
66
58
62
62
62
63
64
61
59
62
58
56
60
62
63
60
61
59
62
62
61
64
61
62
70
54
47
49
55
52
43
49
51
is
70
63
62
65
66
58
62
62
62
63
64
61
59
62
58
56
60
62
63
60
61
59
62
62
61
64
61
62
65
51
44
46
54
50
42
45
47
65
62
60
62
62
55
58
59
59
60
60
57
55
59
55
54
57
58
60
57
58
56
59
59
58
61
58
57
60
48
42
44
52
46
40
42
43
60
60
52
51
51
46
48
50
48
50
50
48
46
49
46
47
49
48
50
48
48
48
49
49
49
50
55
51
5A-31
-------
Site ID
40191011
40191018
40191020
40191028
40191030
40191032
40191034
40213001
40213003
40213007
40217001
40218001
40258033
80013001
80050002
80050006
80130011
80310014
80310025
80350004
80410013
80410016
80450012
80590002
80590005
80590006
80590011
80590013
80677001
80677003
80690007
80690011
80690012
80691004
80770020
80810002
80830006
80830101
81030005
81230009
lat
32.20
32.43
32.05
32.30
31.88
32.17
32.38
33.42
32.95
32.51
33.08
33.29
34.55
39.84
39.57
39.64
39.96
39.75
39.70
39.53
38.96
38.85
39.54
39.80
39.64
39.91
39.74
39.54
37.14
37.10
40.28
40.59
40.64
40.58
39.13
40.51
37.35
37.20
40.04
40.39
long
-110.88
-111.06
-110.77
-110.98
-111.00
-110.98
-111.13
-111.54
-111.76
-111.31
-111.74
-111.29
-112.48
-104.95
-104.96
-104.57
-105.24
-105.03
-105.00
-105.07
-104.82
-104.90
-107.78
-105.10
-105.14
-105.19
-105.18
-105.30
-107.63
-107.87
-105.55
-105.14
-105.28
-105.08
-108.31
-107.89
-108.59
-108.49
-107.85
-104.74
State
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Arizona
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
Colorado
County
Pima
Pima
Pima
Pima
Pima
Pima
Pima
Final
Final
Final
Final
Final
Yavapai
Adams
Arapahoe
Arapahoe
Boulder
Denver
Denver
Douglas
El Paso
El Paso
Garfield
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
La Plata
La Plata
Larimer
Larimer
Larimer
Larimer
Mesa
Moffat
Montezuma
Montezuma
Rio Blanco
Weld
Base Case
57
58
60
57
60
56
56
61
59
61
60
64
63
61
66
62
61
58
58
68
65
67
62
58
64
67
66
62
64
63
64
68
61
60
63
61
61
60
60
67
Baseline
57
58
60
57
60
56
56
61
59
61
60
64
63
61
66
62
61
58
58
68
65
67
62
58
64
67
66
62
64
63
64
68
61
60
63
61
61
60
60
67
70
57
58
60
57
60
56
56
61
59
61
60
64
63
61
66
62
61
58
58
68
65
67
62
58
64
67
66
62
64
63
64
68
61
60
63
61
61
60
60
67
65
53
56
55
54
55
52
53
58
57
59
58
61
62
58
63
58
58
55
55
65
63
65
60
55
61
63
63
59
63
62
61
64
58
57
62
59
60
59
58
64
60
47
50
48
48
50
46
48
49
51
55
51
52
60
48
53
50
48
46
46
55
58
61)
55
46
51
53
52
49
61)
59
53
54
49
47
57
55
57
55
55
54
5A-32
-------
Site ID
320010002
320030022
320030023
320030043
320030071
320030073
320030075
320030538
320030540
320030601
320031019
320032002
320190006
320310016
320310020
320310025
320311005
320312002
320312009
325100002
350010023
350010024
350010027
350010029
350010032
350011012
350011013
350130008
350130023
350171003
350250008
350290003
lat
39.47
36.39
36.81
36.11
36.17
36.17
36.27
36.14
36.14
35.98
35.79
36.19
39.60
39.53
39.47
39.40
39.54
39.25
39.65
39.17
35.13
35.06
35.15
35.02
35.06
35.19
35.19
31.93
32.32
32.69
32.73
32.26
long
-118.78
-114.91
-114.06
-115.25
-115.26
-115.33
-115.24
-115.06
-115.08
-114.85
-115.36
-115.12
-119.25
-119.81
-119.78
-119.74
-119.75
-119.96
-119.84
-119.73
-106.59
-106.58
-106.70
-106.66
-106.76
-106.51
-106.61
-106.63
-106.77
-108.12
-103.12
-107.72
State
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
Nevada
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
County
Churchill
Clark
Clark
Clark
Clark
Clark
Clark
Clark
Clark
Clark
Clark
Clark
Lyon
Washoe
Washoe
Washoe
Washoe
Washoe
Washoe
Carson City
Bernalillo
Bernalillo
Bernalillo
Bernalillo
Bernalillo
Bernalillo
Bernalillo
Dona Ana
Dona Ana
Grant
Lea
Luna
Base Case
53
61
57
67
66
66
65
61
61
65
66
61
61
60
61
60
61
55
61
61
60
61
64
61
58
64
61
61
60
61
60
59
Baseline
53
61
57
67
66
66
65
61
61
65
66
61
61
60
61
60
61
55
61
61
60
61
64
61
58
64
61
61
60
61
60
59
70
53
61
57
67
66
66
65
61
61
65
66
61
61
60
61
60
61
55
61
61
60
61
64
61
58
64
61
61
60
61
60
59
65
53
59
56
65
64
64
63
60
60
64
64
60
60
59
60
60
60
55
60
60
58
59
62
60
57
63
59
60
59
61
60
58
60
52
55
55
57
57
57
56
53
53
60
60
53
59
58
59
58
59
55
58
60
54
55
59
55
52
59
55
59
58
59
58
57
5A-33
-------
Site ID
350431001
350439004
350450009
350450018
350451005
350451233
350490021
350610008
490030003
490037001
490050004
490071003
490110004
490131001
490352004
490353006
490450003
490490002
490495008
490495010
490570002
490571003
560050123
560050456
560070100
560130232
560210100
560350700
560370077
560370200
560370300
560410101
lat
35.30
35.62
36.74
36.81
36.80
36.81
35.62
34.81
41.49
41.95
41.73
39.61
40.90
40.21
40.74
40.74
40.54
40.25
40.43
40.14
41.21
41.30
44.65
44.15
41.39
43.08
41.18
42.49
41.16
41.68
41.75
41.37
long
-106.55
-106.72
-107.98
-107.65
-108.47
-108.70
-106.08
-106.74
-112.02
-112.23
-111.84
-110.80
-111.88
-110.84
-112.21
-111.87
-112.30
-111.66
-111.80
-111.66
-111.98
-111.99
-105.29
-105.53
-107.62
-107.55
-104.78
-110.10
-108.62
-108.02
-109.79
-111.04
State
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
New
Mexico
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Utah
Wyoming
Wyoming
Wyoming
Wyoming
Wyoming
Wyoming
Wyoming
Wyoming
Wyoming
Wyoming
County
Sandoval
Sandoval
San Juan
San Juan
San Juan
San Juan
Santa Fe
Valencia
Box Elder
Box Elder
Cache
Carbon
Davis
Duchesne
Salt Lake
Salt Lake
Tooele
Utah
Utah
Utah
Weber
Weber
Campbell
Campbell
Carbon
Fremont
Laramie
Sublette
Sweetwater
Sweetwater
Sweetwater
Uinta
Base Case
56
59
58
64
56
55
60
58
59
60
59
64
61
63
65
65
64
64
59
63
64
64
60
60
60
61
61
60
60
60
61
58
Baseline
56
59
58
64
56
55
60
58
59
60
59
64
61
63
65
65
64
64
59
63
64
64
60
60
60
61
61
60
60
60
61
58
70
56
59
58
64
56
55
60
58
59
60
59
64
61
63
65
65
64
64
59
63
64
64
60
60
60
61
61
60
60
60
61
58
65
55
59
57
62
54
53
60
57
57
59
57
60
58
62
62
61
61
62
57
61
61
61
58
58
58
60
59
59
58
57
60
57
60
52
57
52
57
50
48
58
52
49
55
54
56
52
59
54
53
53
57
53
57
54
53
53
53
56
58
54
58
55
53
56
54
*The design value from the monitor(s) with the highest projected ozone in each scenario is
shown in bold blue text
5A-34
-------
Table 3A-9.
Site ID
50199991
50350005
51010002
51130003
51190007
51191002
51191008
51430005
200910010
201030003
201070002
201619991
201730001
201730010
201730018
201770013
201910002
201950001
202090021
220050004
220150008
220170001
220190002
220190008
220190009
220330003
220330009
220330013
220470009
220470012
220511001
220550007
220570004
220630002
220710012
220730004
220770001
220870004
220890003
Design Values for Central Region Monitors
lat
34.18
35.20
35.83
34.45
34.76
34.84
34.68
36.18
38.84
39.33
38.14
39.10
37.78
37.70
37.90
39.02
37.48
38.77
39.12
30.23
32.54
32.68
30.14
30.26
30.23
30.42
30.46
30.70
30.22
30.21
30.04
30.22
29.76
30.31
29.99
32.51
30.68
29.94
29.98
long
-93.10
-90.19
-93.21
-94.14
-92.28
-92.26
-92.33
-94.12
-94.75
-94.95
-94.73
-96.61
-97.34
-97.31
-97.49
-95.71
-97.37
-99.76
-94.64
-90.97
-93.75
-93.86
-93.37
-93.28
-93.58
-91.18
-91.18
-91.06
-91.32
-91.13
-90.28
-92.05
-90.77
-90.81
-90.10
-92.05
-91.37
-89.92
-90.41
State
Arkansas
Arkansas
Arkansas
Arkansas
Arkansas
Arkansas
Arkansas
Arkansas
Kansas
Kansas
Kansas
Kansas
Kansas
Kansas
Kansas
Kansas
Kansas
Kansas
Kansas
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
Louisiana
County
Clark
Crittenden
Newton
Polk
Pulaski
Pulaski
Pulaski
Washington
Johnson
Leavenworth
Linn
Riley
Sedgwick
Sedgwick
Sedgwick
Shawnee
Sumner
Trego
Wyandotte
Ascension
Bossier
Caddo
Calcasieu
Calcasieu
Calcasieu
East Baton Rouge
East Baton Rouge
East Baton Rouge
Iberville
Iberville
Jefferson
Lafayette
Lafourche
Livingston
Orleans
Ouachita
Pointe Coupee
St. Bernard
St. Charles
Base Case
56
64
57
65
55
58
57
62
60
59
60
63
55
64
63
63
66
67
56
63
69
67
67
61
66
67
64
59
63
65
63
60
61
62
60
59
62
61
59
Baseline
54
63
55
63
52
55
54
60
59
58
58
62
54
63
62
62
65
67
55
62
67
64
67
61
64
67
63
59
62
65
62
59
61
61
58
59
61
59
58
70
52
63
54
60
51
53
52
58
55
54
55
60
52
61
59
59
63
65
51
58
61
59
64
57
61
63
59
55
58
61
59
57
57
57
55
54
58
57
55
65
48
61
49
56
44
47
46
54
51
50
51
57
49
57
55
57
59
62
47
54
55
53
61
55
57
58
56
52
54
57
55
54
52
53
51
51
54
53
51
60
44
60
46
53
39
42
42
50
46
44
47
55
46
53
51
53
55
60
42
49
49
47
57
51
53
53
51
48
49
52
50
50
47
49
47
46
50
50
47
5A-35
-------
Site ID
220930002
220950002
221030002
221210001
280010004
280110001
280330002
280450003
280470008
280490010
280590006
280750003
280810005
281619991
290030001
290190011
290270002
290370003
290390001
290470003
290470005
290470006
290490001
290770036
290770042
290970004
290990019
291130003
291370001
291570001
291831002
291831004
291860005
291890005
291890014
292130004
295100085
400019009
400159008
400170101
lat
29.99
30.06
30.43
30.50
31.56
33.75
34.82
30.30
30.39
32.39
30.38
32.36
34.26
34.00
39.95
39.08
38.71
38.76
37.69
39.41
39.30
39.33
39.53
37.26
37.32
37.24
38.45
39.04
39.48
37.70
38.87
38.90
37.90
38.49
38.71
36.71
38.66
35.75
35.11
35.48
long
-90.82
-90.61
-90.20
-91.21
-91.39
-90.72
-89.99
-89.40
-89.05
-90.14
-88.53
-88.73
-88.77
-89.80
-94.85
-92.32
-92.09
-94.58
-94.04
-94.27
-94.38
-94.58
-94.56
-93.30
-93.20
-94.42
-90.40
-90.86
-91.79
-89.70
-90.23
-90.45
-90.42
-90.71
-90.48
-93.22
-90.20
-94.67
-98.25
-97.75
State
Louisiana
Louisiana
Louisiana
Louisiana
Mississippi
Mississippi
Mississippi
Mississippi
Mississippi
Mississippi
Mississippi
Mississippi
Mississippi
Mississippi
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Oklahoma
Oklahoma
Oklahoma
County
St. James
St. John the Baptist
St. Tammany
West Baton Rouge
Adams
Bolivar
DeSoto
Hancock
Harrison
Hinds
Jackson
Lauderdale
Lee
Yalobusha
Andrew
Boone
Callaway
Cass
Cedar
Clay
Clay
Clay
Clinton
Greene
Greene
Jasper
Jefferson
Lincoln
Monroe
Perry
Saint Charles
Saint Charles
Sainte Genevieve
Saint Louis
Saint Louis
Taney
St. Louis City
Adair
Caddo
Canadian
Base Case
58
62
63
60
56
63
58
54
58
50
57
52
51
53
60
57
55
58
63
63
62
64
64
57
59
66
64
63
58
61
68
65
60
59
66
58
64
66
64
62
Baseline
57
61
62
59
55
62
57
51
53
49
55
51
51
52
59
56
55
57
61
62
61
63
63
56
58
62
64
62
58
62
67
64
60
58
65
56
63
62
62
62
70
53
57
59
55
53
60
57
51
55
47
55
51
50
51
55
53
52
53
58
58
57
58
58
53
55
60
60
59
55
59
63
61
57
55
61
54
59
61
60
58
65
49
53
56
52
51
57
54
47
47
44
51
48
49
49
51
50
49
50
54
53
52
54
54
49
51
54
57
56
53
58
58
57
54
51
58
50
55
55
55
54
60
45
48
52
47
49
53
52
44
44
41
48
47
48
48
46
46
46
45
50
48
47
48
48
46
47
49
52
53
50
56
53
53
50
47
53
46
49
51
50
49
5A-36
-------
Site ID
400219002
400270049
400310651
400370144
400430860
400719010
400871073
400892001
400979014
401090033
401090096
401091037
401159004
401210415
401359021
401430137
401430174
401430178
401431127
480271047
480290032
480290052
480290059
480391004
480391016
480610006
480850005
481130069
481130075
481130087
481210034
481211032
481390016
481391044
481410029
481410055
481410057
481671034
481830001
482010024
lat
35.85
35.32
34.63
36.11
36.16
36.96
35.16
34.48
36.23
35.48
35.48
35.61
36.92
34.90
35.41
36.36
35.95
36.13
36.20
31.09
29.52
29.63
29.28
29.52
29.04
25.89
33.13
32.82
32.92
32.68
33.22
33.41
32.48
32.18
31.79
31.75
31.67
29.25
32.38
29.90
long
-94.99
-97.48
-98.43
-96.36
-98.93
-97.03
-97.47
-94.66
-95.25
-97.49
-97.30
-97.48
-94.84
-95.78
-94.52
-96.00
-96.00
-95.76
-95.98
-97.68
-98.62
-98.56
-98.31
-95.39
-95.47
-97.49
-96.79
-96.86
-96.81
-96.87
-97.20
-96.94
-97.03
-96.87
-106.32
-106.40
-106.29
-94.86
-94.71
-95.33
State
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Oklahoma
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
County
Cherokee
Cleveland
Comanche
Creek
Dewey
Kay
McClain
McCurtain
Mayes
Oklahoma
Oklahoma
Oklahoma
Ottawa
Pittsburg
Sequoyah
Tulsa
Tulsa
Tulsa
Tulsa
Bell
Bexar
Bexar
Bexar
Brazoria
Brazoria
Cameron
Collin
Dallas
Dallas
Dallas
Denton
Denton
Ellis
Ellis
El Paso
El Paso
El Paso
Galveston
Gregg
Harris
Base Case
66
64
66
65
66
64
63
62
68
66
65
67
64
65
63
67
65
66
67
64
69
71
62
??
65
58
72
71
72
71
74
72
67
63
58
63
63
70
73
74
Baseline
61
62
65
62
65
62
61
60
63
65
64
66
61
63
61
64
61
63
64
62
68
69
60
75
63
57
70
70
71
69
72
70
66
60
58
63
62
69
67
73
70
61
59
62
60
63
60
58
59
63
61
60
62
60
60
59
62
60
60
62
60
65
66
58
70
61
56
65
65
66
65
68
66
62
57
58
62
62
66
62
68
65
54
55
59
54
60
56
54
55
55
58
57
59
54
55
54
55
52
54
55
57
62
63
54
65
57
54
61
61
62
60
63
61
58
53
57
61
61
62
54
64
60
49
50
55
48
56
53
50
52
50
53
52
54
50
50
50
50
46
48
49
54
58
59
51
60
54
52
56
56
57
55
58
56
54
50
56
60
60
59
48
59
5A-37
-------
Site ID
482010026
482010029
482010046
482010047
482010051
482010055
482010062
482010066
482010070
482010075
482010416
482011015
482011034
482011035
482011039
482011050
482030002
482150043
482151048
482210001
482311006
482450009
482450011
482450022
482450101
482450102
482450628
482451035
482510003
482570005
483091037
483390078
483491051
483550025
483550026
483611001
483611100
483670081
483739991
483970001
lat
29.80
30.04
29.83
29.83
29.62
29.70
29.63
29.72
29.74
29.75
29.69
29.76
29.77
29.73
29.67
29.58
32.67
26.23
26.13
32.44
33.15
30.04
29.90
29.86
29.73
29.94
29.87
29.98
32.35
32.56
31.65
30.35
32.03
27.77
27.83
30.09
30.19
32.87
30.70
32.94
long
-95.13
-95.67
-95.28
-95.49
-95.47
-95.50
-95.27
-95.50
-95.32
-95.35
-95.29
-95.08
-95.22
-95.26
-95.13
-95.02
-94.17
-98.29
-97.94
-97.80
-96.12
-94.07
-93.99
-94.32
-93.89
-94.00
-93.96
-94.01
-97.44
-96.32
-97.07
-95.43
-96.40
-97.43
-97.56
-93.76
-93.87
-97.91
-94.67
-96.46
State
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
County
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Harrison
Hidalgo
Hidalgo
Hood
Hunt
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Johnson
Kaufman
McLennan
Montgomery
Navarro
Nueces
Nueces
Orange
Orange
Parker
Polk
Rockwall
Base Case
71
72
70
70
71
72
70
69
69
71
70
68
75
72
76
72
67
56
55
67
62
66
66
64
70
63
65
65
70
64
65
68
64
66
66
66
63
70
62
68
Baseline
69
71
69
69
70
71
69
68
68
69
69
67
73
71
74
71
63
55
55
66
61
64
65
62
69
62
63
63
68
61
63
67
61
64
64
64
60
68
61
66
70
65
67
64
64
65
66
64
63
63
64
64
63
68
66
n
67
58
54
53
62
57
61
61
59
66
58
60
59
64
58
60
63
58
62
62
61
58
64
59
62
65
61
63
60
60
60
61
59
59
59
60
59
59
63
61
65
63
52
52
52
58
54
56
57
55
62
54
56
55
60
53
55
59
53
59
58
57
53
61
56
59
60
56
59
56
55
55
56
54
54
54
55
54
54
58
56
60
59
47
51
50
54
50
52
53
51
58
50
52
51
56
50
52
55
50
56
55
52
49
57
54
54
5A-38
-------
Site ID
484230007
484390075
484391002
484392003
484393009
484393011
484530014
484530020
484690003
484790016
lat
32.34
32.99
32.81
32.92
32.98
32.66
30.35
30.48
28.84
27.51
long
-95.42
-97.48
-97.36
-97.28
-97.06
-97.09
-97.76
-97.87
-97.01
-99.52
State
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
Texas
*The design value from the monitor(s) with
shown in bold blue text
Table 3A-10. Design Values for Midwest
Site ID
170010007
170190007
170191001
170230001
170310001
170310032
170310064
170310076
170311003
170311601
170314002
170314007
170314201
170317002
170436001
170491001
170650002
170831001
170859991
170890005
170971007
171110001
171132003
171150013
lat
39.92
40.24
40.05
39.21
41.67
41.76
41.79
41.75
41.98
41.67
41.86
42.06
42.14
42.06
41.81
39.07
38.08
39.11
42.29
42.05
42.47
42.22
40.52
39.87
long
-91.34
-88.19
-88.37
-87.67
-87.73
-87.55
-87.60
-87.71
-87.79
-87.99
-87.75
-87.86
-87.80
-87.67
-88.07
-88.55
-88.62
-90.32
-90.00
-88.27
-87.81
-88.24
-89.00
-88.93
State
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
County
Smith
Tarrant
Tarrant
Tarrant
Tarrant
Tarrant
Travis
Travis
Victoria
Webb
the highest projected
Monitors
County
Adams
Champaign
Champaign
Clark
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
DuPage
Effingham
Hamilton
Jersey
Jo Daviess
Kane
Lake
McHenry
McLean
Macon
Base Case
67
73
71
??
75
72
65
63
63
60
ozone in each
Base Case
57
58
60
58
63
60
55
63
52
63
57
49
56
54
59
57
62
62
58
61
58
60
58
59
Baseline
64
71
70
75
73
70
64
61
60
59
scenario
Baseline
56
58
59
58
62
59
54
62
51
63
56
49
56
54
58
57
63
62
57
60
58
60
56
58
70
61
66
65
70
69
66
61
59
57
57
is
70
56
58
59
58
62
59
54
62
51
63
56
49
56
54
58
57
63
62
57
60
58
60
56
58
65
56
62
61
65
64
61
57
55
53
56
65
55
55
57
54
58
60
55
58
53
59
56
50
58
57
54
54
60
61
56
56
59
55
53
56
60
52
57
56
60
59
57
54
52
50
54
60
54
53
54
50
54
60
55
54
54
54
56
51
59
59
50
51
56
60
55
52
60
51
51
53
5A-39
-------
Site ID
171170002
171190008
171191009
171193007
171199991
171430024
171431001
171570001
171613002
171630010
171670014
171971011
172012001
180030002
180030004
180110001
180150002
180190008
180350010
180390007
180431004
180550001
180570006
180590003
180630004
180690002
180710001
180810002
180839991
180890022
180890030
180892008
180910005
180910010
180950010
180970050
180970057
180970073
180970078
181090005
lat
39.40
38.89
38.73
38.86
38.87
40.69
40.75
38.18
41.51
38.61
39.83
41.22
42.33
41.22
41.09
40.00
40.54
38.39
40.30
41.72
38.31
38.99
40.07
39.94
39.76
40.96
38.92
39.42
38.74
41.61
41.68
41.64
41.72
41.63
40.00
39.86
39.75
39.79
39.81
39.58
long
-89.81
-90.15
-89.96
-90.11
-89.62
-89.61
-89.59
-89.79
-90.52
-90.16
-89.64
-88.19
-89.04
-85.02
-85.10
-86.40
-86.55
-85.66
-85.25
-85.83
-85.83
-86.99
-85.99
-85.84
-86.40
-85.38
-86.08
-86.15
-87.49
-87.30
-87.49
-87.49
-86.91
-86.68
-85.66
-86.02
-86.19
-86.06
-86.11
-86.48
State
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
County
Macoupin
Madison
Madison
Madison
Madison
Peoria
Peoria
Randolph
Rock Island
Saint Clair
Sangamon
Will
Winnebago
Allen
Allen
Boone
Carroll
Clark
Delaware
Elkhart
Floyd
Greene
Hamilton
Hancock
Hendricks
Huntington
Jackson
Johnson
Knox
Lake
Lake
Lake
LaPorte
LaPorte
Madison
Marion
Marion
Marion
Marion
Morgan
Base Case
57
64
62
63
60
54
61
58
49
62
58
55
57
56
57
60
58
65
55
56
63
68
57
53
56
54
57
57
65
54
58
58
66
59
54
60
59
59
59
56
Baseline
56
63
61
62
59
51
58
57
48
61
57
54
56
56
56
60
57
65
54
55
63
67
57
53
55
54
57
57
65
54
58
58
66
59
54
60
58
59
59
56
70
56
63
61
62
59
51
58
57
48
61
57
54
56
56
56
60
57
65
54
55
63
67
57
53
55
54
57
57
65
54
58
58
66
59
54
60
58
59
59
56
65
55
62
60
61
58
48
55
55
47
60
56
50
53
53
53
56
54
59
51
51
58
62
53
49
52
51
52
53
60
52
56
56
62
56
50
56
54
55
55
52
60
54
49
58
49
57
45
52
53
45
59
55
46
50
49
50
51
50
53
47
48
52
57
49
45
48
48
47
48
55
50
54
54
59
52
46
51
50
50
50
47
5A-40
-------
Site ID
181230009
181270024
181270026
181290003
181410010
181410015
181411007
181450001
181630013
181630021
181670018
181670024
181699991
181730008
181730009
181730011
210130002
210150003
210190017
210290006
210373002
210430500
210470006
210590005
210610501
210670012
210890007
210910012
210930006
211010014
211110027
211110051
211110067
211130001
211390003
211451024
211759991
211850004
211930003
211950002
lat
38.11
41.62
41.51
38.01
41.55
41.70
41.74
39.61
38.11
38.01
39.49
39.56
40.82
38.05
38.19
37.95
36.61
38.92
38.46
37.99
39.02
38.24
36.91
37.78
37.13
38.07
38.55
37.94
37.71
37.87
38.14
38.06
38.23
37.89
37.16
37.06
37.92
38.40
37.28
37.48
long
-86.60
-87.20
-87.04
-87.72
-86.37
-86.21
-86.11
-85.87
-87.54
-87.58
-87.40
-87.31
-85.66
-87.28
-87.34
-87.32
-83.74
-84.85
-82.64
-85.71
-84.47
-82.99
-87.32
-87.08
-86.15
-84.50
-82.73
-86.90
-85.85
-87.46
-85.58
-85.90
-85.65
-84.59
-88.39
-88.57
-83.07
-85.44
-83.21
-82.54
State
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
County
Perry
Porter
Porter
Posey
St. Joseph
St. Joseph
St. Joseph
Shelby
Vanderburgh
Vanderburgh
Vigo
Vigo
Wabash
Warrick
Warrick
Warrick
Bell
Boone
Boyd
Bullitt
Campbell
Carter
Christian
Daviess
Edmonson
Fayette
Greenup
Hancock
Hardin
Henderson
Jefferson
Jefferson
Jefferson
Jessamine
Livingston
McCracken
Morgan
Oldham
Perry
Pike
Base Case
65
56
54
62
52
58
53
60
64
63
55
55
61
63
63
64
52
56
60
60
64
57
62
69
57
58
60
65
58
68
65
68
70
57
58
58
60
66
62
63
Baseline
65
56
54
62
51
57
53
60
63
63
55
55
60
63
62
64
52
56
60
60
64
56
61
68
57
58
60
65
58
68
65
68
10
59
62
63
60
66
62
63
70
65
56
54
62
51
57
53
60
63
63
55
55
60
63
62
64
52
56
60
60
64
56
61
68
57
58
60
65
58
68
65
68
70
59
62
63
60
66
62
63
65
60
54
51
58
48
53
49
56
59
59
51
51
57
59
58
59
49
51
54
56
58
52
58
64
54
54
54
60
53
63
60
62
64
55
58
60
55
61
57
57
60
54
52
47
53
45
49
45
51
54
54
47
47
53
54
53
54
46
46
49
51
52
47
55
58
50
50
49
54
49
58
54
57
58
51
54
57
50
55
52
52
JA-41
-------
Site ID
211990003
212130004
212218001
212219991
212270008
212299991
260050003
260190003
260210014
260270003
260370001
260490021
260492001
260630007
260650012
260770008
260810020
260810022
260910007
260990009
260991003
261010922
261050007
261130001
261210039
261250001
261390005
261470005
261530001
261579991
261610008
261619991
261630001
261630019
261659991
390030009
390071001
390090004
390170004
390170018
lat
37.10
36.71
36.78
36.78
37.04
37.70
42.77
44.62
42.20
41.90
42.80
43.05
43.17
43.84
42.74
42.28
42.98
43.18
42.00
42.73
42.51
44.31
43.95
44.31
43.28
42.46
42.89
42.95
46.29
43.61
42.24
42.42
42.23
42.43
44.18
40.77
41.96
39.31
39.38
39.53
long
-84.61
-86.57
-87.85
-87.85
-86.25
-85.05
-86.15
-86.11
-86.31
-86.00
-84.39
-83.67
-83.46
-82.64
-84.53
-85.54
-85.67
-85.42
-83.95
-82.79
-83.01
-86.24
-86.29
-84.89
-86.31
-83.18
-85.85
-82.46
-85.95
-83.36
-83.60
-83.90
-83.21
-83.00
-85.74
-84.05
-80.57
-82.12
-84.54
-84.39
State
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Kentucky
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Ohio
Ohio
Ohio
Ohio
Ohio
County
Pulaski
Simpson
Trigg
Trigg
Warren
Washington
Allegan
Benzie
Berrien
Cass
Clinton
Gene see
Gene see
Huron
Ingham
Kalamazoo
Kent
Kent
Lenawee
Macomb
Macomb
Manistee
Mason
Missaukee
Muskegon
Oakland
Ottawa
St. Clair
Schoolcraft
Tuscola
Washtenaw
Washtenaw
Wayne
Wayne
Wexford
Allen
Ashtabula
Athens
Butler
Butler
Base Case
53
54
57
58
50
56
69
61
68
63
57
61
60
61
57
60
60
59
61
66
68
60
61
58
66
66
63
64
60
58
62
61
61
69
56
61
62
57
66
66
Baseline
53
53
57
58
50
57
69
61
68
62
56
60
59
61
56
59
60
58
60
65
68
60
60
57
65
66
62
63
60
57
62
60
61
68
55
61
62
57
65
65
70
53
53
57
58
50
57
69
61
68
62
56
60
59
61
56
59
60
58
60
65
68
60
60
57
65
66
62
63
60
57
62
60
61
68
55
61
62
57
65
65
65
49
50
53
54
47
53
64
56
63
58
52
57
55
57
53
56
56
54
57
61
64
55
56
54
60
62
58
60
56
54
58
56
58
64
52
57
57
53
60
60
60
45
46
49
50
44
48
58
52
58
54
49
53
52
54
49
52
51
50
53
57
60
51
52
51
55
57
53
56
52
50
54
53
55
60
49
53
52
48
55
54
5A-42
-------
Site ID
390179991
390230001
390230003
390250022
390271002
390350034
390350060
390350064
390355002
390410002
390479991
390490029
390490037
390490081
390550004
390570006
390610006
390610010
390610040
390810017
390830002
390850003
390850007
390870011
390870012
390890005
390930018
390950024
390950027
390950034
390970007
390990013
391030004
391090005
391130037
391219991
391331001
391351001
391510016
391510022
lat
39.53
40.00
39.86
39.08
39.43
41.56
41.49
41.36
41.54
40.36
39.64
40.08
39.97
40.09
41.52
39.67
39.28
39.21
39.13
40.37
40.31
41.67
41.73
38.63
38.51
40.03
41.42
41.64
41.49
41.68
39.79
41.10
41.06
40.08
39.79
39.94
41.18
39.84
40.83
40.71
long
-84.73
-83.80
-84.00
-84.14
-83.79
-81.58
-81.68
-81.86
-81.46
-83.06
-83.26
-82.82
-82.96
-82.96
-81.25
-83.94
-84.37
-84.69
-84.50
-80.62
-82.69
-81.42
-81.24
-82.46
-82.66
-82.43
-82.10
-83.55
-83.72
-83.31
-83.48
-80.66
-81.92
-84.11
-84.13
-81.34
-81.33
-84.72
-81.38
-81.60
State
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
County
Butler
Clark
Clark
Clermont
Clinton
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Delaware
Fayette
Franklin
Franklin
Franklin
Geauga
Greene
Hamilton
Hamilton
Hamilton
Jefferson
Knox
Lake
Lake
Lawrence
Lawrence
Licking
Lorain
Lucas
Lucas
Lucas
Madison
Mahoning
Medina
Miami
Montgomery
Noble
Portage
Preble
Stark
Stark
Base Case
64
61
60
62
62
59
52
56
58
59
57
66
61
58
60
58
68
64
66
61
59
59
53
55
60
59
54
56
58
61
59
57
57
59
62
52
56
59
61
58
Baseline
63
60
60
62
61
58
51
56
58
59
57
66
61
58
60
57
68
64
66
60
59
59
53
55
60
58
53
55
58
61
58
57
57
59
62
51
56
58
61
58
70
63
60
60
62
61
58
51
56
58
59
57
66
61
58
60
57
68
64
66
60
59
59
53
55
60
58
53
55
58
61
58
57
57
59
62
51
56
58
61
58
65
59
56
55
57
56
58
51
55
57
55
52
61
56
54
56
53
62
58
60
57
55
59
52
50
54
54
53
54
54
59
54
53
53
54
57
48
52
55
57
53
60
54
51
50
51
51
57
51
54
57
51
48
56
52
49
52
48
56
53
54
53
51
58
52
45
49
49
53
53
51
56
49
50
49
50
52
44
48
51
52
49
5A-43
-------
Site ID
391514005
391530020
391550009
391550011
391650007
391670004
391730003
470010101
470090101
470090102
470259991
470370011
470370026
470419991
470651011
470654003
470890002
470930021
470931020
471050109
471210104
471490101
471550101
471550102
471570021
471570075
471571004
471632002
471632003
471650007
471650101
471870106
471890103
540030003
540110006
540219991
540250003
540291004
540390010
540610003
lat
40.93
41.11
41.45
41.24
39.43
39.43
41.38
35.97
35.63
35.60
36.47
36.21
36.15
36.04
35.23
35.10
36.11
36.09
36.02
35.72
35.29
35.73
35.70
35.56
35.22
35.15
35.38
36.54
36.58
36.30
36.45
35.95
36.06
39.45
38.42
38.88
37.91
40.42
38.35
39.65
long
-81.12
-81.50
-80.59
-80.66
-84.20
-81.46
-83.61
-84.22
-83.94
-83.78
-83.83
-86.74
-86.62
-85.73
-85.18
-85.16
-83.60
-83.76
-83.87
-84.34
-84.95
-86.60
-83.61
-83.50
-90.02
-89.85
-89.83
-82.42
-82.49
-86.65
-86.56
-87.14
-86.29
-77.96
-82.43
-80.85
-80.63
-80.58
-81.63
-79.92
State
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Ohio
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
Tennessee
West Virginia
West Virginia
West Virginia
West Virginia
West Virginia
West Virginia
West Virginia
County
Stark
Summit
Trumbull
Trumbull
Warren
Washington
Wood
Anderson
Blount
Blount
Claiborne
Davidson
Davidson
DeKalb
Hamilton
Hamilton
Jefferson
Knox
Knox
Loudon
Meigs
Rutherford
Sevier
Sevier
Shelby
Shelby
Shelby
Sullivan
Sullivan
Sumner
Sumner
Williamson
Wilson
Berkeley
Cabell
Gilmer
Greenbrier
Hancock
Kanawha
Monongalia
Base Case
59
59
57
61
62
60
60
56
61
53
49
51
54
55
57
57
59
54
56
59
56
52
59
58
63
63
61
62
61
59
55
54
55
56
60
53
55
63
63
63
Baseline
58
59
56
61
62
59
60
56
61
53
48
51
54
55
56
56
59
54
56
58
56
52
58
58
62
63
59
62
61
58
54
54
55
55
59
53
55
63
63
62
70
58
59
56
61
62
59
60
56
61
53
48
51
54
55
56
56
59
54
56
58
56
52
58
58
62
63
59
62
61
58
54
54
55
55
59
53
55
63
63
62
65
54
55
53
57
57
55
56
52
56
49
45
47
50
51
53
53
54
50
52
54
53
48
55
56
57
58
56
57
56
54
51
49
51
53
54
48
51
59
57
57
60
50
50
49
53
51
50
52
47
51
44
42
43
46
48
49
49
50
45
47
49
50
44
51
53
53
54
52
52
52
50
47
45
47
50
49
43
47
56
51
52
5A-44
-------
Site ID
540690010
540939991
541071002
550090026
550210015
550250041
550270001
550290004
550350014
550390006
550410007
550550002
550590019
550610002
550630012
550710007
550730012
550790010
550790026
550790085
550870009
550890008
550890009
551010017
551050024
551110007
551170006
551199991
551270005
551330027
lat
40.11
39.09
39.32
44.53
43.32
43.10
43.47
45.24
44.76
43.69
45.56
43.00
42.50
44.44
43.78
44.14
44.71
43.02
43.06
43.18
44.31
43.34
43.50
42.71
42.51
43.44
43.68
45.21
42.58
43.02
long State
-80.
-79.
-81.
-87.
-89.
-89.
-88.
-86.
-91.
-88.
-88.
-88.
-87.
-87.
-91.
-87.
-89.
-87.
-87.
-87.
-88.
-87.
-87.
-87.
-89.
-89.
-87.
-90.
-88.
-88.
70 West Virginia
66 West Virginia
55 West Virginia
9 1 Wisconsin
1 1 Wisconsin
36 Wisconsin
62 Wisconsin
99 Wisconsin
14 Wisconsin
42 Wisconsin
8 1 Wisconsin
82 Wisconsin
8 1 Wisconsin
5 1 Wisconsin
23 Wisconsin
62 Wisconsin
77 Wisconsin
93 Wisconsin
9 1 Wisconsin
90 Wisconsin
40 Wisconsin
92 Wisconsin
8 1 Wisconsin
80 Wisconsin
06 Wisconsin
68 Wisconsin
72 Wisconsin
60 Wisconsin
50 Wisconsin
22 Wisconsin
County
Ohio
Tucker
Wood
Brown
Columbia
Dane
Dodge
Door
Eau Claire
Fond du Lac
Forest
Jefferson
Kenosha
Kewaunee
La Crosse
Manitowoc
Marathon
Milwaukee
Milwaukee
Milwaukee
Outagamie
Ozaukee
Ozaukee
Racine
Rock
Sauk
Sheboygan
Taylor
Walworth
Waukesha
*The design value from the monitor(s) with the highest projected
shown in bold blue text
Table 3A-11. Design Values for Northeast Monitors
Site ID
90010017
90011123
90013007
90019003
90031003
lat long
41.00
41.40
41.15
41.12
41.78
-73.59
-73.44
-73.10
-73.34
-72.63
State
Connecticut
Connecticut
Connecticut
Connecticut
Connecticut
County
Fairfield
Fairfield
Fairfield
Fairfield
Hartford
Base Case Baseline
61
56
57
59
56
55
62
63
52
61
54
58
60
63
54
66
54
56
60
65
59
66
62
57
59
54
71
54
59
57
ozone in
Base
Case
60
56
56
58
55
54
61
63
51
60
53
57
59
63
53
65
53
55
60
64
58
66
62
57
58
53
70
53
58
57
each scenario is
Baseline
70 70
67 67
73 73
74 74
63 63
70
60
56
56
58
55
54
61
63
51
60
53
57
59
63
53
65
53
55
60
64
58
66
62
57
58
53
70
53
58
57
70
66
63
68
69
59
65
56
52
52
55
52
52
58
59
50
57
51
55
60
59
52
60
51
53
57
61
56
62
58
57
55
51
65
52
55
53
65
61
58
63
65
54
60
52
48
47
51
50
49
54
54
49
53
48
52
60
54
51
55
49
51
54
57
53
58
53
56
52
49
60
50
52
50
60
56
54
58
59
50
5A-45
-------
Site ID
90050005
90070007
90090027
90099002
90110124
90131001
90159991
100010002
100031007
100031010
100031013
100032004
100051002
100051003
110010041
110010043
230010014
230052003
230090102
230090103
230112005
230130004
230173001
230194008
230230006
230290019
230290032
230310038
230310040
230312002
240030014
240051007
240053001
240090011
240130001
240150003
240170010
240199991
240210037
lat
41.82
41.55
41.30
41.26
41.35
41.98
41.84
38.98
39.55
39.82
39.77
39.74
38.64
38.78
38.90
38.92
43.97
43.56
44.35
44.38
44.23
43.92
44.25
44.74
44.01
44.53
44.96
43.66
43.59
43.34
38.90
39.46
39.31
38.54
39.44
39.70
38.50
38.45
39.42
long
-73.30
-72.63
-72.90
-72.55
-72.08
-72.39
-72.01
-75.56
-75.73
-75.56
-75.50
-75.56
-75.61
-75.16
-76.95
-77.01
-70.12
-70.21
-68.23
-68.26
-69.79
-69.26
-70.86
-68.67
-69.83
-67.60
-67.06
-70.63
-70.88
-70.47
-76.65
-76.63
-76.47
-76.62
-77.04
-75.86
-76.81
-76.11
-77.38
State
Connecticut
Connecticut
Connecticut
Connecticut
Connecticut
Connecticut
Connecticut
Delaware
Delaware
Delaware
Delaware
Delaware
Delaware
Delaware
D.C
D.C.
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maine
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
County p Baseline
Litchfield
Middlesex
New Haven
New Haven
New London
Tolland
Windham
Kent
New Castle
New Castle
New Castle
New Castle
Sussex
Sussex
D.C.
D.C.
Androscoggin
Cumberland
Hancock
Hancock
Kennebec
Knox
Oxford
Penobscot
Sagadahoc
Washington
Washington
York
York
York
Anne Arundel
Baltimore
Baltimore
Calvert
Carroll
Cecil
Charles
Dorchester
Frederick
58
65
64
72
67
63
58
59
59
61
62
60
61
64
59
62
51
58
59
56
52
56
46
48
50
50
47
50
53
60
64
65
66
63
60
65
62
60
63
58
65
64
72
67
63
58
59
59
61
62
60
61
64
59
62
51
58
59
56
52
56
46
48
50
50
47
50
53
60
64
65
66
63
60
65
62
60
63
70
54
61
60
67
63
59
55
55
55
56
57
55
58
60
55
58
48
55
56
53
49
53
45
45
47
48
45
47
50
57
60
60
61
59
57
61
58
56
59
65 60
50
56
56
63
59
55
51
51
51
52
53
51
54
57
50
53
44
50
53
49
45
49
43
42
44
45
43
44
47
52
55
55
56
54
54
57
54
51
57
46
52
51
57
54
50
47
48
48
48
49
48
51
53
46
49
40
46
50
46
42
46
42
39
40
43
41
41
44
48
51
51
51
50
51
53
51
47
54
5A-46
-------
Site ID
240230002
240251001
240259001
240290002
240313001
240330030
240338003
240339991
240430009
245100054
250010002
250034002
250051002
250070001
250092006
250094005
250095005
250130008
250150103
250154002
250170009
250171102
250213003
250250041
250250042
250270015
250270024
330012004
330050007
330074001
330074002
330090010
330111011
330115001
330131007
330150014
330150016
lat
39.71
39.41
39.56
39.31
39.11
39.06
38.81
39.03
39.57
39.33
41.98
42.64
41.63
41.33
42.47
42.81
42.77
42.19
42.40
42.30
42.63
42.41
42.21
42.32
42.33
42.27
42.10
43.57
42.93
44.27
44.31
43.63
42.72
42.86
43.22
43.08
43.05
long
-79.01
-76.30
-76.20
-75.80
-77.11
-76.88
-76.74
-76.82
-77.72
-76.55
-70.02
-73.17
-70.88
-70.79
-70.97
-70.82
-71.10
-72.56
-72.52
-72.33
-71.36
-71.48
-71.11
-70.97
-71.08
-71.88
-71.62
-71.50
-72.27
-71.30
-71.22
-72.31
-71.52
-71.88
-71.51
-70.75
-70.71
State
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Maryland
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
Massachusetts
New Hampshire
New Hampshire
New Hampshire
New Hampshire
New Hampshire
New Hampshire
New Hampshire
New Hampshire
New Hampshire
New Hampshire
County p Baseline
Garrett
Harford
Harford
Kent
Montgomery
Prince
George's
Prince
George's
Prince
George's
Washington
Baltimore
(City)
Barnstable
Berkshire
Bristol
Dukes
Essex
Essex
Essex
Hampden
Hampshire
Hampshire
Middlesex
Middlesex
Norfolk
Suffolk
Suffolk
Worcester
Worcester
Belknap
Cheshire
Coos
Coos
Grafton
Hillsborough
Hillsborough
Merrimack
Rockingham
Rockingham
60
73
63
62
61
61
64
62
59
62
60
58
60
65
58
57
57
60
53
58
55
55
59
57
50
56
56
52
51
58
51
50
54
58
53
55
55
60
73
63
62
61
61
64
62
59
62
60
58
60
65
58
57
57
60
53
58
55
55
59
57
50
56
56
52
51
58
51
50
54
58
53
55
55
70
59
68
58
57
56
57
59
58
57
57
57
55
57
61
56
54
54
57
50
54
52
52
56
55
48
53
53
50
48
57
50
48
51
55
50
52
52
65 60
58
62
53
53
52
52
55
53
55
52
53
52
53
57
53
51
50
52
46
50
49
48
53
52
46
50
49
47
46
56
49
46
48
51
47
48
48
58
57
49
49
48
48
50
49
54
48
48
49
49
53
50
47
46
48
43
47
45
45
50
48
43
46
46
45
43
55
48
44
45
49
45
44
45
5A-47
-------
Site ID
330150018
340010006
340030006
340071001
340110007
340130003
340150002
340170006
340190001
340210005
340219991
340230011
340250005
340273001
340290006
340315001
340410007
360010012
360050133
360130006
360130011
360150003
360270007
360290002
360310002
360310003
360410005
360430005
360450002
360530006
360551007
360610135
360631006
360650004
360671015
360715001
360750003
360790005
360810124
lat
42
39
40
39
39
40
39
40
40
40
40
40
40
40
40
41
40
42
40
42
42
42
41
42
44
44
43
43
44
42
43
40
43
43
43
41
43
41
40
.86
.46
.87
.68
.42
.72
.80
.67
.52
.28
.31
.46
.28
.79
.06
.06
.92
.68
.87
.50
.29
.11
.79
.99
.37
.39
.45
.69
.09
.73
.15
.82
.22
.30
.05
.52
.28
.46
.74
long
-71.38
-74.45
-73.99
-74.86
-75.03
-74.19
-75.21
-74.13
-74.81
-74.74
-74.87
-74.43
-74.01
-74.68
-74.44
-74.26
-75.07
-73.76
-73.88
-79.32
-79.59
-76.80
-73.74
-78.77
-73.90
-73.86
-74.52
-74.99
-75.97
-75.78
-77.55
-73.95
-78.48
-75.72
-76.06
-74.22
-76.46
-73.71
-73.82
State
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
New
Hampshire
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
Jersey
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
York
County p Baseline
Rockingham
Atlantic
Bergen
Camden
Cumberland
Essex
Gloucester
Hudson
Hunterdon
Mercer
Mercer
Middlesex
Monmouth
Morris
Ocean
Passaic
Warren
Albany
Bronx
Chautauqua
Chautauqua
Chemung
Dutchess
Erie
Essex
Essex
Hamilton
Herkimer
Jefferson
Madison
Monroe
New York
Niagara
Oneida
Onondaga
Orange
Oswego
Putnam
Queens
56
61
64
67
58
64
68
64
63
65
62
66
66
60
67
61
52
57
64
62
62
57
58
61
57
57
57
55
62
55
59
64
64
53
60
56
58
58
71
56
61
64
67
58
64
68
64
63
65
62
66
66
60
67
61
52
57
64
62
62
57
58
61
57
57
57
55
62
55
59
64
64
53
60
56
58
58
71
70
53
57
59
62
54
60
63
59
58
60
58
61
61
56
62
56
48
52
60
61
60
55
54
60
56
55
55
54
60
53
58
61
63
52
58
52
56
54
67
65 60
49
54
56
58
50
55
58
55
55
56
54
56
56
53
57
53
45
49
56
59
59
53
50
60
55
55
54
53
59
51
57
58
62
50
56
48
55
50
65
46
50
52
53
46
52
54
51
51
52
50
52
51
50
53
49
43
46
52
58
57
52
46
60
54
54
52
52
58
49
56
55
58
49
54
44
53
46
60
5A-48
-------
Site ID
360830004
360850067
360870005
360910004
360930003
361010003
361030002
361030004
361030009
361099991
361111005
361173001
361192004
420010002
420019991
420030008
420030010
420030067
420031005
420050001
420070002
420070005
420070014
420110006
420110011
420130801
420170012
420210011
420270100
420279991
420290100
420334000
420430401
420431100
420450002
420479991
420490003
420550001
420590002
lat
42.78
40.60
41.18
43.01
42.80
42.09
40.75
40.96
40.83
42.40
42.14
43.23
41.05
39.93
39.92
40.47
40.45
40.38
40.61
40.81
40.56
40.68
40.75
40.51
40.38
40.54
40.11
40.31
40.81
40.72
39.83
41.12
40.25
40.27
39.84
41.60
42.14
39.96
39.81
long
-73.46
-74.13
-74.03
-73.65
-73.94
-77.21
-73.42
-72.71
-73.06
-76.65
-74.49
-77.17
-73.76
-77.25
-77.31
-79.96
-80.02
-80.17
-79.73
-79.56
-80.50
-80.36
-80.32
-75.79
-75.97
-78.37
-74.88
-78.92
-77.88
-77.93
-75.77
-78.53
-76.85
-76.68
-75.37
-78.77
-80.04
-77.48
-80.27
State
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
New York
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
County p Baseline
Rensselaer
Richmond
Rockland
Saratoga
Schenectady
Steuben
Suffolk
Suffolk
Suffolk
Tompkins
Ulster
Wayne
Westchester
Adams
Adams
Allegheny
Allegheny
Allegheny
Allegheny
Armstrong
Beaver
Beaver
Beaver
Berks
Berks
Blair
Bucks
Cambria
Centre
Centre
Chester
Clearfield
Dauphin
Dauphin
Delaware
Elk
Erie
Franklin
Greene
57
71
62
56
54
57
75
67
72
58
59
56
63
57
59
67
65
65
71
64
63
67
64
59
62
66
66
62
64
65
61
65
59
63
61
56
60
56
59
57
71
62
56
54
57
75
67
72
58
59
56
63
57
59
67
65
65
71
64
63
67
64
59
62
66
66
62
64
65
61
65
59
63
61
56
60
56
59
70
53
67
58
52
51
55
70
63
67
56
56
55
58
53
55
65
63
63
68
61
62
65
63
54
57
61
61
58
60
61
55
61
54
57
57
54
60
53
58
65 60
50
63
53
49
48
54
65
58
62
55
53
53
54
51
52
62
60
62
65
59
62
63
61
51
54
58
57
55
57
58
51
58
51
53
52
52
59
51
57
47
58
49
47
45
53
60
52
57
54
50
52
50
49
50
60
58
60
58
57
56
57
60
48
51
55
53
52
55
55
47
56
48
50
49
51
59
49
57
5A-49
-------
Site ID
420630004
420690101
420692006
420710007
420710012
420730015
420750100
420770004
420791100
420791101
420810100
420850100
420859991
420890002
420910013
420950025
420958000
420990301
421010004
421010024
421011002
421119991
421174000
421250005
421250200
421255001
421290006
421290008
421330008
421330011
440030002
440071010
440090007
500030004
500070007
510030001
510130020
510330001
510360002
lat
40.56
41.48
41.44
40.05
40.04
41.00
40.34
40.61
41.21
41.27
41.25
41.22
41.43
41.08
40.11
40.63
40.69
40.46
40.01
40.08
40.04
39.99
41.64
40.15
40.17
40.45
40.43
40.30
39.97
39.86
41.62
41.84
41.50
42.89
44.53
38.08
38.86
38.20
37.34
long
-78.92
-75.58
-75.62
-76.28
-76.11
-80.35
-76.38
-75.43
-76.00
-75.85
-76.92
-80.48
-80.15
-75.32
-75.31
-75.34
-75.24
-77.17
-75.10
-75.01
-75.00
-79.25
-76.94
-79.90
-80.26
-80.42
-79.69
-79.51
-76.70
-76.46
-71.72
-71.36
-71.42
-73.25
-72.87
-78.50
-77.06
-77.38
-77.26
State
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Rhode Island
Rhode Island
Rhode Island
Vermont
Vermont
Virginia
Virginia
Virginia
Virginia
County p Baseline
Indiana
Lackawanna
Lackawanna
Lancaster
Lancaster
Lawrence
Lebanon
Lehigh
Luzerne
Luzerne
Lycoming
Mercer
Mercer
Monroe
Montgomery
Northampton
Northampton
Perry
Philadelphia
Philadelphia
Philadelphia
Somerset
Tioga
Washington
Washington
Washington
Westmoreland
Westmoreland
York
York
Kent
Providence
Washington
Bennington
Chittenden
Albemarle
Arlington
Caroline
Charles
67
60
58
66
64
61
63
62
55
54
57
62
55
54
63
61
57
59
55
69
67
55
60
61
60
61
63
61
61
62
61
61
64
54
52
54
65
57
61
67
60
58
66
64
61
63
62
55
54
57
62
55
54
63
61
57
59
55
69
67
55
60
61
60
61
63
61
61
62
61
61
64
54
52
54
65
57
61
70
63
56
54
57
57
59
58
57
51
51
53
61
54
50
58
56
52
55
51
65
62
53
57
59
59
60
60
58
53
55
57
58
60
51
51
52
60
53
56
65 60
60
53
51
53
53
57
54
53
47
47
51
60
53
47
55
53
49
53
48
60
58
51
54
57
57
59
57
55
50
51
52
54
56
48
50
50
55
49
51
57
50
49
50
50
55
51
50
44
44
48
59
52
45
51
49
46
51
44
56
54
50
52
56
56
58
55
53
46
47
48
50
51
46
49
49
51
45
46
5A-50
-------
Site ID
510410004
510590030
510610002
510690010
510719991
510850003
510870014
511071005
511130003
511390004
511479991
511530009
511611004
511630003
511650003
511790001
511970002
515100009
516500008
518000004
518000005
lat
37.36
38.77
38.47
39.28
37.33
37.61
37.56
39.02
38.52
38.66
37.17
38.85
37.28
37.63
38.48
38.48
36.89
38.81
37.10
36.90
36.67
long
-77.59
-77.10
-77.77
-78.08
-80.56
-77.22
-77.40
-77.49
-78.44
-78.50
-78.31
-77.63
-79.88
-79.51
-78.82
-77.37
-81.25
-77.04
-76.39
-76.44
-76.73
State
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
Virginia
County p Baseline
Chesterfield
Fairfax
Fauquier
Frederick
Giles
Hanover
Henrico
Loudoun
Madison
Page
Prince Edward
Prince
William
Roanoke
Rockbridge
Rockingham
Stafford
Wythe
Alexandria
City
Hampton City
Suffolk City
Suffolk City
58
65
50
54
51
58
61
60
59
55
52
58
53
51
55
57
57
63
58
59
56
58
65
50
54
51
58
61
60
59
55
52
58
53
51
55
57
57
63
58
59
56
70
52
60
48
52
50
53
55
56
57
54
48
55
52
50
53
53
56
59
55
57
54
65 60
48
55
45
51
49
49
50
53
56
53
46
52
49
48
53
48
55
54
51
53
52
44
51
43
50
49
45
46
49
56
52
44
49
47
47
52
44
55
49
47
49
50
*The design value from the monitor(s) with the highest projected ozone in each scenario is
shown in bold blue text
3A.7 Monitors Excluded from the Quantitative Analysis
There were 1219 ozone monitors with complete ozone data for at least one DV period
covering the years 2009-2013. Of those sites, we quantitatively analyzed 1150 in this analysis.
As discussed in chapter 3, 69 sites were excluded from the quantitative analysis of emissions
reductions needed to reach alternative standard levels. These sites fall into one of three
categories, as discussed in more detail in the following three subsections.
3A. 7.1 Sites without Projections Due to Insufficient Days
Some monitors were excluded from the analysis because no future design value could be
projected at the site. This occurred when there were not enough modeled high ozone days (4 or
JA-51
-------
fewer) at the site to compute a design value according to EPA SIP modeling guidance. A list of
the 36 sites falling into this category is given in Table 3A-12.
Table 3A-12. Monitors Without Projections due to Insufficient High Modeling Days to
Meet EPA Guidance for Projecting Design Values
Site ID
60231004
60450008
60750005
60932001
160230101
230031100
260330901
270052013
270177416
270750005
270834210
271370034
300298001
300490004
311079991
380070002
380130004
380150003
380171004
380250003
380530002
380570004
380650002
410170122
410290201
410591003
460110003
530090013
530330080
530530012
530570020
530730005
550030010
551250001
560390008
lat
40.78
39.15
37.77
41.73
43.46
46.70
46.49
46.85
46.71
47.95
44.44
48.41
48.51
46.85
42.83
46.89
48.64
46.83
46.93
47.31
47.58
47.30
47.19
44.02
42.23
45.83
44.35
48.30
47.57
46.78
48.40
48.95
46.60
46.05
43.67
long
-124.18
-123.20
-122.40
-122.63
-113.56
-68.03
-84.36
-95.85
-92.52
-91.50
-95.82
-92.83
-114.00
-111.99
-97.85
-103.38
-102.40
-100.77
-96.86
-102.53
-103.30
-101.77
-101.43
-121.26
-122.79
-119.26
-96.81
-124.62
-122.31
-121.74
-122.50
-122.55
-90.66
-89.65
-110.60
State
California
California
California
California
Idaho
Maine
Michigan
Minnesota
Minnesota
Minnesota
Minnesota
Minnesota
Montana
Montana
Nebraska
North Dakota
North Dakota
North Dakota
North Dakota
North Dakota
North Dakota
North Dakota
North Dakota
Oregon
Oregon
Oregon
South Dakota
Washington
Washington
Washington
Washington
Washington
Wisconsin
Wisconsin
Wyoming
County
Humboldt
Mendocino
San Francisco
Siskiyou
Butte
Aroostook
Chippewa
Becker
Carlton
Lake
Lyon
Saint Louis
Flathead
Lewis and Clark
Knox
Billings
Burke
Burleigh
Cass
Dunn
McKenzie
Mercer
Oliver
Deschutes
lackson
Umatilla
Brookings
Clallam
King
Pierce
Skagit
Whatcom
Ashland
Vilas
Teton
5A-52
-------
Site ID
560391011
lat
44.56
long
-110.40
State
Wyoming
County
Teton
3A. 7.2 Winter Ozone
As discussed in Chapter 2 of the RIA, high winter ozone concentrations that have been
observed in mountain valleys in the Western U.S. are believed to result from the combination of
strong wintertime inversions, large NOx and VOC emissions from nearby oil and gas operations,
increased UV intensity due to reflection off of snow surfaces and potentially still uncharacterized
sources of free radicals. Current modeling tools are not sufficient to properly characterize ozone
formation for these winter ozone episodes due to 1) the challenging task of capturing complex
local "cold pool" meteorology using a model resolution that is optimized to capture regional and
synoptic scale process, 2) uncertainties in quantifying the local emissions from oil and gas
operations and 3) uncertainties in the chemistry that occurs both in the atmosphere and on snow
surfaces during these episodes. Therefore, it was not appropriate to project ozone design values
at monitors impacted by winter events. To identify sites impacted by winter events, we examined
the ambient data that went into creating the 2009-2013 5-year weighted design value in locations
known to have conditions favorable for winter ozone formation (i.e. all sites in Wyoming, Utah,
and Colorado). At these sites, we evaluated the four highest 8-hr daily maximum ozone values in
each year from 2009-2013 to identify wintertime ozone episodes. A site was categorized as
having a design value impacted by wintertime ozone if at least 20% of the days examined (4 out
20) had ozone values greater than or equal to 75 ppb and occurred during a "winter" month
(November-March). The seven sites identified as being affected by wintertime ozone events are
listed in Table 3A-13.
Table 3A-13. Monitors Determined to Have Design Values Affected by Winter Ozone
Events
Site ID
081030006
560130099
560350097
lat
40.09
42.53
42.98
long
-108.76
-108.72
-110.35
State
Colorado
Wyoming
Wyoming
County
Rio
Blanco
Fremont
Sublette
#of
summer
DV
days*
>=75
0
1
0
# of winter
DV days*
>=75
7
4
3
highest
max
106
93
83
2009-
2013
DV
71
67
64
5A-53
-------
560350099 42.72 -109.75 Wyoming Sublette 0 4 123 77
560350100 42.79 -110.06 Wyoming Sublette 04 84 67
560350101 42.87 -109.87 Wyoming Sublette 04 89 66
560351002 42.37 -109.56 Wyoming Sublette 04 94 68
*DV days defined here are the days with the 4 highest 8-hr daily maximum ozone values in each year from 2009-
2013 (20 days).
3A. 7.3 Monitoring Sites in Rural/Remote Areas of the West and Southwest
As mentioned in chapter 3 of the RIA, model-predicted ozone concentrations at 26 sites in
rural/remote areas in the West and Southwest were excluded from the quantitative analysis (see
list of sites in Table 3A-14). All of these 26 monitoring sites have 2025 baseline concentrations
below 70 ppb, which is the upper end of the NAAQS range being proposed by the EPA.
Therefore, no emissions reductions would be required for these sites for a primary standard of 70
ppb. Furthermore, only 15 of these sites would exceed a standard of 65 ppb. The remaining 11
would only exceed a standard of 60 ppb, which is not within the EPA's proposed range.
These 26 sites have two common characteristics. First, they have small modeled response
to large regional NOx and VOC reductions in 2025 compared to other sites in the region.
Second, these monitors would have DVs that remain above the standard after applying
reductions needed to bring large urban areas in the region into attainment. Figure 3 A-16 shows
the response of design values at all sites in the Southwest region to a 75% NOx reduction in this
region. Although design values at many urban sites drop by more than 10 ppb beyond the 2025
base case values, the response at the more remote and rural sites to modeled NOx reductions is
relatively small. Many of these sites do show response between 2009-2013 DVs and base case
2025 DVs (up to 15 ppb decreases) suggesting that national on-the-books controls and proposed
EPA rules could lower ozone DVs in these areas. However, modeling of additional NOx
reductions within the region provide little incremental benefit suggesting that most of the
regional anthropogenic sources impacting ozone at these locations have already been accounted
for in the 2025 base case scenario.
5A-54
-------
Change in O3 DV
• >10ppb
O 5-10ppb
O 3 - 5 ppb
O 1 -3 ppb
O < 1 ppb
Figure 3A-16.Projected change in 2025 ozone design values with an additional 75%
regional NOx control (Southwest region; stars represent sites identified in Table
3A-14)
A variety of influences including transport from California and other regional transport,
cross-border pollution from Mexico, and exceptional events (e.g., wildfires and stratospheric
intrusions) could contribute to ozone concentrations at the 26 sites. Each of these contributors is
described further below, along with the Clean Air Act provisions that offer varying degrees of
regulatory relief.
We have qualitatively characterized the predominant ozone influence for each site in Table
3 A-14. These qualitative characterizations are based on the modeled response to large regional
NOx reductions in 2025, proximity to the Mexican border (i.e., potential influence from trans-
border pollution) and altitude (e.g., potential influence of ozone transported from the free
troposphere: stratospheric intrusions or long range transport of international anthropogenic
ozone). Figure 3A-17 shows the location of all sites listed in Table 3A-14 and for demonstrative
purposes assigns each site to a category based on the predominant source of ozone in that
location. As the table and figure indicate, all 26 sites have 2025 baseline design values below 70
5A-55
-------
ppb, 15 sites have design values between 65-70 ppb, and 11 sites have design values between 60-
65 ppb. Of the 26 sites, 12 sites are characterized as border sites, 8 sites are characterized as
being strongly influenced by California emissions, and 6 sites are influenced by other ozone
sources.
Table 3A-14. Monitors with Limited Response to Regional NOx and National VOC
Emissions Reductions in the 2025 Baseline
Name Site ID State
Chiricahua NM 40038001 Arizona
Grand Canyon NP 40058001 Arizona
Alamo Lake 40128000 Arizona
Yuma Supersite 40278011 Arizona
El Centro-9th st 60251003 California
Death Valley NM 60270101 California
Yosemite NP 60430003 California
Sequoia and Kings ^^ ^.^^
Canyon NP
Gothic 80519991 Colorado
Weminuche 80671004 ^^
Wilderness Area
Great Basin NP 320330101 Nevada
Sunland Park City 350130Q17 New
Yard Mexico
3 Miles N of El 350130020 New
Paso Mexico
Altitude Monitor
County
(m) Type
Cochise 1570 CASTNET
Coconino 2152 CASTNET
La Paz 376 SLAMS
Yuma 51 SLAMS
Imperial - SLAMS
Non-EPA
Inyo 125 Federal
(NPS)
Mariposa 5265 CASTNET
Non-EPA
Tulare 1890 Federal
(NPS)
Gunnison 2926 CASTNET
Non-EPA
La Plata 2367 Federal
(USFS)
White Pine 2060 CASTNET
Dona Ana - SLAMS
Dona Ana 1250 SLAMS
Predominant
O3 Sources
Mexican border
California +
Other sources
California
Mexican border
+ California
California +
Mexican Border
California +
Other sources
California +
Other sources
California +
Other sources
Other sources
Southwest
region +
Other sources
California +
Other sources
Central region +
Mexican border
Central region +
Mexican border
2009-
2013
DV
72
71
71
75
81
71
77
81
66
72
72
66
67
Baseline
DV
67
66
65
67
66
66
68
67
64
68
66
63
62
5A-56
-------
Altitude Monitor
Name Site ID State County . .
(m) Type
2MlfromMT 350130021 New ^ ^ mg ^^
Cristo Rey Mexico
US-Mexico Border 350130Q22 New Dona Ana 12gQ SLAMS
Crossing Mexico
BLM land near 350151005 New 7gQ SLAMS
Carlsbad Mexico
Big Bend NP 480430101 Texas Brewster 1052 CASTNET
El Paso UTEP 481410037 Texas El Paso 1158 SLAMS
Skyline Park 481410044 Texas El Paso 1158 SLAMS
El PasoChamizal 481410058 Texas El Paso 1201 SLAMS
BLM,. , , , 483819991 Texas Randall 780 SLAMS
Land/Carlsbad
Canyonlands NP 490370101 Utah San Juan 1814 CASTNET
North Lava Flow 490530006 Utah Washington 846 SLAMS
Non-EPA
Zion NP 490530130 Utah Washington 1213 Federal
(NPS)
Centennial 560019991 Wyoming Albany 3178 CASTNET
Pinedale 560359991 Wyoming Sublette 2388 CASTNET
2009-
Predommant
03 Sources Dv
Central region +
Mexican border
Central region +
Mexican border
Central region +
Southwest
region +
Mexican border
Mexican border 70
Central region +
Mexican border
Central region +
Mexican border
Central region +
Mexican border
Central region +
Mexican border 73
+ Other sources
Other sources 68
California 68
California +
Other sources
Other sources 69
Southwest
region + Other 65
sources
Baseline
DV
67
66
66
69
67
65
65
66
64
62
65
65
62
5A-57
-------
v—-s
: * 7
^ Border impacts (12)
Ł California impacts (8)
0: Other sources (6)
* Other sites > 60 ppb in 2025 baseline (119)
» Other sites <= 60 ppb in 2025 baseline (214)
Souces Esri. USGS. NOAA
Figure 3A-17. Location of sites identified in Table 3A-14
In Figure 3 A-17, the colored dots categorize sites by the predominant source of ozone.
Many sites may be influenced by more than one source but are placed in a single category for
illustrative purposes in the Figure. All ozone monitoring sites categorized as not substantially
affected by natural or transported influences in Table 3 A-14 are shown as small diamonds. Gray
diamonds represent sites that had DVs less than or equal to 60 ppb in the 2025 baseline (or post-
2025 baseline for California sites). Black diamonds represent sites that had DVs greater than 60
ppb in the 2025 baseline (or post-2025 baseline for California sites).
In this section we look at examples of how these sites might leverage various Clean Air
Act provisions to comply with requirements for lower alternative standard levels including:
interstate transport provisions, exceptional events demonstrations, rural transport designations,
and requirements for nonattainment areas in international border areas (i.e., section 179B of the
Clean Air Act).
5A-58
-------
Clean Air Act sections 110 and 126 have provisions designed to reduce significant
transport contributions from upwind areas to downwind nonattainment areas. Although the RIA
accounted for the impacts of regional NOx reductions it was not able to account for the impacts
of emissions reductions in California on ozone at downwind sites in other states since California
is likely to have an attainment date that is later than that of other western states. As discussed in
the main RIA, many areas of California will not be required to meet the current ozone standard
until after 2025 (i.e., 2027 or later) and may not be required to meet a new ozone standard until
sometime between 2032 and 203.25 Although California will likely implement some of the
emissions reductions necessary to meet these standards prior to their attainment date, there is a
considerable uncertainty about how to quantify the portion of the emissions reductions that may
occur by 2025. Therefore, we did not account for the benefits of California emissions reductions
on design values in downwind states.
However, it is very likely that reductions in California emissions would lead to
substantial reductions in DVs at some monitoring locations in downwind states. For example, as
shown in Figure 3 A-18, a 90% NOx reduction in California has the potential to substantially
improve ozone concentrations at downwind receptors in Nevada, Utah, and Arizona. Given the
number of monitors in California projected to violate a revised ozone standard in 2025, it is quite
likely that the state will adopt substantial local and regional controls to reach attainment. These
controls would benefit areas outside of California as well. In addition, California and other
states may have obligations to reduce emissions if those emissions are contributing substantially
to interstate transport, as required under sections 110 and 126 of the Clean Air Act. Although no
assessment of state-receptor linkages has been completed for alternative levels of the NAAQS
(70, 65, and 60 ppb) this figure indicates that it is possible that California (and other Western
states) could potentially have significant benefit toward attainment at downwind receptors which
might then be subject to interstate transport provisions of the Clean Air Act.
25 The EPA will likely finalize designations for a revised ozone NAAQS in late 2017. Depending on the precise
timing of the effective date of those designations, nonattainment areas classified as Severe 15 will likely have to
attain sometime between late 2032 and early 2033 and nonattainment areas classified as Extreme will likely have to
attain sometime between late 2037 and early 2038.
5A-59
-------
Change in O3
• > 10 ppb
• 5-10 ppb
O 3 - 5 ppb
O 1 - 3 ppb
O < 1 ppb
Figure 3A-18.Projected change in 2025 ozone design values with an additional 90%
California NOx control (Southwest region; stars represent sites identified in
Table 3A-14)
An air agency can request and the EPA can agree to exclude data associated with event-
influenced exceedances or violations of a NAAQS provided the event meets the statutory
requirements in section 319 of the CAA:
• The event "affects air quality."
• The event "is not reasonably controllable or preventable."
• The event is "caused by human activity that is unlikely to recur at a particular location or
[is] a natural event."26
The EPA's implementing regulations, the 2007 Exceptional Events Rule, further specify that
states must provide evidence that:27
26 A natural event is further described in 40 CFR 50. l(k) as "an event in which human activity
plays little or no direct causal role."
27 See 72 Federal Register 13560 (March 22, 2007), 40 CFR Part 50.1, 40 CFR Part 50.14 and 40
CFR Part 51.930.
5A-60
-------
• "There is a clear causal relationship between the measurement under consideration and
the event that is claimed to have affected the air quality in the area;"
• "The event is associated with a measured concentration in excess of normal historical
fluctuations, including background;" and
• "There would have been no exceedance or violation but for the event."
Once an air agency requests data exclusion by flagging the subject data and submitting
supporting documentation showing that the data have been affected by exceptional events (e.g.,
stratospheric intrusions or wildfires) and the EPA concurs with this request, the event-influenced
data would be excluded from the data set used in regulatory decisions, including determining
whether or not an area is attaining or violating a NAAQS. As an example, Figure 3A-18 shows
five years of daily ozone values at Weminuche Wilderness area in La Plata County, CO. This
figure shows evidence of both episodic and persistent high ozone at this monitoring site between
2009 and 2013. Several short periods of elevated ozone in springtime (March and April) could
potentially be due to stratospheric intrusion, although more analysis would be required to
definitively determine the source(s) contributing to these high concentrations. This figure also
shows more prolonged periods of ozone values above 65 ppb which are unlikely to qualify for
exclusion as exceptional events. In cases where design values are only 1-2 ppb above an
alternative NAAQS level, excluding data from one or two days may be enough to show
compliance with the standard even if there are still some periods with high ozone values that do
not qualify. This is especially true because the standard is based on a three-year average. As can
be seen in Figure 3 A-19, some years (2009, 2012 and 2013) have substantially fewer days above
65 ppb at this site than others (2010 and 2011). Eliminating a few high ozone days in 2010 and
2011 could potentially bring this site's projected design value below a threshold level when
averaged with ozone concentrations from one or more low-ozone years.
JA-61
-------
Daily Max 8-hr Ozone Concentrations for site 080671004
01/01
04/01
07/01
10/01
01/01
Figure 3A-19. Daily 8-hr maximum ozone values at ozone monitor in Weminuche
Wilderness area in La Plata County Colorado from 2009-2013. Horizontal line
provided at 65 ppb.
Other C AA provisions that could provide regulatory relief for air agencies and potential
regulated entities include designation as a rural transport area (182(h)) or a determination that the
area would have attained but for the contribution of international emissions (179B). Rural
transport areas must show that the area does not contain emissions sources that substantially
impact monitored ozone concentrations in the area or in other areas and that they are not in or
adjacent to a Metropolitan Statistical Area (MSA). Section 179B demonstrations are used for
areas that can demonstrate they would have attained the standard but for emissions emanating
from outside the U.S. The EPA has used section 179B authority previously to approve attainment
plans for Mexican border areas in El Paso, TX (Os, PM10, and CO plans); Nogales, AZ (PM10
plan); and Imperial Valley, CA (PM10 plan). The 1-hour Os attainment plan for El Paso, TX was
approved by EPA as sufficient to demonstrate attainment of the NAAQS by the Moderate
classification deadline of November 15, 1996, taking into account "but for" international
emissions sources in Ciudad Juarez, Mexico (69 FR 32450, June 10, 2004). The state's
5A-62
-------
demonstration included airshed modeling using only the U.S. emissions data because emissions
data from Ciudad Juarez were not available.
3A.8 Calculation Methodology for W126 Metric
Calculation of the W126 metric occurs in several steps. The first step is to sum the
weighted hourly ozone concentrations within each calendar month, resulting in monthly index
values. Since plant and tree species are not photosynthetically active during nighttime hours,
only ozone concentrations observed during daytime hours (defined as 8:00 AM to 8:00 PM local
time) are included in the summations. The monthly W126 index values are calculated from the
hourly ozone concentration data as follows:
Monthly W126 = Zg=1ZŁ8^a.,l" ,.., Equation (3A-1)
where TV is the number of days in the month, dis the day of the month (d = 1, 2, ..., N), h is the
hour of the day (h = 0, 1, ..., 23), and Cdh is the hourly ozone concentration observed on day d,
hour h, in parts per million.
Next, the monthly W126 index values are adjusted for missing data. If Nm is defined as the
number of daytime ozone concentrations observed during month m (i.e. the number of terms in
the monthly index summation), then the monthly data completeness rate is Vm = Nm /12 * N.
The monthly index values are adjusted by dividing them by their respective Vm. Monthly index
values are not computed if the monthly data completeness rate is less than 75 percent (Vm<
0.75).
Finally, the annual W126 index values are computed as the maximum sum of their
respective adjusted monthly index values occurring in three consecutive months (i.e., January-
March, February-April, etc.). Three-month periods spanning across two years (i.e., November-
January, December-February) are not considered, because the seasonal nature of ozone makes it
unlikely for the maximum values to occur at that time of year. The annual W126 concentrations
are considered valid if the data meet the annual data completeness requirements for the existing
standard. Three-year W126 index values are calculated by taking the average of annual W126
index values in the same three-month period in three consecutive years.
5A-63
-------
3A.9 References
Gilliam, R. C., W. Appel, and S. Phillips. The Atmospheric Model Evaluation Tool (AMET): Meteorology Module.
Presented at 4th Annual CMAS Models-3 Users Conference, Chapel Hill, NC, September 26 - 28, 2005.
(http://www.cmascenter.org/)
National Research Council (NRC), 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations, Washington, DC: National Academies Press.
Phillips, S., K. Wang, C. Jang, N. Possiel, M. Strum, T. Fox, 2007: Evaluation of 2002 Multi-pollutant Platform: Air
Toxics, Ozone, and Paniculate Matter, 7th Annual CMAS Conference, Chapel Hill, NC, October 6-8, 2008.
Simon, H, Baker, K., Phillips, S., 2012: Compilation and interpretation of photochemical model performance
statistics published between 2006 and 2012. Atmospheric Environment 61, 124-139.
U.S. Environmental Protection Agency (US EPA), 2005; Technical Support Document for the Final Clean Air
Interstate Rule: Air Quality Modeling; Office of Air Quality Planning and Standards; RTF, NC; (CAIR
Docket OAR-2005-0053-2149).
U.S. Environmental Protection Agency (US EPA) 2009; Technical Support Document for the Proposal to Designate
an Emissions Control Area for Nitrogen Oxides, Sulfur Oxides, and Paniculate Matter: EPA-420-R-007,
329pp. (http://www.epa.gov/otaq/regs/nonroad/marine/ci/420r09007.pdf)
U.S. Environmental Protection Agency (US EPA) 2014; Draft Modeling Guidance for demonstrating attainment of
air quality goals for ozone, PM2.5, and regional haze. September 2014, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 27711.
5A-64
-------
CHAPTER 4: CONTROL STRATEGIES AND EMISSIONS REDUCTIONS
Overview
In order to estimate the costs and benefits of alternative ozone standards, the EPA has
analyzed hypothetical control strategies that areas across the country might employ to attain
alternative revised primary ozone standards of 70, 65, and 60 ppb. This chapter documents the
emission control measures EPA applied to simulate attainment with these alternative ozone
standards and the projected emission reductions associated with the measures.
This chapter is organized into four sections. Section 4.1 provides a summary of the steps
used to conduct the control strategy analysis. Section 4.2 describes the emission reductions by
sector in analyzing the baseline controls to meet the current (75 ppb) ozone standard. Section 4.3
discusses control measures and emission reductions applied as part of the alternative standard
analyses. Section 4.4 lists the key limitations and uncertainties associated with the control
strategy analysis. And finally, Section 4.5 lists the references for the chapter.
For the purposes of this discussion, it will be helpful to define some terminology. These
definitions are specific to this analysis:
• Base Case - Emissions projected to the year 2025 reflecting current state and federal
programs28. This does not include control programs specifically for the purpose of
attaining the current ozone standard (75 ppb).
• Baseline - For all areas of the U.S. except California, the base case plus additional
emissions reductions needed to reach attainment of the current ozone standard (75
ppb) as well as emissions resulting from the Clean Power Plan (U.S. EPA, 2014b).
Several areas in California are not required to meet the existing standard by 2025 and
may not be required to meet a revised standard until 2037, thus we conducted
analyses of attainment in California for post-2025. We explain the baseline treatment
for California in more detail later in this chapter.
• Alternative Standard Analysis - Emissions reductions and associated hypothetical
controls needed to reach attainment of the alternative standards. These reductions and
controls are incremental to the baseline.
• Design Value - A metric that is compared to the level of the National Ambient Air
Quality Standard (NAAQS) to determine compliance. Design values are typically
! A complete list of programs included in the 2025 base case emissions is included in the Technical Support
Document: Preparation of Emissions Inventories for the Version 6.1, 2011 Emissions Modeling Platform (U.S.
EPA, 2014d).
4-1
-------
used to classify nonattainment areas, assess progress towards meeting the NAAQS,
and develop control strategies. The design value for the 8-hour ozone standard is
calculated as the 3-year average of the 4th highest 8-hour daily maximum
concentration recorded at each monitoring site.
The EPA analyzed the impact that additional emissions control measures, across
numerous sectors, would have on predicted ambient ozone concentrations incremental to the
baseline. These control measures are based on information available at the time of this analysis,
and include primarily end of pipe controls. Additional emission abatement strategies such as
fuel switching, energy efficiency, and process changes may also be employed to reach emission
reduction targets. The potential impact of some of these types of strategies is discussed in
Chapter 7. Note that we did not conduct this analysis incremental to controls applied as part of
previous NAAQS analyses (e.g., NOX or PIVh.s because the data and modeling on which these
previous analyses were based are now considered outdated and are not compatible with the
current ozone NAAQS analysis. In addition, there were no incremental NOx controls applied in
the PM2.5 NAAQS and therefore there would be little to no impact on the controls selected to
meet the alternative ozone standards analyzed.29 Thus, the analysis for the alternative standards
focuses specifically on incremental improvements beyond the current standard and other existing
and proposed rules. The selection of control strategies is based on a least cost approach selecting
from those controls for which we have adequate information on costs, effectiveness, and
applicability. The hypothetical control strategies presented in this RIA represent illustrative
options for emissions reductions that achieve national attainment of the alternative standards.
The hypothetical control strategies are not recommendations or requirements for how a revised
ozone standard should be implemented, and states will make all final decisions regarding
implementation strategies for a revised ozone NAAQS.
29 There were no additional NOx controls applied in the PM2 5 NAAQS RIA, and therefore there would be little to
no impact on the controls selected as part of this analysis. In addition, the only geographic areas that exceed the
alternative ozone standard levels analyzed in this RIA and in the 2012 PM25 NAAQS RIA are in California. The
attainment dates for a new PM2 5 NAAQS would likely precede attainment dates for a revised ozone NAAQS. While
the 2012 PM25 NAAQS RIA concluded that controls on directly emitted PM25 were the most cost-effective on
a$/ug basis, states may choose to adopt different control options. These options could include NOx controls. It is
difficult to determine the impact on costs and benefits for this RIA because it is highly dependent upon the control
measures that would be chosen and the costs of these measures.
4-2
-------
The NOX control measures dataset for non-electric generating unit (EGU) point sources
applied in these control strategies reflects a number of revisions that EPA made since the
completion of the previous ozone NAAQS RIA. These changes include:
• Removal of incorrect links between control measures and SCCs,
• Updates to cost equations where more recent data was available to improve their
accuracy,
• Inclusion of information that has recently become available concerning known and
emerging technologies for reducing NOX emissions, and
• Revising costs and control efficiencies from control measures in the dataset based on
recently obtained information from industry and multi-jurisdictional organizations
(e.g., Ozone Transport Commissions and Lake Michigan Air Directors Consortium).
These revisions made the NOX control measures dataset for non-EGUs more accurate,
defensible, and up to date. These improvements in this dataset will improve our control strategy
and cost analyses not only for this RIA, but also for other potential rulemakings where control of
NOX from non-EGUs is an important concern.
4.1 Control Strategy Analysis Steps
The primary year of analysis for analyzing the incremental costs and benefits of meeting
a revised ozone standard is 2025. The analysis year was chosen because most areas of the U.S.
will be required to meet a revised ozone standard by 2025. In estimating the incremental costs
and benefits of potential alternative standards, we recognize that there are several areas that are
not required to meet the existing ozone standard by 2025. The Clean Air Act allows areas with
more significant air quality problems to take additional time to reach the existing standard.
Several areas in California are not required to meet the existing standard by 2025 and may not be
required to meet a revised standard until December 31, 2037. Depending on how areas in
California are eventually designated, some areas may have attainment dates earlier than 2037,
but for simplicity we are not distinguishing unique attainment years for different locations within
California. To reflect these differences in required attainment dates, we conducted analyses of
attainment in California for the period post-2025.
-------
While our goal for the California analysis was to reflect 2038, we were not able to project
emissions and air quality beyond 2025 for California. However, we were able to adjust baseline
air quality to reflect mobile source emissions reductions for California that would occur between
2025 and 2030; these emissions reductions were the result of state and federal mobile source
regulations expected to be fully implemented by 2030. For ease of discussion throughout the
analyses we refer to the time periods for potential attainment in California and in other areas of
the U.S. as post-2025 and 2025, respectively. Because we estimate incremental emissions
reductions, costs, and benefits for these two distinct time periods, it is not appropriate to add the
estimates together or to directly compare the estimates.
To conduct the control strategy analyses, we require information on (i) control costs, (ii)
control effectiveness in terms of NOx or VOC emissions reduced, (iii) the sensitivity of ozone
design values to the NOx and VOC emissions reductions, and (iv) design value targets for each
area. For the air quality modeling, the EPA prepared one control scenario for an alternative
standard level of 70 ppb (Step 2 below) because we did not expect to have sufficient known
controls for all locations to reach attainment for all of the alternative standards analyzed. The
control scenario is not really designed to model how areas reach an alternative standard level of
70 ppb, instead it sets up a process for developing, and applying, a list of potentially available
known controls ordered by cost. To develop a sufficient amount of controls for the 65 and 60
ppb alternative standard levels, we also worked to identify known controls in areas or regions
expected to contribute to nonattainment for these alternative standards.
The following steps were taken by the EPA to analyze the impacts and costs of the
control scenario incremental to the base case air quality modeling:
1. Identify geographic areas in the U.S. projected to exceed the alternative standard of
70 ppb in the year 2025 in the base case air quality modeling.
2. Develop a hypothetical control scenario for these areas and generate a control case
2025 emissions inventory for all areas except California; for California develop a
control case post-2025 emissions inventory.
3. Perform air quality modeling to assess the air quality impacts of the hypothetical
control scenario. Additionally, perform a series of emissions sensitivity simulations to
develop average ozone response to across-the-board NOx and VOC emissions
reductions in different areas (see Chapter 3).
4-4
-------
4. Calculate the portion of the hypothetical control scenario emission reductions that are
attributed to meeting the baseline. Estimate any additional emissions reductions
beyond the known controls that are needed to meet the current standard based on
average ozone response factors. These are the baseline emission reductions.
5. Estimate the additional emissions reductions incremental to the baseline that are
needed to meet the alternative standards of 70, 65, and 60 ppb. Costs of controls
incremental to (i.e., over and above) the baseline reductions are attributed to the costs
of meeting the alternative standards. These emissions reductions can come from
specific known controls or emission reductions needed beyond known controls, also
referred to as unknown controls. Potential controls may be categorized as unknown
because these needed emissions reductions come from sectors for which we have not
sufficiently explored emissions abatement opportunities or sectors that might require
non-traditional abatement through measures like energy efficiency or process
changes.
The following sections will discuss in more detail the analysis steps presented above.
4.2 Baseline Control Strategy
Establishing the baseline allows us to estimate the incremental costs and benefits of
attaining the alternative standards. Three steps were used to develop the baseline. First, we
estimated 2025 base case emissions and air quality, reflecting "on the books" regulations (see
Section 3.1.3 Emissions Inventories for a discussion of the rules included in the Base Case for
this analysis).30 Second, we accounted for changes in ozone predicted to occur due to one
potential approach for implementing the Clean Power Plan. Third, we identified additional
controls that could be applied to demonstrate attainment of the current ozone standard of 75 ppb.
Additional control measures were used in three sectors to meet the current ozone standard
in establishing the baseline:31 Non-Electric Generating Unit Point Sources (Non-EGUs), Non-
Point (Area) Sources, and Nonroad Mobile Sources. See Table 4-1 for a summary of controls
30 Among others factors, the baseline for this analysis is also affected by the choice of the future year ~ a year
farther into the future allows for more time for federal measures to work and to attain. This baseline is also affected
by the air quality starting point, potentially reducing the amount of emissions reductions required for attainment -
this analysis started from a standard of 75 ppb, where the analysis in 2008 started from a standard of 84 ppb. In
addition, we have identified additional "known" controls to apply in this analysis, controls which are less expensive
per ton than unknown controls.
31 In establishing the baseline, the U.S. EPA selected a set of cost-effective controls to simulate attainment of the
current ozone standard. These control sets are hypothetical as states will ultimately determine controls as part of
the SIP process.
4-5
-------
applied in the baseline analysis. There were several areas in California that did not reach
attainment of the current standard with known controls. For these geographic areas, we estimated
the additional emissions reductions needed beyond those achieved by identified known controls
for NOX and VOC to attain the current standard.
4-6
-------
Table 4-1. Controls Applied for the Alternative Standard Analyses Control Strategy
Sector
NOx
VOC
Non-EGU Point
LEG (Low Emission Combustion)
Solvent Recovery System
SCR (Selective Catalytic Reduction)
Work Practices, and Material
Reformulation/Substitution
SNCR (Selective Non-Catalytic Reduction)
Low-VOC materials Coatings and Add-
On Controls
NSCR (Non-Selective Catalytic Reduction)
Low VOC Adhesives and Improved
Application Methods
LNB (Low NOX Burner Technology)
Permanent Total Enclosure (PTE)
LNB + SCR
Solvent Substitution, Non-Atomized
Resin Application Methods
LNB + SNCR
Petroleum Wastewater Treatment
Controls
OXY-Firing
Incineration (Thermal, Catalytic, etc) to
Reduce VOC Emissions
Biosolid Injection Technology
LNB + Flue Gas Recirculation
LNB + Over Fire Air
Ignition Retard
Natural Gas Reburn
Ultra LNB
NonPoint
NSCR (Non-Selective Catalytic Reduction)
Process Modification to Reduce Fugitive
VOC Emissions
LEG (Low Emission Combustion)
Reformulation to Reduce VOC Content
LNB (Low NOX Burner Technology)
Incineration (Thermal, Catalytic, etc) to
Reduce VOC Emissions
LNB Water Heaters
Low Pressure/Vacuum (LPV) Relief
Valves in Gasoline Storage Tanks
Reduced Solvent Utilization
Gas Recovery in Landfills
Nonroad
Diesel Retrofits & Engine Rebuilds
A map of the country is presented in Figure 41, which shows the counties projected to
exceed the current ozone standard of 75 ppb in the 2025 base case scenario. This includes 8
projected exceeding counties in California and 3 exceeding counties in Texas. NOx control
4-7
-------
measures were applied in these 11 counties in the baseline analysis to meet the current ozone
standard. In addition, NOx control measures were applied to 40 California counties and 52 Texas
counties adjacent to exceeding counties in order to address transport coming from these adjacent
counties. A map of the areas where control measures were applied to demonstrate attainment of
the current standard and establish the baseline is presented in Figure 4-2.
To construct the post-2025 baseline, we included mobile source NOX and VOC emissions
changes that are projected to occur in California from 2025 - 2030 as a result of California's
current mobile source control programs and projected changes in vehicle miles traveled and
nonroad activity levels. These changes were included because they would result in emission
reductions that would contribute toward attainment of the current ozone standard. No emission
projections were available for other sectors for this time period, and no mobile source emissions
projections were available beyond 2030. Additionally, VOC controls were applied in California
counties highlighted in Figure 4-2. Even with the above mentioned controls, some areas in
California did not reach attainment with known controls. For these areas, we estimated the
additional NOX emission reductions needed beyond identified known controls to attain the
standard.
4-8
-------
Legend
665 counties are projected to be below 0 075 ppm
^^| 11 counties are projected to exceed 0 075 ppm
There are 676 counties with monitors
0 200 400 800 Kilometers
I i i i I i i |
Figure 4-1. Counties Projected to Exceed the Baseline Level of the Current Ozone
Standard (75 ppb) in 2025 Base Case
4-9
-------
Legend
NOx t VOC Emission Reductions
NOx Emission Reductions
0 200 400 800 Kilometers
I i i i I i i i I
Figure 4-2. Counties Where Emissions Reductions Were Applied to Demonstrate
Attainment of the Current Standard for the Baseline Analysis
Tables 4-2 and 4-3 summarize the NOX and VOC emission reductions needed to
demonstrate attainment of the current ozone standard (75 ppb).
4-10
-------
Table 4-2. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Attainment of the Current Standard for the 2025 Baseline - U.S.,
except California (1,000 tons/year)"
Geographic Area
East
Emissions Sector
ECU
Non-EGU Point
Nonpoint
Nonroad
NOx
_
18
26
0.83
voc
_
.
-
-
Onroad
West
Total
ECU
45
_
-
_
Non-EGU Point
Nonpoint
Nonroad
Onroad
Total
a Estimates are rounded to two significant figures.
b For the control strategy and cost analysis, "East" includes the Northeast, Midwest, and Central regions, and "West"
includes the Southwest region. See Chapter 3 for a description of these regions.
Table 4-3. Summary of Emission Reductions (Known and Unknown Controls) Applied to
Demonstrate Attainment in California for the post-2025 Baseline (1,000
tons/year)"
Known Controls
Emissions Sector
ECU
Non-EGU Point
Nonpoint
Nonroad
NOx
_
14
14
4.2
VOC
_
0.61
47
-
Onroad
Unknown Controls
Total
All
Total
31
160
190
48
_
48
a Emission reduction estimates are rounded to two significant figures.
The 2025 baseline for this analysis presents one scenario of future year air quality based
upon specific control measures, additional emission reductions beyond known controls,
promulgated federal rules such as Tier 3, and specific years of initial values for air quality
monitoring and emissions data. This analysis presents one illustrative strategy relying on the
identified federal measures and other strategies that states may employ. States may ultimately
4-11
-------
employ other strategies and/or other federal rules may be adopted that would also help in
achieving attainment with the current standard.
4.3 Alternative Standard Analyses
After identifying the controls in the baseline scenario, additional controls needed to meet
the alternative standards were identified in four sectors: Electric Generating Units (EGUs), Non-
Electric Generating Unit Point Sources (Non-EGUs), Non-Point (Area) Sources, and Nonroad
Mobile Sources. Onroad mobile source controls were not applied because they are largely
addressed in existing rules such as the recent Tier 3 rule. Controls applied for the alternative
standard analyses were the same as were applied for the baseline analysis (see Table 4-1 for a
summary of controls applied in the baseline) with the addition of Selective Catalytic Reduction
applied to EGUs. Other than the addition of EGU controls, the primary difference between the
controls applied for the alternative standards versus the baseline was the geographic areas to
which they were applied.
The EPA performed a national scale air quality modeling analysis to estimate ozone
concentrations for the future base case year of 2025. To accomplish this, we modeled multiple
emissions cases for 2025, including the 2025 base case and twelve 2025 emissions sensitivity
simulations. The twelve emissions sensitivity simulations were used to develop ozone sensitivity
factors (ppb/ton) from the modeled response of ozone to changes in NOx and VOC emissions
from various sources and locations. These ozone sensitivity factors were then used to determine
the amount of emissions reductions needed to reach the 2025 baseline and evaluate potential
alternative standard levels of 70, 65, and 60 ppb incremental to the baseline. As mentioned
previously, only a subset of known controls were included in the modeled control scenarios.
Therefore, any additional emissions reductions may include both known and unknown controls.
In areas of the country outside of California, Texas, and the northeast, total emissions reductions
needed beyond the baseline were based entirely on response factors, i.e., not based on the
hypothetical control scenario used in the air quality modeling.
Figure 4-3 shows the counties projected to exceed the alternative standards analyzed for
the 2025 baseline for areas other than California. For the 70 ppb scenario, emissions reductions
were required for monitors in the Central and Northeast regions (see Chapter 3, Figure 3-3 for a
4-12
-------
depiction of the regions). For the 65 and 60 ppb scenarios, emissions reductions were applied in
all regions with projected baseline DVs above these levels. For the 60 ppb scenario, additional
VOC tons were identified in Chicago because some sites in that area experienced NOX
disbenefits meaning that the regional NOX reductions resulted in ozone DV increases from below
60 ppb to above 60 ppb. Therefore it was not possible to identify a scenario in which regional
NOX reductions and maximum known VOC controls alone resulted in all Midwest monitors
meeting a 60 ppb standard. An iterative approach was used determine a combination of NOX and
VOC emissions reductions that would not lead to over-control at either the NOx-limited monitor
with the highest design value or the VOC-limited monitor with the highest design value.
Because of the regional approach we used for locations other than Texas, we were not able to
geographically fine tune the control strategies, and thus there is some uncertainty in the actual
amounts of emissions reductions estimated for attaining the alternative standards.
Legend
I 9 counties are projected to exceed 70 ppb
59 additional counties are projected to be below 70 but exceed 05 ppb
173 additional counties are projected to be below 66 but exceed 60 ppb
0 200 400
I i i i l
800 Kilometers
I _ I
Figure 4-3. Projected Ozone Design Values in the 2025 Baseline Scenario
4-13
-------
Figure 4-4 shows the counties projected to exceed the alternative standards analyzed for
the post-2025 baseline analysis for California. For the California post-2025 alternative standard
analyses, all known controls were applied in the baseline so incremental reductions are from
unknown controls.
0 50 100 200K«omeBrs
Legend
I 4 counties are projected r exceed 70 ppb
8 additional counties are projected to be below 70 but exceed SS ppb
6 additional counties ars projected to be below 65 but exceed 60 ppb
Figure 4-4. Projected Ozone Design Values in the post-2025 Baseline Scenario
4.3.1 Identifying Known Controls Needed to Meet the Alternative Standards
For the 2025 alternative control strategy analyses of 70, 65 and 60 ppb, known NOX
controls for four sectors were used: EGUs, non-EGU point, nonpoint, and nonroad mobile
sources. In a smaller number of geographic areas, VOC controls were applied to non-EGU point
and nonpoint sources.
4-14
-------
For Texas, reductions were first applied to NOX sources in the areas surrounding Dallas
and Houston. Additional reductions then were applied to VOC sources in the area surrounding
Houston, to NOX sources in other parts of Texas, and to sources in the surrounding states. For the
Northeast, Midwest, and Southwest areas of the country, reductions were applied to NOX sources
within the regions. For regions where additional reductions were needed, controls were applied
to VOC sources in urban areas with the highest ozone design values in the region.32
For California, all known controls were applied in the baseline analysis so there were no
known controls available to apply toward the incremental emissions reductions needed for the
alternative analysis levels for post-2025. Maps of the areas where control measures were applied
to demonstrate attainment of the alternative analysis levels are presented in Figures 4-5 and 4-6.
Note that we do not account for between region transport of ozone, and therefore, especially for
the 65 and 60 ppb alternative standard levels, we may be overstating the amount of emissions
reductions needed for attainment.
32 Texas, California, and the northeast were included in the hypothetical control scenario for an alternative standard
of 70 ppb. Other regions were modeled as part of the 2025 emissions sensitivity simulations.
4-15
-------
Legend
NOx + VOC Emission Reductions
NOx Emission Reductions
0 200 400 800 Kilometers
lii I i i i I
Figure 4-5. Counties Where Emissions Reductions Were Applied to Demonstrate
Attainment with a 70 ppb Ozone Standard in the 2025 Analysis
4-16
-------
Legend
NOx + VOC Emission Reductions
NOx Emission Reductions
200 400
800 Kilometers
I
Figure 4-6. Counties Where Emissions Reductions Were Applied to Demonstrate
Attainment with 65 and 60 ppb Ozone Standards in the 2025 Analyses
Table 4-4 shows the number of exceeding counties and the number of adjacent counties
to which controls were applied for the alternative standards. For a complete list of geographic
areas for the alternative standards see Appendix 4. A.
4-17
-------
Table 4-4. Number of Counties with Exceedances and Number of Additional Counties
Where Reductions Were Applied for the 2025 Alternative Standards Analyses -
U.S., except California
Alternative Standard Number of Counties with Number of Additional Counties Where Reductions
Exceedances Were Applied
70 ppb 15 479
65ppb 115 1,925
60 ppb 289 l,791a
a Number of additional counties where reductions are applied declined for 60 ppb analysis because the
number of overall counties in the analysis remained the same while the number of exceeding counties increased.
Tables 4-5 through 4-7 show the emissions reductions from known controls for the
alternative standards analyzed. No exceedances were projected for the West region, outside of
California, for the 70 ppb alternative standard. For the lower alternative standards of 65 and 60
ppb, similar controls were applied as were used in the 70 ppb analysis, but the geographic area in
which they were applied increased. The largest emission reductions were in the non-EGU point
source and nonpoint sectors. For details regarding emission reductions by control measure see
Appendix 4. A
Table 4-5. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Nationwide Attainment with a 70 ppb Ozone Standard in 2025,
except California (1,000 tons/year)"
Geographic
Area
East
West
Emissions
Sector
ECU
Non-EGU Point
Nonpoint
Nonroad
Onroad
Total
ECU
Non-EGU Point
Nonpoint
Nonroad
Onroad
Total
Baseline Emissions
NOx VOC
884
1,485
1,487
1,235
1,135
6,226
159
226
193
219
197
1,025
Emission Reductions
NOx VOC
25
210 0.
260
5
-
490
-
-
-
-
-
-
-
.98
54
-
-
55
-
-
-
-
-
-
a Emission reduction estimates are rounded to two significant figures.
4-18
-------
Table 4-6. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Nationwide Attainment with a 65 ppb Ozone Standard in 2025 -
except California (1,000 tons/year)"
Geographic Area
Emissions Sector
NOx
VOC
East
ECU
170
Non-EGU Point
410
3.6
Nonpoint
420
95
Nonroad
12
Total
1,000
99
West
ECU
36
Non-EGU Point
38
0.47
Nonpoint
37
6.6
Nonroad
Total
110
a Emission reduction estimates are rounded to two significant figures.
Table 4-7. Summary of Emission Reductions by Sector for Known Controls Applied to
Demonstrate Nationwide Attainment with a 60 ppb Ozone Standard in 2025 -
except California (1,000 tons/year)"
Geographic Area
East
West
Emissions Sector
ECU
Non-EGU Point
Nonpoint
Nonroad
Total
ECU
Non-EGU Point
Nonpoint
Nonroad
Total
NOx
170
410
420
12
1,000
62
48
39
1.3
150
VOC
-
4.2
99
-
100
-
0.47
6.6
-
7
' Emission reduction estimates are rounded to two significant figures.
4.3.2 Known Control Measures Analyzed
Known control measures were applied to electric generating units (EGU), non-EGU
point, nonpoint (area), and nonroad mobile sources for demonstration of attainment with the
current and alternative standards. The applied control measures were identified using the EPA's
Control Strategy Tool (CoST) (U.S. EPA, 2014c), Integrated Planning Model (IPM), and
NONROAD Model. CoST models emissions reductions and engineering costs associated with
control strategies applied to point, area, and mobile sources of air pollutant emissions by
matching control measures to emissions sources using algorithms such as "maximum emissions
4-19
-------
reduction", "least cost", and "apply measures in series". For this analysis, we applied the
maximum emissions reduction algorithm. These controls are described further in Appendix 4. A.
Specific controls were applied in the air quality modeling only for a portion of the analysis (for
California, Texas, and the northeast, areas projected to exceed an alternative standard level of 70
ppb). A majority of the emission reductions needed were identified using the ozone sensitivity
factors developed from the twelve emissions sensitivity simulations (see Chapter 3, Section 3.1
for a discussion of the development of the ozone sensitivity factors).
Nonpoint and nonroad mobile source emissions data are generated at the county level,
and therefore controls for these emissions sectors were applied at the county level. EGU and
non-EGU point source controls are applied to individual point sources. Control measures were
applied to point and nonpoint sources of NOX, including: industrial boilers, commercial and
institutional boilers, reciprocating internal combustion engines in the oil and gas industry and
other industries, glass manufacturing furnaces, and cement kilns. The analysis for nonroad
mobile sources applied NOX controls to diesel engines.
In a portion of the geographic areas where NOx controls were applied, the EPA also
applied control measures to sources of VOC including surface coating, solvents, and fuel storage
tanks. VOC reductions were analyzed in the urban areas with the highest ozone design values in
each region: northern and southern California; Denver in the Southwest; Houston and Dallas in
the Central region; Chicago, Detroit, and Louisville in the Midwest; New York, Baltimore, and
Pittsburgh in the Northeast. Even among these areas, in some cases NOX reductions necessary to
bring the very highest urban areas into attainment made the VOC reductions unnecessary in other
high ozone urban areas within the region.
To more accurately depict available controls, the EPA employed a decision rule in which
controls were not applied to any non-EGU or nonpoint sources with less than 25 tons/year of
emissions per pollutant. This decision rule is more inclusive of sources than the rule we
employed in the previous Os and PIVb.sNAAQS RIAs where we applied a minimum of 50
tons/year for each pollutant. We modified the decision rule for this NAAQS analysis to
recognize the potential for emissions reductions in the large number of sources emitting in the
25-50 tons/year range in order to devise control strategies for full, or closer to full, attainment.
4-20
-------
Historically, the reason for not applying controls to sources emitting less than 25 tons/year has
been that many point sources with emissions below this level already have controls in place. For
the analysis, we applied best engineering judgement to apply controls to select sources.
4.3.3 Emissions Reductions beyond Known Controls Needed to Meet the Alternative Standards
There were several areas where known controls did not achieve enough emissions
reductions to attain the alternative standards of 70, 65, and 60 ppb. To complete the analysis, the
EPA then estimated the additional emissions reductions beyond known controls needed to reach
attainment, also referred to as unknown controls. For information on the methodology used to
develop the emission reductions estimates, see Chapter 3. Table 4-8 shows the emissions
reductions needed from unknown controls in 2025 for the U.S., except California, for the
alternative standards analyzed. Table 4-9 shows the reductions needed from unknown controls
for California for the post-2025 analysis.
Table 4-8. Summary of Emissions Reductions by Alternative Standard for Unknown
Controls for 2025 - except California (1,000 tons/year)3
Alternative Standard Region NOx VOC
70 ppbbEast 150 -
West - -
65ppbc East 750
West - -
60ppbd East 1,900 41
West 350 -
a Estimates are rounded to two significant figures.
b Unknown controls for the 70 ppb alternative standard are needed in the Northeast and Central regions (see Chapter
3 for a description of these regions).
0 Unknown controls for the 65 ppb alternative standard are needed in the Northeast, Central, and Midwest regions
(see Chapter 3 for a description of these regions).
d Unknown controls for the 60 ppb alternative standard are needed in the Northeast, Central, Midwest, and
Southwest regions (see Chapter 3 for a description of these regions).
Table 4-9. Summary of Emissions Reductions by Alternative Level for Unknown Controls
for post-2025 - California (1,000 tons/year)3
Alternative Standard Region NOx VOC
70 ppb CA 53 -
65 ppb CA 110 -
4-21
-------
60ppb
CA
140
a Estimates are rounded to two significant figures.
4.3.4 Summary of Emissions Reductions Needed to Meet the Alternative Standards
Table 4-10 summarizes the known and unknown emissions reductions needed to meet the
alternative standard levels in 2025 for the East and West, except California. In the East for 2025,
the unknown NOx reductions needed as percentage of the total rises from 23 percent to 66
percent as the alternative standard level decreases from 70 ppb to 60 ppb. Meanwhile, no
unknown VOC reductions are needed in the East for the 70 ppb and 65 ppb levels. In the West
(except California) for 2025, unknown NOx reductions are not needed until the 60 ppb level,
when the unknown tons constitute about 70 percent of the total reductions needed. No unknown
VOC reductions are needed in the West (except California) for 2025 for any of the alternative
standard levels.
Table 4-10. Summary of Known and Unknown Emissions Reductions by Alternative
Standard Levels in 2025, Except California (1,000 tons/year)"
Alternative Standard
Geographic Area
East
West
Emissions Reductions
NOx Known
NOx Unknown
% NOx Unknown
VOC Known
VOC Unknown
% VOC Unknown
NOx Known
NOx Unknown
% NOx Unknown
VOC Known
VOC Unknown
% VOC Unknown
70 ppb
490
150
23%
55
0
0%
0
0
N/A
0
0
N/A
65 ppb
1,000
750
43%
99
0
0%
110
0
0%
7
0
0%
60 ppb
1,000
1,900
66%
100
41
29%
150
350
70%
7
0
0%
a Estimates are rounded to two significant figures.
Table 4-11 shows again that there were no known NOx emissions reductions identified
for meeting the alternative standard levels for post-2025 California and that 100 percent of the
NOx tons needed were unknown. Meanwhile, no unknown VOC reductions are needed for the
any of the alternative standard levels for post-025 California.
4-22
-------
Table 4-11. Summary of Known and Unknown Emissions Reductions by Alternative
Standard Levels for post-2025 - California (1,000 tons/year)"
Alternative Standard
Geographic Area
California
Emissions Reductions
NOx Known
NOx Unknown
% NOx Unknown
VOC Known
VOC Unknown
% VOC Unknown
70ppb
0
53
100%
0
0
N/A
65 ppb
0
110
100%
0
0
N/A
60 ppb
0
140
100%
0
0
N/A
a Estimates are rounded to two significant figures.
4.4 Limitations and Uncertainties
EPA's analysis is based on its best judgment for various input assumptions that are
uncertain. As a general matter, the Agency selects the best available information from
engineering studies of air pollution controls and has set up what it believes is the most reasonable
modeling framework for analyzing the cost, emissions changes, and other impacts of regulatory
controls. However, the estimates of emissions reductions associated with our control strategies
above are subject to important limitations and uncertainties. In the following, we discuss the
limitations and uncertainties that are most significant.
• Illustrative control strategy: A control strategy is the set of actions that States
may take to meet a standard, such as which industries should be required to install
end-of-pipe controls or certain types of equipment and technology. The illustrative
control strategy analysis in this RIA presents only one potential pathway to
attainment. The control strategies are not recommendations for how a revised ozone
standard should be implemented, and States will make all final decisions regarding
implementation strategies for the revised NAAQS. We do not presume that the
control strategies presented in this RIA are an exhaustive list of possibilities for
emissions reductions.
• Emissions Inventories and Air Quality Modeling: These serve as a foundation
for the projected ozone values, control strategies and costs in this analysis and thus
limitations and uncertainties for these inputs impact the results, especially for
4-23
-------
issues such as future year emissions projections and information on controls
currently in place at sources. Limitations and uncertainties for these inputs are
discussed in previous chapters devoted to these subject areas. In addition, there are
factors that affect emissions, such as economic growth and the makeup of the
economy (e.g., growth in the oil and natural gas sector), that introduce additional
uncertainty.
Projecting level and geographic scope of exceedances: Estimates of the
geographic areas that would exceed revised alternative levels of the standard in a
future year, and the level to which those areas would exceed, are approximations
based on a number of factors. The actual nonattainment determinations that would
result from a revised standard will likely depend on the consideration of local
issues, changes in source operations between the time of this analysis and
implementation of a new standard, and changes in control technology over time.
Assumptions about the baseline: There is significant uncertainty about the
illustration of the impact of rules, especially the Clean Power Plan because it is a
proposal and because it contains significant flexibility for states to determine how
to choose measures to comply with the standard.
Applicability of control measures: The applicability of a control measure to a
specific source varies depending on a number of process equipment factors such as
age, design, capacity, fuel, and operating parameters. These can vary considerably
from source to source and over time. This analysis makes assumptions across broad
categories of sources nationwide.
Control measure advances over time: The control measures applied do not reflect
potential effects of technological change that may be available in future years and
the effects of "learning by doing" or "learning by researching" are not accounted
for in the emissions reduction estimates. Thus, all estimates of impacts associated
with control measures applied reflect our current knowledge, and not projections,
of the measures' effectiveness. In our analysis, we do not have the necessary data
for cumulative output, fuel sales, or emissions reductions for all sectors included in
order to properly generate control costs that reflect learning-curve impacts or the
4-24
-------
impacts of technological change. We believe the effect of including these impacts
would be to lower our estimates of costs for our projected year control strategies.
• Pollutants to be targeted: Local knowledge of atmospheric chemistry in each
geographic area may result in a different prioritization of pollutants (VOC and
NOX) for control. For the baseline in this analysis, we included only promulgated
or proposed rules, but that there may be additional regulations promulgated in the
future that reduce NOx or VOC emissions. These regulations could reduce the
current baseline levels of emissions.
4.5 References
U.S. Environmental Protection Agency (U.S. EPA). 2014a. Control of Air Pollution from Motor Vehicles: Tier 3
Motor Vehicle Emission and Fuel Standards. Office of Transportation and Air Quality. Available at
http://www.epa.gov/otaq/tier3.htm.
U.S. Environmental Protection Agency (U.S. EPA). 2014b. Proposed Carbon Pollution Guidelines for Existing
Power Plants and Emission Standards for Modified and Reconstructed Power Plants. Available at
http://www2. epa. gov/sites/production/files/2014 -06/documents/20140602ria-clean-power-plan.pdf
U.S. Environmental Protection Agency (U.S. EPA). 2014c. Control Strategy Tool (CoST) Documentation Report.
Office of Air Quality Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/ttnecasl/cost.htnx
U.S. Environmental Protection Agency (2014d) Preparation of Emissions Inventories for the Version 6.1, 2011
Emissions Modeling Platform (http://www.epa.gov/ttn/chief/emc
4-25
-------
APPENDIX 4: CONTROL STRATEGIES AND EMISSIONS REDUCTIONS
Overview
Chapter 4 describes the approach that EPA used in applying control measures to
demonstrate attainment of alternative ozone standard levels of 70, 65, and 60 ppb and estimating
the resulting emissions reductions. This Appendix contains more detailed information about the
control strategy analyses, including the control measures that were applied and the geographic
areas in which they were applied.
4A.I Types of Control Measures
Several types of control measures were applied in the analyses for the baseline and
alternative standard levels. These can be grouped into the following classes:
• Max NOX Reductions - NOX control measures for nonEGU point, nonpoint, and nonroad
sources. For each of these sources, we identified the most effective control (i.e., control
with the highest percent reduction) that could be applied to the source, given the
following constraints:
• the source must emit at least 50 tons/yr of NOX (see description of controls on smaller
sources below);
• any control for nonEGU point sources must result in a reduction of NOX emissions of
at least 5 tons/yr; and
• any replacement control (i.e., a more effective control replacing an existing control)
must achieve at least 10% more reduction than the existing control (e.g., we would
not replace a 60% control with a 65% control).
• NOX Reductions from EGU SCRs - SCRs applied to coal-fired EGUs where no SCR is
currently in place.
• NOX 25-50 TPY Source Reductions - Similar to the Max NOX Reductions above, except
for smaller sources in the 25-50 ton/year NOX emissions range.
• Max VOC Reductions - Similar to Max NOX Reductions described above, except this
includes only VOC controls.
4A.2 Application of Control Measures in Geographic Areas
Control measures were applied to geographic areas including or adjacent to areas that
were projected to exceed the baseline and alternative standards. See Tables 4A-1 to 4A-4 for a
listing of the NOX and VOC control groups and geographic areas to which they were applied.
4A-1
-------
Table 4A-1. Geographic Areas for Application of NOx Controls in
Alternative Standard Analyses - U.S., except California3
Geographic Areas and Control Groups Baseline
EAST
Central Region
Max NOx Reductions within TX buffer x
NOx ECU SCR within TX buffer
NOx 25-50 tpy Source controls within TX buffer
Unknown control NOx Reductions within TX buffer
Max NOx Reductions outside of TX buffer
NOx ECU SCR outside of TX buffer
NOx 25-50 tpy Source controls outside of TX buffer
Unknown control NOx Reductions outside of TX buffer
Northeast Region
Max NOx Reductions within Northeast buffer
NOx ECU SCR within Northeast buffer
NOx 25-50 tpy Source controls within Northeast buffer
Unknown control NOx Reductions within Northeast buffer
Max NOx Reductions outside of Northeast buffer
NOx ECU SCR outside of Northeast buffer
NOx 25-50 tpy Source controls outside of Northeast buffer
Unknown control NOx Reductions outside NE buffer
Midwest Region
Max NOx Reductions in Midwest Region
NOx ECU SCR in Midwest Region
NOx 25-50 tpy Source controls in Midwest Region
Unknown control NOx Reductions in MW Region
WEST
Southwest Region
Max NOx Reductions in Southwest Region
NOx ECU SCR in Southwest Region
NOx 25-50 tpy Source controls in Southwest Region
Unknown control NOx Reductions in SW Region
the Baseline and
70 ppb 65 ppb
x x
X X
X X
U U
X X
X
X
U
X X
X X
X X
U U
X X
X
X
U
X
X
X
U
X
X
60 ppb
x
X
X
U
X
X
X
U
X
X
X
U
X
X
X
U
X
X
X
U
X
X
X
U
a "x" indicates known controls were applied; "U" indicates unknown control reductions.
4A-2
-------
Table 4A-2. Geographic Areas for Application of VOC Controls in the Baseline and
Alternative Standard Analyses - U.S., except California"
Geographic Areas and Control Groups
Baseline 70 ppb 65 ppb 60 ppb
EAST
Central Region
Max VOC Reductions within Houston buffer
Max VOC Reductions within Dallas buffer
x
x
Northeast Region
Max VOC Reductions within CT-NJ-NY buffer
Max VOC Reductions within Baltimore buffer
Midwest Region
Max VOC Reductions in Chicago buffer
Unknown control VOC Reductions in Chicago
WEST
Southwest Region
Max VOC Reductions in Denver buffer
x
U
a "x" indicates known controls were applied; "U" indicates unknown control reductions.
Table 4A-3. Geographic Areas for Application of NOX Controls in the Baseline and
Alternative Standard Analyses - California"
Geographic Areas and Control Groups
California
Max NOx Reductions within California (CA) buffer
NOx 25-50 tpy Source controls within N. CA buffer
Unknown Control NOx Reductions within N. CA buffer
NOx 25-50 tpy Source controls within S. CA buffer
Unknown Control NOx Reductions within S. CA buffer
Baseline
x
x
U
x
U
70 ppb
U
U
65 ppb
U
U
60 ppb
U
U
a "x" indicates known controls were applied; "U" indicates unknown control reductions.
4A-3
-------
Table 4A-4. Geographic Areas for Application of VOC Controls in the Baseline and
Alternative Standard Analyses - California"
Geographic Areas and Control Groups Baseline 70 ppb 65 ppb 60 ppb
California
Max VOC Reductions within N. California buffer x
Max VOC Reductions within S. California buffer x
a "x" indicates known controls were applied.
4A.3 NOX Control Measures for NonEGU Point Sources
Several types of NOX control technologies exist for non-EGU point sources: selective
catalytic reduction (SCR), selective noncatalytic reduction (SNCR), natural gas reburn (NGR),
coal reburn, and low-NOx burners (LNB). In some cases, LNB accompanied by flue gas
recirculation (FGR) is applicable, such as when fuel-borne NOX emissions are expected to be of
greater importance than thermal NOX emissions. When circumstances suggest that combustion
controls do not make sense as a control technology (e.g., sintering processes, coke oven batteries,
sulfur recovery plants), SNCR or SCR may be an appropriate choice. Finally, SCR can be
applied along with a combustion control such as LNB with overfire air (OFA) to further reduce
NOX emissions. All of these control measures are available for application on industrial boilers.
Besides industrial boilers, other non-EGU point source categories covered in this RIA
include petroleum refineries, kraft pulp mills, cement kilns, stationary internal combustion
engines, glass manufacturing, combustion turbines, and incinerators. NOX control measures
available for petroleum refineries, particularly process heaters at these plants, include LNB,
SNCR, FGR, and SCR along with combinations of these technologies. NOX control measures
available for kraft pulp mills include those available to industrial boilers, namely LNB, SCR,
SNCR, along with water injection. NOX control measures available for cement kilns include
those available to industrial boilers, namely LNB, SCR, and SNCR. Non-selective catalytic
reduction (NSCR) can be used on stationary internal combustion engines. OXY-firing, a
technique to modify combustion at glass manufacturing plants, can be used to reduce NOX at
such plants. LNB, SCR, and SCR plus steam injection (SI) are available measures for
combustion turbines. Finally, SNCR is an available control technology at incinerators.
Tables 4A-5 through 4 A-12 contain lists of the NOX and VOC control measures applied in
these analyses for non-EGU point sources,EGUs, nonpoint sources, and nonroad sources. The
4A-4
-------
table also presents the associated emission reductions for the baseline and alternative standard
analyses. The number of geographic areas in which they were applied expanded as the level of
the alternative standard analyzed became more stringent.
Table 4A-5. NOX Control Measures Applied in the Baseline Analysis
Reductions
NOx Control Measure (tons/yr)
Adjust Air to Fuel Ratio and Ignition Retard - Gas Fired 1C Engines 694
Biosolid Injection Technology - Cement Kilns 1,021
Episodic Ban - Open Burning 570
Ignition Retard - 1C Engines 30
Low Emission Combustion - Gas Fired Lean Burn 1C Engines 1,552
Low NOx Burner - Commercial/Institutional Boilers & 1C Engines 6,699
Low NOx Burner - Industr/Commercial/Institutional (ICI) Boilers 165
Low NOx Burner - Industrial Combustion 270
Low NOx Burner - Lime Kilns 309
Low NOx Burner - Natural Gas-Fired Turbines 4,735
Low NOx Burner - Residential Furnaces 5,381
Low NOx Burner - Residential Water Heaters & Space Heaters 6,605
Low NOx Burner and Flue Gas Recirculation - Coke Oven/Blast Furnace 36
Low NOx Burner and Flue Gas Recirculation - Iron & Steel Mills - Reheating 42
Low NOx Burner and SCR - Coal-Fired ICI Boilers 1,174
Low NOx Burner and SCR - Industr/Commercial/Institutional Boilers 5,286
Natural Gas Reburn - Natural Gas-Fired ECU Boilers 79
Nonroad Diesel Retrofits & Engine Rebuilds - e.g., Construction Equipment 4,998
Non-Selective Catalytic Reduction (NSCR) - 4 Cycle Rich Burn 1C Engines 20,008
OXY-Firing - Glass Manufacturing 5,429
Selective Catalytic Reduction (SCR) - Cement Kilns 2,982
Selective Catalytic Reduction (SCR) - Coal Fired ECU Boilers 76
Selective Catalytic Reduction (SCR) - Fluid Catalytic Cracking Units 945
Selective Catalytic Reduction (SCR) - 1C Engines, Diesel 2,041
Selective Catalytic Reduction (SCR) - ICI Boilers 3,218
Selective Catalytic Reduction (SCR) - Industrial Incinerators 1,062
Selective Catalytic Reduction (SCR) - Petroleum Refinery Gas-Fired Process Heaters 456
Selective Catalytic Reduction (SCR) - Sludge Incineration 366
Selective Catalytic Reduction (SCR) - Utility Boilers 55
Selective Non-Catalytic Reduction (SNCR) - Comm/Inst. Incinerators 16
Selective Non-Catalytic Reduction (SNCR) - Industrial Incinerators 130
Selective Non-Catalytic Reduction (SNCR) - Utility Boilers 158
4A-5
-------
Table 4A-6. VOC Control Measures Applied in the Baseline Analysis
Reductions
VOC Control Measure (tons/yr)
Control of Fugitive Releases - Oil & Natural Gas Production 39
Control Technology Guidelines - Wood Furniture Surface Coating 777
Gas Recovery - Municipal Solid Waste Landfill 2,064
Improved Work Practices, Material Substitution, Add-On Controls - Cleaning Solvents 155
Improved Work Practices, Material Substitution, Add-On Controls - Printing 77
Incineration - Other 181
Incineration - Surface Coating 6,362
Low-VOC Coatings and Add-On Controls - Surface Coating 945
LPV Relief Valve - Underground Tanks 353
Permanent Total Enclosure (PTE) - Surface Coating 100
Process Modification - Oil and Natural Gas Production 4,311
RACT - Graphic Arts 3,423
Reduced Solvent Utilization - Surface Coating 342
Reformulation - Aerosol Paints 12
Reformulation - Architectural Coatings 1,912
Reformulation - Industrial Adhesives 5,076
Reformulation-Process Modification - Automobile Refmishing 4,665
Reformulation-Process Modification - Cold Cleaning 5,877
Reformulation-Process Modification - Cutback Asphalt 6,478
Reformulation-Process Modification - Open Top Degreasing 47
Reformulation-Process Modification - Surface Coating 4,225
Wastewater Treatment Controls- POTWs 237
Table 4A-7. NOx Control Measures Applied in the 70 ppb Alternative Standard Analysis
Reductions
NOx Control Measure (tons/yr)
Adjust Air to Fuel Ratio and Ignition Retard - Gas Fired 1C Engines 8,091
Biosolid Injection Technology - Cement Kilns 1,315
Episodic Ban - Open Burning 2,086
Ignition Retard - 1C Engines 486
Low Emission Combustion - Gas Fired Lean Burn 1C Engines 83,046
Low NOx Burner - Coal Cleaning 255
Low NOx Burner - Commercial/Institutional Boilers & 1C Engines 20,004
Low NOx Burner - Industr/Commercial/Institutional (ICI) Boilers 2,290
Low NOx Burner - Industrial Combustion 495
Low NOx Burner - Lime Kilns 1,870
Low NOx Burner - Natural Gas-Fired Turbines 14,408
Low NOx Burner - Residential Furnaces 12,056
Low NOx Burner - Residential Water Heaters & Space Heaters 18,843
Low NOx Burner and Flue Gas Recirculation - Coke Oven/Blast Furnace 350
4A-6
-------
Reductions
NOx Control Measure (tons/yr)
Low NOx Burner and Flue Gas Recirculation - Iron & Steel Mills - Reheating 388
Low NOx Burner and Over Fire Air - Utility Boilers 178
Low NOx Burner and SCR - Coal-Fired ICI Boilers 7,197
Low NOx Burner and SCR - Industr/Commercial/Institutional Boilers 22,632
Low NOx Burner and SNCR - Industr/Commercial/Institutional Boilers 153
Low Sulfur Fuel - Miscellaneous 2,892
Natural Gas Reburn - Natural Gas-Fired ECU Boilers 262
Nonroad Diesel Retrofits & Engine Rebuilds - e.g., Construction Equipment 4,984
Non-Selective Catalytic Reduction (NSCR) - 4 Cycle Rich Burn 1C Engines 189,563
Non-Selective Catalytic Reduction (NSCR) - Nitric Acid Mfg 659
OXY-Firing - Glass Manufacturing 17,155
SCR and Flue Gas Recirculation - Fluid Catalytic Cracking Units 163
SCR and Flue Gas Recirculation - ICI Boilers 307
Selective Catalytic Reduction (SCR) - Ammonia Mfg 4,476
Selective Catalytic Reduction (SCR) - Cement Kilns 18,260
Selective Catalytic Reduction (SCR) - Fluid Catalytic Cracking Units 2,995
Selective Catalytic Reduction (SCR) - 1C Engines, Diesel 1,675
Selective Catalytic Reduction (SCR) - ICI Boilers 6,049
Selective Catalytic Reduction (SCR) - Industrial Combustion 1,111
Selective Catalytic Reduction (SCR) - Industrial Incinerators 717
Selective Catalytic Reduction (SCR) - Petroleum Refinery Gas-Fired Process Heaters 4,757
Selective Catalytic Reduction (SCR) - Process Heaters 19
Selective Catalytic Reduction (SCR) - Sludge Incineration 7,991
Selective Catalytic Reduction (SCR) - Utility Boilers 30,482
Selective Non-Catalytic Reduction (SNCR) - Coke Mfg 1,543
Selective Non-Catalytic Reduction (SNCR) - CommVInst. Incinerators 1,082
Selective Non-Catalytic Reduction (SNCR) - ICI Boilers 154
Selective Non-Catalytic Reduction (SNCR) - Industrial Incinerators 482
Selective Non-Catalytic Reduction (SNCR) - Miscellaneous 31
Selective Non-Catalytic Reduction (SNCR) - Municipal Waste Combustors 304
Ultra-Low NOx Burner - Process Heaters 229
4A-7
-------
Table 4A-8. VOC Control Measures Applied in the 70 ppb Alternative Standard Analysis
Reductions
VOC Control Measure (tons/yr)
Control of Fugitive Releases - Oil & Natural Gas Production 16
Control Technology Guidelines - Wood Furniture Surface Coating 1,184
Flare - Petroleum Flare 110
Gas Recovery - Municipal Solid Waste Landfill 242
Improved Work Practices, Material Substitution, Add-On Controls - Cleaning Solvents 10
Improved Work Practices, Material Substitution, Add-On Controls - Printing 148
Incineration - Other 350
Incineration - Surface Coating 14,071
Low-VOC Coatings and Add-On Controls - Surface Coating 209
LPV Relief Valve - Underground Tanks 4,011
Permanent Total Enclosure (PTE) - Surface Coating 446
RACT - Graphic Arts 3,054
Reduced Solvent Utilization - Surface Coating 2,541
Reformulation - Architectural Coatings 17,678
Reformulation - Industrial Adhesives 1,793
Reformulation-Process Modification - Automobile Refmishing 3,209
Reformulation-Process Modification - Cold Cleaning 1,600
Reformulation-Process Modification - Cutback Asphalt 817
Reformulation-Process Modification - Surface Coating 3,571
Solvent Recovery System - Printing/Publishing 31
Wastewater Treatment Controls- POTWs 217
Table 4A-9. NOx Control Measures Applied in the 65 ppb Alternative Standard Analysis
Reductions
NOx Control Measure (tons/yr)
Adjust Air to Fuel Ratio and Ignition Retard - Gas Fired 1C Engines 27,057
Biosolid Injection Technology - Cement Kilns 6,423
Episodic Ban - Open Burning 4,423
Ignition Retard - 1C Engines 761
Low Emission Combustion - Gas Fired Lean Burn 1C Engines 174,033
Low NOx Burner - Coal Cleaning 518
Low NOx Burner - Commercial/Institutional Boilers & 1C Engines 40,691
Low NOx Burner - Industr/Commercial/Institutional (ICI) Boilers 4,226
Low NOx Burner - Industrial Combustion 1,578
Low NOx Burner - Lime Kilns 5,273
Low NOx Burner - Miscellaneous Sources 21
Low NOx Burner - Natural Gas-Fired Turbines 26,982
Low NOx Burner - Residential Furnaces 16,660
Low NOx Burner - Residential Water Heaters & Space Heaters 57,314
Low NOx Burner - Steel Foundry Furnaces 294
Low NOx Burner - Surface Coating Ovens 26
Low NOx Burner and Flue Gas Recirculation - Coke Oven/Blast Furnace 420
4A-8
-------
Reductions
NOx Control Measure (tons/yr)
Low NOx Burner and Flue Gas Recirculation - Iron & Steel Mills - Reheating 892
Low NOx Burner and Over Fire Air - Utility Boilers 333
Low NOx Burner and SCR - Coal-Fired ICI Boilers 30,790
Low NOx Burner and SCR - Industr/Commercial/Institutional Boilers 37,948
Low NOx Burner and SNCR - Industr/Commercial/Institutional Boilers 263
Low Sulfur Fuel - Miscellaneous 3,194
Natural Gas Reburn - Natural Gas-Fired ECU Boilers 480
Nonroad Diesel Retrofits & Engine Rebuilds - e.g., Construction Equipment 12,863
Non-Selective Catalytic Reduction (NSCR) - 4 Cycle Rich Burn 1C Engines 323,763
Non-Selective Catalytic Reduction (NSCR) - Nitric Acid Mfg 927
OXY-Firing - Glass Manufacturing 29,546
SCR and Flue Gas Recirculation - Fluid Catalytic Cracking Units 185
SCR and Flue Gas Recirculation - ICI Boilers 318
SCR and Flue Gas Recirculation - Process Heaters 548
Selective Catalytic Reduction (SCR) - Ammonia Mfg 5,151
Selective Catalytic Reduction (SCR) - Cement Kilns 36,013
Selective Catalytic Reduction (SCR) - Fluid Catalytic Cracking Units 4,108
Selective Catalytic Reduction (SCR) - 1C Engines, Diesel 8,905
Selective Catalytic Reduction (SCR) - ICI Boilers 18,284
Selective Catalytic Reduction (SCR) - Industrial Combustion 4,428
Selective Catalytic Reduction (SCR) - Industrial Incinerators 1,006
Selective Catalytic Reduction (SCR) - Iron Ore Processing 1,195
Selective Catalytic Reduction (SCR) - Petroleum Refinery Gas-Fired Process Heaters 7,691
Selective Catalytic Reduction (SCR) - Process Heaters 19
Selective Catalytic Reduction (SCR) - Sludge Incineration 9,007
Selective Catalytic Reduction (SCR) - Space Heaters 272
Selective Catalytic Reduction (SCR) - Utility Boilers 211,200
Selective Non-Catalytic Reduction (SNCR) - Coke Mfg 2,399
Selective Non-Catalytic Reduction (SNCR) - Comm/Inst. Incinerators 1,260
Selective Non-Catalytic Reduction (SNCR) - ICI Boilers 170
Selective Non-Catalytic Reduction (SNCR) - Industrial Combustion 69
Selective Non-Catalytic Reduction (SNCR) - Industrial Incinerators 1,502
Selective Non-Catalytic Reduction (SNCR) - Miscellaneous 132
Selective Non-Catalytic Reduction (SNCR) - Municipal Waste Combustors 1,351
Selective Non-Catalytic Reduction (SNCR) - Utility Boilers 329
Ultra-Low NOx Burner - Process Heaters 300
4A-9
-------
Table 4A-10. VOC Control Measures Applied in the 65 ppb Alternative Standard Analysis
Reductions
VOC Control Measure (tons/yr)
Control of Fugitive Releases - Oil & Natural Gas Production 31
Control Technology Guidelines - Wood Furniture Surface Coating 1,928
Flare - Petroleum Flare 110
Gas Recovery - Municipal Solid Waste Landfill 332
Improved Work Practices, Material Substitution, Add-On Controls - Cleaning Solvents 245
Improved Work Practices, Material Substitution, Add-On Controls - Printing 564
Incineration - Other 379
Incineration - Surface Coating 25,785
Low VOC Adhesives and Improved Application Methods - Industrial Adhesives 223
Low-VOC Coatings and Add-On Controls - Surface Coating 1,267
LPV Relief Valve - Underground Tanks 7,317
Permanent Total Enclosure (PTE) - Surface Coating 1,554
Petroleum and Solvent Evaporation - Surface Coating Operations 159
RACT - Graphic Arts 5,988
Reduced Solvent Utilization - Surface Coating 2,796
Reformulation - Architectural Coatings 39,057
Reformulation - Industrial Adhesives 1,698
Reformulation-Process Modification - Automobile Refmishing 5,264
Reformulation-Process Modification - Cutback Asphalt 3,058
Reformulation-Process Modification - Surface Coating 6,913
Solvent Recovery System - Printing/Publishing 842
Solvent Substitution and Improved Application Methods - Fiberglass Boat Mfg 14
Wastewater Treatment Controls- POTWs 242
Table 4A-11. NOx Control Measures Applied in the 60 ppb Alternative Standard Analysis
Reductions
NOx Control Measure (tons/yr)
Adjust Air to Fuel Ratio and Ignition Retard - Gas Fired 1C Engines 27,547
Biosolid Injection Technology - Cement Kilns 6,423
Episodic Ban - Open Burning 4,561
Ignition Retard - 1C Engines 821
Low Emission Combustion - Gas Fired Lean Burn 1C Engines 178,146
Low NOx Burner - Coal Cleaning 518
Low NOx Burner - Commercial/Institutional Boilers & 1C Engines 40,876
Low NOx Burner - Industr/Commercial/Institutional (ICI) Boilers 4,319
Low NOx Burner - Industrial Combustion 1,578
Low NOx Burner - Lime Kilns 5,273
Low NOx Burner - Miscellaneous Sources 35
Low NOx Burner - Natural Gas-Fired Turbines 29,155
Low NOx Burner - Residential Furnaces 17,103
Low NOx Burner - Residential Water Heaters & Space Heaters 57,726
Low NOx Burner - Steel Foundry Furnaces 294
4 A-10
-------
Reductions
NOx Control Measure (tons/yr)
Low NOx Burner - Surface Coating Ovens 26
Low NOx Burner and Flue Gas Recirculation - Coke Oven/Blast Furnace 420
Low NOx Burner and Flue Gas Recirculation - Iron & Steel Mills - Reheating 892
Low NOx Burner and Over Fire Air - Utility Boilers 333
Low NOx Burner and SCR - Coal-Fired ICI Boilers 30,817
Low NOx Burner and SCR - Industr/Commercial/Institutional Boilers 38,350
Low NOx Burner and SNCR - Industr/Commercial/Institutional Boilers 263
Low Sulfur Fuel - Miscellaneous 3,194
Natural Gas Reburn - Natural Gas-Fired ECU Boilers 502
Nonroad Diesel Retrofits & Engine Rebuilds - e.g., Construction Equipment 12,863
Non-Selective Catalytic Reduction (NSCR) - 4 Cycle Rich Burn 1C Engines 325,380
Non-Selective Catalytic Reduction (NSCR) - Nitric Acid Mfg 958
OXY-Firing - Glass Manufacturing 29,546
SCR and Flue Gas Recirculation - Fluid Catalytic Cracking Units 185
SCR and Flue Gas Recirculation - ICI Boilers 318
SCR and Flue Gas Recirculation - Process Heaters 548
Selective Catalytic Reduction (SCR) - Ammonia Mfg 5,151
Selective Catalytic Reduction (SCR) - Cement Kilns 36,013
Selective Catalytic Reduction (SCR) - Fluid Catalytic Cracking Units 4,135
Selective Catalytic Reduction (SCR) - 1C Engines, Diesel 9,288
Selective Catalytic Reduction (SCR) - ICI Boilers 18,284
Selective Catalytic Reduction (SCR) - Industrial Combustion 4,428
Selective Catalytic Reduction (SCR) - Industrial Incinerators 1,006
Selective Catalytic Reduction (SCR) - Iron Ore Processing 1,195
Selective Catalytic Reduction (SCR) - Petroleum Refinery Gas-Fired Process Heaters 7,691
Selective Catalytic Reduction (SCR) - Process Heaters 19
Selective Catalytic Reduction (SCR) - Sludge Incineration 9,007
Selective Catalytic Reduction (SCR) - Space Heaters 335
Selective Catalytic Reduction (SCR) - Utility Boilers 236,736
Selective Non-Catalytic Reduction (SNCR) - Coke Mfg 2,399
Selective Non-Catalytic Reduction (SNCR) - CommVInst. Incinerators 1,260
Selective Non-Catalytic Reduction (SNCR) - ICI Boilers 170
Selective Non-Catalytic Reduction (SNCR) - Industrial Combustion 92
Selective Non-Catalytic Reduction (SNCR) - Industrial Incinerators 1,502
Selective Non-Catalytic Reduction (SNCR) - Miscellaneous 132
Selective Non-Catalytic Reduction (SNCR) - Municipal Waste Combustors 1,351
Selective Non-Catalytic Reduction (SNCR) - Utility Boilers 329
Ultra-Low NOx Burner - Process Heaters 300
4 A-11
-------
Table 4A-12. VOC Control Measures Applied in the 60 ppb Alternative Standard Analysis
Reductions
VOC Control Measure (tons/yr)
Control of Fugitive Releases - Oil & Natural Gas Production 33
Control Technology Guidelines - Wood Furniture Surface Coating 2,063
Flare - Petroleum Flare 110
Gas Recovery - Municipal Solid Waste Landfill 372
Improved Work Practices, Material Substitution, Add-On Controls - Cleaning Solvents 265
Improved Work Practices, Material Substitution, Add-On Controls - Printing 564
Incineration - Other 379
Incineration - Surface Coating 26,109
Low VOC Adhesives and Improved Application Methods - Industrial Adhesives 237
Low-VOC Coatings and Add-On Controls - Surface Coating 1,523
LPV Relief Valve - Underground Tanks 7,610
Permanent Total Enclosure (PTE) - Surface Coating 1,857
Petroleum and Solvent Evaporation - Surface Coating Operations 237
RACT - Graphic Arts 6,273
Reduced Solvent Utilization - Surface Coating 2,886
Reformulation - Architectural Coatings 40,866
Reformulation - Industrial Adhesives 1,698
Reformulation-Process Modification - Automobile Refinishing 5,633
Reformulation-Process Modification - Cutback Asphalt 3,571
Reformulation-Process Modification - Surface Coating 7,072
Solvent Recovery System - Printing/Publishing 888
Solvent Substitution and Improved Application Methods - Fiberglass Boat Mfg 14
Wastewater Treatment Controls- POTWs 242
4A.4 VOC Control Measures for Non-EGU Point Sources
VOC controls were applied to a number of non-EGU point sources. Some examples are
permanent total enclosures (PTE) applied to paper and web coating operations and fabric
operations, and incinerators or thermal oxidizers applied to wood products and marine surface
coating operations. A PTE confines VOC emissions to a particular area where they can be
destroyed or used in a way that limits emissions to the outside atmosphere, and an incinerator or
thermal oxidizer destroys VOC emissions through exposure to high temperatures (2,000 degrees
Fahrenheit or higher). Another control is petroleum and solvent evaporation applied to printing
and publishing sources as well as to surface coating operations.
4A.5 NOX Control Measures for Nonpoint (Area) and Nonroad Sources
The nonpoint source sector of the emissions inventory is composed of sources that are
generally too small and/or numerous to estimate emissions on an individual source basis (e.g.,
4 A-12
-------
dry cleaners, residential furnaces, woodstoves, fireplaces, backyard waste burning, etc). Instead,
we estimate their emissions for each county as a whole, often using an emissions factor that is
applied to a surrogate of activity such as population or number of houses.
Control measures for nonpoint sources are also applied at the county level, i.e., to the
county level emissions as a whole. Several control measures were applied to NOX emissions from
nonpoint sources. One is low NOX burner technology to reduce NOX emissions. This control is
applied to industrial oil, natural gas, and coal combustion sources. Other nonpoint source
controls include the installation of low-NOx space heaters and water heaters in commercial and
institutional sources, and episodic bans on open burning. The open burning control measure
applied to yard waste and land clearing debris. It consists of periodic daily bans on burning such
waste, as the predicted ozone levels indicate that such burning activities should be postponed.
This control measure is not applied to any prescribed burning activities.
Retrofitting diesel nonroad equipment can provide NOX and HC benefits. The retrofit
strategies included in the RIA nonroad retrofit measure are:
• Installation of emissions after-treatment devices called selective catalytic reduction
("SCRs")
• Rebuilding engines ("rebuild/upgrade kit")
We chose to focus on these strategies due to their high NOX emissions reduction potential
and widespread application.
4A.6 VOC Control Measures for Nonpoint (Area) Sources
Some VOC controls for nonpoint sources are for the use of low or no VOC materials for
graphic art sources. Other controls involve the application of limits for adhesive and sealant
VOC content in wood furniture and solvent source categories. The OTC solvent cleaning rule
establishes hardware and operating requirements for specified vapor cleaning machines, as well
as solvent volatility limits and operating practices for cold cleaners. The Low Pressure/Vacuum
Relief Valve control measure is the addition of low pressure/vacuum (LP/V) relief valves to
gasoline storage tanks at service stations with Stage II control systems. LP/V relief valves
4 A-13
-------
prevent breathing emissions from gasoline storage tank vent pipes. Another control based on a
California South Coast Air Quality Management District (SQAQMD) establishes VOC content
limits for metal coatings along with application procedures and equipment requirements.
Switching to Emulsified Asphalts is a generic control measure replacing VOC-containing
cutback asphalt with VOC-free emulsified asphalt. The Reformulation control measures include
switching to and/or encouraging the use of low-VOC materials.
4 A-14
-------
CHAPTER 5: HUMAN HEALTH BENEFITS ANALYSIS APPROACH AND RESULTS
5.1 Synopsis
This chapter presents the estimated human health benefits for the proposed range of
National Ambient Air Quality Standards (NAAQS) for ozone. In this chapter, we quantify the
health-related benefits of the ozone air quality improvements resulting from the illustrative
emission control scenarios that reduce emissions of the ozone precursor pollutants nitrogen
oxides (NOX) and volatile organic compounds (VOCs) to reach the set of alternative ozone
NAAQS levels being considered. This chapter also estimates the health co-benefits of the fine
particulate matter (PM2.5)-related air quality improvements that would occur as a result of
reducing NOx emissions.33 The EPA Administrator is proposing to revise the level of the
primary ozone standard to within a range of 65 to 70 ppb and is soliciting comment on
alternative standard levels below 65 ppb, and as low as 60 ppb. In the Regulatory Impact
Analysis (RIA) we analyze the following alternative standard levels: 70, 65 and 60 ppb.
We selected 2025 as the primary year of analysis because the Clean Air Act requires
most areas of the U.S. to meet a revised ozone standard by 2025. Benefits are estimated
incremental to attainment of the existing standard of 75 ppb. In estimating the incremental costs
and benefits of potential alternative standards, we recognize that there are several areas that the
Act does not require to meet the existing ozone standard of 75 ppb by 2025. The Clean Air Act
provides areas with more significant air quality problems with additional time to reach the
existing standard. Several areas in California are not expected to meet the existing standard by
2025 and may not be required to meet a revised standard until December 31, 2037.
We estimated the benefits of California attaining a revised standard in 2038 to account
for the fact that many locations in this state must attain a revised standard at a later date than the
rest of the U.S. We assume that projected nonattainment areas everywhere in the U.S. excluding
California will be designated such that they attain a revised standard by 2025, and we develop
our projected baseline emissions, air quality, and population estimates for 2025. We also assume
33 VOC reductions associated with simulated attainment of alternative ozone standards also have the potential to
impact PM.2.5 concentrations, but we are not able to model those effects at this time.
5-1
-------
that the projected nonattainment areas in California will be designated such that they reach
attainment by approximately 2038.
We were not able to project baseline emissions and air quality levels beyond 2025 for
California for sectors other than mobile sources. We account for changes in mobile source
precursor emissions expected to occur between 2025 and 2030. While there is uncertainty about
the precise timing of emissions reductions and related costs for California, we assume costs
occur through 2038. We also model benefits for California accounting for population growth to
2038 (see section 5.6.1). Because we were unable to account for the change in emissions from
other sectors our projected baseline may under- or over-estimate the post-2025 ozone levels in
California. In this analysis, we refer to estimates of nationwide benefits of attaining an
alternative standard everywhere in the U.S. except California as the 2025 scenario. The post-
202 5 scenario refers to estimates of nationwide benefits of attaining an alternative standard just
in California.
Because we estimate incremental costs and benefits for these two distinct scenarios
reflecting attainment in different years, it is not appropriate to either sum, or directly compare,
the estimates. Consequently, in presenting both incidence and dollar benefit estimates in this
chapter, we present and discuss the 2025 scenario and post-2025 scenario in separate sections
(see sections 5.7.1 and 5.7.2, respectively).
Benefits estimated for the 2025 and post-2025 scenarios are relative to an analytical
baseline in which the nation attains the current primary ozone standard (i.e., 4th highest daily
maximum 8-hour ozone concentration of 75 ppb) and incorporates promulgated national
regulations and illustrative emission controls to simulate attainment with 75 ppb. Table 5-1
summarizes the estimated monetized benefits (total and ozone only) of attaining alternative
ozone standards of 70 ppb, 65 ppb, and 60 ppb for the 2025 scenario (i.e., nationwide benefits of
attaining everywhere in the U.S. but California). Table 5-2 presents the same types of benefit
estimates for the post-2025 scenario (i.e., nationwide benefits of attaining just in California).
These estimates reflect the sum of the economic value of estimated morbidity and mortality
effects related to changes in exposure to ozone and fine particulate matter (PIVh.s). However, it is
important to emphasize that it is not appropriate to compare the ozone-only benefits to total
5-2
-------
costs. There are additional unquantified benefits which are described in Section 5.2. The
estimated benefits for attaining the proposed standards are incremental to the substantial benefits
estimated for several recent implementation rules (e.g., U.S. EPA, 201 le, 2014a).
Table 5-1. Estimated Monetized Benefits of Attainment of the Alternative Ozone Standards
for the 2025 Scenario (nationwide benefits of attaining each alternative standard
everywhere in the U.S. except California) - Full Attainment (billions of 2011$) a
Total Benefits
Discount
Rate
3%
7%
70ppb
$6.9to$14+B
$6.4to$13+B
65 ppb
$20to$41+B
$19 to $38 +B
60 ppb
$37 to $75 +B
$34 to $70 +B
Ozone-only Benefits (range
reflects Smith etal, 2009 and b $2.0 to $3.4+B $6.4to$ll+B $12to$20+B
Zanobetti and Schwartz, 2008)
PM2.s Co-benefits (range reflects 3o/o $4.8 to $11 $14 to $31 $25 to $56
Krewski et al., 2009 and Lepeule
etal, 2012) 7% $4.3 to $9.7 $12 to $28 $22 to $50
a Rounded to two significant figures. It was not possible to quantify all benefits in this analysis due to data
limitations. "B" is the sum of all unquantified health and welfare benefits. These estimates reflect the economic
value of avoided morbidities and premature deaths using risk coefficients from the studies noted.
b Ozone-only benefits reflect short-term exposure impacts and as such are assumed to occur in the same year as
ambient ozone reductions. Consequently, social discounting is not applied to the benefits for this category.
Table 5-2. Estimated Monetized Benefits of Attainment of the Alternative Ozone Standards
for the Post-2025 Scenario (nationwide benefits of attaining each alternative
standard just in California) - Full Attainment (billions of 2011$) a
Discount „„ , ,_ , ,„ ,
Rate 70 ppb 65 ppb 60 ppb
~ ~ ~ 3% $l.lto$2.0+B $2.3 to $4.2 +B $3.4 to $6.2+B
Total Benefits 7% $1.1 to $2.0+B $2.2 to $4.1+B $3.2 to $5.9+B
Ozone-only Benefits (range reflects
Smith et al., 2009 and Zanobetti and b en ^ t ci i ci A t co A co i t c^ f.
„ , . ^nox $0.66 to $1.1 $1.4 to $2.4 $2.1 to $3.6
Schwartz, 2008)
PM2.s Co-benefits (range reflects 3o/o $0.42 to $0.95 $0.83 to $1.9 $1.1 to $2.6
Krewski et al., 2009 and Lepeule et
al 2012) 7% $0.38 to $0.86 $0.75 to $1.7 $1.0 to $2.3
a Rounded to two significant figures. It was not possible to quantify all benefits in this analysis due to data
limitations. "B" is the sum of all unquantified health and welfare benefits. These estimates reflect the economic
value of avoided morbidities and premature deaths using risk coefficients from the studies noted.
b Ozone-only benefits reflect short-term exposure impacts and as such are assumed to occur in the same year as
ambient ozone reductions. Consequently, social discounting is not applied to the benefits for this category.
In addition to ozone and PIVb.s benefits, implementing emissions controls to reach some
of the alternative ozone standards would reduce other ambient pollutants, such as VOCs and
5O
-------
NO2. However, because the method used in this analysis to simulate attainment does not account
for changes in ambient concentrations of other pollutants, we were not able to quantify the co-
benefits of reduced exposure to these pollutants. In addition, due to data and methodology
limitations, we were unable to estimate some anticipated health benefits associated with
exposure to ozone and PIVb.s.
5.2 Overview
This chapter presents estimated health benefits for three alternative ozone standards (70,
65 and 60ppb) that the EPA could quantify, given the available resources, data and methods.
Separate set of benefits are presented for the 2025 scenario, representing nationwide benefits of
attaining an alternative standard everywhere in the U.S. but California in 2025 and the post-2025
scenario, representing nationwide benefits of attaining an alternative standard just in California
in 2037. This chapter characterizes the benefits of implementing new ozone standards by
answering three key questions:
1. What health effects are avoided by reducing ambient ozone levels to attain a
revised ozone standard?
2. What is the economic value of these effects?
3. What are the co-benefits of reductions in ambient PIVb.s associated with
reductions in emissions of ozone precursors (specifically NOx)?
In this analysis, we quantify an array of adverse health impacts attributable to ozone and
PM2.5. The Integrated Science Assessment for Ozone and Related Photochemical Oxidants
("ozone ISA") (U.S. EPA, 2013a) identifies the human health effects associated with ozone
exposure, which include premature death and a variety of illnesses associated with acute (days-
long) and chronic (months to years-long) exposures. Similarly, the Integrated Science
Assessment for Paniculate Matter ("PM ISA") (U.S. EPA, 2009b) identifies the human health
effects associated with ambient particles, which include premature death and a variety of
illnesses associated with acute and chronic exposures. Air pollution can affect human health in a
variety of ways, and in Table 5-3 we summarize the "categories" of effects and describe those
that we could quantify in our "core" benefits estimates and those we were unable to quantify
because we lacked the data, time or techniques.
5-4
-------
This list of unquantified benefit categories is not exhaustive and we are not always able
to quantify each effect completely. Endpoints that the ozone and PM ISAs classified as causal or
likely causal we quantified with confidence. We excluded from quantification effects not
identified as having at least a causal or likely causal relationship with the affected pollutants.
Selecting endpoints in this way should not imply that these pollutants are unrelated to other
human health and environmental effects. Following this criterion, we excluded some effects that
were identified in previous lists of unquantified benefits in other RIAs (e.g., UVb exposure), but
are not identified in the most recent ISA as having a causal or likely causal relationship with
ozone. In designing this benefits analysis, including the identification of endpoints to include in
the core estimate, we also considered the design of the Health Risk and Exposure Assessment
(HREA) completed as part of this ozone NAAQS review (USEPA, 2014b). The design and
implementation of the HREA was subjected to rigorous peer review by the Science Advisory
Board's (SAB's) Clean Air Scientific Advisory Committee (CAS AC) with the results of that
review being presented in letter form (Frey, and Samet 2012 for the first draft and Frey, 2014 for
the second draft of the HREA). The overall design of the HREA, including the health endpoints
selected for modeling was supported by the CAS AC.34
This benefits analysis relies on an array of data inputs—including emissions estimates,
modeled ozone air quality, health impact functions and valuation estimates among others—
which are themselves subject to uncertainty and may in turn contribute to the overall uncertainty
in this analysis. We employ several techniques to characterize this uncertainty, which are
described in detail in sections 5.5 and 5.7.3.
Table 5-3. Human Health Effects of Pollutants Potentially Affected by Strategies to Attain
the Primary Ozone Standards (endpoints included in the core analysis are identified
with a red checkmark)
Benefits Category
Specific Effect
Effect Has
Been
Quantified
Effect Has
Been
Monetized
More
Information
Improved Human Health
34 The CAS AC expressed their support for the overall design of the HREA, including endpoints selected and
epidemiological studies used in supplying the effect estimates used to model those endpoints (Samet and Frey, 2012,
p. 15 and Frey, 2014, p. 9).
5-5
-------
Benefits Category
Specific Effect
Effect Has
Been
Quantified
Effect Has
Been
Monetized
More
Information
Reduced incidence
of premature
mortality from
exposure to ozone
Premature mortality based on short-term
exposure (all ages)
Premature respiratory mortality based on
long-term exposure (age 30-99)
Reduced incidence
of morbidity from
exposure to ozone
Hospital admissions—respiratory causes
(age > 65)
Emergency department visits for asthma
(all ages)
Asthma exacerbation (age 6-18)
Minor restricted-activity days (age 18-65)
School absence days (age 5-17)
Decreased outdoor worker productivity
(age 18-65)
Other respiratory effects (e.g.,
medication use, pulmonary
inflammation, decrements in lung
functioning)
Cardiovascular (e.g., hospital admissions,
emergency department visits)
Reproductive and developmental effects
(e.g., reduced birthweight, restricted
fetal growth)
Section 5.6
ozone ISA'
Reduced incidence
of premature
mortality from
exposure to Plvh.s
Adult premature mortality based on
cohort study estimates and expert
elicitation estimates (age >25 or age >30)
Infant mortality (age <1)
Reduced incidence
of morbidity from
exposure to
Non-fatal heart attacks (age > 18)
Hospital admissions—respiratory (all
ages)
Hospital admissions—cardiovascular (age
>20)
Emergency department visits for asthma
(all ages)
Acute bronchitis (age 8-12)
Lower respiratory symptoms (age 7-14)
Upper respiratory symptoms (asthmatics
age 9-11)
Asthma exacerbation (asthmatics age 6-
18)
Lost work days (age 18-65)
Minor restricted-activity days (age 18-65)
Chronic Bronchitis (age >26)
Emergency department visits for
cardiovascular effects (all ages)
Strokes and cerebrovascular disease (age
50-79)
Other cardiovascular effects (e.g., other
ages)
See section
5.6 and
Appendix 5D
PMISAC
5-6
-------
Benefits Category
Specific Effect
Effect Has
Been
Quantified
Effect Has
Been
Monetized
More
Information
Other respiratory effects (e.g., pulmonary
function, non-asthma ER visits, non-
bronchitis chronic diseases, other ages
and populations)
Reproductive and developmental effects
(e.g., low birth weight, pre-term births,
etc.)
Cancer, mutagenicity, and genotoxicity
effects
PMISAc-d
Reduced incidence Asthma hospital admissions (all ages)
of morbidity from Chronic lung disease hospital admissions
exposure to NO2 (age > 65)
Respiratory emergency department visits
(all ages)
Asthma exacerbation (asthmatics age 4-
18)
Acute respiratory symptoms (age 7-14)
Premature mortality
Other respiratory effects (e.g., airway
hyperresponsiveness and inflammation,
lung function, other ages and
populations)
NO2ISA'
a Due to concerns over translating incidence estimates into dollar benefits, for long-term ozone exposure-related
respiratory mortality, we included estimates of reduced incidence as part of the core analysis, but included
associated dollar benefits as a sensitivity analysis (see section 5.3).
b We are in the process of considering an update to the worker productivity analysis for ozone based on more recent
literature (see section 5.6.3.4). As noted in section 5.6.3.4 we are requesting public comment on the approach we
present for modeling this endpoint and will consider that input in determining whether to proceed with an updated
simulation of this endpoint.
0 We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
d We assess these benefits qualitatively because current evidence is only suggestive of causality or there are other
significant concerns over the strength of the association.
e We assess these benefits qualitatively due to time and resource limitations for this analysis.
As described in Chapter 1 of this RIA, there are important differences worth noting in the
design and analytical objectives of NAAQS RIAs compared to RIAs for rules that implement
technology standards, such as Tier 3 (U.S. EPA, 2014a). The NAAQS RIAs illustrate the
potential costs and benefits of attaining a revised air quality standard nationwide. These analyses
simulate an array of strategies to reduce emissions at different sources and may model well-
established emission control technologies for sectors and emission controls for which the control
technology has not yet been developed (i.e., "unknown" controls). This type of RIA accounts for
existing regulations and controls needed to attain the current standards and so estimated benefits
and costs are incremental to attaining the current standard. In short, NAAQS RIAs hypothesize,
5-7
-------
but do not predict, the emission reduction strategies that States may enact when implementing a
revised NAAQS. Setting a NAAQS does not result directly in costs or benefits. By contrast, the
emission reductions from implementation rules are generally for specific, well-characterized
sources, such as the recent MATS rule addressing emissions from coal and oil-fired electricity
generating units (U.S. EPA, 201 le). In general, the EPA is more confident in the magnitude and
location of the emission reductions for implementation rules. As such, emission reductions
achieved under promulgated implementation rules such as MATS have been reflected in the
baseline of this NAAQS analysis (the full set of rules reflected in baseline are presented in
section 3.1.3). Subsequent implementation rules will be reflected in the baseline for the next
ozone NAAQS review. For this reason, the benefits estimated provided in this RIA and all other
NAAQS RIAs should not be added to the benefits estimated for implementation rules.
5.3 Updated Methodology Presented in this RIA
The benefits analysis presented in this chapter incorporates an array of policy and
technical changes that the Agency has adopted since the previous review of the ozone standards
in 2008 and the proposed reconsideration in 2010. Below we note the aspects of this analysis that
differ from the reconsideration RIA (U.S. EPA, 2010d):
1. The population demographic data in BenMAP-CE (U.S. EPA, 2014d) reflects the 2010
Census and future projections based on economic forecasting models developed by
Woods and Poole, Inc. (Woods and Poole, 2012). These data replace the earlier
demographic projection data from Woods and Poole (2007). This update was introduced
in the final PM NAAQS RIA (U.S. EPA, 2012b).
2. The baseline incidence rates used to quantify air pollution-related hospital admissions
and emergency department visits and the asthma prevalence rates were updated to replace
the earlier rates. This update was introduced in the final CSAPR (U.S. EPA, 201 Id).
3. We updated the median wage data in the cost-of-illness studies. This update was
introduced in the final PM NAAQS RIA (U.S. EPA, 2012b).
4. Updates for ozone-related effects:
a. Incorporated new mortality studies. We include two new multi-city studies to
estimate deaths attributable to short-term exposure for the core analysis (Smith et al,
2009 and Zanobetti and Schwartz 2008). We also estimate long-term respiratory
deaths using Jerrett et al. (2009). While we believe the evidence supports including
long-term respiratory deaths in the core analysis, limitations in our ability to specify a
lag between exposure and the onset of death (i.e. the cessation lag that is required for
5-S
-------
valuing these deaths) prevents us from estimating dollar benefits in the core analysis.
Both the new short-term and long-term mortality studies were included in the HREA
completed in support of this NAAQS review with the overall design of that HREA
(including inclusion of these new studies) being subjected to rigorous review by
CAS AC (Frey, and Samet 2012; Frey, 2014).
b. Incorporated new morbidity studies. The ozone ISA (U.S. EPA, 2013a) identifies
several new epidemiological studies examining the association between short-term
ozone exposure and respiratory hospitalizations, respiratory emergency department
visits, and exacerbated asthma. Upon carefully evaluating this new literature, we
added several new studies to our health impact assessment. Several of these studies
were also included in the HREA, which as noted earlier, underwent rigorous review
by CASAC.
c. Expanded uncertainty assessment. We added a comprehensive, qualitative assessment
of the various uncertain parameters and assumptions within the benefits analysis and
expanded the evaluation of air quality benchmarks for ozone-related mortality. We
introduce this expanded assessment in this RIA (see sections 5.5 and 5.7.3).
5. Updates for PIVh.s-related effects
a. Incorporated new mortality studies. We updated the American Cancer Society cohort
study to Krewski et al. (2009) and updated the Harvard Six Cities cohort study to
Lepeule et al. (2012). The effect coefficient for Krewski et al. (2009) is identical to
the previous coefficient, and the Lepeule et al. (2012) is roughly similar to the
previous coefficient. Both studies show narrower confidence intervals. The update for
the American Cancer Society cohort was introduced in the proposal RIA for the PM
NAAQS review (U.S. EPA, 2012b) and the update for the Harvard Six Cities cohort
was introduced in the final RIA for the PM NAAQS review (U.S. EPA, 2012b).
b. Incorporated new morbidity studies. The epidemiological literature has produced
several recent studies examining the association between short-term PIVh.s exposure
and respiratory and cardiovascular hospitalizations, respiratory and cardiovascular
emergency department visits, and stroke. Upon careful evaluation of new literature in
the PM ISA and Provisional Assessment, we added several new studies and health
endpoints to our health impact assessment. These updates were introduced in the
proposal (U.S. EPA, 2012) and final RIAs for the PM NAAQS review (U.S. EPA,
2012b).
c. Updated the survival rates for non-fatal acute myocardial infarctions. Based on recent
data from Agency for Healthcare Research and Quality's Healthcare Utilization
Project National Inpatient Sample database (AHRQ, 2009), we identified death rates
for adults hospitalized with acute myocardial infarction stratified by age. These rates
replaced the survival rates from Rosamond et al. (1999). This update was introduced
in the final RIA for the PM NAAQS review (U.S. EPA, 2012b).
5-9
-------
d. Expanded uncertainty assessment. We clarified the comprehensive assessment of the
various uncertain parameters and assumptions within the benefits analysis and
expanded the evaluation of air quality benchmarks. This update was introduced in the
proposed CSAPR RIA (U.S. EPA, 2010g) and refined in the final PM NAAQS RIA
(U.S. EPA, 2012b).
Although the list above identifies the major changes implemented since the 2010 ozone
reconsideration RIA, the EPA has also updated several additional components of the benefits
analysis since the 2008 ozone NAAQS RIA (U.S. EPA, 2008a), which were reflected in the
reconsideration RIA. In the Portland Cement NESHAP proposal RIA (U.S. EPA, 2009a), the
Agency no longer assumed a concentration threshold in the concentration-response function for
PM2.s-related health effects and began estimating the benefits derived from the two major cohort
studies of PM2.5 and mortality as the core benefits estimates, while still including a range of
sensitivity estimates based on the EPA's PIVh.s mortality expert elicitation. In the NO2 NAAQS
proposal RIA (U.S. EPA, 2009a), we revised the estimate used for the value-of-a-statistical life
to be consistent with Agency guidance.
5.4 Human Health Benefits Analysis Methods
We follow a "damage-function" approach in calculating total benefits of the modeled
changes in environmental quality.35 This approach estimates changes in individual health
endpoints (specific effects that can be associated with changes in air quality) and assigns values
to those changes assuming independence of the values for those individual endpoints. Total
benefits are calculated simply as the sum of the values for all non-overlapping health endpoints.
The "damage-function" approach is the standard method for assessing costs and benefits of
environmental quality programs and has been used in several recent published analyses (Levy et
al, 2009; Fann et al, 2012a; Tagaris et al, 2009).
To assess economic values in a damage-function framework, the changes in
environmental quality must be translated into effects on people or on the things that people
value. In some cases, the changes in environmental quality can be directly valued, as is the case
for changes in visibility. In other cases, such as for changes in ozone and PM, an impact analysis
35 The damage function approach is a more comprehensive method of estimating total benefits than the hedonic
price approach applied to housing prices, which requires homebuyers to be knowledgeable of the full magnitude
of health risks associated with their home purchase.
5-10
-------
must first be conducted to convert air quality changes into effects that can be assigned dollar
values. For the purposes of this RIA, the health impacts analysis (HIA) is limited to those health
effects that are directly linked to ambient levels of air pollution and specifically to those linked to
ozone and PIVb.s.
We note at the outset that the EPA rarely has the time or resources to perform extensive
new research to measure directly either the health outcomes or their values for regulatory
analyses. Thus, similar to Kunzli et al. (2000) and other, more recent health impact analyses, our
estimates are based on the best available methods of benefits transfer. Benefits transfer is the
science and art of adapting primary research from similar contexts to obtain the most accurate
measure of benefits for the environmental quality change under analysis. Adjustments are made
for the level of environmental quality change, the socio-demographic and economic
characteristics of the affected population, and other factors to improve the accuracy and
robustness of benefits estimates.
Benefits estimates for ozone were generated using the damage function approach outlined
above wherein potential changes in ambient ozone levels (associated with future attainment of
alternative standard levels) were explicitly modeled and then translated into reductions in the
incidence of specific health endpoints. In generating ozone benefits estimates for the two
scenarios considered in the RIA (2025 and post-2025), we actually utilized three distinct benefits
simulations including one completed for 2025 and two completed for 2038. The way in which
these three benefits simulations were used to generate estimates for the two time periods reported
in the RIA (2025 and post-2035) is described in section 5.4.3.
In contrast to ozone, we used a benefit-per-ton (reduced form) approach in modeling PIVh.s
co-benefits (see section 5.4.4 for additional detail). With this approach, we utilize the results of
previous benefits analysis simulations focusing on PIVh.s to derive benefits-per-ton estimates for
NOx.36 We then combine these dollar-per-ton estimates with projected reductions in NOx
associated with meeting a given alternative standard to project co-benefits associated with PIVh.s.
We acknowledge increased uncertainty associated with the dollar-per-ton approach for PIVh.s,
36 In addition to dollar-per-ton estimates for NOx, we also utilized incidence-per-ton values (also for NOx) for
specific health endpoints in order to generate incidence reduction estimates associated with the dollar benefits.
5-11
-------
relative to explicitly modeling benefits using gridded PIVh.s surfaces specific to the baseline and
alternative scenarios being considered in this review (see sections 5.4.4 and 5.7.3 and Appendix
5A, Table 5A-1 for additional discussion).
Sections 5.4.1 and 5.4.2 describe respectively, the underlying basis for the health and
economic valuation estimates. Section 5.4.3 describes the procedure used to combine the three
benefits simulations referenced above in order to generate benefits for the two time periods
considered in the RIA (2025 and post-2025). Finally, section 5.4.4 provides an overview of the
benefit-per-ton estimates used to estimate the PIVh.s co-benefits from NOx emission reductions in
this RIA.
5.4.1 Health Impact Assessment
The health impact assessment (HIA) quantifies the changes in the incidence of adverse
health impacts resulting from changes in human exposure to PIVh.s and ozone air quality. HIAs
are a well-established approach for estimating the retrospective or prospective change in adverse
health impacts expected to result from population-level changes in exposure to pollutants (Levy
et al, 2009). PC-based tools such as the environmental Benefits Mapping and Analysis Program
- Community Edition (BenMAP-CE) can systematize health impact analyses by applying a
database of key input parameters, including health impact functions and population projections—
provided that key input data are available, including air quality estimates and risk coefficients
(U.S. EPA, 2014d). Analysts have applied the HIA approach to estimate human health impacts
resulting from hypothetical changes in pollutant levels (Hubbell et al., 2005; Tagaris et al., 2009;
Fann et al., 2012a). The EPA and others have relied upon this method to predict future changes
in health impacts expected to result from the implementation of regulations affecting air quality
(e.g., U.S. EPA, 2014d). For this assessment, the HIA is limited to those health effects that are
directly linked to ambient ozone and PIVb.s concentrations. There may be other indirect health
impacts associated with implementing emissions controls, such as occupational health exposures.
5-12
-------
The HIA approach used in this analysis involves three basic steps: (1) utilizing
projections of ozone air quality37 and estimating the change in the spatial distribution of the
ambient air quality; (2) determining the subsequent change in population-level exposure; (3)
calculating health impacts by applying concentration-response relationships drawn from the
epidemiological literature to this change in population exposure (Hubbell et al, 2009).
A typical health impact function might look as follows:
Ay = 1 - (e^A*)y0 • Pop (5.1)
where yo is the baseline incidence rate for the health endpoint being quantified (for
example, a health impact function quantifying changes in mortality would use the baseline, or
background, mortality rate for the given population of interest); Pop is the population affected by
the change in air quality; Ax is the change in air quality; and P is the effect coefficient drawn
from the epidemiological study. Figure 5-1 provides a simplified overview of this approach.
37 Projections of ambient ozone concentrations for this analysis were generated by applying emissions reductions
described in chapters 3 and 4 to gridded surfaces of recent-year ozone concentrations. The full methodology
which incorporates information both from ambient measurements and from photochemical modeling simulations
is described in section 3.4.1 of this RIA.
5-13
-------
Baseline Air Quality
Post-Policy Scenario Air Quality
Incremental Air
Quality
Improvement
Background
Incidence
Rate
>. Effect -
Estimate
Mortality
Reduction
Figure 5-1. Illustration of BenMAP-CE Approach
5.4.2 Economic Valuation of Health Impacts
After quantifying the change in adverse health impacts, the final step is to estimate the
economic value of these avoided impacts. The appropriate economic value for a change in a
health effect depends on whether the health effect is viewed ex ante (before the effect has
occurred) or ex post (after the effect has occurred). Reductions in ambient concentrations of air
pollution generally lower the risk of future adverse health effects by a small amount for a large
population. The appropriate economic measure is therefore ex ante willingness to pay (WTP) for
changes in risk. Epidemiological studies generally provide estimates of the relative risks of a
particular health effect for a given increment of air pollution (often per 10 ppb ozone). These
relative risks can be used to develop risk coefficients that relate a unit reduction in ozone or
PM2.5 to changes in the incidence of a health effect. In order to value these changes in incidence,
WTP for changes in risk need to be converted into WTP per statistical incidence. This measure is
5-14
-------
calculated by dividing individual WTP for a risk reduction by the related observed change in
risk.
For some health effects, such as hospital admissions, WTP estimates are generally not
available. In these cases, we use the costs of treating or mitigating the effect, which generally
understate the true value of reductions in risk of a health effect because they exclude the value of
avoided pain and suffering from the health effect.
We use the BenMAP-CE version 1.0.8 (U.S. EPA, 2014d) to estimate the health impacts
and monetized health benefits for the proposed standard range. Figure 5-2 shows the data inputs
and outputs for the BenMAP-CE program.38
Census
Population Data
Modeled Baseline
and Post-Control
Ambient Ozone
2025 Population
Projections
Woods & Poole
Population
Projections
Ozone Health
Functions
Economic
Valuation
Functions
Ozone Incremental
Air Quality Change
Ozone-Related
Health Impacts
Background
Incidence and
Prevalence Rates
Monetized Ozone-
related Benefits
Blue identifies a user-selected input within the BenMAP-CE program
Green identifies a data input generated outside of the BenMAP-CE program
38 The environmental Benefits Mapping and Analysis Program—Community Edition (BenMAP-CE) is an open-
source PC-based tool that quantifies the number and economic value of air pollution-related deaths and illnesses. As
compared to the version that it replaces, BenMAP v4, the BenMAP-CE tool uses the same computational algorithms
and input data to calculate incidence counts and dollar values—for a given air quality change, both versions report
the same estimates, within rounding. BenMAP-CE differs from the legacy version of BenMAP in two important
ways: (1) it is open-source and the uncompiled code is available to the public; (2) it is written in C#, which is both
more broadly used and modern than the code it replaces (Delphi). BenMAP-CE was last used to support the Ozone
Health Risk and Exposure Assessment completed in support of the current review.
5-15
-------
Figure 5-2. Data Inputs and Outputs for the BenMAP-CE Program
5.4.3 Estimating Benefits for the 2025 and Post-2025 Scenarios
As described in section 5.1, we estimated benefit for two scenarios: 2025 and post-2025.
The need for these two distinct time periods reflects the fact that, while most of the U.S. will
have attained both the current and any alternative standard by 2025, there are portions of the
country with more significant air quality problems (including several areas in California) that
may not be required to meet an alternative standard until as late as December 31, 2037.
Consequently, for each alternative standard we model a 2025 scenario reflecting the nationwide
benefits of attaining that standard everywhere in the U.S. except California. We then model a
post-2025 scenario, which represents nationwide benefits from attaining that same standard in
California. Due to the temporal disconnect between these two scenarios, we do not attempt to
sum these two estimates, but instead present each estimate in separate sections of this document
(sections 5.7.1 and 5.7.2, respectively).
Our approach for estimating the benefits of attaining alternate ozone standards post-2025
is illustrated in Figure 5-3; in this figure, Simulation A represents our approach for estimating the
benefits of attaining alternate ozone standards in every state except California in 2025. We first
estimated the benefits occurring in 2038 from all areas (including California) attaining each
alternative standard (Simulation C). Next, we simulated the nationwide benefits of attaining each
alternate ozone standard in 2038 for every state except California (Simulation B). Subtracting
Simulation B from Simulation C calculates the benefits of attaining each alternate ozone
standard after 2025— that is, the nationwide benefits from California alone attaining the standard
in 2038. There are important caveats associated with this approach mentioned in Section 5.1 and
discussed further in section 5.7.3.
5-16
-------
Procedure for Deriving 2025 and Pos1>202S Benefits Using 2025 and 2037
Benefits Simulations
BenMflP Benefits Simulations
2025 scenario: Nationwide benefits in 2025
resulting from ail in the U.S. attaining
the alternative standard under consideration
excluding California (here 70ppb). This is
A
Figure 5-3. Procedure for Generating Benefits Estimates for the 2025 and Post-2025
Scenarios
5.4.4 Benefit-per-ton Estimates for PM2.s
We used a "benefit-per-ton" approach to estimate the PIVb.s co-benefits in this RIA. EPA
has applied this approach in several previous RIAs (e.g., U.S. EPA, 2014a). These benefit-per-
ton estimates provide the total monetized human health co-benefits (the sum of premature
mortality), of reducing one ton of NOx(as a PIVh.s precursor) from a specified source.39 In
general, these estimates apply the same benefits methods (e.g., health impact assessment then
economic valuation), which are described further below, for all PIVh.s impacts attributable to a
sector, and these benefits are then divided by the tons of a PIVh.s precursor (e.g., NOx) from that
sector. As discussed below, we acknowledge that this approach has greater uncertainty relative to
explicitly modeling benefits for PIVh.s based on application of gridded surfaces specifically
39 In generating these estimates, we first use incidence-per-ton values to generate estimates of reductions in
morbidity and mortality incidence for core endpoints (see Table 5-3). Then in estimating dollar values associated
with these reductions in incidence, we use dollar-per-ton values for mortality only, noting that this is likely to
provide coverage for upwards of 97% of the total dollar benefits (i.e., morbidity endpoints provide less than 3% of
total benefits - see RIA from last PM review, USEPA, 2012, Table 5-19).
5-17
-------
generated for the baseline and alternative standard levels being considered in this ozone NAAQS
review. However, resource and time constraints prevented us from completing detailed PIVh.s
modeling as part of this review.
We used a method to calculate the regional benefit-per-ton estimates that is a slightly
modified version of the national benefit-per-ton estimates described in the TSD: Estimating the
Benefit per Ton of'ReducingPIVh.s Precursors from 17 Sectors (U.S. EPA, 2013b). The national
estimates used in this NAAQS review were derived using the approach published in Fann et al.
(2012c), but they have since been updated to reflect the epidemiology studies and Census
population data first applied in the final PM NAAQS RIA (U.S. EPA, 2012b). The approach in
Fann et al. (2012c) is similar to the work previously published by Fann et al. (2009), but the
newer study includes improvements that provide more refined estimates of PIVh.s-related health
benefits for emissions reductions in the various sectors. Specifically, the air quality modeling
data reflect industrial sectors that are more narrowly defined. In addition, the updated air quality
modeling data reflects more recent emissions data — a 2005 baseline projected to 2016 rather
than 2001 baseline projected to 2015 — and has higher spatial resolution (12 km rather than 36
km grid cells).40
In Section 5.6 below, we describe all of the data inputs used in deriving the dollar-per-ton
values for each sector, including the demographic data, baseline incidence, and valuation
functions. The specification of effect estimates (including selection of epidemiology studies)
used in the derivation of the benefit-per-ton values for PIVh.s is described in detail in Appendix
5D. The benefit-per-ton estimates (by sector) that resulted from this modeling as well as the NOx
reductions for each alternative standard level used in generating the PIVh.s cobenefit estimates are
presented in Appendix 5E. Additional information on the source apportionment modeling for
each of the sectors can be found in Fann et al. (2012c) and the TSD (U.S. EPA, 2013b).
40 Sector-level estimates of PM2s are modeled using CAMx version 5.30. Specifically, the paniculate source
apportionment technology (PS AT) incorporated into CAMx generates estimates of the contribution from specific
emission source groups to primary emitted and secondarily formed PM2.5, PSAT uses reactive tracers in generating
these fractional estimates in order to capture nonlinear formation and removal processes related to PM2.5.
Contributions from each sector are modeled at the 12km level, while boundary conditions are represented using a
36km grid resolution (additional detail on modeling can be found in Fann et al., 2012)
5-18
-------
Specifically for this analysis, we applied the benefit-per-ton estimates for 2025 and 2030 in
generating PIVh.s co-benefit estimates for the 2025 and post-2025 time periods, respectively (both
sets of benefit-per-ton estimates are presented in Appendix 5E).41
As discussed in greater detail in section 5.7.3 and Appendix 5A, Table 5A-1, we recognize
uncertainty associated with application of the benefit-per-ton approach used in modeling PIVh.s
cobenefits. The benefit-per-ton estimates used here reflect specific geographic patterns of
emissions reductions and specific air quality and benefits modeling assumptions associated with
the derivation of those estimates. Consequently, these estimates may not reflect local variability
in factors associated with PIVh.s-realted health impacts (e.g., population density, baseline health
incidence rates) since air quality modeling that could have shed light on local conditions was not
performed for this RIA. Therefore, use of these benefit-per-ton values to estimate co-benefits
may lead to higher or lower benefit estimates than if co-benefits were calculated based on direct
air quality modeling.
5.5 Characterizing Uncertainty
In any complex analysis using estimated parameters and inputs from numerous models,
there are likely to be many sources of uncertainty. This analysis is no exception. As outlined both
in this and preceding chapters, this analysis includes many data sources as inputs, including
emission inventories, air quality data from models (with their associated parameters and inputs),
population data, population estimates, health effect estimates from epidemiology studies,
economic data for monetizing benefits, and assumptions regarding the future state of the world
(i.e., regulations, technology, and human behavior). Each of these inputs may be uncertain and
would affect the benefits estimate. When the uncertainties from each stage of the analysis are
compounded, even small uncertainties can have large effects on the total quantified benefits.
After reviewing the EPA's approach, the National Research Council (NRC) (2002, 2008),
which is part of the National Academies of Science, concluded that the EPA's general
methodology for calculating the benefits of reducing air pollution is reasonable and informative
in spite of inherent uncertainties. The NRC also highlighted the need to conduct rigorous
41 We do not have benefit-per-ton estimates for 2038. The last year available is 2030, which is an underestimate of
the 2038 benefits because the population grows and ages over time.
5-19
-------
quantitative analyses of uncertainty and to present benefits estimates to decision makers in ways
that foster an appropriate appreciation of their inherent uncertainty. Since the publication of these
reports, the EPA has continued work to improve the characterization of uncertainty in both
health incidence and benefits estimates. In response to these recommendations, we have
expanded our previous analyses to incorporate additional quantitative and qualitative
characterizations of uncertainty. Although we have not yet been able to make as much progress
towards a full, probabilistic uncertainty assessment as envisioned by the NAS as we had hoped,
we have added a number of additional quantitative and qualitative analyses to highlight the
impact that uncertain assumptions may have on the benefits estimates. These additional analyses
focus primarily on uncertainty related to the mortality endpoint (for both ozone and PIVh.s) since
mortality is the driver for dollar benefits. In addition, for some inputs into the benefits analysis,
such as the air quality data, it is difficult to address uncertainty probabilistically due to the
complexity of the underlying air quality models and emission inputs. Therefore, we decline to
construct alternative assumptions simply for the purpose of probabilistic uncertainty
characterization when there is no scientific literature to support those alternate assumptions.
To characterize uncertainty and variability, we follow an approach that combines
elements from two recent analyses by the EPA (U.S. EPA, 201 Ob; 201 Ib), and uses a tiered
approach developed by the World Health Organization (WHO) for characterizing uncertainty
(WHO, 2008). We present this tiered assessment as well as an assessment of the potential impact
and magnitude of each aspect of uncertainty in Appendix 5 A (results of these assessments are
summarized in section 5.7.3). Data limitations prevent us from treating each source of
uncertainty quantitatively and from reaching a full-probabilistic simulation of our results, but we
were able to consider the influence of uncertainty in the risk coefficients and economic valuation
functions by incorporating several quantitative analyses described in more detail below:
1. P± Monte Carlo assessment that accounts for random sampling error and between study
variability in the epidemiological and economic valuation studies for ozone-related health
effects. See section 5.5.1 for additional detail on the Monte Carlo assessment.
2. A series of sensitivity analyses primarily focused on the mortality endpoint (for both
ozone and PIVh.s). We focus on mortality in conducting sensitivity analyses reflecting the
important role that this endpoint plays in driving both ozone-related and PIVb.s (co-
benefit) related dollar benefits. These sensitivity analyses address factors related to (a)
estimating incidence (e.g., multiple epidemiology studies providing alternative effect
5-20
-------
estimates, shape of the C-R function including potential for thresholds) and (b) estimating
associated dollar benefits (e.g., income elasticity related to willingness to pay functions
and uncertainty in specifying lag structures for long-term exposure-related mortality). See
section 5.5.2 for additional detail on the set of sensitivity analyses completed for this
RIA.
3. Supplemental analyses which allow us to consider additional factors related to the
benefits analysis. These include an assessment of the age-related differentiation of short-
term ozone exposure-related mortality (including life year saved and how estimates of
avoided mortality are distributed across age ranges). In addition, we also looked at the
relationship between estimates of mortality and the underlying baseline ambient air levels
used in their derivation. These analyses allow us to consider which range of baseline
levels drive the benefits estimates. Finally, we considered the fraction of benefit estimates
(for the 2025 scenario) which are associated with application of (higher-confidence)
known emissions controls. See section 5.5.3 for additional detail on the supplemental
analyses completed for this RIA.
5.5.1 Monte Carlo Assessment
Similar to other recent RIAs, we used Monte Carlo methods for characterizing random
sampling error associated with the concentration response functions from epidemiological
studies and random effects modeling to characterize both sampling error and variability across
the economic valuation functions. The Monte Carlo simulation in the BenMAP-CE software
randomly samples from a distribution of incidence and valuation estimates to characterize the
effects of uncertainty on output variables. Specifically, we used Monte Carlo methods to
generate confidence intervals around the estimated health impact and monetized benefits. The
reported standard errors in the epidemiological studies determined the distributions for individual
effect estimates for endpoints estimated using a single study. For endpoints estimated using a
pooled estimate of multiple studies, the confidence intervals reflect both the standard errors and
the variance across studies. The confidence intervals around the monetized benefits incorporate
the epidemiology standard errors as well as the distribution of the valuation function. These
confidence intervals do not reflect other sources of uncertainty inherent within the estimates,
such as baseline incidence rates, populations exposed and transferability of the effect estimate to
diverse locations. As a result, the reported confidence intervals and range of estimates give an
incomplete picture about the overall uncertainty in the benefits estimates.
5-21
-------
In this RIA, we provide confidence intervals for ozone-related benefits, but we are unable
to provide confidence intervals for PIVh.s-related co-benefits due to the use of benefit-per-ton
estimates.
5.5.2 Sensitivity Analysis Addressing Both Incidence and Dollar Benefit Valuation
We assign the greatest economic value to the reduction in mortality risk. Therefore, it is
particularly important to characterize to a reasonable extent the uncertainties associated with
reductions in premature mortality, including both incidence estimation and the translation of
reduced mortality into equivalent dollar benefits. Each of the sensitivity analyses completed for
this RIA are briefly described below. The reader is referred to section 5.7.3.1 for discussion of
the results and observations stemming from these sensitivity analyses.
• Alternative C-R functions for short-term ozone exposure-related mortality:
Alternative concentration-response functions are useful for assessing uncertainty beyond
random statistical error, including uncertainty in the functional form of the model or
alternative study designs. For ozone we have included two multi-city studies (Smith et
al, 2009 and Zanobetti and Schwartz 2008) in our core estimate of the range for short-
term exposure-related mortality. For the sensitivity analysis addressing this endpoint, we
have included additional multi-city and meta-analysis studies utilized in RIAs completed
for previous ozone NAAQS reviews (Bell et al., 2004 and 2005, Huang, 2005, Ito et al.,
2005 and Levy et al., 2005), as well as alternative model specifications from the Smith et
al (2009) study. The selection of studies for the core and sensitivity analyses, reflects
consideration for recommendations made both by the NAS in relation to modeling ozone
benefits (p. 80, NRC, 2008) and the CASAC in their review of the HREA completed in
support of this ozone NAAQS review. We also considered information and
recommendations provided in the latest ozone ISA. Additional detail on the selection and
use of studies in modeling short-term exposure-related mortality for the core analysis is
presented in section 5.6.3.1.
• Impact of potential thresholds on the modeling of long-term ozone exposure-related
respiratory mortality: Consistent with the HREA, we estimate counts of respiratory
deaths from long-term exposure to ozone in our core analysis. As discussed in detail in
section 5.6.3.1, the Jerrett et al., 2009 study from which the mortality effect estimate was
derived included an exploration of potential thresholds in the concentration-response
function. To provide a more comprehensive picture of potential benefits associated with
long term ozone exposures, we use the results of the threshold analysis conducted by
Jerrett et al, 2009 to conduct a sensitivity analysis evaluating models with a range of
potential thresholds in addition to a non-threshold model (see section 5.7.3.1 and
Appendix 5B, section 5B.1).
5-22
-------
• Considering alternative C-R functions in estimating long-term PMi.s exposure-
related mortality: In modeling co-benefits related to reductions in long-term exposure to
PM2.5, our dollar-per-ton approach relies on estimates based on two studies (Krewski et
al., 2009 and Lepeule et al, 2012 - see Appendix 5D, section 5D.1). To better understand
the concentration-response relationship between PIVh.s exposure and premature mortality,
the EPA conducted an expert elicitation in 2006 (Roman et al., 2008; lEc, 2006).42 In
general, the results of the expert elicitation support the conclusion that the benefits of
PM2.5 control are very likely to be substantial. Using alternate relationships between
PM2.5 and premature mortality supplied by experts, higher and lower benefits estimates
are plausible, but most of the expert-based estimates of the mean PIVh.s effect on
mortality fall between the two epidemiology-based estimates (Roman et al., 2008).
Application of the expert elicitation-based effect estimates (as part of characterizing
uncertainty) is covered in section 5.7.3.1 and Appendix 5B, section 5B.2. In addition to
these studies, we have included a discussion of other recent multi-state cohort studies
conducted in North America, but we have not estimated benefits using the effect
coefficients from these studies (see Appendix 5D, section 5D.1).
• Specifying the cessation lag for long-term PMi.s exposure-related respiratory
mortality: As discussed in section 5.1 and 5.6.4.1, uncertainty in projecting the cessation
lag for long-term PIVh.s exposure-related respiratory mortality prevents us from
estimating dollar benefits associated with projected reductions in mortality. In the
absence of clear evidence pointing to a particular lag structure, we have decided to use
two lag structures (the 20 year segmented lag used for PIVh.s and an assumption of zero
lag - see section 5.6.4.1). The range of dollar benefits that result have been included as
sensitivity analyses and not in the core analysis.
• Income elasticity in the specification of willingness to pay (WTP) functions used for
mortality and morbidity endpoints: There is uncertainty in specifying the degree to
which the WTP function used in valuing mortality and some morbidity endpoints tracks
projected increase in income over time (i.e., the income elasticity for WTP). We
completed a sensitivity analysis to evaluate the potential impact of this factor on dollar
estimates generated for mortality (and a subset of morbidity endpoints - see section
5.6.4.1).
Even these multiple estimates (including confidence intervals in the case of estimates
generated for ozone) cannot account for the role of other input variables in contributing to
overall uncertainty, including emissions and air quality modeling, baseline incidence rates, and
population exposure estimates. Furthermore, the approach presented here does not yet include
methods for addressing correlation between input parameters and the identification of reasonable
upper and lower bounds for input distributions characterizing uncertainty in additional model
42 Expert elicitation is a formal, highly structured and well documented process whereby expert judgments, usually
of multiple experts, are obtained (Ayyub, 2002).
5-23
-------
elements. As a result, the reported confidence intervals and range of estimates give an
incomplete picture about the overall uncertainty in the estimates. Thus, confidence intervals
reported for individual endpoints and for total benefits should be interpreted within the context of
the larger uncertainty surrounding the entire analysis.
5.5.3 Supplemental Analyses
We have also conducted a number of supplemental analyses designed to provide
additional perspectives on the core mortality estimates generated for this RIA. These analyses
(the results of which are described in section 5.7.3.2) include:
• Age group-differentiated aspects of short-term ozone exposure-related mortality:
We examined several risk metrics intended to characterize how mortality risk reductions
are distributed across different age ranges. These include (a) estimated reduction in life
years lost, (b) distribution of mortality incidence reductions across age ranges and (c)
estimated reductions in baseline mortality incidence rates by age group.
• Analysis of baseline ozone levels used in modeling short-term ozone exposure-
related mortality: We assess the relationship between short-term exposure-related
mortality for ozone and the distribution of baseline (i.e., reflecting attainment of the
current standard) 8hr max daily values used in deriving those estimates (see section
5.7.3.2 and Appendix 5C, section 5C.2). This analysis allows us to explore how estimates
of ozone- attributable mortality are distributed with regard to projected ambient ozone
levels, including the fraction of overall mortality that falls within specific ozone ranges.
We note that, while the latest ozone ISA did not provide support for a threshold in
relation to short-term exposure-related mortality, it did note that there is reduced
confidence in specifying the nature of the concentration-response function at lower ozone
levels (in the range of 20ppb and below) (ozone ISA, U.S. EPA, 2013a, section 2.5.4.4).
We use the distribution of short-term mortality across ozone levels to determine the
fraction of mortality reductions (i.e., benefits) that fall within this lower confidence
range.43
• Analysis of baseline PMi.s levels used in modeling short-term ozone exposure-
related mortality: We also include a similar plot of the baseline annual PIVh.s levels used
in modeling long-term PIVh.s exposure-related mortality (see section 5.7.3.2 and
Appendix 5C, section 5C.2). However, we are using a reduced form dollar-per-ton
43 However, care must be taken in interpreting this range of reduced confidence since benefits estimates are based on
the average daily 8hr max across the ozone season and not on a true daily time series of 8hr metrics within each grid
cell. The use of a seasonal mean 8hr max (rather than the more temporally differentiated daily time series) has been
shown to generate nearly identical benefit estimates at the national-level due to underlying linearity in the benefits
model being used. However, the use of the seasonal average 8hr metric, rather than a full daily time series, does
decrease overall temporal variability in a plot of mortality versus ozone level which introduces uncertainty into the
interpretation of these plots.
5-24
-------
approach in modeling PIVh.s cobenefits, we do not have spatially differentiated PIVb.s
values (and associated mortality estimates) with which to derive this type of distributional
plot specifically for this RIA and consequently, we have reproduced a plot from the
earlier analysis used to generate the benefit-per-ton values.
• Fraction of core ozone and PMi.s (cobenefit) estimates associated with application of
known emissions controls for the 2025 scenario: This analysis estimates the fraction of
core incidence and associated dollar benefits that are associated with application of the
set of known (higher confidence) emissions control measures. Note that this analysis is
only completed for the 2025 scenario since application of controls in California
(associated with the post-2025 scenario) exclusively involves application of unknown
controls.
5.5.4 Qualitative Assessment of Uncertainty and Other Analysis Limitations
Although we strive to incorporate as many quantitative assessments of uncertainty as
possible, there are several aspects we are only able to address qualitatively. These aspects are
important factors to consider when evaluating the relative benefits of the emission reduction
strategies for the proposed and alternative standards.
The total monetized benefits presented in this chapter are based on our interpretation of
the best available scientific literature and methods and supported by the EPA's independent SAB
(Health Effects Subcommittee of the Advisory Council on Clean Air Compliance Analysis)
(SAB-HES) (U.S. EPA- SAB, 2010a) and the National Academies of Science (NAS) (NRC,
2002, 2008). The benefits estimates are subject to a number of assumptions and uncertainties.
To more fully address all these uncertainties including those we cannot quantify, we
apply a four-tiered approach using the WHO uncertainty framework (WHO, 2008), which
provides a means for systematically linking the characterization of uncertainty to the
sophistication of the underlying risk assessment. The EPA has applied similar approaches in
previous analyses (U.S. EPA, 2010b, 201 Ib, U.S. EPA, 2012a - the HREA). Using this
framework, we summarize the key uncertainties in the health benefits analysis, including our
assessment of the direction of potential bias, magnitude of impact on the monetized benefits,
degree of confidence in our analytical approach, and our ability to assess the source of
uncertainty. More information on this approach and the uncertainty characterization are available
in section 5.7.3 and Appendix 5A.
5-25
-------
As previously described, we strive to monetize as many of the benefits anticipated from
the proposed and alternative standards as possible given data and resource limitations, but the
monetized benefits estimated in this RIA inevitably only reflect a portion of the total health
benefits. Data and methodological limitations prevented the EPA from quantifying or monetizing
the benefits from several important health benefit categories from emission reduction strategies
to attain the alternative ozone standards analyzed in this RIA, including potential co-benefits
from reducing NCh exposure (see section 5.6.3.6 for more information) and reductions in VOC
exposures.
5.6 Benefits Analysis Data Inputs
In Figure 5-2 above, we summarized the key data inputs to the health impact and
economic valuation estimate. Below we summarize the data sources for each of these inputs,
including demographic projections, incidence and prevalence rates, effect coefficients, and
economic valuation. We indicate where we have updated key data inputs since the benefits
analysis conducted for the 2008 ozone NAAQS RIA (U.S. EPA, 2008a) and the 2010 ozone
NAAQS Reconsideration RIA (U.S. EPA, 2010d).
5.6.1 Demographic Data
Quantified and monetized human health impacts depend on the demographic
characteristics of the population, including age, location, and income. We use population
projections based on economic forecasting models developed by Woods and Poole, Inc. (Woods
and Poole, 2012). The Woods and Poole (WP) database contains county-level projections of
population by age, sex, and race out to 2040, relative to a baseline using the 2010 Census data;
the 2008 proposal RIA incorporated WP projections relative to a baseline using 2000 Census
data. Projections in each county are determined simultaneously with every other county in the
United States to take into account patterns of economic growth and migration. The sum of
growth in county-level populations is constrained to equal a previously determined national
population growth, based on Bureau of Census estimates (Hollman et al, 2000). According to
WP, linking county-level growth projections together and constraining to a national-level total
growth avoids potential errors introduced by forecasting each county independently. County
projections are developed in a four-stage process:
5-26
-------
• First, national-level variables such as income, employment, and populations are
forecasted.
• Second, employment projections are made for 179 economic areas defined by the Bureau
of Economic Analysis (U.S. BEA, 2004), using an "export-base" approach, which relies
on linking industrial-sector production of non-locally consumed production items, such
as outputs from mining, agriculture, and manufacturing with the national economy. The
export-based approach requires estimation of demand equations or calculation of
historical growth rates for output and employment by sector.
• Third, population is projected for each economic area based on net migration rates
derived from employment opportunities and following a cohort-component method based
on fertility and mortality in each area.
• Fourth, employment and population projections are repeated for counties, using the
economic region totals as bounds. The age, sex, and race distributions for each region or
county are determined by aging the population by single year of age by sex and race for
each year through 2040 based on historical rates of mortality, fertility, and migration.
5.6.2 Baseline Incidence and Prevalence Estimates
Epidemiological studies of the association between pollution levels and adverse health
effects generally provide a direct estimate of the relationship of air quality changes to the relative
risk of a health effect, rather than estimating the absolute number of avoided cases. For example,
a typical result might be that a 5 ppb decrease in 8hr max daily ozone levels might be associated
with a decrease in hospital admissions of three percent. The baseline incidence of the health
effect is necessary to convert this relative change into a number of cases. A baseline incidence
rate is the estimate of the number of cases of the health effect per year in the assessment location,
as it corresponds to baseline pollutant levels in that location. To derive the total baseline
incidence per year, this rate must be multiplied by the corresponding population number. For
example, if the baseline incidence rate is the number of cases per year per million people, that
number must be multiplied by the millions of people in the total population.
Table 5-4 summarizes the sources of baseline incidence rates and provides average
incidence rates for the endpoints included in the analysis. For both baseline incidence and
prevalence data, we used age-specific rates where available. We applied concentration-response
functions to individual age groups and then summed over the relevant age range to provide an
estimate of total population benefits. In most cases, we used a single national incidence rate, due
to a lack of more spatially disaggregated data. Whenever possible, the national rates used are
5-27
-------
national averages, because these data are most applicable to a national assessment of benefits.
For some studies, however, the only available incidence information comes from the studies
themselves; in these cases, incidence in the study population is assumed to represent typical
incidence at the national level. County, state and regional incidence rates are available for
hospital admissions, and county-level data are available for premature mortality.
We projected mortality rates such that future mortality rates are consistent with our
projections of population growth (Abt Associates, 2012). To perform this calculation, we began
first with an average of 2004-2006 cause-specific mortality rates. Using Census Bureau
projected national-level annual mortality rates stratified by age range, we projected these
mortality rates to 2050 in 5-year increments (Abt Associates, 2012; U.S. Bureau of the Census
2002).
The baseline incidence rates for hospital admissions and emergency department visits
reflect the updated rates first applied in the CSAPR RIA (U.S. EPA, 201 Id). In addition, we
have updated the baseline incidence rates for acute myocardial infarction. These updated rates
(AHRQ, 2007) provide a better representation of the rates at which populations of different ages,
and in different locations, visit the hospital and emergency department for air pollution-related
illnesses. Also, the new baseline incidence rates are more spatially refined. For many locations
within the U.S., these data are resolved at the county- or state-level, providing a better
characterization of the geographic distribution of hospital and emergency department visits than
the previous national rates. Lastly, these rates reflect unscheduled hospital admissions only,
which represents a conservative assumption that most air pollution-related visits are likely to be
unscheduled. If air pollution-related hospital admissions are scheduled, this assumption would
underestimate these benefits.
For the set of endpoints affecting the asthmatic population, in addition to baseline
incidence rates, prevalence rates of asthma in the population are needed to define the applicable
population. Table 5-5 lists the prevalence rates used to determine the applicable population for
asthma symptoms. Note that these reflect current asthma prevalence and assume no change in
prevalence rates in future years. We updated these rates in the CSAPR RIA (U.S. EPA, 201 Id).
5-28
-------
Table 5-4. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population
Rates
Endpoint
Parameter
Value
Source
Mortality
Hospitalizations
ER Visits
Nonfatal Myocardial
Infarction (heart
attacks)
Asthma Exacerbations
Acute Bronchitis
Lower Respiratory
Symptoms
Upper Respiratory
Symptoms
Work Loss Days
School Loss Days
Minor Restricted-
Activity Days
Daily or annual mortality
rate projected to 2025a
Daily hospitalization rate
Daily ER visit rate for asthma
and cardiovascular events
Daily nonfatal myocardial
infarction incidence rate per
person, 18+
Incidence among asthmatic
African-American children
daily wheeze
daily cough
daily shortness of breath
Annual bronchitis incidence
rate, children
Daily lower respiratory
symptom incidence among
children c
Daily upper respiratory
symptom incidence among
asthmatic children
Daily WLD incidence rate per
person (18-65)
Aged 18-24
Aged 25-44
Aged 45-64
Rate per person per year,
assuming 180 school days
per year
Daily MRAD incidence rate
per person
Age-, cause-, and
county-specific rate
Age-, region-, state-,
county- and cause-
specific rate
Age-, region-, state-,
county- and cause-
specific rate
Age-, region-, state-,
and county-specific
rate
0.173
0.145
0.074
0.043
0.0012
0.3419
0.00540
0.00678
0.00492
9.9
0.02137
CDC WONDER (2004-2006)
U.S. Census bureau, 2000
2007 HCUP data files"
2007 HCUP data files'
2007 HCUP data files" adjusted by
0.93 for probability of surviving
after 28 days (Rosamond et al.,
1999)
Ostro et al. (2001)
American Lung Association (2002,
Table 11)
Schwartz et al. (1994, Table 2)
Pope etal. (1991, Table 2)
1996 HIS (Adams, Hendershot, and
Marano, 1999, Table 41); U.S.
Census Bureau (2000)
National Center for Education
Statistics (1996) and 1996 HIS
(Adams et al., 1999, Table 47);
Ostro and Rothschild (1989,
p. 243)
a Mortality rates are only available at 5-year increments.
b Healthcare Cost and Utilization Program (HCUP) database contains individual level, state and regional-level
hospital and emergency department discharges for a variety of ICD codes (AHRQ, 2007).
0 Lower respiratory symptoms are defined as two or more of the following: cough, chest pain, phlegm, and wheeze.
Table 5-5. Asthma Prevalence Rates
Population Group
All Ages
<18
Value
0.0780
0.0941
Asthma Prevalence Rates
Source
American Lung Association (2010, Table 7)
5-29
-------
5-17 0.1070
18-44 0.0719
45-64 0.0745
65+ 0.0716
African American, 5-17 0.1776 American Lung Association (2010, Table 9)
African American, <18 0.1553 American Lung Association a
a Calculated by ALA for U.S. EPA, based on NHIS data (CDC, 2008).
5.6.3 Effect Coefficients
The modeling of incidence and benefits for each pollutant employ distinct and separate
sets of effect estimates since these are obtained from epidemiological studies specific to a given
endpoint/pollutant combination. In the case of PIVh.s, the dollar-per-ton approach being
employed reflects application of effect estimates (for all endpoints) which have been used by the
EPA in previous RIA's (e.g., PM NAAQS, U.S. EPA, 2012b). Consequently, while we identify
the studies and effect estimates reflected in the dollar-per-ton values used for PIVh.s (see Table 5-
8), we do not present a detailed discussion of those effect estimates in this section and instead
present that more detailed discussion in Appendix 5D. By contrast, with ozone, we conducted
detailed benefits modeling using an updated set of effect estimates that reflects a combination of
values used in previous RIAs together with updated values. For that reason, we provide a
detailed discussion of effect estimates for ozone (including the rationale for their selection) in
this section.
The first step in selecting effect coefficients is to identify the health endpoints to be
quantified. We base our selection of health endpoints on consistency with the EPA's IS As
(which replace previous "Criteria Documents"), with input and advice from the HES, a scientific
review panel specifically established to provide advice on the use of the scientific literature in
developing benefits analyses for the EPA 's Report to Congress on The Benefits and Costs of the
Clean Air Act 1990 to 2020 (U.S. EPA, 2011 a). In addition, we have included more recent
epidemiology studies from the ozone ISA (U.S. EPA, 2013a), PM ISA (U.S. EPA, 2009b), and
the PM Provisional Assessment (U.S. EPA, 2012c).44 In selecting health endpoints for ozone, we
also considered the suite of endpoints included in core modeling for the HREA, which was
supported by CAS AC (Frey, and Samet 2012; Frey, 2014). In general, we follow a weight of
evidence approach, based on the biological plausibility of effects, availability of concentration-
' The peer-reviewed studies in the Provisional Assessment have not yet undergone external review by the SAB.
5-30
-------
response functions from well conducted peer-reviewed epidemiological studies, cohesiveness of
results across studies, and a focus on endpoints reflecting public health impacts (like hospital
admissions) rather than physiological responses (such as changes in clinical measures like
Forced Expiratory Volume [FEV1]).
There are several types of data that can support the determination of types and magnitude
of health effects associated with air pollution exposures. These sources of data include
toxicological studies (including animal and cellular studies), human clinical trials, and
observational epidemiology studies. All of these data sources provide important contributions to
the weight of evidence surrounding a particular health impact. However, only epidemiology
studies provide direct concentration-response relationships that can be used to evaluate
population-level impacts of reductions in ambient pollution levels in a health impact assessment.
For the data-derived estimates, we relied on the published scientific literature to ascertain
the relationship between ozone and PIVb.s and adverse human health effects. We evaluated
epidemiological studies using the selection criteria summarized in Table 5-6. These criteria
include consideration of whether the study was peer-reviewed, the match between the pollutant
studied and the pollutant of interest, the study design and location, and characteristics of the
study population, among other considerations. In general, the use of concentration-response
functions from more than a single study can provide a more representative distribution of the
effect estimate. However, there are often differences between studies examining the same
endpoint, making it difficult to pool the results in a consistent manner. For example, studies may
examine different pollutants or different age groups. For this reason, we consider very carefully
the set of studies available examining each endpoint and select a consistent subset that provides a
good balance of population coverage and match with the pollutant of interest. In many cases,
either because of a lack of multiple studies, consistency problems, or clear superiority in the
quality or comprehensiveness of one study over others, a single published study is selected as the
basis of the effect estimate.
5-31
-------
When several effect estimates for a pollutant and a given health endpoint (with the
exception of mortality)45 have been selected, they are quantitatively combined or pooled to
derive a more robust estimate of the relationship. The BenMAP Manual Technical Appendices
for an earlier version of the program provides details of the procedures used to combine multiple
impact functions (Abt Associates, 2012). In general, we used fixed or random effects models to
pool estimates from different single city studies of the same endpoint. Fixed effect pooling
simply weights each study's estimate by the inverse variance, giving more weight to studies with
greater statistical power (lower variance). Random effects pooling accounts for both within-study
variance and between-study variability, due, for example, to differences in population
susceptibility. We used the fixed effect model as our null hypothesis and then determined
whether the data suggest that we should reject this null hypothesis, in which case we would use
the random effects model.46 Pooled impact functions are used to estimate hospital admissions
and asthma exacerbations. When combining evidence across multi-city studies (e.g.,
cardiovascular hospital admission studies), we use equal weights pooling. The effect estimates
drawn from each multi-city study are themselves pooled across a large number of urban areas.
For this reason, we elected to give each study an equal weight rather than weighting by the
inverse of the variance reported in each study. For more details on methods used to pool
incidence estimates, see the BenMAP Manual Appendices (Abt Associates, 2012).
Effect estimates selected for a given health endpoint were applied consistently across all
locations nationwide. This applies to both impact functions defined by a single effect estimate
and those defined by a pooling of multiple effect estimates. Although the effect estimate may, in
fact, vary from one location to another (e.g., because of differences in population susceptibilities
or differences in the composition of PM), location-specific effect estimates are generally not
available.
45 In the case of mortality, when we have multiple studies providing effect estimates for the core analysis, we
include the range of the resulting incidence and dollar benefit estimates rather than pooling them in order to provide
additional characterization of overall confidence associated with this key endpoint. However, for morbidity
endpoints we do pool estimates as described here.
46 EPA recently changed the algorithm BenMAP uses to calculate study variance, which is used in the pooling
process. Prior versions of the model calculated population variance, while the version used here calculates sample
variance.
5-32
-------
Table 5-6. Criteria Used When Selecting C-R Functions
Consideration
Comments
Peer-Reviewed
Research
Study Type
Study Period
Seasonality
Population Attributes
Study Size
Peer-reviewed research is preferred to research that has not undergone the peer-
review process.
Prospective vs. cohort: Among studies that consider chronic exposure (e.g., over a year
or longer), prospective cohort studies are preferred over ecological studies because
they control for important individual-level confounding variables that cannot be
controlled for in ecological studies.
Multi-city vs. pooled/meta-analysis: In recommending approaches for modeling ozone-
related mortality, the NAS notes a number of advantages multi-city time series studies
as compared to meta-analyses. Multi-city studies utilize a consistent model structure
and can include factors that explain differences between effect estimates among the
cities. By contrast, with meta-analyses given the aggregation of large sets of studies, the
construct definitions can become imprecise and the results difficult to interpret. In
addition, meta-analyses can suffer from publication bias which can result in high-biased
effect estimates. Ultimately, the NAS recommends that the greatest emphasis be placed
on estimates based on systematic new multi-city analyses without excluding
consideration of meta-analyses (NRC, 2008). Reflecting these observations by the NAS,
in modeling ozone benefits, we have included several newer multi-city studies in the
core analysis and a combination of multi-city and meta analyses in the accompanying
sensitivity analysis. Placing emphasis on these two multi-city studies was also supported
by the CASAC in the context of the HREA (Frey, and Samet 2012; Frey, 2014).
Studies examining a relatively longer period of time (and therefore having more data)
are preferred, because they have greater statistical power to detect effects. Studies that
are more recent are also preferred because of possible changes in pollution mixes,
medical care, and lifestyle over time. However, when there are only a few studies
available, studies from all years will be included.
While the measurement of PM is typically collected across the full year, ozone
monitoring seasons can vary substantially across different regions of the country. Given
modeling constraints, we were not able to consider variation in ozone seasons in
modeling benefits and instead, had to select a standard ozone season for the entire
country (May 1st through September 31st). Consequently, in selecting effect estimates,
we favored those values that reflected ozone seasons close to this fixed ozone season
(i.e., we did not include effect estimates based on a full year of ozone monitoring data).
The most technically appropriate measures of benefits would be based on impact
functions that cover the entire sensitive population but allow for heterogeneity across
age or other relevant demographic factors. In the absence of effect estimates specific to
age, sex, preexisting condition status, or other relevant factors, it may be appropriate to
select effect estimates that cover the broadest population to match with the desired
outcome of the analysis, which is total national-level health impacts. When available,
multi-city studies are preferred to single city studies because they provide a more
generalizable representation of the concentration-response function.
Studies examining a relatively large sample are preferred because they generally have
more power to detect small magnitude effects. A large sample can be obtained in
several ways, including through a large population or through repeated observations on
a smaller population (e.g., through a symptom diary recorded for a panel of asthmatic
children).
5-33
-------
Consideration
Comments
Study Location
Pollutants Included in
Model
Measure of PM
Economically Valuable
Health Effects
Non-overlapping
Endpoints
U.S. studies are more desirable than non-U.S. studies because of potential differences in
pollution characteristics, exposure patterns, medical care system, population behavior,
and lifestyle. National estimates are most appropriate when benefits are nationally
distributed; the impact of regional differences may be important when benefits only
accrue to a single area.
An important factor affecting the specification of co-pollutant models for ozone and PM
is sampling frequency. While ozone is typically measured every hour of each day during
the ozone season for a specific location, PM is typically measured every 3rd or 6th day.
For this reason, when modeling the PM effect, epidemiological models specifying co-
pollutants are preferred because this approach controls for the potential ozone effect
while not diminishing the effective sample size available for specifying the PM effect.
However, when modeling the ozone effect, the use of copollutants modeling (with PM)
can substantially reduce sample size (by 1/3 to 1/6) since only days with both ozone and
PM can be used. While these copollutants models may control for potential PM effects,
they also result in a substantially less robust characterization of the ozone effect due to
the reduced number of ozone measurements. For this reason, while we favor
copollutants models in modeling PM benefits, for ozone we favor single pollutant
models for the core estimate and reserve copollutants models for sensitivity analyses.
For this analysis, impact functions based on PM2.s are preferred to PMio because of the
focus on reducing emissions of PMzs precursors, and because air quality modeling was
conducted for this size fraction of PM. Where PMz.s functions are not available, PMio
functions are used as surrogates, recognizing that there will be potential downward
(upward) biases if the fine fraction of PMio is more (less) toxic than the coarse fraction.
Some health effects, such as forced expiratory volume and other technical
measurements of lung function, are difficult to value in monetary terms. These health
effects are not quantified in this analysis.
Although the benefits associated with each individual health endpoint may be analyzed
separately, care must be exercised in selecting health endpoints to include in the overall
benefits analysis because of the possibility of double-counting of benefits.
The specific studies from which effect estimates for the core analysis related to ozone
exposure are drawn are included in Table 5-7. Table 5-8 identifies studies reflected in the dollar-
per-ton analysis for PIVh.s. We highlight in red those studies that have been added since the
benefits analysis conducted for the ozone reconsideration (U.S. EPA, 2010d) or the ozone
NAAQS RIA (U.S. EPA, 2008a). In all cases where effect estimates are drawn directly from
epidemiological studies, standard errors are used as a partial representation of the uncertainty in
the size of the effect estimate. Table 5-9 summarizes those health endpoints and studies we have
included as sensitivity analyses for ozone.
5-34
-------
Table 5-7. Health Endpoints and Epidemiological Studies Used to Quantify Ozone-Related
Health Impacts in the Core Analysis a
Endpoint
„ . Study
Study „ , .
Population
Relative Risk or Effect Estimate (P)
(with 95th Percentile Confidence
Interval orSE, respectively)
Premature Mortality
Premature
mortality—short-term
Smith etal. (2009)
Zanobetti and Schwartz
(2008)
Premature respiratory
mortality-long-term Jerrett et al. (2009)
(incidence only)
All ages
>29 years
P = 0.00032 (0.00008)
P = 0.00051 (0.00012)
P = 0.003971 (0.00133)
Hospital Admissions
Respiratory
Pooled estimate:
Katsouyanni etal. (2009)
> 65 years
P = 0.00061 (0.00041) natural splines
P = 0.00064 (0.00040) penalized splines
Asthma-related
emergency
department visits
Pooled estimate:
Glad etal. (2012)
Ito et al. (2007)
Mar and Koenig (2010)
Peel et al. (2005)
Sarnatetal. (2013)
Wilson etal. (2005)
0-99 years
P = 0.00306 (0.00117)
P = 0.00521 (0.00091)
P = 0.01044 (0.00436) (0-17 yr olds)
P = 0.00770 (0.00284) (18-99 yr olds)
P = 0.00087 (0.00053)
P = 0.00111 (0.00028)
RR = 1.022 (0.996 - 1.049) per 25
Other Health Endpoints
Asthma exacerbations
School loss days
Acute respiratory
symptoms (MRAD)
Pooled estimate:b
Mortimer etal. (2002)
O'Connor etal. (2008)
Schildcroutetal. (2006)
Pooled estimate:
Chen et al. (2000)
Gilliland etal. (2001)
Ostro and Rothschild (1989)
6-18 years0
5-17 years
18-65
years
P = 0.00929 (0.00387)
P = 0.00097 (0.00299)
P = 0.00222 (0.00282)
P = 0.015763 (0.004985)
P = 0.007824 (0.004445)
P = 0.002596 (0.000776)
a Studies highlighted in red represent updates incorporated since the 2008 ozone NAAQS RIA (U.S. EPA, 2008a).
b As discussed in sections 5.3 and 5.6.3.1, while we believe that available evidence supports inclusion of estimates of
long-term exposure-related respiratory mortality incidence in the core analysis, uncertainty in specifying a lag
structure for this endpoint (a key factor in valuation) prevents us from including dollar benefits in the core analysis
and instead, these are included as part of the sensitivity analysis exploring this endpoint.
0 The original study populations were 5 to 12yrs for the O'Conner et al., (2008) and Schildcrout et al., (2006) and 5-
9yrs for the Mortimer et al., (2002) study. Based on advice from the SAB-HES, we extended the applied population
to 6-18yrs for all three studies, reflecting the common biological basis for the effect in children in the broader age
group. See: U.S. EPA-SAB (2004a) andNRC (2002)
5-35
-------
Table 5-8. Health Endpoints and Epidemiological Studies Used to Quantify PMi.s-Related
Health Impacts in the Core Analysis a
Endpoint
Study
Study Population
Relative Risk or Effect Estimate (P)
(with 95th Percentile Confidence
Interval orSE, respectively)
Premature Mortality
Premature mortality-
cohort study, all-cause
Premature mortality—
all-cause
Krewski et al. (2009)
Lepeule et al. (2012)
Woodruff etal. (1997)
> 29 years RR = 1.06 (1.04-1.06) per 10 u.g/m3
> 24 years RR = 1.14 (1.07-1.22) per 10 u.g/m3
Infant (< 1 OR = 1.04 (1.02-1.07) per 10 u.g/m3
year)
Chronic Illness
Nonfatal heart attacks
Peters et al. (2001)
Pooled estimate:
Pope et al. (2006)
Sullivan et al. (2005)
Zanobetti et al. (2009)
Zanobetti and Schwartz (2006)
Adults(>18
years)
OR = 1.62 (1.13-2.34) per 20 u.g/m3
P = 0.00481 (0.00199)
P = 0.00198 (0.00224)
P = 0.00225 (0.000591)
P = 0.0053 (0.00221)
Hospital Admissions
Respiratory
Cardiovascular
Asthma-related
emergency department
visits
Zanobetti et al. (2009)—ICD
460-519 (All respiratory)
Kloog et al. (2012)-ICD 460-
519 (All Respiratory
Moolgavkar (2000)-ICD 490-
496 (Chronic lung disease)
Babinetal. (2007)-ICD493
(asthma)
Sheppard(2003)-ICD493
(asthma)
Pooled estimate:
Zanobetti et al. (2009)—ICD
390-459 (all cardiovascular)
Peng etal. (2009)-ICD 426-
427; 428; 430-438; 410-414;
429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)
Peng etal. (2008)-ICD 426-
427; 428; 430-438; 410-414;
429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)
Bell etal. (2008a)-ICD 426-
427; 428; 430-438; 410-414;
429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)
Moolgavkar (2000)-ICD 390-
429 (all cardiovascular)
Pooled estimate:
Mar etal. (2010)
Slaughter etal. (2005)
Glad etal. (2012)
> 64 years p=0.00207 (0.00446)
P=0.0007 (0.000961)
18-64 years 1.02 (1.01-1.03) per 36 |jg/m3
< 19 years p=0.002 (0.004337)
> 64 years
RR = 1.04 (1.01-1.06) per 11.8 u.g/m3
P=0.00189 (0.000283)
P=0.00068
(0.000214)
P=0.00071
(0.00013)
P=0.0008
(0.000107)
20-64 years RR=1.04 (t statistic: 4.1) per 10 u.g/m3
All ages
RR = 1.04 (1.01-1.07) per 7 u.g/m3
RR = 1.03 (0.98-1.09) per 10 u.g/m3
P=0.00392 (0.002843)
Other Health Endpoints
5-36
-------
Acute bronchitis
Dockeryetal. (1996)
8-12 years OR = 1.50 (0.91-2.47) per 14.9 u.g/m3
Asthma exacerbations
Pooled estimate:
Ostro et al. (2001) (cough,
wheeze, shortness of breath)b
Mar et al. (2004) (cough,
shortness of breath)
6-18 years b OR = 1.03 (0.98-1.07)
OR =1.06 (1.01-1.11)
OR = 1.08 (1.00-1.17) per 30 u.g/m3
RR= 1.21 (1-1.47) per
RR = 1.13 (0.86-1.48) per 10 u.g/m3
Work loss days
Acute respiratory
symptoms (MRAD)
Upper respiratory
symptoms
Lower respiratory
symptoms
Ostro (1987) 18-65 years
Ostro and Rothschild (1989) „„ rr
.... . . . . .. .. . ' 18-65 years
(Minor restricted activity days)
, ,„„„„> Asthmatics,
Popeetal. 1991 „ ..
9-11 years
Schwartz and Neas (2000) 7-14 years
P=0.0046 (0.00036)
P=0.00220 (0.000658)
1.003 (1-1.006) per 10 u.g/m3
OR = 1.33 (1.11-1.58) per 15 u.g/m3
a Studies highlighted in red represent updates incorporated since the ozone NAAQS RIA (U.S. EPA, 2008a). These
updates were introduced in the PM NAAQS RIA (U.S. EPA, 2012b).
b The original study populations were 8 to 13 for the Ostro et al. (2001) study and 7 to 12 for the Mar et al. (2004)
study. Based on advice from the SAB-HES, we extended the applied population to 6-18, reflecting the common
biological basis for the effect in children in the broader age group. See: U.S. EPA-SAB (2004a,b) and NRC (2002).
Table 5-9. Health Endpoints and Epidemiological Studies Used to Quantify Ozone-Related
Health Impacts in the Sensitivity Analysis a
Endpoint
Study
Study
Population
Relative Risk or Effect Estimate (P)
(with 95th Percentile Confidence
Interval orSE, respectively)
Premature Mortality
Premature respiratory
mortality—long-term
Premature
mortality—short-term
Jerrett et al. (2009)-based models:
- non-threshold ozone only (86
cities)
- non-threshold ozone only (96
cities)
- threshold 40 ppbb > 29 years p=
- threshold 45 ppb p=
- threshold 50 ppb p=
-threshold 55 ppb p=
- threshold 56 ppb p=
- threshold 60 ppb p=
Smith et al. (2009) (copollutant
model with PMio)
Bell etal. (2005)
Levy et al. (2005)
Bell et al. (2004) All ages
Ito et al. (2005)
Schwartz et al. (2005)
Huang et al. (2005)
(cardiopulmonary)
P=0.002664 (0.000969)
P=0.00286 (0.000942)
:0.00312 (0.00096)
:0.00336 (0.001)
:0.00356 0.00106)
:0.00417 (0.00118)
:0.00432 (0.00121)
:0.00402 (0.00137)
P=0.00026 (0.00017)
:0.00080 (0.00021)
:0.00112 (0.00018)
:0.00026 (0.00009)
:0.00117 (0.00024)
:0.00043 (0.00015)
P=0.00026 (0.00009)
a Studies highlighted in red represent updates incorporated since the 2008 ozone NAAQS RIA (U.S. EPA, 2008b).
b All threshold models are ozone-only and based on the full 96 city dataset.
5-37
-------
5.6.3.1 Ozone Premature Mortality Effect Coefficients
Core Mortality Effect Coefficients for Short-term ozone Exposure. The overall body
of evidence indicates that there is likely to be a causal relationship between short-term ozone
exposure and premature death. The 2013 ozone ISA states that:
"The evaluation of new multi-city studies that examined the association between
short-term ozone exposure and mortality found evidence which supports the
conclusions of the 2006 ozone AQCD. These new studies reported consistent
positive associations between short-term ozone exposure and all-cause (non-
accidental) mortality, with associations persisting or increasing in magnitude during
the warm season, and provide additional support for associations between ozone
exposure and cardiovascular and respiratory mortality" (ozone ISA section 6.6.3,
USEPA2013).
The ISA concludes by stating that, "Although some uncertainties still remain, the collective body
of evidence is sufficient to conclude there is likely to be a causal relationship between short-term
ozone exposure and total mortality." (ozone ISA section 6.6.3, USEPA 2013a). Regarding
potential confounding of the ozone mortality effect by PM, the ISA states, "Overall, across
studies, the potential impact of PM indices on ozone-mortality risk estimates tended to be much
smaller than the variation in ozone-mortality risk estimates across cities suggesting that ozone
effects are independent of the relationship between PM and mortality." (ozone ISA section 6.3.3,
USEPA 2013 a). However, the ISA does note that the interpretation of the potential confounding
effects of PM on ozone-mortality risk estimates requires caution. This caution reflects in part, the
every-3rd- and every-6th-day PM sampling schedule (in most cities) which limits the overall
sample size available for evaluating potential confounding of the ozone effect by PM. (ozone
ISA section 6.3.3, USEPA 2013a).
In their review of the HREA, the SAB's CAS AC Ozone Review Panel expressed support
for epidemiological studies and corresponding concentration-response functions used in the
HREA, which included effect estimates obtained from Smith et al, (2009) and Zanobetti and
Schwartz (2008b) (Frey and Samet, 2012, p. 17-18 and Frey, 2014, p, 9). Furthermore, the
CAS AC specifically noted support for the use of multi-city studies, where available in modeling
5-38
-------
the health endpoints included in the analysis (Frey and Samet, 2012, p. 15). In addition, they
expressed support for modeling total risk, and de-emphasized the importance of estimating risk
associated with ozone concentrations above the lowest measure level (LML) from contributing
epidemiological studies (Frey and Samet, 2012, p. 10 and 18).
In 2006, the EPA requested an NAS study to answer four key questions regarding ozone-
related mortality: (1) how did the epidemiological literature to that point improve our
understanding of the size of the ozone-related mortality effect? (2) How best can EPA quantify
the level of ozone-related mortality impacts from short-term exposure? (3) How might EPA
estimate the change in life expectancy? (4) What methods should EPA use to estimate the
monetary value of changes in ozone-related mortality risk and life expectancy?
In 2008, the NAS (NRC, 2008) issued a series of recommendations to the EPA regarding
the quantification and valuation of ozone-related short-term mortality. Chief among these was
that"... short-term exposure to ambient ozone is likely to contribute to premature deaths" and the
committee recommended that "ozone-related mortality be included in future estimates of the
health benefits of reducing ozone exposures..." The NAS also recommended that".. .the greatest
emphasis be placed on the multi-city and NMMAPS studies without exclusion of the meta-
analyses" (NRC, 2008). In addition, NAS recommended that EPA "should give little or no
weight to the assumption that there is no causal association between estimated reductions in
premature mortality and reduced ozone exposure" (NRC, 2008). In 2010, the Health Effects
Subcommittee of the Advisory Council on Clean Air Compliance Analysis, while reviewing
EPA's The Benefits and Costs of the Clean Air Act 1990 to 2020 (U.S. EPA, 2011 a), also
confirmed the NAS recommendation to include ozone mortality benefits (U.S. EPA-SAB,
2010a).
In view of the findings of the ozone ISA, the NAS panel, the HES panel, and the CAS AC
panel, we include ozone-related premature mortality for short-term exposure in the core health
effects analysis using effect coefficients from the Smith et al. (2009) NMMAPS analysis and the
Zanobetti and Schwartz (2008) multi-city study. As discussed below, we also include several
additional studies as sensitivity analyses. This approach with an emphasis on newer multi-city
studies is consistent with recommendations provided by the NAS in their ozone mortality report
5-39
-------
(NRC, 2008), "The committee recommends that the greatest emphasis be placed on estimates
from new systematic multi-city analyses that use national databases of air pollution and
mortality, such as in the NMMAPS, without excluding consideration of meta-analyses of
previously published studies." In selecting the Smith et al. (2009) and Zanobetti and Schwartz
(2008b) studies, we point both to CASAC support for the use of these two studies in the context
of the HREA completed for this NAAQS review (Samet and Frey, 2012, p. 15 and Frey, 2014, p.
9) and the fact that both of these studies are multi-city studies published more recently (as
compared with other multi-city studies or meta-analyses included in the sensitivity analyses - see
discussion below).
The Smith et al., (2009) study is a reanalysis of the NMMAPS data set focused on
evaluating the relationship between short-term ozone exposure and mortality. While this
reproduces the core national-scale estimates presented in Bell et al., (2004), it also explores the
sensitivity of the mortality effect to different model specifications including (a) regional versus
national Bayes-based adjustment,47 (b) co-pollutants models considering PMio, (c) all-year
versus ozone-season based estimates, and (d) consideration of a range of ozone metrics,
including the daily 8hr max, which is of particular interest in the context of the RIA given that is
the metric that is used in the form of the ozone standard. In addition, the Smith et al. (2009)
study does not use the trimmed mean approach employed in the Bell et al. (2004) study in
preparing ozone monitor data, which is another advantage.48 In selecting effect estimates from
Smith et al. (2009) for use in the core analysis, we focused on an ozone-only estimate for non-
accidental mortality based on the 8hr max metric for the warmer ozone season. In addition, for
the sensitivity analysis, we included a copollutants model (ozone and PMio) from Smith et al.
(2009) for all-cause mortality which also used the 8hr max ozone metric for the ozone season. As
47 In Bayesian modeling, effect estimates are "updated" from an assumed prior value using observational data. In
the Smith et al (2009) approach, the prior values are either a regional or national mean of the individual effect
estimates obtained for each individual city. The Bayesian adjusted city-specific effect estimates are then
calculated by updating the selected prior value based on the relative precision of each city-specific estimate and
the variation observed across all city-specific individual effect estimates. City-specific estimates are pulled
towards the prior value if they have low precision and/or there is low overall variation across estimates. City-
specific estimates are given less adjustment if they are precisely estimated and/or there is greater overall
variation across estimates.
48 There are a number of concerns regarding the trimmed mean approach including (1) the potential loss of temporal
variation in the data when the approach is used (this could impact the size of the effect estimate) and (2) a lack of
complete documentation for the approach which prevents a full reviewing or replication of the technique.
5-40
-------
noted in Table 5-6, the decision to use a single pollutant model for the core analysis and reserve
the copollutants model for the sensitivity analysis reflects our concern that the reduced sampling
frequency for days with copollutants measurements (1/3 and 1/6) can impact characterization of
the ozone effect which is the focus of this assessment. In addition, as noted earlier in this section,
the latest ozone ISA states that ozone effects are likely to be independent of the relationship
between PM and mortality, which further supports the approach of favoring single pollutant
models in the core analysis.
The Zanobetti and Smith (2008) study evaluated the relationship between ozone exposure
(using an 8hr mean metric for the warm season June-August) and all-cause mortality in 48 U.S.
cities using data collected between 1989 and 2000. The study presented single pollutant
concentration-response functions based on shorter (0-3 day) and longer (0-20) day lag structures,
with the comparison of effects based on these different lag structures being a central focus of the
study. For the core analysis, we used the 0-3 day lag based concentration-response function since
this had the strongest effect and tighter confidence interval. Note that, because the RIA utilizes
the 8hr max ozone metric, we had to covert the effect estimate from Zanobetti and Smith (2008)
which is based on an 8hr mean metric to an equivalent effect estimate based on an 8hr max. To
do this, we used the ozone metric approach wherein the original effect estimate (and standard
error) is multiplied by the appropriate ozone metric adjustment ratio.49
Core Mortality Effect Coefficient for Long-term ozone Exposure. We also estimated
long-term ozone exposure-related respiratory mortality incidence in the core analysis. The
available evidence did not allow us to characterize how long-term exposure to ozone related to
the year of onset of mortality (i.e., a cessation lag), which is a necessary input to quantifying the
discounted dollar benefits. For this reason, we report the dollar benefits of avoided long-term
exposure-related respiratory deaths as a sensitivity analyses (see section 5.7.3.1). Support for
49 These adjustment ratios are created by (a) obtaining summary air quality (composite monitor values) for each
urban study area/ozone season combination reflected in the original epidemiology study, (b) calculating the ratio of
the 8hr max to the study-specific air metric (for each of the urban study areas), and (c) taking the average of these
urban-study area ratios. Ratio adjustment of the effect estimate does introduce uncertainty into the benefits estimates
generated using these adjusted effect estimates, however, adjustments of relatively similar metrics (e.g., 8hr max and
8hr mean), as is the case with the Zanobetti and Schwartz (2008) study, are likely to introduce less uncertainty than
adjustments for more disparate ratios (e.g., 24hr or Ihr max ratios to 8hr max equivalents).
5-41
-------
modeling long-term exposure-related mortality incidence comes from the final ozone ISA as well
as recommendations provided by CAS AC in their review of the HREA completed for the current
ozoneNAAQS review (Frey, 2014, p. 3 and 9).
The final ozone ISA references long-term respiratory mortality in section 7.2.1 (USEPA
2013 a) where they state, "The positive results from various designs and locations support a
relationship between long-term exposure to ambient ozone concentrations and respiratory health
effects and mortality." Later in that chapter, the ISA states that: "The strongest evidence for an
association between long-term exposure to ambient ozone concentrations and mortality is
derived from associations reported in the Jerrett et al. (2009) study for respiratory mortality that
remained robust after adjusting for PIVh.s concentrations." (Section 7.7.1, USEPA, 2013a). In that
same section, the authors also state that:
"Coherence and biological plausibility for this observation [the association between
long-term exposure and respiratory mortality] is provided by evidence from
epidemiologic, controlled human exposure, and animal toxicological studies for the
effects of short- and long-term exposure to ozone on respiratory effects (see Sections
6.2 and 7.2). Respiratory mortality is a relatively small portion of total mortality
[about 7.6% of all deaths in 2010 were due to respiratory causes (Murphy et al.,
2012)], thus it is not surprising that the respiratory mortality signal may be difficult
to detect in studies of cardiopulmonary or total mortality."
While the ozone ISA concludes that evidence is suggestive of a causal association between total
mortality and long-term ozone exposure (section 7.7.1), specifically with regard to respiratory
health effects (including mortality), the ISA concludes that there is likely to be a causal
association (section 7.2.8).
In their review of the HREA completed for the ozone NAAQS review and specifically
modeling of the long-term exposure-related respiratory mortality endpoint, the CAS AC states
that, "The basis for estimating long-term mortality (respiratory) risks relies on a single study,
Jerrett et al. (2009), and the HREA should acknowledge the uncertainty and confidence in
modeling results from the use of a single study, albeit a good one." And, while they comment on
the size of the chronic obstructive pulmonary disease (COPD) mortality effect attributable to
5-42
-------
ozone (page 7-68) they go on to state that, "the CASAC concurs that Jerrett et al. (2009) is an
appropriate study to use at this time as the basis for the long-term mortality risk estimates given
its adequacy and the lack of alternative data." (Frey, C., 2014). This advice supersedes previous
advice provided by the HES to include long-term ozone exposure-related mortality only as a
sensitivity analysis.
The Jerrett et al. (2009) study was the first to explore the relationship between long-term
ozone exposure and respiratory mortality (rather than focusing on cardiopulmonary mortality).
Jerrett et al. (2009) exhibits a number of strengths including (a) the study is based on the 1.2
million participant American Cancer Society cohort drawn from all 50 states, DC, and Puerto
Rico (included ozone data from 1977 [5 years before enrollment in the cohort began] to 2000);
(b) it includes copollutants models that controlled for PIVh.s; and (c) it explored the potential for
a threshold concentration associated with the long-term mortality endpoint. However there are
attributes to this study that affect how we interpret the long-term exposure-related respiratory
mortality estimates. First, while CASAC notes that Jerrett et al. (2009) is well designed, it is a
single study—and so provides the only quantitative basis for estimating this endpoint. By
comparison, we estimate short-term exposure-related mortality risk using several studies.
There is also the potential existence and location of a threshold in the C-R function
relating mortality and long-term ozone concentrations. That uncertainty could greatly influence
our quantitative risk estimates (we address the potential for a threshold in a sensitivity analysis
below). The CASAC did address the use of the zero threshold versus threshold based models in
the context of the FIREA, stating that, "The EPA examined the threshold analysis contained in
Jerrett et al. (2009) and found that the mortality model including a threshold at 56 ppb had the
lowest log likelihood value of all models examined. However, it is not clear whether the 56 ppb
threshold model is a better predictor of respiratory mortality than when using a linear model for
the Jerrett et al. data. Different, but valid statistical tests produced different conclusions about the
threshold versus linear models. The less stringent test judged the 56 ppb threshold model to be
superior to the linear model, but the confidence interval indicates the threshold could occur
anywhere from 0 to 60 ppb. Using the more stringent statistical test, none of the threshold
models produce better predictions than the linear model. Given these results, the CASAC
concurs with the EPA's planned approach [as stated in the 2nd draft FIREA] to conduct a
5-43
-------
sensitivity analysis evaluating potential thresholds in the C-R functions that relate long-term
ozone exposures with respiratory mortality and to not make the threshold models the core
analytical procedure in the PA." (Frey, 2014, p. 13-14) Here, the CASAC clearly states their
support for inclusion of the zero threshold (linear model) as the core approach, while treating the
threshold models as sensitivities.
Reflecting this advice provided by CASAC, we have generated the core benefit estimate
for long-term exposure-related respiratory mortality using a non-threshold co-pollutant model
(with PIVh.s) obtained from Jerrett et al. (2009) (see Tables 5-7 and 5-9). Using a co-pollutant
model is consistent with the fact that this study applied seasonal average metrics that are
insensitive to co-pollutant monitoring for PIVh.s; we explore the influence of co-pollutant models
and thresholds in a sensitivity analysis below.50 The effect estimates used to model long-term
ozone-attributable mortality are calculated using a seasonal average of peak (1-hr maximum)
measurements. These long-term exposure metrics can be viewed as long-term exposures to daily
peak ozone over the warmer months, as compared with annual average levels such as are used in
long-term PM exposure calculations. This increases the need for care in attempting to combine
estimates of long-term ozone-attributable mortality and short-term ozone-attributable mortality
estimates, in order to avoid double counting. It is also important to keep in mind that our
estimates of short-term ozone- attributable mortality are for all-causes, while estimates of long-
term ozone-attributable mortality are focused on respiratory-related mortality. This further limits
the ability to compare estimates of long-term and short-term exposure related mortality.
50 See Table 5-6 for additional discussion of the issue involving reduced sampling frequency and modeling of short-
term ozone exposure-related mortality.
5-44
-------
Sensitivity Analysis: Alternate Mortality Effect Coefficients for Short-term Ozone
Exposure. In addition to the ozone-related studies we use for the core estimates, we also
evaluate several alternative studies to characterize uncertainty in the core estimates. These
alternative studies include a mix of meta-analyses and multi-city studies which have been
included in RIAs completed for previous ozone NAAQS reviews, including the review
completed in 2008 as well as the Reconsideration completed in 2010 (USEPA, 2008a and
USEPA 2010d, respectively). The decision to include a mixture of meta-analyses and multi-city
studies in the sensitivity analysis reflects the recommendation from the NRC reference earlier,
that in modeling short-term exposure-related mortality for ozone, emphasis be placed on more
recent multi-city studies without excluding consideration for meta-analyses.
In selecting effect estimates from each study, we followed the criteria presented in
Table 5-6. Consequently, we favored effect estimates reflecting the warmer ozone monitoring
period, if available. We also favored effect estimates based on the 8hr max air metric if available.
And finally, while we considered multi-pollutant models (providing some coverage for potential
confounding by PM), we placed primary emphasis on single-pollutant models since these would
typically have a significantly larger dataset with which to specify the ozone mortality effect. The
decision to emphasize single-pollutant models and deemphasize copollutants models
(specifically in modeling short-term endpoints) reflected observations in the current ozone ISA.
In relation to short-term mortality, the authors note limitations of co-pollutant models due to the
reduced sampling frequency association with PM in most cities. However, they also note that,
"Together, these co-pollutant-adjusted findings across respiratory endpoints provide support for
the independent effects of short-term exposures to ambient ozone." (ozone ISA section 6.2.9,
USEPA, 2013a). The set of multi-city and meta-analysis studies selected for inclusion in the
sensitivity analysis (including effect estimates and standard errors) is presented in Table 5-9.
Studies and effect estimates selected for the core analysis are also included in the table for
completeness. Detailed discussions of each of these studies can be found in the current ozone
ISA as well as the RIA for the ozone NAAQS completed in 2008.
Threshold-Based Effect Coefficients for Long-term Ozone Exposure. As discussed in
the HREA completed as part of this NAAQS review (U.S. EPA, 2014b), the exploration of
5-45
-------
potential thresholds for long-term exposure-related respiratory mortality (as discussed in the
Jerrett et al, 2009 study) deserves additional discussion.51 In their memo clarifying the results of
their study (see Sasser, 2014), the authors note that in terms of goodness of fit, long-term health
risk models including ozone clearly performed better than models without ozone, indicating the
improved predictions of respiratory mortality when ozone is included. In exploring different
functional forms, they report that the model including a threshold at 56 ppb had the lowest log-
likelihood value of all models evaluated (i.e., linear models and models including thresholds
ranging from 40-60 ppb), and thus provided the best overall statistical fit to the data. However,
they also note that it is not clear whether the 56 ppb threshold model is a better predictor of
respiratory mortality than when using a linear (no-threshold) model for this dataset. Using one
statistical test, the model with a threshold at 56 ppb was determined to be statistically superior to
the linear model. Using another, more stringent test, none of the threshold models considered
were statistically superior to the linear model. Under the less stringent test, although the
threshold model produces a statistically superior prediction than the linear model, there is
uncertainty about the specific location of the threshold, if one exists. This is because the
confidence intervals on the model predictions indicate that a threshold could exist anywhere
from 0 to 60 ppb. The authors conclude that considerable caution should be exercised in using
any specific threshold, particularly when the more stringent statistical test indicates there is no
significantly improved prediction.
Based on this additional information from the authors (Sasser, 2014), we have chosen to
reflect the uncertainty about the existence and location of a potential threshold by estimating
mortality attributable to long-term ozone exposures using a range of threshold-based effect
estimates as sensitivity analyses (see section 5.7.3.1 and Appendix 5B, section 5B.1).
Specifically, we generate additional long-term risk results using unique risk models that include
51 The approach we developed to explore the potential for thresholds related to long-term exposure-related mortality
was presented in a memorandum to CASAC which was also released to the public (Sasser, 2014). That
memorandum describes additional data obtained from the authors of Jerrett et al. (2009) to support modeling of
potential thresholds and also lays out our proposed approach for exploring the impact of potential thresholds on
estimates of long-term exposure-related mortality (including presentation of the threshold-based results as a
sensitivity analysis and inclusion of the non-threshold model-based results as the core analysis). This plan, including
these details related to presentation of the threshold and non-threshold based estimates were supported by CASAC
(Frey, 2014, p. 13-14).
5-46
-------
a range of thresholds from 40 ppb to 60 ppb in 5 ppb increments, while also including a model
with a threshold equal to 56 ppb, which had the lowest log likelihood value for all models
examined.52 In addition, to exploring the impact of potential thresholds, as part of the sensitivity
analysis we also explore the impact of using ozone-only (non-threshold) models in estimating
long-term exposure-related respiratory mortality.53
Ozone Exposure Metric. Both the NMMAPS analysis and the individual time series studies
upon which the meta-analyses were based use the 24-hour average or 1-hour maximum ozone
concentrations as exposure metrics. The 24-hour average is not the most relevant ozone exposure
metric to characterize population-level exposure. Given that the majority of the people tend to be
outdoors during the daylight hours and concentrations are highest during the daylight hours, the
24-hour average metric is not appropriate. Moreover, the 1 -hour maximum metric uses an
exposure window different than that used for the current ozone NAAQS. Together, this means
that the most biologically relevant metric, and the one used in the ozone NAAQS since 1997 is
the maximum daily 8-hour average ozone. Thus, we have converted ozone mortality health
impact functions that use a 24-hour average or 1-hour maximum ozone metric to maximum 8-
hour average ozone concentration using standard conversion functions.
This practice is consistent with the form of the current ozone standard. This conversion
also does not affect the relative magnitude of the health impact function from a mathematical
standpoint. An equivalent change in the 24-hour average, 1-hour maximum and 8-hour
maximum will provide the same overall change in incidence of a health effect.54 The conversion
ratios are based on observed relationships between the 24-hour average and 8-hour maximum
52 There is a separate effect estimate (and associated standard error) for each of the fitted threshold models estimated
in Jerrett et al. (2009). As a result, the sensitivity of estimated mortality attributable to long-term ozone
concentrations is affected by both the assumed threshold level (below which there is no effect of ozone) and the
effect estimate applied to ozone concentrations above the threshold.
53 The set of ozone-only non-threshold effect estimates include (a) a value based on the 86 cities for which there are
copollutants monitoring data for both ozone and PM2 5 (this best compared with the core estimate based on the
copollutants non-threshold model) and (b) a value based on the 96 cities for which there is PM2 5 data (these 96
cities were used in developing the threshold-based effect estimates used in the analysis).
54 However, it is important to note that different ozone metrics may not be well-correlated (from either a spatial or
temporal standpoint) within a given geographic area which means that application of ratio-converted effect estimates
for the same endpoint can result in different incidence estimates for the same location under certain conditions. This
introduces uncertainty into the use of these ratio-adjusted effect estimates (see Appendix 5 A).
5-47
-------
ozone values. For example, in the Bell et al., 2004 analysis of ozone-related premature mortality,
the authors found that the relationship between the 24-hour average, the 8-hour maximum, and
the 1 -hour maximum was 2:1.5:1, so that the derived health impact effect estimate based on the
1-hour maximum should be half that of the effect estimate based on the 24-hour values (and the
8-hour maximum three-quarters of the 24-hour effect estimate).
In the sensitivity analyses for this benefits analysis, we apply national effect estimates
based on the pooled multi-city results reported in Bell et al (2004) and the three meta-analysis
studies. Bell et al (2004), Bell et al (2005), Levy et al (2005), and Ito et al (2005) all provide
national conversion ratios between daily average and 8-hour and 1-hour maxima, based on
national data.
5.6.3.2 Hospital Admissions and Emergency Department Visits
Because of the availability of detailed hospital admission and discharge records, there is
an extensive body of literature examining the relationship between hospital admissions and air
pollution. For this reason, we pool together the incidence estimates using several different
studies for many of the hospital admission endpoints. In addition, some studies have examined
the relationship between air pollution and emergency department (ED) visits. Since most
emergency department visits do not result in an admission to the hospital (i.e., most people going
to the emergency department are treated and return home), we treat hospital admissions and
emergency department visits separately, taking account of the fraction of emergency department
visits that are admitted to the hospital. Specifically, within the baseline incidence rates, we parse
out the scheduled hospital visits from unscheduled ones as well as the hospital visits that
originated in the emergency department.
With regard to short-term hospital admissions and ED visits, the current ozone ISA states
that, "[c]ompared with studies reviewed in the 2006 ozone AQCD, a larger number of recent
studies examined hospital admissions and ED visits for specific respiratory outcomes. Although
limited in number, both single- and multi-city studies consistently found positive associations
between short-term ozone exposures and asthma and COPD hospital admissions and ED visits,
with more limited evidence for pneumonia. Consistent with the conclusions of the 2006 ozone
AQCD, in studies that conducted seasonal analyses, risk estimates were elevated in the warm
5-48
-------
season compared to cold season or all-season analyses, specifically for asthma and COPD."
(ozone ISA section 6.2.9, USEPA, 2013a). In this same section, the ISA also addresses potential
thresholds in effect: "Although the C-R relationship has not been extensively examined,
preliminary examinations found no evidence of a threshold between short term ozone exposure
and asthma hospital admissions and pediatric asthma ED visits..." Regarding the potential for
confounding by other pollutants including PM, the ISA observes that, "Several epidemiologic
studies of respiratory morbidity and mortality evaluated the potential confounding effects of
copollutants, in particular, PMio, PIVb.s, or NO2. In most cases, effect estimates remained robust
to the inclusion of copollutants." (ozone ISA section 6.2.9, USEPA 2013a).
Based on consideration for these observation from the ISA, and a thorough review of
available epidemiological studies, for the core analysis, we model respiratory hospital
admissions (for 65-99yr olds) using effect estimates obtained from Katsouyanni et al, 2009 and
asthma-related emergency room visits (for all ages) using several single-city studies. The
Katsouyanni et al., 2009 study is for all respiratory hospital admissions, and thus to avoid double
counting, we do not provide separate estimates for specific subcategories of respiratory
admissions such as asthma. The Katsouyanni et al., 2009 study provides effect estimates specific
to the summer season, which is an advantage, however it also utilizes the Ihr max metric, which
required adjustment using air metric ratios to generate equivalent 8hr max effect estimates for
use in the RIA.55 The study provides summer season single pollutant effect estimates based both
on natural and penalized splines. We used both of these effect estimates and pooled the results
using equal-weight averaging. It is also important to note that, while the Katsouyanni et al., 2009
study did include a set of effect estimates based on copollutants modeling (with PMio), these
were based on the full year rather than the summer season. Given our focus on warmer ozone
season-based models, we only considered the single pollutant models in the RIA.
A number of studies are available to model respiratory ED visits. However, at this time
we do not have a valuation function for this endpoint. Since we do have a valuation function
available for the narrower category of asthma-related ED visits, for the core estimate, we have
55 Given that the Katsouyanni et al., 2009 study included a larger number of cities (14), rather than constructing an
air metric adjustment ratio based on this set of urban study areas, we used a national ratio to adjust effect estimates
to represent the 8hr metric used in the RIA.
5-49
-------
focused on asthma-related ED visits (for which we can estimate the economic value), using a set
of single city studies together with random-effects pooling to generate a single pooled estimate.
The set of single city studies used in this calculation include: Peel et al., (2005) and Sarnat et al,
(2013) both for Atlanta, Wilson et al., (2005) and Mar and Koenig (2009) for Seattle, Wilson et
al., (2005) for Portland ME, Ito et al., (2007) for New York City, and Glad et al., (2012) for
Pittsburgh. We note that of these single city studies, only the Ito et al., (2007) study included a
co-pollutant model (for PM2.s).56 In addition, two of the studies required adjustments of their
betas to reflect the 8hr max air metric. Specifically, Glad et al., (2012) utilizes the Ihr max air
metric, while Sarnat et al., 2013 utilized the 24hr average metric. Each required the use of air
metric ratios to adjust their betas. In generating a single pooled benefit estimate for this endpoint,
we used random/fixed effects pooling to combine estimates across these single city studies.
5.6.3.3 Acute Health Events and School/Work Loss Days
In addition to mortality, chronic illness, and hospital admissions, a number of acute
health effects not requiring hospitalization are associated with exposure to ozone and PM2.5. The
sources for the effect estimates used to quantify these effects are described below.
Asthma exacerbations. For this RIA, we have followed the SAB-HES recommendations
regarding asthma exacerbations in developing the core estimate (U.S. EPA-SAB, 2004a).
Although certain studies of acute respiratory events characterize these impacts among only
asthmatic populations, others consider the full population, including both asthmatics and non-
asthmatics. For this reason, incidence estimates derived from studies focused only on asthmatics
cannot be added to estimates from studies that consider the full population—to do so would
double-count impacts. To prevent such double-counting, we estimated the exacerbation of
asthma among children and excluded adults from the calculation. Asthma exacerbations
occurring in adults are assumed to be captured in the general population endpoints such as work
loss days and minor restricted activity days (MRADs). Finally, we note the important distinction
56 While we have included copollutants models as sensitivity analyses for mortality, given the reduced role of
morbidity endpoints in driving overall dollar benefits, we have not included separate copollutants models for any of
the morbidity endpoints as sensitivity analyses. In the case of the Ito et al., (2007) study, we do note, that while the
copollutants model does result in a somewhat smaller effect estimate for asthma ED visits as compared with the
single pollutant (ozone-only) model, the SE for the copollutants model is also larger (as would be expected).
5-50
-------
between the exacerbation of asthma among asthmatic populations, and the onset of asthma
among populations not previously suffering from asthma; in this RIA, we quantify the
exacerbation of asthma among asthmatic populations and not the onset of new cases of asthma.
Based on advice from the SAB-HES (EPA-SAB 2004a), regardless of the age ranges
included in the source epidemiology studies, we extend the applied population to ages 6 to 18,
reflecting the common biological basis for the effect in children in the broader age group. This
age range expansion is also supported by NRC (2002, pp. 8, 116).
To characterize asthma exacerbations in children from exposure to ozone, we selected
three multi-city studies (Mortimer et al, 2002, O'Connor et al, 2008, and Schildcrout et al,
2006). Of these three, one of the studies (O'Connor et al., 2008) only included a multi-pollutant
model (for PIVh.s and ISTCh) and consequently, that effect estimate was used in the core analysis.
All three of these studies required the application of air metric ratios to adjust effect estimates to
represent the 8hr metric used in the RIA.57 To combine these three estimates into a single pooled
estimate, we used equal weights.
Acute Respiratory Symptoms. We estimate one type of acute respiratory symptom related to
ozone exposure - MRAD. Minor restricted activity days result when individuals reduce most
usual daily activities and replace them with less strenuous activities or rest, yet not to the point of
missing work or school. For example, a mechanic who would usually be doing physical work
most of the day will instead spend the day at a desk doing paper work and phone work because
of difficulty breathing or chest pain.
For ozone, we modeled MRADs using Ostro and Rothschild (1989). This study provides
a copollutants model (with PM2.s) based on a national sample of 18-64yr olds. The original study
used a 24hr average metric and included control for PM2.5, which necessitated the use of an air
metric ratio to convert the effect estimate to an 8hr max equivalent.
57 Mortimer et al., (2002) had effect estimates based on an 8hr mean metric, O'Connor et al., (2008) utilized a 24hr
metric and Schildcrout et al., (2006) was based on a Ihr max metric. Consequently, all three studies required the
application of air metric ratios to produce effect estimates reflecting an 8hr max metric.
5-51
-------
School loss days (absences). Children may be absent from school due to respiratory or other
acute diseases caused, or aggravated by, exposure to air pollution. Several studies have found a
significant association between ozone levels and school absence rates. We use two studies
(Gilliland et al, 2001; Chen et al, 2000) to estimate changes in school absences resulting from
changes in ozone levels. The Gilliland et al. study estimated the incidence of new periods of
absence, while the Chen et al. study examined daily absence rates. We converted the Gilliland et
al. estimate to days of absence by multiplying the absence periods by the average duration of an
absence. We estimated 1.6 days as the average duration of a school absence, the result of
dividing the average daily school absence rate from Chen et al. (2000) and Ransom and Pope
(1992) by the episodic absence duration from Gilliland et al. (2001). Thus, each Gilliland et al.
period of absence is converted into 1.6 absence days.
Following advice from the National Research Council (NRC, 2002), we calculated
reductions in school absences for the full population of school age children, ages five to 17. This
is consistent with recent peer-reviewed literature on estimating the impact of ozone exposure on
school absences (Hall et al., 2003). We estimated the change in school absences using both Chen
et al. (2000) and Gilliland et al. (2001) and then pooled the results using the random effects
pooling procedure.
5.6.3.4 Unqualified Human Health Effects
The illustrative emission reduction strategies to reach the proposed and alternative
standards described in Chapter 4 would reduce emissions of NOx and VOCs. Although we have
quantified many of the health benefits associated with reducing exposure to ozone and PIVh.s, as
shown in Table 5-3, we are unable to quantify the health benefits associated with reducing the
potential for NCh or VOC exposures due to the absence of air quality modeling data for these
pollutants in this analysis. In addition, we are unable to quantify the effects of VOC reductions
on ambient PIVh.s and associated health effects. Although the method we applied simulated the
impact of attaining the proposed and alternative standards on ambient levels of ozone, this
method does not simulate how the illustrative emission reductions would affect ambient levels of
NO2 or VOC. Below we provide a qualitative description of these health benefits. In general,
previous analyses have shown that the monetized value of these additional health benefits is
much smaller than ozone and PM2.5-related benefits (U.S. EPA, 2010a, 2010c, 2010d).
5-52
-------
Epidemiological researchers have associated NCh exposure with adverse health effects in
numerous lexicological, clinical and epidemiological studies, as described in the Integrated
Science Assessment for Oxides of Nitrogen—Health Criteria (NO2 ISA) (U.S. EPA, 2008b). The
NO2 ISA provides a comprehensive review of the current evidence of health and environmental
effects of NO2. The NCh ISA concluded that the evidence "is sufficient to infer a likely causal
relationship between short-term NCh exposure and adverse effects on the respiratory system."
These epidemiologic and experimental studies encompass a number of endpoints including
emergency department visits and hospitalizations, respiratory symptoms, airway
hyperresponsiveness, airway inflammation, and lung function. Effect estimates from
epidemiologic studies conducted in the United States and Canada generally indicate a 2-
20 percent increase in risks for ED visits and hospital admissions and higher risks for respiratory
symptoms. The NCh ISA concluded that the relationship between short-term NCh exposure and
premature mortality was "suggestive but not sufficient to infer a causal relationship" because it is
difficult to attribute the mortality risk effects to NCh alone. Although the NO2 ISA stated that
studies consistently reported a relationship between NO2 exposure and mortality, the effect was
generally smaller than that for other pollutants such as PM. We did not quantify these benefits
due to data constraints.
The EPA last quantified the value of ozone-related worker productivity in the final
Regulatory Impact Analysis supporting the Transport Rule (USEPA, 201 Id). That analysis
applied information reported in Crocker and Horst (1981) to relate changes in ground-level
ozone to changes in the productivity of outdoor citrus workers. That study found that a 10
percent reduction in ozone translated to a 1.4 increase in income among outdoor citrus
workers. Concerned that this study might not adequately characterize the relationship between
ground-level ozone and the productivity of agricultural workers because of the vintage of the
underlying data, the Agency subsequently omitted this endpoint.
In 2012, Graff Zivin and Neidell published "The impact of pollution on worker
productivity" in the American Economic Review. That study combined data on individual-level
daily harvest rates for Outdoor Agricultural Workers (OWAs) with ground-level ozone pollution
to characterize changes in worker productivity. The authors used data on harvest rates from a
500-acre farm in the Central Valley of California. That farm produced three crops (blueberries
5-53
-------
and two types of grapes) and the harvesting laborers were paid through piece rate contracts. The
analyses in the paper were based on 2009 and 2010 California growing seasons. The analyses
were not affected by: (i) endogenous ozone exposure (because there were limited local sources of
ozone precursors); (ii) avoidance behavior (because the work has to be performed outdoors); and
(iii) shirking (due to the nature of the piece rate contract).
Table 3 in Graff Zivin and Neidell (2012) reports the main result: A 10 ppb increase in
work-day ozone concentration (represented by hourly measurements averaged between 6am and
3pm) will result in a decline of 0.143 (with a standard error of 0.068) in standardized hourly
pieces collected on a given work day. The standardized hourly pieces were "the average hourly
productivity minus the minimum number of pieces per hour required to reach the piece rate
regime, divided by the standard deviation of productivity for each crop" (Graff Zivin and
Neidell, 2012; p. 3665). The range of ozone concentrations in the sample was between 10.50 ppb
and 86.0 ppb (Table 1 in Graff Zivin and Neidell, 2012). This result is significant and robust
under different model specifications designed to test modeling assumptions. Based on the effect
estimate and individual-level information in their dataset, the authors estimated the effect of an
increase in ozone concentration on the worker productivity, as measured by the average number
of pieces collected per hour during a given work day (rather than by standardized hourly piece
rate that was used in regression modeling). They found a decline of 5.5% in worker productivity
due to a 10 ppb increase in average work-day ozone concentration.
While Graff Zivin and Neidell (2012) report the information needed to quantify ozone-
related worker productivity, we are still evaluating whether and how to most appropriately apply
the limited evidence from this study in a national benefits assessment. An important issue is the
generalizability of the results to the appropriate population. We are considering the
appropriateness of applying the results of this study to estimate the benefits of increased worker
productivity as part of the final ozone RIA, and seek public comment on this approach as an
important input to our consideration.
5-54
-------
5.6.4 Economic Valuation Estimates
Reductions in ambient concentrations of air pollution generally lower the risk of future
adverse health effects for a large population. Therefore, the appropriate economic measure is
willingness-to-pay (WTP) for changes in risk of a health effect rather than WTP for a health
effect that would occur with certainty (Freeman, 1993). Epidemiological studies generally
provide estimates of the relative risks of a particular health effect that is avoided because of a
reduction in air pollution. We converted those changes in risk to units of avoided statistical
incidence for ease of presentation. We calculated the value of avoided statistical incidences by
dividing individual WTP for a risk reduction by the related observed change in risk. For
example, suppose a measure is able to reduce the risk of premature mortality from 2 in 10,000 to
1 in 10,000 (a reduction of 1 in 10,000). If individual WTP for this risk reduction is $100, then
the WTP for an avoided statistical premature mortality amounts to $1 million ($100/0.0001
change in risk). Using this approach, the size of the affected population is automatically taken
into account by the number of incidences predicted by epidemiological studies applied to the
relevant population. The same type of calculation can produce values for statistical incidences of
other health endpoints.
WTP estimates generally are not available for some health effects, such as hospital
admissions. In these cases, we instead used the cost of treating or mitigating the effect to
estimate the economic value. Cost-of-illness (COI) estimates generally (although not necessarily
in all cases) understate the true value of reducing the risk of a health effect, because they reflect
the direct expenditures related to treatment, but not the value of avoided pain and suffering
(Harrington and Portney, 1987; Berger, 1987).
We provide unit values for health endpoints (along with information on the distribution
of the unit value) in Table 5-10.58 All values are in constant year 2011$, adjusted for growth in
real income for WTP estimates out to 2024 using projections provided by Standard and Poor's,
58 We note that a number of the endpoints included in Table 5-10 and discussed in this section were only modeled
for PM2s and consequently could be moved to Appendix 5C (as was done with the discussion of effect estimates
specific to PM2 5). However, given that many of the valuation functions discussed are shared between the two
pollutants and given the relatively shorter length of discussions associated with these valuation functions, we have
included coverage of all valuation functions (for ozone and PM2 5 -related endpoints) in this section.
5-55
-------
which is discussed in further detail below.59 Economic theory argues that WTP for most goods
(such as environmental protection) will increase if real income increases. Several of the valuation
studies used in this analysis were conducted in the late 1980s and early 1990s, and we are in the
process of reviewing the literature to update these unit values. The discussion below provides
additional details on valuing specific PIVh.s-related related endpoints.
5.6.4.1 Mortality Valuation
Following the advice of the SAB's Environmental Economics Advisory Committee
(SAB-EEAC), the EPA currently uses the value of statistical life (VSL) approach in calculating
the core estimate of mortality benefits, because we believe this calculation provides the most
reasonable single estimate of an individual's willingness to trade off money for reductions in
mortality risk (U.S. EPA-SAB, 2000). The VSL approach is a summary measure for the value of
small changes in mortality risk experienced by a large number of people. For a period of time
(2004-2008), the Office of Air and Radiation (OAR) valued mortality risk reductions using a
VSL estimate derived from a limited analysis of some of the available studies. OAR arrived at a
VSL using a range of $1 million to $10 million (2000$) consistent with two meta-analyses of the
wage-risk literature. The $1 million value represented the lower end of the interquartile range
from the Mrozek and Taylor (2002) meta-analysis of 33 studies. The $10 million value
represented the upper end of the interquartile range from the Viscusi and Aldy (2003) meta-
analysis of 43 studies. The mean estimate of $5.5 million (2000$) was also consistent with the
mean VSL of $5.4 million estimated in the Kochi et al. (2006) meta-analysis. However, the
Agency neither changed its official guidance on the use of VSL in rule-makings nor subjected
the interim estimate to a scientific peer-review process through SAB or other peer-review group.
59 Income growth projections are only currently available in BenMAP through 2024, so both the 2025 and 2038
estimates use income growth only through 2024 and are therefore likely underestimates. Currently, BenMAP does
not have an inflation adjustment to 2011$. We ran BenMAP for a currency year of 2010$ and calculated the benefit-
per-ton estimates in 2010$. We then adjusted the resulting benefit-per-ton estimates to 2011$ using the Consumer
Price Index (CPI-U, all items). This approach slightly underestimates the inflation for medical index and wage index
between 2010 and 2011, which affects COI estimates and wage-based estimates.
5-56
-------
Table 5-10. Unit Values for Economic Valuation of Health Endpoints (2011$)
Central Estimate of Value Per Statistical Incidence
Health Endpoint
Premature Mortality (Value of a
Statistical Life)
Nonfatal Myocardial Infarction
(heart attack)
3% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-64
Age 65 and over
7% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-64
Age 65 and over
1990 Income Level
$8,300,000
$100,000
$110,000
$120,000
$210,000
$100,000
$100,000
$110,000
$120,000
$190,000
$100,000
2024 Income Level
$10,000,000
$100,000
$110,000
$120,000
$210,000
$100,000
$100,000
$110,000
$120,000
$190,000
$100,000
Derivation of Distributions of Estimates
The EPA currently recommends a central VSL of $4.8 million (1990$,
1990 income) based on a Weibull distribution fitted to 26 published
VSL estimates (5 contingent valuation and 21 labor market studies).
The underlying studies, the distribution parameters, and other useful
information are available in Appendix B of the EPA's Guidelines for
Preparing Economic Analyses (U.S. EPA, 2010e).
No distributional information available. Age-specific cost-of-illness
values reflect lost earnings and direct medical costs over a 5-year
period following a nonfatal Ml. Lost earnings estimates are based on
Cropper and Krupnick (1990). Direct medical costs are based on
simple average of estimates from Russell et al. (1998) and Wittels
et al. (1990).
Lost earnings:
Cropper and Krupnick (1990). Present discounted value of 5 years of
lost earnings in 2000$:
age of onset: at 3% at 7%
25-44 $9,000 $8,000
45-54 $13,000 $12,000
55-65 $77,000 $69,000
Direct medical expenses (2000$): An average of:
1. Wittels et al. (1990) ($100,000— no discounting)
2. Russell et al. (1998), 5-year period ($22,000 at 3% discount rate;
$21,000 at 7% discount rate)
(continued)
-------
Table 5-10. Unit Values for Economic Valuation of Health Endpoints (2011$) a (continued)
Central Estimate of Value Per Statistical Incidence
Health Endpoint
2000 Income Level
2024 Income Level
Derivation of Distributions of Estimates
Hospital Admissions
Chronic Lung Disease (18-64)
$22,000
$22,000
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g., average hospital care costs, average length of
hospital stay, and weighted share of total chronic lung illnesses)
reported in Agency for Healthcare Research and Quality (2007)
(www.ahrq.gov).
LT\
OO
Asthma Admissions (0-64)
$16,000
$16,000
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g., average hospital care costs, average length of
hospital stay, and weighted share of total asthma category illnesses)
reported in Agency for Healthcare Research and Quality (2007)
(www.ahrq.gov).
All Cardiovascular
Age 18-64
Age 65-99
$44,000
$42,000
$44,000
$42,000
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g., average hospital care costs, average length of
hospital stay, and weighted share of total cardiovascular category
illnesses) reported in Agency for Healthcare Research and Quality
(2007) (www.ahrq.gov).
All respiratory (ages 65+)
$37,000
$37,000
No distributions available. The COI point estimates (lost earnings
plus direct medical costs) are based on ICD-9 code level information
(e.g., average hospital care costs, average length of hospital stay,
and weighted share of total respiratory category illnesses) reported
in Agency for Healthcare Research and Quality, 2007
(www.ahrq.gov).
Emergency Department Visits
for Asthma
$440
$440
No distributional information available. Simple average of two unit
COI values (2000$):
(1) $310, from Smith et al. (1997) and
(2) $260, from Stanford et al. (1999).
(continued)
-------
Table 5-10. Unit Values for Economic Valuation of Health Endpoints (2011$) a (continued)
Central Estimate of Value Per Statistical Incidence
Health Endpoint
2000 Income Level
2024 Income Level
Derivation of Distributions of Estimates
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms
(URS)
$35
$32
Combinations of the three symptoms for which WTP estimates are
available that closely match those listed by Pope et al. result in seven
different "symptom clusters," each describing a "type" of URS. A
dollar value was derived for each type of URS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. In the absence of information
surrounding the frequency with which each of the seven types of
URS occurs within the URS symptom complex, we assumed a
uniform distribution between $9.2 and $43 (2000$).
Lower Respiratory Symptoms
(LRS)
$22
$21
Combinations of the four symptoms for which WTP estimates are
available that closely match those listed by Schwartz et al. result in
11 different "symptom clusters," each describing a "type" of LRS. A
dollar value was derived for each type of LRS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for LRS is the
average of the dollar values for the 11 different types of LRS. In the
absence of information surrounding the frequency with which each
of the 11 types of LRS occurs within the LRS symptom complex, we
assumed a uniform distribution between $6.9 and $25 (2000$).
Asthma Exacerbations
$56
$60
Asthma exacerbations are valued at $45 per incidence, based on the
mean of average WTP estimates for the four severity definitions of a
"bad asthma day," described in Rowe and Chestnut (1986). This
study surveyed asthmatics to estimate WTP for avoidance of a "bad
asthma day," as defined by the subjects. For purposes of valuation,
an asthma exacerbation is assumed to be equivalent to a day in
which asthma is moderate or worse as reported in the Rowe and
Chestnut (1986) study. The value is assumed to have a uniform
distribution between $16 and $71 (2000$).
(continued)
-------
Table 5-10. Unit Values for Economic Valuation of Health Endpoints (2011$) a (continued)
Central Estimate of Value Per Statistical Incidence
Health Endpoint
2000 Income Level
2024 Income Level
Derivation of Distributions of Estimates
Respiratory Ailments Not Requiring Hospitalization (continued)
Acute Bronchitis
$460
$500
Assumes a 6-day episode, with the distribution of the daily value
specified as uniform with the low and high values based on those
recommended for related respiratory symptoms in Neumann et al.
(1994). The low daily estimate of $10 is the sum of the mid-range
values recommended by lEc (1994) for two symptoms believed to be
associated with acute bronchitis: coughing and chest tightness. The
high daily estimate was taken to be twice the value of a minor
respiratory restricted-activity day, or $110 (2000$).
Work Loss Days (WLDs)
Variable Variable
(U.S. median = $150) (U.S. median = $150)
No distribution available. Point estimate is based on county-specific
median annual wages divided by 52 and then by 5—to get median
daily wage. U.S. Year 2000 Census, compiled by Geolytics, Inc.
(Geolytics, 2002)
o\
o
School Loss Days
$98
$98
No distribution available. Based on (1) the probability that, if a
school child stays home from school, a parent will have to stay home
from work to care for the child, and (2) the value of the parent's lost
productivity.
Minor Restricted Activity Days $64
(MRADs)
$68
Median WTP estimate to avoid one MRAD from Tolley et al. (1986).
Distribution is assumed to be triangular with a minimum of $22 and
a maximum of $83, with a most likely value of $52 (2000$). Range is
based on assumption that value should exceed WTP for a single mild
symptom (the highest estimate for a single symptom—for eye
irritation—is $16) and be less than that for a WLD. The triangular
distribution acknowledges that the actual value is likely to be closer
to the point estimate than either extreme.
a All estimates are rounded to two significant digits. Unrounded estimates in 2000$ are available in the Appendix J of the BenMAP user manual (Abt Associates,
2012). Income growth projections are only currently available in BenMAP through 2024, so both the 2025 and 2038 estimates use income growth only through
9094 °TlH °1"P thprpfnrP lilcplv llTlHprPQtiTTlfltPQ Currently, BenMAP does not have an inflation adjustment to 2011$. We ran BenMAP for a currency year of 2010$ and calculated the benefit-per-ton estimates in 2010$. We then
adjusted the resulting benefit-per-ton estimates to 2011$ using the Consumer Price Index /'(^TDT T T oil itPTTlQV This approach slightly underestimates the inflation for medical index and wage index between 2010 and 2011
affects COI estimates and wage-based estimates
-------
During this time, the Agency continued work to update its guidance on valuing mortality
risk reductions, including commissioning a report from meta-analytic experts to evaluate
methodological questions raised by the EPA and the SAB on combining estimates from the
various data sources. In addition, the Agency consulted several times with the SAB-EEAC on the
issue. With input from the meta-analytic experts, the SAB-EEAC advised the Agency to update
its guidance using specific, appropriate meta-analytic techniques to combine estimates from
unique data sources and different studies, including those using different methodologies (i.e.,
wage-risk and stated preference) (U.S. EPA-SAB, 2007).
Until updated guidance is available, the Agency determined that a single, peer-reviewed
estimate applied consistently best reflects the SAB-EEAC advice it has received. Therefore, the
Agency has decided to apply the VSL that was vetted and endorsed by the SAB in the Guidelines
for Preparing Economic Analyses (U.S. EPA, 2000)60 while the Agency continues its efforts to
update its guidance on this issue. This approach calculates a mean value across VSL estimates
derived from 26 labor market and contingent valuation studies published between 1974 and
1991. The mean VSL across these studies is $4.8 million (1990$) or $6.3 million (2000$).61 The
Agency is committed to using scientifically sound, appropriately reviewed evidence in valuing
mortality risk reductions and has made significant progress in responding to the SAB-EEAC's
specific recommendations. In the process, the Agency has identified a number of important
issues to be considered in updating its mortality risk valuation estimates. These are detailed in a
white paper on "Valuing Mortality Risk Reductions in Environmental Policy," which underwent
review by the SAB-EEAC. A meeting with the SAB on this paper was held on March 14, 2011
and formal recommendations were transmitted on July 29, 2011 (U.S. EPA-SAB, 2011). EPA is
taking SAB's recommendations under advisement.
The economics literature concerning the appropriate method for valuing reductions in
premature mortality risk is still developing. The adoption of a value for the projected reduction in
60 In the updated Guidelines for Preparing Economic Analyses (U.S. EPA, 2010e), EPA retained the VSL endorsed
by the SAB with the understanding that further updates to the mortality risk valuation guidance would be
forthcoming in the near future.
61 In this analysis, we adjust the VSL to account for a different currency year (2011$) and to account for income
growth to 2024. After applying these adjustments to the $6.3 million value, the VSL is $10 million. Income
growth projections are only currently available in BenMAP through 2024, so both the 2025 and 2038 estimates
use income growth only through 2024 and are therefore likely underestimates.
5-61
-------
the risk of premature mortality is the subject of continuing discussion within the economics and
public policy analysis community. The EPA strives to use the best economic science in its
analyses. Given the mixed theoretical finding and empirical evidence regarding adjustments to
VSL for risk and population characteristics (e.g., Smith et al, 2004; Alberini et al, 2004; Aldy
and Viscusi, 2008), we use a single VSL for all reductions in mortality risk.
Although there are several differences between the labor market studies the EPA uses to
derive a VSL estimate and the ozone and PIVh.s air pollution context addressed here, those
differences in the affected populations and the nature of the risks imply both upward and
downward adjustments. Table 5-11 lists some of these differences and the expected effect on the
VSL estimate for air pollution-related mortality. In the absence of a comprehensive and balanced
set of adjustment factors, the EPA believes it is reasonable to continue to use the $4.8 million
(1990$) value adjusted for inflation and income growth over time while acknowledging the
significant limitations and uncertainties in the available literature.
Table 5-11. Influence of Applied VSL Attributes on the Size of the Economic Benefits of
Reductions in the Risk of Premature Death (U.S. EPA, 2006a)
Attribute Expected Direction of Bias
Age Uncertain, perhaps overestimate
Life Expectancy/Health Status Uncertain, perhaps overestimate
Attitudes Toward Risk Underestimate
Income Uncertain
Voluntary vs. Involuntary Uncertain, perhaps underestimate
Catastrophic vs. Protracted Death Uncertain, perhaps underestimate
The SAB-EEAC has reviewed many potential VSL adjustments and the state of the
economics literature. The SAB-EEAC advised the EPA to "continue to use a wage-risk-based
VSL as its primary estimate, including appropriate sensitivity analyses to reflect the uncertainty
of these estimates," and that "the only risk characteristic for which adjustments to the VSL can
be made is the timing of the risk" (U.S. EPA-SAB, 2000). In developing our core estimate of the
benefits of premature mortality reductions, we have followed this advice.
For PM2.s-related premature mortality, we assume that there is a "cessation" lag between
exposures and the total realization of changes in health effects. For PIVh.s, we assumed that some
of the incidences of premature mortality related to PIVh.s exposures occur in a distributed fashion
5-62
-------
over the 20 years following exposure and discounted over the period between exposure and
premature mortality. Although the structure of the lag is uncertain, the EPA follows the advice of
the SAB-HES to assume a segmented lag structure characterized by 30 percent of mortality
reductions in the first year, 50 percent over years 2 to 5, and 20 percent over the years 6 to 20
after the reduction in PIVh.s (U.S. EPA-SAB, 2004c). To take this into account in the valuation of
reductions in premature mortality, we discount the value of premature mortality occurring in
future years using rates of 3 percent and 7 percent.62 Changes in the cessation lag assumptions do
not change the total number of estimated deaths but rather the timing of those deaths. As such,
the monetized PIVb.s co-benefits using a 7 percent discount rate are only approximately 10
percent less than the monetized benefits using a 3 percent discount rate. Further discussion of
this topic appears in the EPA's Guidelines for Preparing Economic Analyses (U.S. EPA, 2010e).
For ozone, we acknowledge substantial uncertainty associated with specifying the lag for
long-term respiratory mortality. As stated earlier, it is this uncertainty related to specifying a lag
structure which prevents us from including this endpoint as a monetary benefit estimate within
the core analysis.63 In presenting dollar benefit estimates as part of the sensitivity analysis, we
include both an assumption of zero lag and a lag structure matching that used for the core PIVh.s
estimate (the SAB 20 year segmented lag). Inclusion of the zero lag reflects consideration for the
possibility that the long-term respiratory mortality estimate captures primarily, an accumulation
of short-term mortality effects across the ozone season.64 The use of the 20 year segmented lag
62 The choice of a discount rate, and its associated conceptual basis, is a topic of ongoing discussion within the
federal government. To comply with OMB Circular A-4, EPA provides monetized benefits using discount rates of
3% and 7% (OMB, 2003). A 3% discount reflects reliance on a "social rate of time preference" discounting
concept. A 7% rate is consistent with an "opportunity cost of capital" concept to reflect the time value of
resources directed to meet regulatory requirements.
63 Recall however, that we consider the estimate of reduced incidence of respiratory mortality associated with long-
term ozone exposure to have sufficient support in the literature to be included as part of the core estimate.
64 In presenting risk estimates associated with modeling long-term ozone-related respiratory mortality in the HREA,
we noted that: "The effect estimates used in modeling long-term Os-attributable mortality, utilize a seasonal average
of peak (1-hr maximum) measurements. These long-term exposure metrics can be viewed as long-term exposures to
daily peak Os over the warmer months, as compared with annual average levels such as are used in long-term PM
exposure calculations. This increases the need for care in interpreting these long-term Os-attributable mortality
estimates together with the short-term Os-attributable mortality estimates, in order to avoid double counting."
(USEPA, 2014b). This statement was included in the 2nd draft of the HREA that CAS AC commented extensively
on (p. 7-21) (Frey 2014). While the CASAC made recommendations on specific aspects of our approach for
modeling this endpoint (specifically the need to include consideration for potential thresholds as a sensitivity
analysis), they did not criticize this observation regarding the potential that this metric could actually represent an
5-63
-------
reflects consideration for advice provided by the HES (USEPA-SAB, 2010a), where they state
that, "[i]f Alternative Estimates are derived using cohort mortality evidence, there is no evidence
in the literature to support a different cessation lag between ozone and particulate matter. The
HES therefore recommends using the same cessation lag structure and assumptions as for
particulate matter when utilizing cohort mortality evidence for ozone." Dollar benefit estimates
generated using both lag assumptions are presented as sensitivity analyses (see section 5.7.3.1).
Uncertainties Specific to Premature Mortality Valuation. The economic benefits
associated with reductions in the risk of premature mortality are the largest category of
monetized benefits in this RIA. In addition, in prior analyses, the EPA identified valuation of
mortality-related benefits as the largest contributor to the range of uncertainty in monetized
benefits (Mansfield et al, 2009).65 Because of the uncertainty in estimates of the value of
reducing premature mortality risk, it is important to adequately characterize and understand the
various types of economic approaches available for valuing reductions in mortality risk. Such an
assessment also requires an understanding of how alternative valuation approaches reflect that
some individuals may be more susceptible to air pollution-induced mortality or reflect
differences in the nature of the risk presented by air pollution relative to the risks studied in the
relevant economics literature.
The health science literature on air pollution indicates that several human characteristics
affect the degree to which mortality risk affects an individual. For example, some age groups
appear to be more susceptible to air pollution than others (e.g., the elderly and children). Health
status prior to exposure also affects susceptibility. An ideal benefits estimate of mortality risk
reduction would reflect these human characteristics, in addition to an individual's WTP to
improve one's own chances of survival plus WTP to improve other individuals' survival rates.
accumulation of short-term daily peak exposures and that care needed to be taken to avoid double-counting of
mortality incidence. Our inclusion of a zero lag reflects the potential that the estimate of long-term exposure-related
respiratory mortality could (to a significant extent) capture an accumulation of short-term effects. Note, that we
include the zero threshold model together with a 20 year segmented lag model in order to capture a potential range
of lag effect and both are given equal coverage in generating the dollar benefit estimates included as a sensitivity
analysis for this endpoint.
65 This conclusion was based on an assessment of uncertainty based on statistical error in epidemiological effect
estimates and economic valuation estimates. Additional sources of model error such as those examined in the
PM2 5 mortality expert elicitation (Roman et al., 2008) may result in different conclusions about the relative
contribution of sources of uncertainty.
5-64
-------
The ideal measure would also take into account the specific nature of the risk reduction
commodity that is provided to individuals, as well as the context in which risk is reduced. To
measure this value, it is important to assess how reductions in air pollution reduce the risk of
dying from the time that reductions take effect onward and how individuals value these changes.
Each individual's survival curve, or the probability of surviving beyond a given age, should shift
as a result of an environmental quality improvement. For example, changing the current
probability of survival for an individual also shifts future probabilities of that individual's
survival. This probability shift will differ across individuals because survival curves depend on
such characteristics as age, health state, and the current age to which the individual is likely to
survive.
Although a survival curve approach provides a theoretically preferred method for valuing
the benefits of reduced risk of premature mortality associated with reducing air pollution, the
approach requires a great deal of data to implement. The economic valuation literature does not
yet include good estimates of the value of this risk reduction commodity. As a result, in this
study we value reductions in premature mortality risk using the VSL approach.
Other uncertainties specific to premature mortality valuation include the following:
• Across-study variation: There is considerable uncertainty as to whether the available
literature on VSL provides adequate estimates of the VSL for risk reductions from air
pollution reduction. Although there is considerable variation in the analytical designs and
data used in the existing literature, the majority of the studies involve the value of risks to
a middle-aged working population. Most of the studies examine differences in wages of
risky occupations, using a hedonic wage approach. Certain characteristics of both the
population affected and the mortality risk facing that population are believed to affect the
average WTP to reduce the risk. The appropriateness of a distribution of WTP based on
the current VSL literature for valuing the mortality-related benefits of reductions in air
pollution concentrations therefore depends not only on the quality of the studies (i.e., how
well they measure what they are trying to measure), but also on the extent to which the
risks being valued are similar and the extent to which the subjects in the studies are
similar to the population affected by changes in pollution concentrations.
• Level of risk reduction: The transferability of estimates of the VSL from the wage-risk
studies to the context of this analysis rests on the assumption that, within a reasonable
range, WTP for reductions in mortality risk is linear in risk reduction. For example,
suppose a study provides a result that the average WTP for a reduction in mortality risk
of 1/100,000 is $50, but that the actual mortality risk reduction resulting from a given
pollutant reduction is 1/10,000. If WTP for reductions in mortality risk is linear in risk
5-65
-------
reduction, then a WTP of $50 for a reduction of 1/100,000 implies a WTP of $500 for a
risk reduction of 1/10,000 (which is 10 times the risk reduction valued in the study).
Under the assumption of linearity, the estimate of the VSL does not depend on the
particular amount of risk reduction being valued. This assumption has been shown to be
reasonable provided the change in the risk being valued is within the range of risks
evaluated in the underlying studies (Rowlatt et al, 1998).
• Voluntariness of risks evaluated: Although job-related mortality risks may differ in
several ways from air pollution-related mortality risks, the most important difference may
be that job-related risks are incurred voluntarily, or generally assumed to be, whereas air
pollution-related risks are incurred involuntarily. Some evidence suggests that people will
pay more to reduce involuntarily incurred risks than risks incurred voluntarily (e.g.,
Lichtenstein and Slovic, 2006). If this is the case, WTP estimates based on wage-risk
studies may understate WTP to reduce involuntarily incurred air pollution-related
mortality risks.
• Sudden versus protracted death: A final important difference related to the nature of
the risk may be that some workplace mortality risks tend to involve sudden, catastrophic
events, whereas air pollution-related risks tend to involve longer periods of disease and
suffering prior to death. Some evidence suggests that WTP to avoid a risk of a protracted
death involving prolonged suffering and loss of dignity and personal control is greater
than the WTP to avoid a risk (of identical magnitude) of sudden death (e.g., Tsuge et al.,
2005; Alberini and Scasny, 2011). To the extent that the mortality risks addressed in this
assessment are associated with longer periods of illness or greater pain and suffering than
are the risks addressed in the valuation literature, the WTP measurements employed in
the present analysis would reflect a downward bias.
• Self-selection and skill in avoiding risk: Recent research (Shogren and Stamland, 2002)
suggests that VSL estimates based on hedonic wage studies may overstate the average
value of a risk reduction. This is based on the fact that the risk-wage trade-off revealed in
hedonic studies reflects the preferences of the marginal worker (i.e., that worker who
demands the highest compensation for his risk reduction for a given job). This worker
must have either a higher workplace risk than the average worker in a given occupation, a
lower risk tolerance than the average worker in that occupation, or both. Conversely, the
marginal worker should have a higher risk tolerance than workers employed in less-risky
sectors. However, the risk estimate used in hedonic studies is generally based on average
risk, so the VSL may be biased, in an ambiguous direction, because the wage differential
and risk measures do not match.
• Baseline risk and age: Recent research (Smith, Pattanayak, and Van Houtven, 2006)
finds that because individuals reevaluate their baseline risk of death as they age, the
marginal value of risk reductions does not decline with age as predicted by some lifetime
consumption models. This research supports findings in recent stated preference studies
that suggest only small reductions in the value of mortality risk reductions with
increasing age (e.g., Alberini et al., 2004).
5-66
-------
5.6.4.2 Hospital Admissions and Emergency Department Valuation
In the absence of estimates of societal WTP to avoid hospital visits/admissions for
specific illnesses, we derive COI estimates for use in the benefits analysis. The International
Classification of Diseases (ICD) (WHO, 1977) code-specific COI estimates used in this analysis
consist of estimated hospital charges and the estimated opportunity cost of time spent in the
hospital (based on the average length of a hospital stay for the illness). We based all estimates of
hospital charges and length of stays on statistics provided by the Agency for Healthcare Research
and Quality's Healthcare Utilization Project National Inpatient Sample (NIS) database (AHRQ,
2007). We estimated the opportunity cost of a day spent in the hospital as the value of the lost
daily wage, regardless of whether the hospitalized individual is in the workforce. To estimate the
lost daily wage, we divided the median weekly wage reported by the 2007 American Community
Survey (ACS) by five and deflated the result to the correct currency year using the CPI-U "all
items" (Abt Associates, 2012). The resulting national average lost daily wage is $150 (2011$).
The total cost-of-illness estimate for an ICD code-specific hospital stay lasting n days, then, was
the mean hospital charge plus daily lost wage multiplied by n. In general, the mean length of stay
has decreased since the 2000 database used in the previous version of BenMAP, while the mean
hospital charge has increased. We provide the rounded unit values in 2011$ for the COI
functions used in this analysis in Table 5-12.
Table 5-12. Unit Values for Hospital Admissions a
Age Range Mean Hospital
End Point
HA, Chronic Lung Disease
HA, Asthma
HA, All Cardiovascular
HA, All Cardiovascular
HA, All Respiratory
ICD Codes
490-496
493
390-429
390-429
460-519
min.
18
0
18
65
65
max.
64
64
64
99
99
Charge
(2011$)
$20,000
$15,000
$41,000
$38,000
$32,000
Mean
Length of
Stay (days)
3.9
3.0
4.1
4.9
6.1
Total Cost of
Illness (unit
value in 2011$)
$22,000
$16,000
$44,000
$42,000
$37,000
a All estimates rounded to two significant digits. Unrounded estimates in 2000$ are available in Appendix J of the
BenMAP user manual (Abt Associates, 2012).
To value asthma emergency department visits, we used a simple average of two estimates
from the health economics literature. The first estimate comes from Smith et al. (1997), who
reported approximately 1.2 million asthma-related emergency department visits in 1987, at a
total cost of $186 million (1987$). The average cost per visit that year was $155; in 2011$, that
cost was $480 (using the CPI-U for medical care to adjust to 2011$). The second estimate comes
5-67
-------
from Stanford et al. (1999), who reported the cost of an average asthma-related emergency
department visit based on 1996-1997 data at $400 (using the CPI-U for medical care to adjust to
2011$). A simple average of the two estimates yields a unit value of $440 (2011$).
5.6.4.3 Nonfatal Myocardial Infarctions Valuation
We were not able to identify a suitable WTP value for reductions in the risk of nonfatal
heart attacks.66 Instead, we use a COI unit value with two components: the direct medical costs
and the opportunity cost (lost earnings) associated with the illness event. Because the costs
associated with a myocardial infarction extend beyond the initial event itself, we consider costs
incurred over several years. Using age-specific annual lost earnings estimated by Cropper and
Krupnick (1990) and a 3% discount rate, we estimated a rounded present discounted value in lost
earnings (in 2000$) over 5 years due to a myocardial infarction of $8,800 for someone between
the ages of 25 and 44, $13,000 for someone between the ages of 45 and 54, and $75,000 for
someone between the ages of 55 and 65. The rounded corresponding age-specific estimates of
lost earnings (in 2000$) using a 7% discount rate are $7,900, $12,000, and $67,000, respectively.
Cropper and Krupnick (1990) do not provide lost earnings estimates for populations under 25 or
over 65. As such, we do not include lost earnings in the cost estimates for these age groups.
We found three possible sources in the literature of estimates of the direct medical costs
of myocardial infarction, which provide significantly different values (see Table 5-13):
• Wittels et al. (1990) estimated expected total medical costs of myocardial infarction over
5 years to be $51,000 (rounded in 1986$) for people who were admitted to the hospital
and survived hospitalization. (There does not appear to be any discounting used.) This
estimated cost is based on a medical cost model, which incorporated therapeutic options,
projected outcomes, and prices (using "knowledgeable cardiologists" as consultants). The
model used medical data and medical decision algorithms to estimate the probabilities of
certain events and/or medical procedures being used. The authors note that the average
length of hospitalization for acute myocardial infarction has decreased over time (from an
average of 12.9 days in 1980 to an average of 11 days in 1983). Wittels et al. used 10
days as the average in their study. It is unclear how much further the length of stay for
myocardial infarction may have decreased from 1983 to the present. The average length
66 We note that this endpoint was only modeled as part of the cobenefits analysis for PM2 5 and is not included in the
ozone-related benefits analysis. As such, we could have moved this discussion to Appendix 5D (as was done with
the discussion of PM2 5-related effect estimates). However, since this was the only broader discussion related to
valuation which is exclusively related to PM2 5 we left it in this section.
5-68
-------
of stay for ICD code 410 (myocardial infarction) in the year-2000 Agency for Healthcare
Research and Quality (AHRQ) HCUP database is 5.5 days (AHRQ, 2000). However, this
may include patients who died in the hospital (not included among our nonfatal
myocardial infarction cases), and whose length of stay was therefore substantially shorter
than it would be if they had not died.
• Eisenstein et al. (2001) estimated 10-year costs of $45,000 in rounded 1997$ (using a 3%
discount rate) for myocardial infarction patients, using statistical prediction (regression)
models to estimate inpatient costs. Only inpatient costs (physician fees and hospital costs)
were included.
Table 5-13. Alternative Direct Medical Cost of Illness Estimates for Nonfatal Heart
Attacks a
Study Direct Medical Costs (2011$) Over an x-Year Period, for x =
Wittelsetal. (1990) $ 170,000 b 5
Russell et al. (1998) $34,000 c 5
Average (5-year) costs $100,000 5
Eisenstein etal. (2001) $76,000c 10
a All estimates rounded to two significant digits. Unrounded estimates in 2000$ are available in appendix J of the
BenMAP user manual (Abt Associates, 2012).
b Wittels et al. (1990) did not appear to discount costs incurred in future years.
0 Using a 3% discount rate. Discounted values as reported in the study.
As noted above, the estimates from these three studies are substantially different, and we
have not adequately resolved the sources of differences in the estimates. Because the wage-
related opportunity cost estimates from Cropper and Krupnick (1990) cover a 5-year period, we
used estimates for medical costs that similarly cover a 5-year period (i.e., estimates from Wittels
et al. (1990) and Russell et al. (1998). We used a simple average of the two 5-year estimates, or
rounded to $85,000, and added it to the 5-year opportunity cost estimate. The resulting estimates
are given in Table 5-14.
Table 5-14. Estimated Costs Over a 5-Year Period of a Nonfatal Myocardial Infarction
(in 2011$) a
Age Group Opportunity Cost Medical Cost b Total Cost
0-24 $0 $100,000 $100,000
25-44 $12,000c $100,000 $110,000
45-54 $18,000c $100,000 $120,000
55-65 $100,000c $100,000 $210,000
>65 $0 $100,000 $100,000
a All estimates rounded to two significant digits, so estimates may not sum across columns. Unrounded estimates in
2000$ are available in appendix J of the BenMAP user manual (Abt Associates, 2012).
b An average of the 5-year costs estimated by Wittels et al. (1990) and Russell et al. (1998).
0 From Cropper and Krupnick (1990), using a 3% discount rate for illustration.
5-69
-------
5.6.4.4 Valuation of Acute Health Events
Asthma exacerbation. Several respiratory symptoms in asthmatics or characterizations of an
asthma episode have been associated with exposure to air pollutants. All of these can generally
be taken as indications of an asthma exacerbation when they occur in an asthmatic. Therefore,
we apply the same set of unit values for all of the variations of "asthma exacerbation".
Specifically, we use a unit value based on the mean WTP estimates for a "bad asthma day,"
described in Rowe and Chestnut (1986). This study surveyed asthmatics to estimate WTP for
avoidance of a "bad asthma day," as defined by the subjects.
Minor Restricted Activity Days Valuation. No studies are reported to have estimated WTP to
avoid a minor restricted activity day. However, Neumann et al. (1994) derived an estimate of
willingness to pay to avoid a minor respiratory restricted activity day, using estimates from
Tolley et al. (1986) of WTP for avoiding a combination of coughing, throat congestion and
sinusitis. This estimate of WTP to avoid a minor respiratory restricted activity day is $38
(1990$), or about $71 (2011$). Although Ostro and Rothschild (1989) statistically linked ozone
and minor restricted activity days, it is likely that most MRADs associated with ozone and PIVh.s
exposure are, in fact, minor respiratory restricted activity days. For the purpose of valuing this
health endpoint, we used the estimate of mean WTP to avoid a minor respiratory restricted
activity day.
School Loss Days Valuation. To value a school absence, we: (1) estimated the probability that
if a school child stays home from school, a parent will have to stay home from work to care for
the child; and (2) valued the lost productivity at the parent's wage. To do this, we estimated the
number of families with school-age children in which both parents work, and we valued a
school-loss day as the probability that such a day also would result in a work-loss day. We
calculated this value by multiplying the proportion of households with school-age children by a
measure of lost wages.
We used this method in the absence of a preferable WTP method. However, this
approach suffers from several uncertainties. First, it omits willingness to pay to avoid the
symptoms/illness that resulted in the school absence; second, it effectively gives zero value to
school absences that do not result in work-loss days; and third, it uses conservative assumptions
5-70
-------
about the wages of the parent staying home with the child. Finally, this method assumes that
parents are unable to work from home. If this is not a valid assumption, then there would be no
lost wages.
For this valuation approach, we assumed that in a household with two working parents,
the female parent will stay home with a sick child. From the Statistical Abstract of the United
States (U.S. Census Bureau, 2001), we obtained: (1) the numbers of single, married and "other"
(widowed, divorced or separated) working women with children; and (2) the rates of
participation in the workforce of single, married and "other" women with children. From these
two sets of statistics, we calculated a weighted average participation rate of 72.85 percent. Our
estimate of daily lost wage (wages lost if a mother must stay at home with a sick child) is based
on the year 2000 median weekly wage among women ages 25 and older (U.S. Census Bureau,
2001). This median weekly wage is $551 (2000$). Dividing by five gives an estimated median
daily wage of $103 (2000$). To estimate the expected lost wages on a day when a mother has to
stay home with a school-age child, we first estimated the probability that the mother is in the
workforce then multiplied that estimate by the daily wage she would lose by missing a workday:
72.85 percent times $103, for a total loss of $75 (2000$). This valuation approach is similar to
that used by Hall et al. (2003).
Work Loss Days Valuation. Work loss days are valued at a day's wage. BenMAP-CE
calculates county-specific median daily wages from county-specific annual wages (by dividing
the annual wage by 52 weeks multiplied by 5 work days per week), on the theory that a worker's
vacation days are valued at the same daily rate as work days.
Upper and Lower respiratory symptoms. Lower and upper respiratory symptoms are each
considered a complex of symptoms. A dollar value was derived for clusters of these symptoms
that most closely match the studies used to calculate incidence (Schwartz and Neas, 2000; Pope
et al, 1991) based on mid-range estimates from each cluster (lEc, 1994).
5.6.4.5 Growth in WTP Reflecting National Income Growth over Time
Our analysis accounts for expected growth in real income over time. This is a distinct
concept from inflation and currency year. Economic theory argues that WTP for most goods
(such as environmental protection) will increase if real incomes increase. There is substantial
5-71
-------
empirical evidence that the income elasticity67 of WTP for health risk reductions is positive,
although there is uncertainty about its exact value. Thus, as real income increases, the WTP for
environmental improvements also increases. Although many analyses assume that the income
elasticity of WTP is unit elastic (i.e., a 10% higher real income level implies a 10% higher WTP
to reduce risk changes), empirical evidence suggests that income elasticity is substantially less
than one and thus relatively inelastic. As real income rises, the WTP value also rises but at a
slower rate than real income.
The effects of real income changes on WTP estimates can influence benefits estimates in
two different ways: through real (national average) income growth between the year a WTP
study was conducted and the year for which benefits are estimated, and through differences in
income between study populations and the affected populations at a particular time. The SAB-
EEAC advised the EPA to adjust WTP for increases in real income over time but not to adjust
WTP to account for cross-sectional income differences "because of the sensitivity of making
such distinctions, and because of insufficient evidence available at present" (U.S. EPA-SAB,
2000). An advisory by another committee associated with the SAB, the Advisory Council on
Clean Air Compliance Analysis (SAB-Council), has provided conflicting advice. While agreeing
with "the general principle that the willingness to pay to reduce mortality risks is likely to
increase with growth in real income" and that "[t]he same increase should be assumed for the
WTP for serious nonfatal health effects," they note that "given the limitations and uncertainties
in the available empirical evidence, the Council does not support the use of the proposed
adjustments for aggregate income growth as part of the primary analysis" (U.S. EPA-SAB,
2004b). Until these conflicting advisories can be reconciled, the EPA will continue to adjust
valuation estimates to reflect income growth using the methods described below, while providing
sensitivity analyses for alternative income growth adjustment factors.
Based on a review of the available income elasticity literature, we adjusted the valuation
of human health benefits upward to account for projected growth in real U.S. income. Faced with
a dearth of estimates of income elasticities derived from time-series studies, we applied estimates
derived from cross-sectional studies in our analysis. Details of the procedure can be found in
67 Income elasticity is a common economic measure equal to the percentage change in WTP for a 1% change in
income.
5-72
-------
Kleckner and Neumann (1999). We note that the literature has evolved since the publication of
this memo and that an array of newer studies identifying potentially suitable income elasticity
estimates are available (lEc, 2012). The EPA anticipates seeking an SAB review of these studies,
and its approach to adjusting WTP estimates to account for changes in personal income, in the
near future. As such, these newer studies have not yet been incorporated into the benefits
analysis. An abbreviated description of the procedure we used to account for WTP for real
income growth between 1990 and 2024 is presented below. Income growth projections are only
currently available in BenMAP through 2024, so both the 2025 and 2038 estimates use income
growth only through 2024 and are therefore likely underestimates.
Reported income elasticities suggest that the severity of a health effect is a primary
determinant of the strength of the relationship between changes in real income and WTP. As
such, we use different elasticity estimates to adjust the WTP for minor health effects, severe and
chronic health effects, and premature mortality. Note that because of the variety of empirical
sources used in deriving the income elasticities, there may appear to be inconsistencies in the
magnitudes of the income elasticities relative to the severity of the effects (a priori one might
expect that more severe outcomes would show less income elasticity of WTP). We have not
imposed any additional restrictions on the empirical estimates of income elasticity. One
explanation for the seeming inconsistency is the difference in timing of conditions. WTP for
minor illnesses is often expressed as a short-term payment to avoid a single episode. WTP for
major illnesses and mortality risk reductions are based on longer-term measures of payment
(such as wages or annual income). Economic theory suggests that relationships become more
elastic as the length of time grows, reflecting the ability to adjust spending over a longer time
period (U.S. EPA, 2010e, p. A-9). Based on this theory, it would be expected that WTP for
reducing long-term risks would be more elastic than WTP for reducing short-term risks. The
relative magnitude of the income elasticity of WTP for visibility compared with those for health
effects suggests that visibility is not as much of a necessity as health, thus, WTP is more elastic
with respect to income. The elasticity values used to adjust estimates of benefits in 2024 are
presented in Table 5-15.68
68
We expect that the WTP for improved visibility in Class 1 areas would also increase with growth in real income.
5-73
-------
Table 5-15. Elasticity Values Used to Account for Projected Real Income Growth a
Benefit Category Central Elasticity Estimate
Minor Health Effect 0.14
Severe and Chronic Health Effects 0.45
Premature Mortality 0.40
a Derivation of estimates can be found in Kleckner and Neumann (1999). COI estimates are not adjusted for income
growth.
In addition to elasticity estimates, projections of real gross domestic product (GDP) and
populations from 1990 to 2024 are needed to adjust benefits to reflect real per capita income
growth. For consistency with the emissions and benefits modeling, we used national population
estimates for the years 1990 to 1999 based on U.S. Census Bureau estimates (Hollman, Mulder,
and Kalian, 2000). These population estimates are based on application of a cohort-component
model applied to 1990 U.S. Census data projections (U.S. Bureau of Census, 2000). For the
years between 2000 and 2024, we applied growth rates based on the U.S. Census Bureau
projections to the U.S. Census estimate of national population in 2000. We used projections of
real GDP provided in Kleckner and Neumann (1999) for the years 1990 to 2010.69 We used
projections of real GDP (in chained 1996 dollars) provided by Standard and Poor's (2000) for
the years 2010 to 2024.70
Using the method outlined in Kleckner and Neumann (1999) and the population and
income data described above, we calculated WTP adjustment factors for each of the elasticity
estimates listed in Table 5-16. Benefits for each of the categories (minor health effects, severe
and chronic health effects, premature mortality, and visibility) are adjusted by multiplying the
unadjusted benefits by the appropriate adjustment factor. For premature mortality, we applied the
income adjustment factor specific to the analysis year, but we do not adjust for income growth
over the 20-year cessation lag. Our approach could underestimate the benefits for the later years
of the lag.
69 U.S. Bureau of Economic Analysis, Table 2A—Real Gross Domestic Product (1997) and U.S. Bureau of
Economic Analysis, The Economic and Budget Outlook: An Update, Table 4—Economic Projections for
Calendar Years 1997 Through 2007 (1997). Note that projections for 2007 to 2010 are based on average GDP
growth rates between 1999 and 2007.
70 In previous analyses, we used the Standard and Poor's projections of GDP directly. This led to an apparent
discontinuity in the adjustment factors between 2010 and 2011. We refined the method by applying the relative
growth rates for GDP derived from the Standard and Poor's projections to the 2010 projected GDP based on the
Bureau of Economic Analysis projections.
5-74
-------
There is some uncertainty regarding the total costs of illness in the future. Specifically,
the nature of medical treatment is changing, including a shift towards more outpatient treatment.
Although we adjust the COI estimates for inflation, we do not have data to project COI estimates
for the cost of treatment in the future or income growth over time, which leads to an inherent
though unavoidable inconsistency between COI- and WTP-based estimates. This approach may
under predict benefits in future years because it is likely that increases in real U.S. income would
also result in increased COI (due, for example, to increases in wages paid to medical workers)
and increased cost of work loss days and lost worker productivity (reflecting that if worker
incomes are higher, the losses resulting from reduced worker production would also be higher).
In addition, cost-of-illness estimates do not include sequelae costs or pain and suffering, the
value of which would likely increase in the future. To the extent that costs would be expected to
increase over time, this increase may be partially offset by advancement in medical technology
that improves the effectiveness of treatment at lower costs. For these reasons, we believe that the
cost-of-illness estimates in this RIA may underestimate (on net) the total economic value of
avoided health impacts.
Table 5-16. Adjustment Factors Used to Account for Projected Real Income Growth a
Benefit Category 2024
Minor Health Effect 1.07
Severe and Chronic Health Effects 1.22
Premature Mortality 1.20
a Based on elasticity values reported in Table 5-15, U.S. Census population projections, and projections of real GDP
per capita.
5.7 Benefits Results
As stated in section 5.1 and described in detail in section 5.4.3, we have estimated
nationwide benefits for 2025 associated with attainment of alternative ozone standards across the
U.S. with the exception of California. We have also estimated the nationwide benefits of
attaining in California for 2038. Because of the temporal disconnect between these two
scenarios, benefit estimates for each are not totaled and instead, are presented separately (section
5.7.1 for 2025 and section 5.7.2 for post-2025).
In addition to these core incidence and benefits estimates, we also present a number of
additional analyses which are intended to inform interpretation of these core benefit estimates
(see section 5.7.3). In completing these additional analyses, in many cases, we did not have to
5-75
-------
generate separate assessments for both scenarios since observations from one scenario could be
readily applied to the other (in these cases, we tended to model the 2025 scenario and then
discuss application of those observations to the post-2025 scenario).
5.7.1 Benefits of the Proposed and Alternative Annual Primary Ozone Standards for the 2025
Scenario
This section presents incidence reductions and associated dollar benefit estimates
associated with the 2025 scenario (i.e., every state apart from California - see section 5.4.3).
Applying the impact and valuation functions described previously in this chapter to the estimated
changes in ozone yields estimates of the changes in physical damages (e.g., premature
mortalities, cases of hospital admissions) and the associated monetary values for those changes.
Similarly applying the incidence per ton and dollar per ton values to the estimates of NOx
reductions produces estimates of changes in PM-related health effect incidence and associated
dollar benefits. Not all known ozone and PM health effects could be quantified or monetized.
The monetized value of these unquantified effects is represented by adding an unknown "B" to
the aggregate total. The estimate of total monetized health benefits is thus equal to the subset of
monetized ozone and PM-related health benefits plus B, the sum of the non-monetized health
benefits and welfare co-benefits; this B represents both uncertainty and a bias in this analysis, as
it reflects those benefits categories that we are unable to monetize in this analysis.
We follow our standard rounding conventions in presenting these benefits results. After
reviewing the presentation of EPA's benefits results, NRC (2002) concluded, "EPA should strive
to present the results of the analyses in ways that avoid conveying an unwarranted degree of
certainty. Such ways include rounding to few significant digits, increasing the use of graphs, and
placing less emphasis on single numbers and greater emphasis on ranges" (p. 161). Following
this advice, we round all benefits estimates to two significant digits, and all rounding occurs after
final summing of unrounded estimates. As such, totals may not sum across columns or rows. In
addition, all incidence estimates are rounded to whole numbers with a maximum of two
significant digits.
Table 5-17 shows the population-weighted air quality change for the alternative standards
averaged across the continental U.S. Table 5-18 summarizes the tons of VOC and NOx
5-76
-------
emissions required to simulate attainment of each alternative standard (further differentiated by
geographic region including east, west and California). Tables 5-19 through 5-23 present the
benefits results for the proposed and alternative ozone standards. Table 5-24 summarizes total
benefits by geographic region (including east, west minus California, and California). Note that
in presenting estimates related to reductions in ozone (Tables 5-19 and 5-20), we include the full
set of core estimates together with a subset of sensitivity analysis results (specifically, alternative
estimates for both short-term and long-term mortality). The benefit estimates presented are
relative to a 2025 analytical baseline reflecting attainment nationwide (excluding California) of
the current primary ozone standards (i.e., 75 ppb) that includes promulgated national regulations
and illustrative emissions controls to simulate attainment with 75 ppb.
Table 5-17. Population-Weighted Air Quality Change for the Proposed and Alternative
Annual Primary Ozone Standards Relative to Analytical Baseline for 2025a
Population-Weighted Summer Season
Standard Ozone Concentration Change (8hr max)b
70 ppb 0.5285
65 ppb 1.6317
60 ppb 3.0222
a Because we used benefit-per-ton estimates for the PM2s co-benefits, population-weighted PM25 changes are not
available.
b Population weighting based on all ages (demographic used in modeling short-term exposure-related mortality for
ozone) for 2025.
Table 5-18. Emission Reductions in Illustrative Emission Reduction Strategies for the
Proposed and Alternative Annual Primary Ozone Standards, by Pollutant and
Region Relative to Analytical Baseline - Full Attainment (tons)"
70 ppb
65 ppb
60 ppb
NOx
East
West
CA
600
0
53
1,700
110
110
2,800
500
140
VOC
East
West
CA
55
0
0
99
7
0
150
7
0
a See Chapter 4 for more information on the illustrative emission reduction strategies. The emissions in this table
reflect both known and unknown controls. Several recent rules such as Tier 3 will have substantially reduced ozone
concentrations by 2025 in the East, thus few additional controls would be needed to reach 70 ppb.
5-77
-------
Table 5-19. Estimated Number of Avoided Ozone-Only Health Impacts for the Proposed
and Alternative Annual Ozone Standards (Incremental to the Analytical
Baseline) for the 2025 Scenario (nationwide benefits of attaining each alternative
standard everywhere in the U.S. except California) a'b
Proposed and Alterative Standards
(95th percentile confidence intervals)
Health Effectb 70 ppb
65 ppb
60 ppb
Avoided Short-Term Mortality - Core Analysis
200
, . . Smith et al. (2009) (all ages)
multi-city v 'v & ' (97 to 300)
studies Zanobetti and Schwartz (2008) (all 340
ages) (180 to 490)
Avoided Long-term Respiratory Mortality - Core Analysis
multi-city Jerrett etal. (2009) (30-99y is) 680
study copollutants model (PM2.5) (230 to 1,100)
Avoided Short-Term Mortality - Sensitivity Analysis
Smith et al. (2009) (all ages) 160
copollutants model (PMio) (-44 to 360)
250
,. . Schwartz (2005) (all ages)
multi-city v ' & ' (77 to 420)
studies Huang et al. (2005) 240
(cardiopulmonary) (88 to 380)
160
Bell etal. (2004) (all ages) (54to2JO)
520
Bell et al. (2005) (all ages)
meta- 730
, Ito etal. (2005) (all ages)
analyses v ;v & ' (440 to 1,000)
740
Levy et al. (2005) (all ages)
Avoided Long-term Respiratory Mortality - Sensitivity Analysis
Jerrett et al. (2009) (age 30-99) (86 460
cities) (ozone-only) (130 to 790)
Jerrett et al. (2009) (age 30-99) (96 500
cites) (ozone-only) (170 to 810)
, . . Jerrett et al (2009) - copollutant (PM2.5 ) model with: °
stud 60 ppb threshold 520
56 ppb threshold 410
55 ppb threshold 120
50 ppb threhsold 6
45 ppb threshold 3
40 ppb threshold <1
Avoided Morbidity - Core Analysis
Hospital admissions - respiratory 360
(age 65+) (-97 to 820)
Emergency department visits for 1,100
asthma (all ages) (100 to 3,400)
, • , 300,000
Asthma exacerbation (age 6-18)
v B ' (-440,000 to 900,000)
Minor restricted-activity days (age 930,000
18-65) (380,000 to 1,500,000)
330 000
School Loss Days (age 5-17) (120,000to 730,000)
630
(310 to 940)
1,000
(560 to 1,500)
2,100
(710 to 3,500)
500
(-140 to 1,100)
780
(240 to 1,300)
740
(280 to 1,200)
510
(170 to 850)
1,600
(780 to 2,500)
2,300
(1,400 to 3,200)
2,300
(1,600 to 3,000)
1,400
(410 to 2,500)
1,500
(550 to 2,500)
1,600
1,100
280
58
47
10
1,100
(-310 to 2,600)
3,500
(330 to 11,000)
1,100
(560 to 1,700)
1,900
(1,000 to 2,800)
3,900
(1,300 to 6,400)
920
(-250 to 2, 100)
1,400
(440 to 2,400)
1,400
(510 to 2,200)
930
(310 to 1,600)
3,000
(1,400 to 4,600)
4,200
(2,500 to 5,800)
4,200
(2,900 to 5,600)
2,600
(760 to 4,500)
2,800
(1,000 to 4,600)
2,900
1,700
510
140
105
13
2,100
(-560 to 4,700)
6,600
(610 to 20,000)
910,000 1,700,000
(-l,300,000to 2,700,000) (-2,500,000 to 5,000,000)
2,900,000
(1,200,000 to 4,500,000)
1,000,000
(360,000 to 2,200,000)
5,300,000
(2,200,000 to 8,300,000)
1,900,000
(660,000 to 4,500,000)
a All incidence estimates are rounded to whole numbers with a maximum of two significant digits.
b All incidence estimates are based on ozone-only models unless otherwise noted.
0 See Appendix 5B, section 5B. 1 for additional detail on the threshold-based sensitivity analysis.
5-78
-------
Table 5-20. Total Monetized Ozone-Only Benefits for the Proposed and Alternative
Annual Ozone Standards (Incremental to the Analytical Baseline) for the 2025
Scenario (nationwide benefits of attaining each alternative standard everywhere
in the U.S. except California) (millions of 2011) a'b
Proposed and Alterative Standards
(95th percentile confidence intervals)
Health Effectb
70ppb
65ppb
60ppb
Avoided Short-Term Mortality - Core Analysis
multi-city
studies
Smith et al. (2009) (all ages)
Zanobetti and Schwartz (2008) (all
ages)
$2,000
($180 to $5,800)
3,400
($300 to $9,600)
$6,400
($560 to $18,000)
11,000
($950 to $30,000)
$12,000
($1,000 to $33,000)
20,000
($1,700 to $55,000)
Avoided Short-Term Mortality - Sensitivity Analysis
multi-city
studies
meta-
analyses
Smith et al. (2009) (all ages)
copollutants model (PMio)
Schwartz (2005) (all ages)
Huang et al. (2005)
(cardiopulmonary)
Bell etal. (2004) (all ages)
Bell etal. (2005) (all ages)
Ito etal. (2005) (all ages)
Levy etal. (2005) (all ages)
Avoided Long-term Respiratory Mortality - Sensitivity
multi-city
study
Jerrett et al. (2009) (age 30-99)
copollutants model(PM2.5)no lag °
Jerrett et al. (2009) (age 30-99)
copollutants model (PM2.5) 20 yr
j
segmented lag
$1,600
(-$390 to $5,900)
$2,500
($200 to $7,700)
$2,400
($200 to $7,000)
1,700
($130 to $4,900)
5,300
($470 to $15,000)
$7,400
($680 to $2 1,000)
$7,500
($700 to $20,000)
Analysis
$6,900
($560 to $2 1,000)
$5,600 to $6300
($460 to $19,000)
$5,100
(-$1,200 to $19,000)
$7,900
($630 to $24,000)
$7,500
($620 to $22,000)
5,200
($420 to $15,000)
17,000
($1,500 to $48,000)
$23,000
($2, 100 to $64,000)
$24,000
($2,200 to $64,000)
$22,000
($1,700 to $64,000)
$18,000 to $20,000
($1,400 to $58,000)
$9,400
(-$2,200 to $34,000)
$15,000
($1,200 to $44,000)
$14,000
($1,100 to $41,000)
9,500
($760 to $28,000)
31,000
($2,700 to $88,000)
$42,000
($3,900 to $120,000)
$43,000
($4,000 to $120,000)
$40,000
($3,200 to $120,000)
$32,000 to $36,000
($2,600 to $110,000)
a All benefits estimates are rounded to whole numbers with a maximum of two significant digits. The monetized
value of the ozone-related morbidity benefits are included in the estimates shown in this table for each mortality
study (and when combined account for from 4-6% of the total benefits, depending on the total mortality estimate
compared against. Note that asthma exacerbations accounts for «1% of the total).
b The sensitivity analysis for long-term exposure-related mortality included an assessment of potential thresholds
however this assessment was implemented using incidence estimates (see Table 5-19). Observations from that
analysis (in terms of fractional impacts on incidence can be directly applied to these benefit results) (see Appendix
5B, section 5B.1)
0 A single central-tendency value is provided in each cell, since the zero-lag model used here did not require
application of a 3% and 7% discount rates (see footnote d below) (note, however that as with all other entries in this
study, we do include a 95th percentile confidence interval range - these are the values within parentheses).
dThe range (outside of the parentheses) within each cell results from application of a 7% and 3% discount rates in
the context of applying the 20year segmented lag (with the 7% resulting in the lower estimate and the 3% the higher
estimate). The range presented within the parentheses reflects consideration for the 95*% confidence interval
generated for each of these estimates and ranges from a low value (2.5*% CI for the 7% discount-based dollar
benefit) to an upper value (97.5*% of the 3% discount-based dollar benefit).
5-79
-------
Table 5-21. Estimated Number of Avoided PM2.s-Related Health Impacts for the
Proposed and Alternative Annual Ozone Standards (Incremental to the
Analytical Baseline) for the 2025 Scenario (nationwide benefits of attaining each
alternative standard everywhere in the U.S. except California) a
Proposed and Alterative Standards
Health Effectb
70ppb
65ppb
60ppb
Avoided PM2.5-related Mortality
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
Woodruff et al. (1997) (infant mortality)
510
1,100
1
1,400
3,300
2
2,600
6,000
5
Avoided PM2.5-related Morbidity
Non-fatal heart attacks
Peters et al. (2001) (age >18)
Pooled estimate of 4 studies (age >18)
Hospital admissions — respiratory (all ages)
Hospital admissions — cardiovascular (age > 18)
Emergency department visits for asthma (all ages)
Acute bronchitis (ages 8-12)
Lower respiratory symptoms (ages 7-14)
Upper respiratory symptoms (asthmatics ages 9-11)
Asthma exacerbation (asthmatics ages 6-18)
Lost work days (ages 18-65)
Minor restricted-activity days (ages 18-65)
600
64
150
180
280
790
10,000
14,000
17,000
65,000
380,000
1,700
180
430
530
790
2,300
29,000
41,000
51,000
180,000
1,100,000
3,100
330
780
950
1,400
4,100
53,000
75,000
100,000
340,000
2,000,000
a All incidence estimates are rounded to whole numbers with a maximum of two significant digits. Because these
estimates were generated using benefit-per-ton estimates, confidence intervals are not available. In general, the 95th
percentile confidence interval for the health impact function alone ranges from approximately ±30 percent for
mortality incidence based on Krewski et al. (2009) and ±46 percent based on Lepeule et al. (2012).
Table 5-22. Monetized PM2.s-Related Health Co-Benefits for the Proposed and
Alternative Annual Ozone Standards (Incremental to Analytical Baseline) for the 2025
5-80
-------
Scenario (nationwide benefits of attaining each alternative standard everywhere in the
U.S. except California) (Millions of 2011)a'b'c
Proposed and Alterative Standards
Monetized Benefits —
70 ppb 65 ppb 60 ppb
3% Discount Rate
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
$4,800
$11,000
$14,000
$31,000
$25,000
$56,000
7% Discount Rate
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
$4,300
$9,700
$12,000
$28,000
$22,000
$50,000
a All estimates are rounded to two significant digits. Because these estimates were generated using benefit-per-ton
estimates, confidence intervals are not available. In general, the 95th percentile confidence interval for monetized
PM2.5 benefits ranges from approximately -90 percent to +180 percent of the central estimates based on Krewski et
al. (2009) and Lepeule et al. (2012). Estimates do not include unquantified health benefits noted in Table 5-2 or
Section 5.6.3.6 or welfare co-benefits noted in Chapter 6.
b The reduction in premature fatalities each year accounts for over 98% of total monetized benefits in this analysis.
Mortality risk valuation assumes discounting over the SAB-recommended 20-year segmented lag structure.
Table 5-23. Estimate of Monetized Ozone and PMi.s Benefits for Proposed and
Alternative Annual Ozone Standards Incremental to the Analytical Baseline for
the 2025 Scenario (nationwide benefits of attaining each alternative standard
everywhere in the U.S. except California) - Full Attainment (billions of 2011$) a
Ozone-only Benefits (range reflects
Smith et al., 2009 and Zanobetti and
Schwartz, 2008)
PMi.s Co-benefits (range reflects
Krewski et al., 2009 and Lepeule et
al., 2012)
Total Benefits
Discount
Rate
b
3%
7%
3%
7%
70 ppb
$2.0 to $3.4 +B
$4.8 to $11
$4.3 to $9.7
$6.9to$14+B
$6.4to$13+B
65 ppb
$6.4to$ll+B
$14 to $31
$12 to $28
$20to$41+B
$19 to $38 +B
60 ppb
$12 to $20 +B
$25 to $56
$22 to $50
$37 to $75 +B
$34 to $70 +B
a Rounded to two significant figures. The reduction in premature fatalities each year accounts for over 98% of total
monetized benefits in this analysis. Mortality risk valuation for PM2 5 assumes discounting over the SAB-
recommended 20-year segmented lag structure. Not all possible benefits are quantified and monetized in this
analysis. B is the sum of all unquantified health and welfare co-benefits. Data limitations prevented us from
quantifying these endpoints, and as such, these benefits are inherently more uncertain than those benefits that we
were able to quantify. These estimates reflect the economic value of avoided morbidities and premature deaths using
risk coefficients from the studies noted.
b Ozone-only benefits reflect short-term exposure impacts and as such are assumed to occur in the same year as
ambient ozone reductions. Consequently, social discounting is not applied to the benefits for this category.
5-81
-------
Table 5-24. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
2025 Scenario (nationwide benefits of attaining each alternative standard
everywhere in the U.S. except California) - Full Attainment a
Proposed and Alterative Standards
70 ppb 65 ppb 60 ppb
East" 99% 96% 92%
California 0% 0% 0%
Rest of West 1% 4% 7%
a Because we use benefit-per-ton estimates to calculate the PM2 5 co-benefits, a regional breakdown for the co-
benefits is not available. Therefore, this table only reflects the ozone benefits.
b Includes Texas and those states to the north and east. Several recent rules such as Tier 3 will have substantially
reduced ozone concentrations by 2025 in the East, thus few additional controls would be needed to reach 70 ppb.
5.7.2 Benefits of the Proposed and Alternative Annual Primary Ozone Standards for the post-
2025 Scenario
This section presents incidence reductions and associated dollar benefit estimates
associated with the post-2025 scenario (i.e., nationwide benefits estimates reflecting attainment
of alternative standards in California - see section 5.4.3). The same rounding conventions
described in section 5.7.1 (for the 2025 estimates) were applied in generating these estimates. As
with estimates generated for the 2025 scenario, total monetized health benefits are equal to the
subset of monetized ozone and PM-related health benefits plus B, the sum of the non-monetized
health benefits and welfare co-benefits. Organization and general content of tables presented
here for the post-2025 scenario mirrors that of tables presented in section 5.7.1 for the 2025
scenario and the reader is referred there for further clarification.
Table 5-25. Population-Weighted Air Quality Change for the Proposed and Alternative
Annual Primary Ozone Standards Relative to Analytical Baseline for post-2025
Population-Weighted Ozone Season Ozone Concentration Change (8hr
Standard max)b
70 ppb 0.1526
65 ppb 0.3311
60 ppb 0.5007
a Because we used benefit-per-ton estimates for the PM25 co-benefits, population-weighted PM25 changes are not
available.
b Population weighting based on all ages (demographic used in modeling short-term exposure-related mortality for
ozone) for 2025.
5-82
-------
Table 5-26. Estimated Number of Avoided Ozone-Only Health Impacts for the Proposed
and Alternative Annual Ozone Standards (Incremental to the Analytical
Baseline) for the Post-2025 Scenario (nationwide benefits of attaining each
alternative standard just in California) a'b
Proposed and Alterative Standards
(95th percentile confidence intervals)
Health Effectb 70 ppb
65 ppb
60 ppb
Avoided Short-Term Mortality - Core Analysis
65
u. v Smith etal. (2009) (all ages)
multi-city v ^ B ' (31 to 97)
studies Zanobetti and Schwartz (2008) (all 110
ages) (57 to 160)
Avoided Long-term Respiratory Mortality - Core Analysis
multi-city Jerrett et al. (2009) (30-99yrs) 260
study copollutants model (PM2 5) (88 to 430)
Avoided Short-Term Mortality - Sensitivity Analysis
Smith etal. (2009) (all ages) 52
copollutants model (PMio) (-14 to 120)
80
u. v Schwartz (2005) (all ages)
multi-city \ i\ t i ps to 140)
studies Huang et al. (2005) 88
(cardiopulmonary) (33 to 140)
52
Bell et al. (2004) (all ages) ~
170
Bell etal. (2005) (all ages)
^ n B ; (80to250)
meta- 230
Ito etal. (2005) (all ages)
analyses v Jy B ' (140 to 330)
240
Levy et al. (2005) (all ages)
Avoided Long-term Respiratory Mortality - Sensitivity Analysis c
Jerrett et al. (2009) (age 30-99) (86 1 80
multi-city cities) (ozone-only) (51 to 300)
study Jerrett et al. (2009) (age 30-99) (96 1 90
cites) (ozone-only) (68 to 310)
Avoided Morbidity - Core Analysis
Hospital admissions -respiratory 120
(age 65+) (-32 to 270)
Emergency department visits for 320
asthma (all ages ) (29 to 980)
97000
Asthma exacerbation (age 6-18) '
Minor restricted-activity days (age 290,000
18-65) (120,000 to 460,000)
ov, IT T^ , ei-TN 110>000
SchoolLoss Davs (aae 5-17)
^iiooii^ udy !> (age ,) (38^OOQ to 24ao()0)
140
(68 to 210)
230
(120 to 340)
560
(190 to 930)
110
(-31 to 250)
170
(54 to 290)
190
(71 to 310)
110
(38 to 190)
360
(170 to 550)
510
(300 to 710)
510
(350 to 670)
380
(110 to 660)
410
(150 to 680)
260
(-69 to 580)
690
(64 to 2,100)
210,000
(-3 10,000 to 620,000)
630,000
(260,000 to 990,000)
230,000
(8 1,000 to 500,000)
210
(100 to 320)
350
(190 to 5 10)
840
(290 to 1,400)
170
(-46 to 380)
260
(81 to 440)
290
(110 to 470)
170
(57 to 280)
550
(260 to 830)
770
(460 to 1,100)
770
(530 to 1,000)
570
(160 to 980)
620
(220 to 1,000)
390
(-100 to 880)
1,000
(97 to 3,200)
310,000
(-460,000 to 930,000)
950,000
(390,000 to 1,500,000)
350,000
(120,000 to 830,000)
a All incidence estimates are rounded to whole numbers with a maximum of two significant digits.
b All incidence estimates are based on ozone-only models unless otherwise noted.
0 The sensitivity analysis for long-term exposure-related mortality included an assessment of potential thresholds,
which was completed for the 2025 scenario (see Table 5-19). Care should be taken in applying the results of that
sensitivity analysis to the post-2025 scenario, although general patterns of impact across the thresholds may apply.
5-83
-------
Table 5-27. Total Monetized Ozone-Only Benefits for the Proposed and Alternative
Annual Ozone Standards (Incremental to the Analytical Baseline) for the post-
2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) (millions of 2011) a'b
Health Effectb
70ppb
Proposed and Alter atiw Standards
(95th percentile confidence intervals)
65ppb
60ppb
Awided Short-Term Mortality - Core Analysis
multi-city
studies
Smith et al. (2009) (all ages)
Zanobetti and Schwartz (2008) (all
ages)
$660
($58 to $1,900)
1,100
($96 to $3,100)
$1,400
($130 to $4,100)
2,400
($210 to $6,700)
$2,100
($190 to $6,100)
3,600
($320 to $10,000)
Awided Short-Term Mortality - Sensitivity Analysis
multi-city
studies
meta-
analyses
Smith et al. (2009) (all ages)
copollutants model (PM10)
Schwartz (2005) (all ages)
Huang et al. (2005)
(cardiopulmonary)
Bell et al. (2004) (all ages)
Bell et al. (2005) (all ages)
Ito et al. (2005) (all ages)
Levy et al. (2005) (all ages)
$530
(-$120 to $1,900)
$820
($65 to $2,500)
$900
($74 to $2,600)
530
($43 to $1,600)
1,700
($150 to $4,900)
$2,400
($220 to $6,600)
$2,400
($220 to $6,500)
$1,100
(-$270 to $4,200)
$1,800
($140 to $5,400)
$1,900
($160 to $5,700)
1,200
($93 to $3,500)
3,700
($320 to $11,000)
$5,200
($470 to $14,000)
$5,200
($480 to $14,000)
$1,700
(-$410 to $6,300)
$2,700
($210 to $8,100)
$2,900
($240 to $8,600)
1,700
($140 to $5,200)
5,600
($490 to $16,000)
$7,800
($710 to $22,000)
$7,800
($730 to $21,000)
Awided Long-term Respiratory Mortality - Sensitivity Analysis
multi-city
study
Jerrett et al. (2009) (age 30-99)
copollutants model (PM2.5) no lag °
Jerrett et al. (2009) (age 30-99)
copollutants model (PM2.5) 20 yr
segmented las
$2,700
($220 to $7,900)
$2,200 to $2,400
($180 to $7,200)
$5,700
($470 to $17,000)
$4,700 to $5,200
($380 to $15,000)
$8,600
($700 to $25,000)
$7,000 to $7,800
($570 to $23,000)
a All benefits estimates are rounded to whole numbers with a maximum of two significant digits. The monetized
value of the ozone-related morbidity benefits are included in the estimates shown in this table for each mortality
study (and when combined account for from 4-6% of the total benefits, depending on the total mortality estimate
compared against. Note that asthma exacerbations accounts for «1% of the total).
b The sensitivity analysis for long-term exposure-related mortality included an assessment of potential thresholds
however this assessment was implemented using incidence estimates (see Table 5-19). Observations from that
analysis (in terms of fractional impacts on incidence can be directly applied to these benefit results) (see Appendix
5B, section 5B.1)
0 A single central-tendency value is provided in each cell, since the zero-lag model used here did not require
application of a 3% and 7% discount rates (see footnote d below) (note, however that as with all other entries in this
study, we do include a 95th percentile confidence interval range - these are the values within parentheses).
dThe range (outside of the parentheses) within each cell results from application of a 7% and 3% discount rates in
the context of applying the 20year segmented lag (with the 7% resulting in the lower estimate and the 3% the higher
estimate). The range presented within the parentheses reflects consideration for the 95*% confidence interval
generated for each of these estimates and ranges from a low value (2.5*% CI for the 7% discount-based dollar
benefit) to an upper value (97.5*% of the 3% discount-based dollar benefit).
5-84
-------
Table 5-28. Estimated Number of Avoided PM2.s-Related Health Impacts for the
Proposed and Alternative Annual Ozone Standards (Incremental to the
Analytical Baseline) for the post-2025 Scenario (nationwide benefits of attaining
each alternative standard just in California) a
Proposed and Alterative Standards
Health Effectb
70ppb
65ppb
60ppb
Avoided PM2.5-related Mortality
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
Woodruff et al. (1997) (infant mortality)
45
100
<1
89
200
<1
120
280
<1
Avoided PM2.5-related Morbidity
Non-fatal heart attacks
Peters et al. (2001) (age >18)
Pooled estimate of 4 studies (age >18)
Hospital admissions — respiratory (all ages)
Hospital admissions — cardiovascular (age > 18)
Emergency department visits for asthma (all ages)
Acute bronchitis (ages 8-12)
Lower respiratory symptoms (ages 7-14)
Upper respiratory symptoms (asthmatics ages 9-11)
Asthma exacerbation (asthmatics ages 6-18)
Lost work days (ages 18-65)
Minor restricted-activity days (ages 18-65)
54
6
14
16
24
67
860
1,200
1,900
5,500
32,000
110
11
27
32
46
130
1,700
2,400
3,800
11,000
64,000
140
16
37
45
64
180
2,300
3,300
5,200
15,000
88,000
a All incidence estimates are rounded to whole numbers with a maximum of two significant digits. Because these
estimates were generated using benefit-per-ton estimates, confidence intervals are not available. In general, the 95th
percentile confidence interval for the health impact function alone ranges from approximately ±30 percent for
mortality incidence based on Krewski et al. (2009) and ±46 percent based on Lepeule et al. (2012).
5-85
-------
Table 5-29. Monetized PM2.s-Related Health Co-Benefits for the Proposed and
Alternative Annual Ozone Standards (Incremental to Analytical Baseline) for the post-
2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) (Millions of 2011) a'b
Proposed and Alterative Standards
Monetized Benefits
70 ppb 65 ppb 60 ppb
3% Discount Rate
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
7% Discount Rate
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
$420
$950
$380
$860
$830
$1,900
$750
$1,700
$1,100
$2,600
$1,000
$2,300
a All estimates are rounded to two significant digits. Because these estimates were generated using benefit-per-ton
estimates, confidence intervals are not available. In general, the 95th percentile confidence interval for monetized
PM2.5 benefits ranges from approximately -90 percent to +180 percent of the central estimates based on Krewski et
al. (2009) and Lepeule et al. (2012). Estimates do not include unquantified health benefits noted in Table 5-3 or
Section 5.6.3.6 or welfare co-benefits noted in Chapter 6.
b The reduction in premature fatalities each year accounts for over 98% of total monetized benefits in this analysis.
Mortality risk valuation assumes discounting over the SAB-recommended 20-year segmented lag structure.
Table 5-30. Estimate of Monetized Ozone and PM2.s Benefits for Proposed and Alternative
Annual Ozone Standards Incremental to the Analytical Baseline for the post-
2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) - Full Attainment (billions of 2011$) a
Ozone-only Benefits (range reflects
Smith et al., 2009 and Zanobetti and
Schwartz, 2008)
PMi.s Co-benefits (range reflects
Krewski et al., 2009 and Lepeule et
al., 2012)
Total Benefits
Discount
Rate
b
3%
7%
3%
7%
70 ppb
$0.66 to $1.1
$0.42 to $0.95
$0.38 to $0.86
$l.lto$2.0+B
$l.lto$2.0+B
65 ppb
$1.4 to $2.4
$0.83 to $1.9
$0.75 to $1.7
$2.3 to $4.2 +B
$2.2to$4.1+B
60 ppb
$2.1 to $3. 6
$1.1 to $2.6
$1.0 to $2.3
$3. 4 to $6.2 +B
$3.2 to $5.9 +B
a Rounded to two significant figures. The reduction in premature fatalities each year accounts for over 98% of total
monetized benefits in this analysis. Mortality risk valuation for PM2 5 assumes discounting over the SAB-
recommended 20-year segmented lag structure. Not all possible benefits are quantified and monetized in this
analysis. B is the sum of all unquantified health and welfare co-benefits. Data limitations prevented us from
quantifying these endpoints, and as such, these benefits are inherently more uncertain than those benefits that we
were able to quantify. These estimates reflect the economic value of avoided morbidities and premature deaths using
risk coefficients from the studies noted.
b Ozone-only benefits reflect short-term exposure impacts and as such are assumed to occur in the same year as
ambient ozone reductions. Consequently, social discounting is not applied to the benefits for this category.
5-86
-------
Table 5-31. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the
post-2025 Scenario (nationwide benefits of attaining each alternative standard
just in California) - Full Attainment a
Proposed and Alterative Standards
70 ppb 65 ppb 60 ppb
East 0% 0% 0%
California 93% 94% 94%
Rest of West 6% 6% 6%
a Because we use benefit-per-ton estimates to calculate the PM2 5 co-benefits, a regional breakdown for the co-
benefits is not available. Therefore, this table only reflects the ozone benefits.
5.7.3 Uncertainty in Benefits Results (including Discussion of Sensitivity Analyses and
Supplemental Analyses)
The dollar value of avoided ozone and PIVh.s related premature deaths account for 94% to
98% of the total monetized benefits. This is true in part because we are unable to quantify many
categories of benefits. The next largest benefit is for reducing the incidence of nonfatal heart
attacks. The remaining categories each account for a small percentage of total benefit; however,
they represent a large number of avoided incidences affecting many individuals. Comparing an
incidence table to the monetary benefits table reveals that the number of incidences avoided and
the dollar value for that endpoint do not always closely correspond. For example, for ozone we
estimate almost 1,000 times more asthma exacerbations would be avoided than premature
mortalities, yet asthma exacerbations account for only a very small fraction («1%) of total
monetized benefits (see Tables 5-20). This reflects the fact that many of the less severe health
effects, while more common, are valued at a lower level than the more severe health effects.
Also, some effects, such as hospital admissions, are valued using a proxy measure of WTP. As
such, the true value of these effects may be higher than that reported in the tables above.
Sources of uncertainty associated with both the modeling of ozone-related benefits and
PM2.s-related cobenefits are discussed qualitatively in Appendix A. Key assumptions and
uncertainties related to the modeling of ozone are presented below.
• We assume that short-term exposure to ozone is associated with mortality and that this
relationship holds across the full range of exposure. Furthermore, we assume that long-
term exposure to ozone is associated with respiratory mortality and while we favor a no-
threshold linear C-R function, we consider the potential impact of thresholds ranging
from 40 ppb to 60 ppb.
5-87
-------
• In modeling long-term exposure-related respiratory mortality, we acknowledge
uncertainty in specifying the nature of the cessation lag and have consequently included
two alternative lag structures including a 20 year segmented lag and a zero lag (the zero
lag reflects the potential that estimates of long-term exposure-related mortality could
actually be capturing the accumulation of short-term exposure-related mortality events).
In addition, we acknowledge the value in exploring the impact of potential thresholds in
effect on the core incidence and benefits estimates.
PM2.5 mortality co-benefits represent a substantial proportion of total monetized benefits
(over 98% of the co-benefits), and these estimates have the following key assumptions and
uncertainties.
• We assume that all fine particles, regardless of their chemical composition, are equally
potent in causing premature mortality. This is an important assumption, because PIVh.s
produced varies considerably in composition across sources, but the scientific evidence is
not yet sufficient to allow differential effects estimates by particle type. The PM ISA,
which was twice reviewed by SAB-CASAC, concluded that "many constituents of PIVb.s
can be linked with multiple health effects, and the evidence is not yet sufficient to allow
differentiation of those constituents or sources that are more closely related to specific
outcomes" (U.S. EPA, 2009b).
• We assume that the health impact function for fine particles is log-linear without a
threshold in this analysis. Thus, the estimates include health benefits from reducing fine
particles in areas with varied concentrations of PIVh.s, including both areas that do not
meet the fine particle standard and those areas that are in attainment, down to the lowest
modeled concentrations.
• We assume that there is a "cessation" lag between the change in PM exposures and the
total realization of changes in mortality effects. Specifically, we assume that some of the
incidences of premature mortality related to PIVh.s exposures occur in a distributed
fashion over the 20 years following exposure based on the advice of the SAB-HES (U.S.
EPA-SAB, 2004c), which affects the valuation of mortality benefits at different discount
rates.
• We recognize uncertainty associated with application of the benefit-per-ton approach
used in modeling PIVh.s cobenefits. The benefit-per-ton estimates used here reflect
specific geographic patterns of emissions reductions and specific air quality and benefits
modeling assumptions associated with the derivation of those estimates (see the TSD
describing the calculation of the national benefit-per-ton estimates (U.S. EPA, 2013b)
and Fann et al. (2012c)). Consequently, these estimates may not reflect local variability in
population density, meteorology, exposure, baseline health incidence rates, or other local
factors associated with the current ozone NAAQS review. Therefore, use of these benefit-
per-ton values to estimate co-benefits may lead to higher or lower benefit estimates than
if co-benefits were calculated based on direct air quality modeling.
5-88
-------
In order to evaluate the quantitative impact of specific sources of uncertainty on the core
risk estimates, we completed a number of sensitivity analyses, some of which have already been
described in presenting the core risk estimates and some of which are presented in Appendix 5B.
A brief overview of these sensitivity analyses, including key observations resulting from those
observations are presented below in section 5.7.3.1. In addition to these sensitivity analyses, we
have also included several supplemental analyses intended to provide additional perspectives on
the core incidence and benefits analyses. These supplemental analyses are presented in detail in
Appendix 5C and are also briefly summarized below in section 5.7.3.2.
5.7.3.1 Sensitivity Analyses
A number of sensitivity analyses have been completed as part of this RIA. Given the
importance of mortality in driving benefits estimates (both for ozone and PIVh.s cobenefits), the
sensitivity analyses completed have been focused largely on the mortality endpoint. Each
sensitivity analysis is briefly described below (including an overview of the approach used in
conducting the analysis and key observations). As identified below, several of the sensitivity
analyses have been presented earlier parallel to presentation of the core estimates, while others
are described in Appendix B.71
• Short-term ozone-exposure related mortality (alternative epidemiological studies
and C-R functions): As described in section 5.6.3.1, in addition to the two core effect
estimates we estimated benefits using seven additional effect estimates including four
multi-city studies and three meta-analysis studies. This sensitivity analysis showed that
the two core incidence and benefits estimates fall within (and towards the lower end of)
the broader range resulting from application of the seven alternative effect estimates (see
Tables 5-19 and 5-20). This increased our overall confidence in the two core studies,
particularly with regard to epidemiological study design and the characterization of the
relationship between short-term exposure and mortality.
• Long-term ozone-exposure related respiratory mortality dollar benefits: As
discussed in section 5.1, we could not specify a cessation lag structure specific to long-
term ozone exposure-related respiratory mortality. While we included incidence estimates
as a part of the core analysis, we present the associated dollar benefits as a sensitivity
analysis (see Tables 5-20 and 5-27). Given uncertainty related to the lag structure, we
also included two different lags in modeling these benefits as a sensitivity analysis - a 20
71 Generally, we gave greater emphasis to those sensitivity analyses examining sources of potential uncertainty
directly associated with estimates of ozone-related mortality and have included summaries of those sensitivity
analyses results parallel to presentation of core incidence and benefits estimates in sections 5.7.1 and 5.7.2.
5-89
-------
year segment lag (as used for PIVh.s) and a zero lag (see section 5.6.4.1 for additional
detail on the lag structures used). The sensitivity analysis suggests that if included in the
core benefit estimate, long-term ozone exposure-related mortality could add substantially
to the overall benefits (see Table 5-20 and 5-27). Additionally, use of a 20 year segment
lag can reduce benefits by 10-20% (relative to a zero lag) depending on the discount rate
applied.
Long-term ozone-exposure related respiratory mortality and potential thresholds:
As discussed in section 5.6.3.1, while we favor the no-threshold model in estimating
benefits related to long-term ozone exposure-related respiratory mortality, as a sensitivity
analysis, we evaluated the impact of thresholds ranging from 40-60ppb.72 These results
are included alongside core incidence estimates in Table 5-19 (see Appendix 5B, section
5B1 for additional details). This sensitivity analysis suggested that a threshold of SOppb
or greater could have a substantial impact on estimated benefits, while thresholds below
this range have a relatively minor impact.
Long-term PMi.s exposure-related mortality and alternative C-R functions (based
on expert elicitation): As discussed in Appendix 5D (section 5D.1) in 2006 we
conducted an expert elicitation to help better characterize uncertainty associated with
long-term PIVh.s exposure-related mortality and specifically the C-R functions used in
modeling that endpoint, including the shape of the functions and potential for thresholds
in effect. As part of the sensitivity analysis for the current ozone RIA, we applied the set
of expert elicitation-based functions to generate an alternative set of PIVb.s incidence and
benefit estimates (see Appendix 5B, section 5B.2). That sensitivity analysis showed that
the two core incidence estimate fall within the range of alternative C-R function based
estimates obtained through expert elicitation (see Table 5B-2). This increases overall
confidence in the core estimates with regard to the form of the functions and magnitude
of the effect estimates.
Income elasticity and the willingness-to-pay (WTP) values used for mortality and
certain morbidity endpoints: As described in Appendix 5B (section 5B.3) we examined
the impact of alternative assumptions regarding income elasticity (i.e., the degree to
which WTP changes as income changes) and the degree of impact on WTP functions
used for mortality and for morbidity endpoints. That sensitivity analysis, suggests that
alternative assumptions regarding income elasticity could result in a moderate impact on
mortality benefits (values ranging from -90% to -130% of the core estimate depending
on the assumption regarding elasticity). Income elasticity was found to have a far more
modest impact on morbidity endpoints modeled using WTP functions.
72 As noted in section 5.6.3.1, our decision to include benefit estimates based on the no-threshold model in the core
analysis reflects recommendations on the HREA provided by CASAC (Frey, 2014).
5-90
-------
5.7.3.2 Supplemental Analyses
In addition to the sensitivity analyses described in section 5.7.3.1, we also included a
number of supplemental analyses intended to examine other attributes of the core risk estimates
(these are briefly described below with additional detail found in Appendix 5C).
• Consideration for age group-differentiated aspects of short-term ozone exposure-
related mortality (including total avoided incidence, life years gained and percent
reduction in baseline mortality): These analyses expand on the basic mortality
incidence estimates presented in section 5.7 by considering (a) estimates of the reduction
in mortality incidence differentiated by age range (i.e., how the avoided deaths map to
different age ranges) and (b) estimates of life years gained by age range and (c) percent
reduction in baseline mortality by age range. By comparing the projected age distribution
of the avoided premature deaths with the age distribution of life years gained, we
observed that about half of the ozone-attributable deaths occur in populations age 75-99
(see Appendix 5C, section 5C.1 Table 5C-1), but half of the life years would occur in
populations younger than 65 (see Appendix 5C, section 5C.1, Table 5C-2). This is
because the younger populations have the potential to lose more life years per death than
older populations based on changes in ozone exposure for the 2025 scenario. Results
presented in Table 5C-3 highlight that when reductions in ozone-attributable mortality (in
going from baseline to an alternative standard level) are considered as a percentage of
total all-cause baseline mortality, the estimates are relatively small and are fairly constant
across age ranges. However, it is important to point out that estimates of total ozone-
attributable mortality represent a substantially larger fraction of all-cause baseline
mortality than the increment attributable to the simulated reduction in ozone associated
with an alternative standard level.
• Evaluation of mortality impacts relative to the baseline pollutant concentrations
(used in generating those mortality estimates) for both short-term ozone exposure-
related mortality and long-term PMi.5 exposure-related mortality: In this
supplemental analysis, we begin by comparing the distribution of short-term ozone
exposure-related mortality against the ozone season-averaged 8hr max values used in
deriving those estimates (for the three alternative standards modeled for the 2025
scenario) (see section 5C.2). In making this comparison, we point out that, rather than
using a full daily time series of 8hr max values for each grid cell, we used an ozone
season-average 8hr max value for incidence modeling. While this simplification did not
impact the overall core incidence estimate for this mortality endpoint, it does impact
efforts to compare the distribution of mortality estimates against associated daily values
(by reducing temporal variation associated with each grid cell calculation). Key
observations resulting from this supplemental analysis are that (a) the vast majority of
reductions in short-term exposure-related mortality for ozone occur in grid cells with
mean 8hr max baseline levels (across the ozone season) between 35 and 55ppb and (b)
virtually all of the mortality reductions are associated with ozone levels above the
<20ppb range identified within the Ozone ISA as being associated with less confidence in
specifying the nature of the C-R function for ozone mortality (ozone ISA, section
5-91
-------
2.5.4.4). The other portion of this supplemental analysis (which compares the distribution
of PM2.5 mortality estimates relative to associated PIVh.s levels used in modeling) relies
on the concept of the lowest measured level (LML) associated with the epidemiological
studies providing the effect estimates used in morality modeling. Specifically, we observe
that there is reduced confidence in specifying the nature of the C-R function at exposure
levels below the range used in the epidemiological study (i.e., below the LML). In the
supplemental analysis, we found that, depending on the mortality study, between 67%
and 93% of the mortality estimate is based on modeling involving PIVh.s levels above the
LML (i.e., levels at which we have increased confidence in specifying the PM2.5-
mortality relationship) (see Appendix 5C.2).
Core Incidence and Dollar Benefits Estimates Reflecting Application of Known
Controls for the 2025 Scenario: In Appendix 5C, section 5C.3, we present a subset of
core incidence and benefits for the 2025 scenario reflecting only application of known
controls in modeling reductions in ozone to attain each alternative standard level. These
known-control based estimates include both ozone-related and PM2.5 co-benefit estimates
as well as total benefits. As expected, the percent of benefits reflecting application of
known controls decreases as you consider more stringent alternative standards. Estimates
presented in Appendix 5C.3 suggest that 77%, 59% 34% of total benefits are associated
with application of known controls (for the 70, 65 and 60 ppb alternative standards,
respectively).
5.8 Discussion
The analysis in this Chapter demonstrates the potential for significant health benefits of
the illustrative emissions controls applied to simulate attainment with the alternative primary
ozone standards. We estimate that by 2025, the emissions reductions to reach the alternative
standards everywhere except California, would have reduced the number of ozone- and PM2.5-
related premature mortalities and produce substantial non-mortality benefits. Furthermore,
emissions reductions required to meet alternative standards in California post-2025 are also
likely to produce substantial reductions in these same endpoints. This proposed rule also
promises to yield significant welfare impacts as well (see Chapter 6). Even considering the
quantified and unquantified uncertainties identified in this chapter, we believe that implementing
the alternative standards would have substantial public health benefits that are likely to outweigh
the costs for the three alternative standards analyzed (see Chapter 7).
Inherent in any complex RIA such as this one are multiple sources of uncertainty. Some
of these we characterized through our quantification of statistical error in the concentration-
response relationships and our use of alternate mortality functions. Others, including the
5-92
-------
projection of atmospheric conditions and source-level emissions, the projection of baseline
morbidity rates, incomes and technological development are unquantified. When evaluated
within the context of these uncertainties, the health impact and monetized benefits estimates in
this RIA can provide useful information regarding the public health benefits associated with the
proposed and alternative primary standards.
There are important differences worth noting in the design and analytical objectives of
NAAQS RIAs compared to RIAs for implementation rules, such as the Tier 3 (U. S. EPA,
2014c). The NAAQS RIAs illustrate the potential costs and benefits of a revised air quality
standard nationwide based on an array of emission reduction strategies for different sources,
incremental to implementation of existing regulations and controls needed to attain the current
standards. In short, NAAQS RIAs hypothesize, but do not predict, the emission reduction
strategies that States may choose to enact when implementing a revised NAAQS. The setting of
a NAAQS does not directly result in costs or benefits, and as such, NAAQS RIAs are merely
illustrative and are not intended to be added to the costs and benefits of other regulations that
result in specific costs of control and emission reductions. By contrast, the emission reductions
from implementation rules are generally for specific, well-characterized sources, such as the
recent MATS rule (U.S. EPA, 201 le). In general, the EPA is more confident in the magnitude
and location of the emission reductions for implementation rules. As such, emission reductions
achieved under promulgated implementation rules such as Tier 3 have been reflected in the
baseline of this NAAQS analysis. Subsequent implementation rules will be reflected in the
baseline for the next ozone NAAQS review. For this reason, the benefits estimated provided in
this RIA and all other NAAQS RIAs should not be added to the benefits estimated for
implementation rules.
In setting the NAAQS, the EPA considers that ozone concentrations vary over space and
time. While the standard is designed to limit concentrations at the highest monitor in an area, it is
understood that emission controls put in place to meet the standard of the highest monitor will
simultaneously result in lower ozone concentrations throughout the entire area. In fact, the
Health Risk and Exposure Assessment for Ozone (HREA) (U.S. EPA, 2014b) shows how
different standard levels would affect the entire distribution of ozone concentrations, and thus
5-93
-------
people's exposures and risk, across a selected set of urban areas. For this reason, it is
inappropriate to use the NAAQS level as a bright line for health effects.
The NAAQS are not set at levels that eliminate the risk of air pollution completely.
Instead, the Administrator sets the NAAQS at a level requisite to protect public health with an
adequate margin of safety, taking into consideration effects on susceptible populations based on
the scientific literature. The risk analysis prepared in support of this ozone NAAQS reported
risks below these levels, while acknowledging that the confidence in those effect estimates is
higher at levels closer to the standard (U.S. EPA, 2014b). While benefits occurring below the
standard may be somewhat more uncertain than those occurring above the standard, the EPA
considers these to be legitimate components of the total benefits estimate. Though there are
greater uncertainties at lower ozone and PIVh.s concentrations, there is no evidence of a threshold
in short-term ozone or PIVh.s-related health effects in the epidemiology literature.73 Given that
the epidemiological literature in most cases has not provided estimates based on threshold
models, there would be additional uncertainties imposed by assuming thresholds or other non-
linear concentration-response functions for the purposes of benefits analysis.
The estimated benefits for the proposed and alternative standards are in addition to the
substantial benefits estimated for several recent implementation rules (U.S. EPA, 2009a, 201 Id,
2014c). Rules such as Tier 3 and other emission reductions will have substantially reduced
ambient ozone concentrations by 2025 in the East, such that few additional controls would be
needed to reach 70 ppb in the East beyond the analytical baseline. These rules that have already
been promulgated have tremendous combined benefits that explain why the number of avoided
premature deaths associated with this NAAQS revision are smaller than were estimated in the
previous ozone NAAQS RIA (U.S. EPA, 2006) for the year 2020 and even smaller than the
mortality risks estimated for the current year in the ozone HREA (U.S. EPA, 2014b).
73 As discussed 5.7.3.1, our modeling of long-term ozone exposure-related respiratory mortality did include
consideration of potential thresholds as a sensitivity analysis. However, key points need to be emphasized in the
context of interpreting this endpoint: (a) based on recommendations from CASAC the core incidence reduction
estimates was based on a non-threshold model (Frey, 2014) and (b) while the incidence estimates for this endpoint
were included as core estimates, associated dollar benefit estimates were only included as sensitivity analyses and
consequently did not factor in to the core dollar benefit estimates generated.
5-94
-------
5.9 References
Abt Associates Inc. 1996. A Paniculate Matter Risk Assessment for Philadelphia and Los Angeles. Prepared for
Office of Air Quality Planning and Standards. Research Triangle Park, NC: EPA Office of Air and Radiation,
OAQPS, July 1996. Available at: .
Abt Associates, Inc. 2010. Environmental Benefits and Mapping Program (Version 4.0). Prepared for U.S.
Environmental Protection Agency, Bethesda, MD. Research Triangle Park, NC: EPA Office of Air Quality
Planning and Standards. Available on the Internet at: .
Abt Associates, Inc. 2012. BenMAP User's Manual Appendices. Prepared for U.S. Environmental Protection
Agency Office of Air Quality Planning and Standards. Research Triangle Park, NC. September. Available at
.
Adams PF, Hendershot GE, Marano MA. 1999. "Current Estimates from the National Health Interview Survey,
1996." Vital and Health Statistics 10(200): 1-212.
Agency for Healthcare Research and Quality (AHRQ). 2000. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available at .
Agency for Healthcare Research and Quality (AHRQ). 2007. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available at .
Agency for Healthcare Research and Quality (AHRQ). 2009. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available at .
Alberini A, Cropper M, Krupnick A, Simon NB. 2004. "Does the value of a statistical life vary with age and health
status? Evidence from the US and Canada." Journal of Environmental Economics and Management 48(1): 769-
792.
Alberini, Anna & Milan Scasny. 2011. "Context and the VSL: Evidence from a Stated Preference Study in Italy and
the Czech Republic." Environmental & Resource Economics 49(4), pages 511-538.
Aldy JE, Viscusi WK. 2008. "Adjusting the value of a statistical life for age and cohort effects." The Review of
Economics and Statistics 90(3):573-581.
American Lung Association (ALA). 2002. Trends in Asthma Morbidity and Mortality. American Lung Association,
Best Practices and Program Services, Epidemiology and Statistics Unit. Available at
.
American Lung Association (ALA). 2010b. Trends in Asthma Morbidity and Mortality. American Lung Association
Epidemiology and Statistics Unit Research and Program Services Division. February. Table 7. Based on NHIS
data (CDC, 2008). Available at http://www.lungusa.org/fmding-cures/our-research/trend-reports/asthma-trend-
report.pdf.
Ayyub, B.M. 2002. Elicitation of Expert Opinions for Uncertainty and Risk. CRC Press, Florida.
Babin, S. M., H. S. Burkom, et al. 2007. "Pediatric Patient Asthma-related Emergency Department Visits and
Admissions in Washington, DC, from 2001-2004, and Associations with Air Quality, Socio-economic Status
and Age Group." Environmental Health 6: 9.
Bell, ML; McDermott, A; Zeger, SL; Samet, JM; Dominici, F. (2004). Ozone and short-term mortality in 95 US
urban communities, 1987-2000. JAMA 292: 2372-2378. http://dx.doi.org/10.1001/jama.292.19.2372
5-95
-------
Bell, ML; Dominici, F; Samet, JM. (2005). Ameta-analysis of time-series studies of ozone and mortality with
comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
http://dx.doi.org/10.1097/01.ede.0000165817.40152.85
Bell, M. L., K. Ebisu, et al. 2008. "Seasonal and Regional Short-term Effects of Fine Particles on Hospital
Admissions in 202 US Counties, 1999-2005." American Journal of 'Epidemiology 168(11): 1301-1310.
Bell, M.L. and F. Dominici. 2008. Effect modification by community characteristics on the short-term effects of
ozone exposure and mortality in 98 U.S. communities. American Journal of Epidemiology. 167:986-997.
Bell, M.L.; A. McDermott; S.L. Zeger; J.M. Samet; F. Dominici. 2004. Ozone and short-term mortality in 95 U.S.
urban communities, 1987-2000. JAMA. 292:2372-2378.
Berger, M.C., G.C. Blomquist, D. Kenkel, and G.S. Tolley. 1987. "Valuing Changes in Health Risks: A Comparison
of Alternative Measures." The Southern Economic Journal 53:977-984.
Centers for Disease Control and Prevention (CDC). 2003. Health, United States. Table 30. Years of Potential Life
Lost Before Age 75 for Selected Causes of Death, According to Sex, Race and Hispanic Origin: United States,
Selected Years 1980- 2000. Available at .
Centers for Disease Control and Prevention (CDC). 2008. National Center for Health Statistics. National Health
Interview Survey, 1999-2008.
Centers for Disease Control and Prevention (CDC). 2010. Table Cl Adult Self-Reported Current Asthma Prevalence
Rate (Percent) and Prevalence (Number) by State or Territory. Centers for Disease Control and Prevention,
Behavioral Risk Factor Surveillance System (BRFSS). Available at:
.
Centers for Disease Control and Prevention (CDC). 2011. United States Life Tables, 2007 National Vital Statistics
Reports. Volume 59, Number 9. September. Available at
.
Chen L, Jennison BL, Yang W, and Omaye ST. 2000. "Elementary school absenteeism and air pollution."
Inhalation Toxicolology 12(11):997-1016.
Cropper, M. L. and A. J. Krupnick. 1990. The Social Costs of Chronic Heart and Lung Disease. Resources for the
Future. Washington, DC. Discussion Paper QE 89-16-REV.
Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis, J.H. Ware, M. Raizenne,
and F.E. Speizer. 1996. "Health Effects of Acid Aerosols On North American Children-Respiratory
Symptoms." Environmental Health Perspectives 104(5):500-505.
Eisenstein, E.L., L.K. Shaw, K.J. Anstrom, C.L. Nelson, Z. Hakim, V. Hasselblad and D.B. Mark. 2001. "Assessing
the Clinical and Economic Burden of Coronary Artery Disease: 1986-1998." Medical Care 39(8):824-35.
Farm N, Lamson A, Wesson K, Risley D, Anenberg SC, Hubbell BJ. 2012a. "Estimating the National Public Health
Burden Associated with Exposure to Ambient PM2s and ozone. Risk Analysis," Risk Analysis 32(1): 81-95.
Farm N, Roman HR, Fulcher CF, Gentile M, Wesson K, Hubbell BJ, Levy JI. 2012b. "Maximizing Health Benefits
and Minimizing Inequality: Incorporating Local Scale Data in the Design and Evaluation of Air Quality
Policies." Risk Analysis 32(6): 908—922.
Farm, N.; K.R. Baker and CM. Fulcher. 2012c. "Characterizing the PM2 5-related health benefits of emission
reductions for 17 industrial, area and mobile emission sectors across the U.S." Environment International 49
41-151.
5-96
-------
Freeman (III), AM. 1993. The Measurement of Environmental and Resource Values: Theory and Methods.
Washington, DC: Resources for the Future.
Frey, C. 2014. CASAC Review of the EPA's Health Risk and Exposure Assessment for Ozone (Second External
Review Draft - February, 2014). U.S. Environmental Protection Agency Science Advisory Board. EPA-
CASAC-14-XXX.Gent, IF.; E.W. Triche; T.R. Holford; K. Belanger; M.B. Bracken; W.S. Beckett, et al. 2003.
Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. Journal of
the American Medical Association. 290(14): 1859-1867.
Frey, C. and J. Samet. 2012. CASAC Review of the EPA's Health Risk and Exposure Assessment for Ozone (First
External Review Draft - Updated August 2012) and Welfare Risk and Exposure Assessment for Ozone (First
External Review Draft - Updated August 2012). U.S. Environmental Protection Agency Science Advisory
Board. EPA-CASAC-13-002.
GeoLytics Inc. 2002. GeoLytics CensusCD® 2000 Short Form Blocks. CD-ROM Release 1.0. GeoLytics, Inc. East
Brunswick, NJ. Available: http://www.geolytics.com/ [accessed 29 September 2004].
Gilliland FD, Berhane K, Rappaport EB, Thomas DC, Avol E, Gauderman WJ, et al. 2001. "The effects of ambient
air pollution on school absenteeism due to respiratory illnesses." Epidemiology 12(l):43-54.
Glad, J.A., L.L. Brink, E.G. Talbott, P.C. Lee, X. Xu, M. Saul, and J. Rager. 2012. "The Relationship of Ambient
ozone and PM2 5 Levels and Asthma Emergency Department Visits: Possible Influence of Gender and
Ethnicity." Archives of Environmental & Occupational Health 62 (2): 103-108.
Graff Zivin, J., Neidell, M. (2012). "The impact of pollution on worker productivity." American Economic Review,
102, 3652-3673.
Hall JV, Brajer V, Lurmann FW. 2003. "Economic Valuation of ozone-related School Absences in the South Coast
Air Basin of California." Contemporary Economic Policy 11 (4):407-417.
Harrington, W., and P.R. Portney. 1987. "Valuing the Benefits of Health and Safety Regulation." Journal of Urban
Economics 22:101-112.
Hollman, F.W., T.J. Mulder, and J.E. Kalian. 2000. Methodology and Assumptions for the Population Projections of
the United States: 1999 to 2100. Population Division Working Paper No. 38, Population Projections Branch,
Population Division, U.S. Census Bureau, Department of Commerce. January. Available at
.
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.
Hubbell BL. 2006. "Implementing QALYs in the analysis of air pollution regulations." Environmental and Resource
Economics 34:365-384.
Hubbell, B., A. Hallberg, D.R. McCubbin, andE. Post. 2005. "Health-Related Benefits of Attaining the 8-Hr ozone
Standard." Environmental Health Perspectives 113:73-82. ICD, 1979
Industrial Economics, Incorporated (ffic). 2006. Expanded Expert Judgment Assessment of 'the Concentration-
Response Relationship Between PM2s Exposure and Mortality. Prepared for: Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC. September. Available at
.
Industrial Economics, Incorporated (ffic). 2012. Updating BenMAP Income Elasticity Estimates—Literature Review.
Prepared for: Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research
5-97
-------
Triangle Park, NC. September. Available at
Ito, K.; G.D. Thurston and R.A. Silverman. 2007. Characterization of PM2.5, gaseous pollutants, and meteorological
interactions in the context of time-series health effects models. Journal of Exposure Science and Environmental
Epidemiology. 17(S2):S45-60.
Jerrett M, Burnett RT, Pope CA, III, et al. 2009. "Long-Term Ozone Exposure and Mortality." New England
Journal of Medicine 360:1085-95.
Katsouyanni, K.; J.M. Samet; H.R. Anderson; R. Atkinson; A.L. Tertre; S. Medina, et al. 2009. Air Pollution and
Health: A European and North American Approach (APHENA). Health Effects Institute.
Kleckner, N., and J. Neumann. 1999. Recommended Approach to Adjusting WTP Estimates to Reflect Changes in
Real Income. Memorandum to Jim DeMocker, U.S. EPA/OPAR. June 3. Available at
.
Kloog, I., B.A. Coull, A. Zanobetti, P. Koutrakis, J.D. Schwartz. 2012. "Acute and Chronic Effects of Particles on
Hospital Admissions in New England." PLoS ONE. Vol 7 (4): 1-8.
Kochi, I., B. Hubbell, and R. Kramer. 2006. "An Empirical Bayes Approach to Combining Estimates of the Value of
Statistical Life for Environmental Policy Analysis." Environmental and Resource Economics. 34: 385-406.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi, Y, et al. 2009. Extended follow-up and spatial analysis of
the American Cancer Society study linking particulate air pollution and mortality. HEI Research Report, 140,
Health Effects Institute, Boston, MA.
Krupnick, A.J., and M.L. Cropper. 1992. "The Effect of Information on Health Risk Valuations." Journal of Risk
and Uncertainty 5(2):29-48.
Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Kerry, F. Horak Jr., V. Puybonnieux-
Texier, P. Quenel, J. Schneider, R. Seethaler, J-C Vergnaud, and H. Somrner. 2000. "Public-Health Impact of
Outdoor and Traffic-Related Air Pollution: A European Assessment." The Lancet 356:795-801.
Lepeule J, Laden F, Dockery D, Schwartz J. 2012. "Chronic Exposure to Fine Particles and Mortality: An Extended
Foliow-Up of the Harvard Six Cities Study from 1974 to 2009." Environmental Health Perspectives 120
(7):965-70.
Levy JI, Baxter LK, Schwartz J. 2009. "Uncertainty and variability in health-related damages from coal-fired power
plants in the United States." Risk Analysis 29(7) 1000-1014.
Lichtenstein, S. and P. Slovic, eds. 2006. The construction of preference. New York: Cambridge University Press.
Mansfield, Carol and Sumeet Patil, 2006. Peer Review of Expert Elicitation. Memo to Ron Evans, EPA. September
29, 2006. RTI International. Available at
.
Mansfield, Carol; Paramita Sinha; Max Henrion. 2009. Influence Analysis in Support of Characterizing Uncertainty
in Human Health Benefits Analysis: Final Report. Prepared for U.S. EPA, Office of Air Quality Planning and
Standards. November. Available at
.
Mar, TF; Koenig, JQ. (2009). Relationship between visits to emergency departments for asthma and ozone exposure
in greater Seattle, Washington. Ann Allergy Asthma Immunol 103: 474-479.http://dx.doi.org/10.1016/S1081-
1206(10)60263-3
5-98
-------
Mar, T. F., J. Q. Koenig and J. Primomo. 2010. "Associations between asthma emergency visits and paniculate
matter sources, including diesel emissions from stationary generators in Tacoma, Washington." Inhalation
Toxicology 22 (6): 445-8.
Mar, T. F., T. V. Larson, et al. 2004. "An analysis of the association between respiratory symptoms in subjects with
asthma and daily air pollution in Spokane, Washington." Inhalation Toxicology 16(13): 809-15.
Medina-Ramon, M.; A. Zanobetti and J. Schwartz. 2006. The effect of ozone and PM10 on hospital admissions for
pneumonia and chronic obstructive pulmonary disease: a national multi-city study. American Journal of
Epidemiology. 163(6):579-588.
Moolgavkar, S.H. 2000. "Air Pollution and Hospital Admissions for Diseases of the Circulatory System in Three
U.S. Metropolitan Areas." Journal of the Aif and Waste Management Association 50:1199-1206.
Mortimer, KM; Neas, LM; Dockery, DW; Redline, S; Tager, IB. (2002). The effect of air pollution on innercity
children with asthma. Eur Respir J 19: 699-705.
Mrozek, J.R., and L.O. Taylor. 2002. "What Determines the Value of Life? A Meta-Analysis." Journal of Policy
Analysis and Management 21(2):253-270.
National Center for Education Statistics (NCHS). 1996. The Condition of Education 1996, Indicator 42: Student
Absenteeism and Tardiness. U.S. Department of Education. Washington, DC.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press. Washington, DC.
National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. National Academies Press. Washington, DC.
Neumann, J.E., M.T. Dickie, and R.E. Unsworth. 1994. Linkage between Health Effects Estimation and Morbidity
Valuation in the Section 812 Analysis—Draft Valuation Document. Industrial Economics Incorporated (IEc)
Memorandum to Jim DeMocker, U.S. Environmental Protection Agency, Office of Air and Radiation, Office of
Policy Analysis and Review. March 31. Available at
.
NRC. 2009. Science and Decisions, Advancing Risk Assessment. Committee on Improving Risk Analysis
Approaches. Washington, DC: The National Academies Press, National Research Council.
O'Connor, GT; Neas, L; Vaughn, B; Kattan, M; Mitchell, H; Crain, EF; III, ER; Gruchalla, R; Morgan, W; Stout, J;
Adams, GK; Lippmann, M. (2008). Acute respiratory health effects of air pollution on children with asthma in
US inner cities. J Allergy Clin Immunol 121: 1133-1139.
Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory Analysis. Washington, DC. Available at
Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens, andM. White. 2001. "Air Pollution and Exacerbation of Asthma
in African-American Children in Los Angeles." Epidemiology 12(2):200-208.
Ostro, B.D. 1987. "Air Pollution and Morbidity Revisited: A Specification Test." Journal of Environmental
Economics Management 14:87-98.
Ostro, B.D. and S. Rothschild. 1989. "Air Pollution and Acute Respiratory Morbidity: An Observational Study of
Multiple Pollutants." Environmental Research 50:238-247.
5-99
-------
Ozkaynak, H., and G.D. Thurston. 1987. "Associations between 1980 U.S. Mortality Rates and Alternative
Measures of Airborne Particle Concentration." Risk Analysis 7(4):449-461.
Peel, J.L., P.E. Tolbert, M. Klein et al. 2005. Ambient air pollution and respiratory emergency department visits.
Epidemiology. Vol. 16 (2): 164-74.
Peng, R. D., H. H. Chang, et al. 2008. "Coarse paniculate matter air pollution and hospital admissions for
cardiovascular and respiratory diseases among Medicare patients." Journal of the American Medical
Association 299(18): 2172-9.
Peng, R. D., M. L. Bell, et al. 2009. "Emergency admissions for cardiovascular and respiratory diseases and the
chemical composition of fine particle air pollution." Environmental Health Perspectives 117(6): 957-63.
Peters A. 2005. "Paniculate matter and heart disease: evidence from epidemiological studies." Toxicology and
Applied Pharmacology Sept 1;207(2 Suppl):477-82.
Peters, A., D. W. Dockery, J. E. Muller and M. A. Mittleman. 2001. "Increased paniculate air pollution and the
triggering of myocardial infarction." Circulation 103 (23): 2810-5.
Poloniecki, J.D., R.W. Atkinson., A.P. de Leon., andH.R. Anderson. 1997. "Daily Time Series for Cardiovascular
Hospital Admissions and Previous Day's Air Pollution in London, UK." Occupational and Environmental
Medicine 54(8):535-540
Pope, C. A., Ill, J. B. Muhlestein, et al. 2006. "Ischemic heart disease events triggered by short-term exposure to
fine paniculate air pollution." Circulation 114(23): 2443-8.
Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. "Respiratory Health and PM10 Pollution:
A Daily Time Series Analysis." American Review of Respiratory Diseases 144:668-674.
Roman, Henry A., Katherine D. Walker, Tyra L. Walsh, Lisa Conner, Harvey M. Richmond, Bryan J. Hubbell, and
Patrick L. Kinney. 2008. "Expert Judgment Assessment of the Mortality Impact of Changes in Ambient Fine
Paniculate Matter in the U.S." Environmental Science & Technology 42(7):2268-2274.
Rosamond, W., G. Broda, E. Kawalec, S. Rywik, A. Pajak, L. Cooper, and L. Chambless. 1999. "Comparison of
Medical Care and Survival of Hospitalized Patients with Acute Myocardial Infarction in Poland and the United
States." American Journal of Cardiology 83:1180-1185
Rowe, R.D., and L.G. Chestnut. 1986. "Oxidants and Asthmatics in Los Angeles: A Benefits Analysis—Executive
Summary." Prepared by Energy and Resource Consultants, Inc. Report to the U.S. Environmental Protection
Agency, Office of Policy Analysis. EPA- 230-09-86-018. Washington, DC.
Rowlatt, P., Spackman, M., Jones, S., Jones-Lee, M., Loonies, G. 1998. Valuation of deaths from air pollution. A
Report for the Department of Environment, Transport and the Regions and the Department of Trade and
Industry. National Economic Research Associates (NERA), London.
Russell, M.W., D.M. Huse, S. Drowns, B.C. Hamel, and S.C. Hartz. 1998. "Direct Medical Costs of Coronary
Artery Disease in the United States." American Journal of Cardiology 81(9): 1110-1115.
Sarnat, J. A., Sarnat, S. E., Flanders, W. D., Chang, H. H., Mulholland, J., Baxter, L., Ozkaynak, H. (2013).
Spatiotemporally resolved air exchange rate as a modifier of acute air pollution-related morbidity in Atlanta. J
Expo Sci EnvironEpidemiol. doi: 10.1038/jes.2013.32
Sasser, E. 2014. Response to Comments Regarding the Potential Use of a Threshold Model in Estimating the
Mortality Risks from Long-term Exposure to Ozone in the Health Risk and Exposure Assessment for Ozone,
5-100
-------
Second External Review Draft. Memorandum to Holly Stallworth, Designated Federal Officer, Clean Air
Scientific Advisory Committee from EPA/OAQPS Health and Environmental Impacts Division.
Schildcrout, J. S., Sheppard, L., Lumley, T., Slaughter, J. C., Koenig, J. Q., & Shapiro, G. G. (2006). Ambient air
pollution and asthma exacerbations in children: an eight-city analysis. Am J Epidemiol, 164(6), 505-517.
Schwartz J, Coull B, Laden F. 2008. "The Effect of Dose and Timing of Dose on the Association between Airborne
Particles and Survival." Environmental Health Perspectives 116: 64-69.
Schwartz J, Dockery DW, Neas LM, Wypij D, Ware JH, Spengler JD, Koutrakis P, Speizer FE, Ferris BG Jr. 1994.
"Acute effects of summer air pollution on respiratory symptom reporting in children." American Journal of
Respiratory and Critical Care Medicine 150(5 Pt 1): 1234-42.
Schwartz, J. 1993. "Particulate Air Pollution and Chronic Respiratory Disease." Environment Research 62:7-13.
Schwartz, J., and L.M. Neas. 2000. "Fine Particles are More Strongly Associated than Coarse Particles with Acute
Respiratory Health Effects in Schoolchildren." Epidemiology 11:6-10.
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.
Shogren, I, andT. Stamland. 2002. "Skill and the Value of Life." Journal of Political Economy 110:1168-1197.
Silverman, R.A.; and K. Ito. 2010. Age-related association of fine particles and ozone with severe acute asthma in
New York City. Journal of Allergy Clinical Immunology. 125(2):367-373.
Slaughter, James C et al. 2005. "Association between paniculate matter and emergency room visits, hospital
admissions and mortality in Spokane, Washington." Journal of Exposure Analysis and Environmental
Epidemiology 15, 153-159.
Smith et al. 2004. "Do the Near Elderly Value Mortality Risks Differently?" The Review of Economics and Statistics
86(1): 423-429.
Smith, D.H., D.C. Malone, K.A. Lawson, L.J. Okamoto, C. Battista, and W.B. Saunders. 1997. "A National
Estimate of the Economic Costs of Asthma." American Journal of Respiratory and Critical Care Medicine
156(3 Pt l):787-793.
Smith, K., Pattanayak SK, Van Houtven G. 2006. "Structural benefit transfer: An example using VSL estimates."
Ecological Economics 60: 361-371.
Smith, R.L.; B. Xu and P. Switzer. 2009. Reassessing the relationship between ozone and short- term mortality in
U.S. urban communities. Inhalation Toxicology. 21:37-61.
Standard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter. Available at
.
Stanford, R., T. McLaughlin and L. J. Okamoto. 1999. "The cost of asthma in the emergency department and
hospital." American Journal of Respiratory and Critical Care Medicine 160 (1): 211-5.
Strickland, M.J.; L.A. Darrow; M. Klein; W.D. Flanders; J.A. Sarnat; L.A. Waller, et al. 2010. Short-term
associations between ambient air pollutants and pediatric asthma emergency department visits. American
Journal of Respiratory Critical Care Medicine. 182:307-316.
5-101
-------
Sullivan, I, L. Sheppard, et al. 2005. "Relation between short-term fine-paniculate matter exposure and onset of
myocardial infarction." Epidemiology 16(1): 41-8.
Tagaris E, Liao KJ, Delucia AJ, et al. 2009. "Potential impact of climate change on air-pollution related human
health effects." Environmental Science & Technolology 43: 4979-4988.
Tolbert, P. E., M. Klein, et al. 2007. "Multipollutant modeling issues in a study of ambient air quality and
emergency department visits in Atlanta." Journal of Exposure Science and Environmental Epidemiology 17
Suppl 2: S29-35.
Tolbert, P.E.; M. Klein; J.L. Peel; S.E. Sarnat and J.A. Sarnat. 2007. Multipollutant modeling issues in a study of
ambient air quality and emergency department visits in Atlanta. Journal of Exposure Science and
Environmental Epidemiology. 17(S2):S29-35.
Tolley, G.S. et al. 1986. Valuation of Reductions in Human Health Symptoms and Risks. University of Chicago.
Final Report for the U.S. Environmental Protection Agency, Office of Policy Analysis. January. Available at
.
Tsuge T, Kishimot A, Takeuchi K. 2005. "A choice experiment approach to the valuation of mortality." Journal of
Risk and Uncertainty 3 l(l):73-95.
U.S. Bureau of Economic Analysis (U.S. BEA). 1997. Real Gross Domestic Product—Table 2A. U.S. Department
of Commerce.
U.S. Bureau of Economic Analysis (U.S. BEA). 1997. The Economic and Budget Outlook: An Update, Table 4—
Economic Projections for Calendar Years 1997 through 2007. September. U.S. Department of Commerce.
U.S. Bureau of Economic Analysis (U.S. BEA). 2004. BEA Economic Areas (EAs). U.S. Department of Commerce.
Available at .
U.S. Bureau of the Census. 2001. Statistical Abstract of the United States, 2001, Section 12: Labor Force,
Employment, and Earnings, Table No. 521. Washington, DC.
U.S. Bureau of the Census. 2002a. Modified Race Data Summary File 2000 Technical Documentation. Census of
Population and Housing: Washington DC. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 1999a. The Benefits and Costs of the Clean Air Act, 1990-2010.
Prepared for U. S. Congress by U. S. EPA, Office of Air and Radiation/Office of Policy Analysis and Review,
Washington, DC, November; EPA report no. EPA-410-R-99-001. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2000. Guidelines for Preparing Economic Analyses. EPA 240-
R-00-003. National Center for Environmental Economics, Office of Policy Economics and Innovation.
Washington, DC. September. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2001. Risk Assessment Guidance for Superfund. Vol. Ill, Part
A. Process for Conducting Probabilistic Risk Assessment (RAGS 3 A). Washington, DC: EPA. (EPA document
number EPA 540-R-02-002; OSWER 9285.7-45; PB2002 963302). Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2004a. Air Quality Criteria for Particulate Matter. National
Center for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection
5-102
-------
Agency, Research Triangle Park, NC EPA/600/P-99/002bF. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2004b. EPA's Risk Assessment Process for Air Toxics: History
and Overview. In: Air Toxics Risk Assessment Reference Library, Technical Resource Manual, Vol. 1, pp. 3-1
- 3-30. (EPA document number EPA-453-K-04-001A). Washington, DC: EPA. Available at:
.
U.S. Environmental Protection Agency, 2006. Final Regulatory Impact Analysis: PM2.5 NAAQS. Prepared by
Office of Air and Radiation. Available: http://www.epa.gov/ttn/ecas/ria.html [accessed 10 March 2008].
U.S. Environmental Protection Agency (U.S. EPA). 2007. Ozone Health Risk Assessment for Selected Urban Areas.
Research Triangle Park, NC: EPA Office of Air Quality Planning and Standards. (EPA document number EPA
452/R-07-009). Available at: .
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis, 2008 National Ambient Air
Quality Standards for Ground-level ozone, Chapter 6. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. March. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2008b. Integrated Science Assessment for Oxides of Nitrogen—
Health Criteria (Final Report). National Center for Environmental Assessment, Research Triangle Park, NC.
July. Available at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=194645
U.S. Environmental Protection Agency (U.S. EPA). 2009a. Regulatory Impact Analysis: National Emission
Standards for Hazardous Air Pollutants from the Portland Cement Manufacturing Industry. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. April. Available at
.
U. S. Environmental Protection Agency (U. S. EPA). 2009b. Integrated Science Assessment for Particulate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division.
December. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2010a. Final Regulatory Impact Analysis (RIA) for the NO2
National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning and Standards, Research
Triangle Park, NC. January. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010b. Quantitative Health Risk Assessment for Particulate
Matter—Final Report. EPA-452/R-10-005. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. September. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010c. Final Regulatory Impact Analysis (RIA) for the SO2
National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning and Standards, Research
Triangle Park, NC. June. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2010d. Regulatory Impact Analysis (RIA) for the Proposed
Reconsideration of the ozone National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning
and Standards, Research Triangle Park, NC. January. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010e. Guidelines for Preparing Economic Analyses. EPA 240-
R-10-001. National Center for Environmental Economics, Office of Policy Economics and Innovation.
Washington, DC. December. Available at .
5-103
-------
U.S. Environmental Protection Agency (U.S. EPA). 2010f. Technical Support Document: Summary of Expert
Opinions on the Existence of a Threshold in the Concentration-Response Function for YM^.s-related Mortality.
Research Triangle Park, NC. June. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2010g. Proposed Regulatory Impact Analysis (RIA)for the
Transport Rule. Office of Air Quality Planning and Standards, Research Triangle Park, NC. January. Available
at.
U.S. Environmental Protection Agency (U.S. EPA). 201 la. Ozone National Ambient Air Quality Standards: Scope
and Methods Plan for Health Risk and Exposure Assessment. Research Triangle Park, NC: EPA. (EPA
document number EPA-452/P-11-001).
U.S. Environmental Protection Agency (U.S. EPA). 2011b. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March.
Available at .
U.S. Environmental Protection Agency (U.S. EPA). 201 Ic. Policy Assessment for the Review of the Particulate
Matter National Ambient Air Quality Standards. EPA-452/D-11-003. April. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011d. Regulatory Impact Analysis for the Federal
Implementation Plans to Reduce Interstate Transport of Fine Particulate Matter and ozone in 27 States;
Correction of SIP Approvals for 22 States. June. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011e. Regulatory Impact Analysis for the Final Mercury and
Air Toxics Standards. EPA-452/R-11-011. December. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 201 If. Regulatory Impact Results for the Reconsideration
Proposal for National Emission Standards for Hazardous Air Pollutants for Industrial, Commercial, and
Institutional Boilers and Process Heaters at Major Sources. December. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for the Proposed Revisions
to the National Ambient Air Quality Standards for Particulate Matter. EPA-452/R-12-003. Office of Air Quality
Planning and Standards, Health and Environmental Impacts Division. December. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/PMRIACombinedFile_Bookmarked.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2012a. Health Risk and Exposure Assessment for Ozone. First
External Review Draft. Research Triangle Park, NC: U.S. Environmental Protection Agency, Research Triangle
Park, NC. (EPA document number EPA 452/P-12-001).
U.S. Environmental Protection Agency (U.S. EPA). 2012b. Regulatory Impact Analysis for the Final Revisions to
the National Ambient Air Quality Standards for Particulate Matter. EPA-452/R-12-003. Office of Air Quality
Planning and Standards, Health and Environmental Impacts Division. December. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2012c. Provisional Assessment of Recent Studies on Health
Effects of Particulate Matter Exposure. EPA/600/R-12/056A National Center for Environmental Assessment—
RTF Division. December.
U.S. Environmental Protection Agency (U.S. EPA). 2013a. Integrated Science Assessment for Ozone and Related
Photochemical Oxidants: Final. Research Triangle Park, NC: U.S. Environmental Protection Agency. (EPA
document number EPA/600/R-10/076F).
5-104
-------
U.S. Environmental Protection Agency (U.S. EPA). 2013b. Technical Support Document: Estimating the Benefit per
ton of Reducing PM2 5 Precursors from 17 sectors. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. February. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2013c. Environmental Benefits Mapping Analysis Program
(BenMAP v4.0). Posted January, 2013. < http://www.epa.gov/auYbenmap/download.html />.
U.S. Environmental Protection Agency (U.S. EPA). 2014a. Control of Air Pollution from Motor Vehicles: Tier 3
Motor Vehicle Emission and Fuel Standards Final Rule: Regulatory Impact Analysis. EPA-420-R-14-005.
Office of Transportation and Air Quality, Assessment and Standards Division. March. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2014a. Environmental Benefits Mapping and Analysis Program
- Community Edition (Version 1.0.8), 2014. Research Triangle Park, NC. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2014b. Health Risk and Exposure Assessment for Ozone. Final
Report. Research Triangle Park, NC: U.S. Environmental Protection Agency, Research Triangle Park, NC.
(EPA document number EPA-452/R-14-004a).
U.S. Environmental Protection Agency (U.S. EPA). 2014c. Regulatory Impact Analysis for the Proposed Carbon
Pollution Guidelines for Existing Power Plants and Emission Standards for Modified and Reconstructed Power
Plants. Research Triangle Park, NC: Office of Air Quality Planning and Standards. (EPA document number
EPA-542/R-14-002, June). Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 1999. An SAB Advisory on the
Clean Air Act Section 812 Prospective Study Health and Ecological Initial Studies. Prepared by the Health and
Ecological Effects Subcommittee (HEES) of the Advisory Council on the Clean Air Compliance Analysis,
Science Advisory Board, U.S. Environmental Protection Agency. Washington DC. EPA-SAB-COUNCIL-
ADV-99-005. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2000. An SAB Report on EPA's
White Paper Valuing the Benefits of Fatal Cancer Risk Reduction. EPA-SAB-EEAC-00-013. July. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2004a. Advisory on Plans for
Health Effects Analysis in the Analytical Plan for EPA 's Second Prospective Analysis—Benefits and Costs of
the Clean Air Act, 1990-2020. EPA-SAB-COUNCIL-ADV-04-002. March. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2004b. Review of the Draft
Analytical Plan for EPA 's Second Prospective Analysis—Benefits and Costs of the Clean Air Act, 1990-2020:
An Advisory by the Advisory Council for Clean Air Compliance Analysis. EPA-SAB-COUNCIL-ADV-04-004.
May. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2004c. Advisory Council on
Clean Air Compliance Analysis Response to Agency Request on Cessation Lag. EPA-COUNCIL-LTR-05-001.
December. Available at
5-105
-------
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2007. SAB Advisory on EPA's
Issues in Valuing Mortality Risk Reduction. EPA-SAB-08-001. October. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2008. Characterizing
Uncertainty in Particulate Matter Benefits Using Expert Elicitation. EPA-COUNCIL-08-002. July. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2009a. Review of EPA 's
Integrated Science Assessment for Particulate Matter (First External Review Draft, December 2008). EPA-
COUNCIL-09-008. May. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2009b. Review of Integrated
Science Assessment for Particulate Matter (Second External Review Draft, July 2009). EPA-CASAC-10-001.
November. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2009c. Consultation on EPA 's
Particulate Matter National Ambient Air Quality Standards: Scope and Methods Plan for Health Risk and
Exposure Assessment. EPA-COUNCIL-09-009. May. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010a. Review of EPA 's DRAFT
Health Benefits of the Second Section 812 Prospective Study of the Clean Air Act. EPA-COUNCIL-10-001.
June. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010b. Review of Risk
Assessment to Support the Review of the Particulate Matter (PM) Primary National Ambient Air Quality
Standards—External Review Draft (September 2009). EPA-CASAC-10-003. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010c. CASACReview of
Quantitative Health Risk Assessment for Particulate Matter—Second External Review Draft (February 2010).
EPA-CASAC-10-008. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010d. CASAC Review of Policy
Assessment for the Review of the PM NAAQS—Second External Review Draft (June 2010). EPA-CASAC-10-
015. Available at
.
5-106
-------
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2011. Review of Valuing
Mortality Risk Reductions for Environmental Policy: A White Paper (December 10, 2010). EPA-SAB-11-Oil
July. Available at
.
World Health Organization (WHO). 1977. International Classification of Diseases, 9th Revision (ICD-9). Geneva:
WHO.
Zanobetti A. and Schwartz, J. 2006. "Air pollution and emergency admissions in Boston, MA." Journal of
Epidemiology and Community Health 60(10): 890-5.
Zanobetti, A; J. Schwartz. 2008. Mortality displacement in the association of ozone with mortality: an analysis of 48
cities in the United States. American Journal of Respiratory and Critical Care Medicine. 177:184-189.
5-107
-------
APPENDIX 5A: COMPREHENSIVE CHARACTERIZATION OF UNCERTAINTY IN
OZONE BENEFITS ANALYSIS
Overview
As noted in Chapter 5, the benefits analysis relies on an array of data inputs—including air
quality modeling, health impact functions and valuation estimates among others—which are
themselves subject to uncertainty and may also in turn contribute to the overall uncertainty in this
analysis. The RIA employs a variety of analytic approaches designed to reduce the extent of the
uncertainty and/or characterize the impact that uncertainty has on the final estimate. We strive to
incorporate as many quantitative assessments of uncertainty as possible (e.g., Monte Carlo
assessments, sensitivity analyses); however, there are some aspects we are only able to characterize
qualitatively.
To more comprehensively and systematically address these uncertainties, including those we
cannot quantify, we adapt the World Health Organization (WHO) uncertainty framework (WHO,
2008), which provides a means for systematically linking the characterization of uncertainty to the
sophistication of the underlying health impact assessment. EPA has applied similar approaches in peer-
reviewed analyses ofPlVh.s-related impacts (U.S. EPA, 2010b, 2011, 2012) and ozone-related impacts
(U.S. EPA, 2014). EPA's Science Advisory Board (SAB) has supported using a tabular format to
qualitatively assess the uncertainties inherent in the quantification and monetization of health impacts,
including identifying potential bias, potential magnitude, confidence in our approach, and the level of
quantitative assessment of each uncertainty (U.S. EPA-SAB, 1999, 2001, 2004a, 2004b, 201 la,
201 Ib). The assessments presented here are largely consistent with those previous peer-reviewed
assessments.
This appendix focuses on uncertainties inherent in the ozone benefits estimates. For more
information regarding the uncertainties inherent in the PM benefits estimates, please see the 2012 PM
NAAQS RIA (U.S. EPA, 2012).
5A.1 Description of Classifications Applied in the Uncertainty Characterization
Table 5A-1 catalogs the most significant sources of uncertainty in the ozone benefits analysis
and then characterizes four dimensions of that uncertainty (listed below). The first two dimensions
focus on the nature of the uncertainty. The third and fourth dimensions focus on the extent to which the
5A-1
-------
analytic approach chosen in the benefits analysis either minimizes the impact of the uncertainty or
quantitatively characterizes its impact.
1) The direction of the bias that a given uncertainty may introduce into the benefits
assessment if not taken into account in the analysis approach;
2) The magnitude of the impact that uncertainty is likely to have on the benefits estimate if
not taken into account in the analysis approach;
3) The extent to which the analytic approach chosen is likely to minimize the impact of
that uncertainty on the benefits estimate; and
4) The extent to which EPA has been able to quantify the residual uncertainty after the
preferred analytic approach has been incorporated into the benefits model.
5A. 1.1 Direction of Bias
The "direction of bias" column in Table 5A-1 is an assessment of whether, if left unaddressed,
an uncertainty would likely lead to an underestimate or overestimate the total monetized benefits. In
some cases we indicate that there are reasons why the bias might go either direction, depending upon
the true nature of the underlying relationship. Where available, we base the classification of the
"direction of bias" on the analysis in the Integrated Science Assessment for Ozone and Related
Photochemical Oxidants (hereafter, "Os ISA") (U.S. EPA, 2013a). Additional sources of information
include advice from SAB and the National Academies of Science (NAS), as well as studies from the
peer-reviewed literature. In some cases we indicate that there is not sufficient information to estimate
whether the uncertainty would likely lead to under or overestimation of benefits; these cases are
identified as "unable to determine."
5A. 1.2 Magnitude of Impact
The "magnitude of impact" column in Table 5A-1 is an assessment of how much plausible
alternative assumptions about the underlying relationship about which we are uncertain could influence
the overall monetary benefits. EPA has applied similar classifications in previous risk and benefit
analyses (U.S. EPA, 201 Ob, 2011, 2014), but we have slightly revised the category names and the cut-
offs here.74 The definitions used here are provided below.
74 In The Benefits and Costs of the Clean Air Act from 1990 to 2020 (U.S. EPA, 2011), EPA applied a classification of
"potentially major" if a plausible alternative assumption or approach could influence the overall monetary benefit
estimate by five percent or more and "probably minor" if an alternative assumption or approach is likely to change the
total benefit estimate by less than five percent. In the Quantitative Health Risk Assessment for Particulate Matter (U.S.
5A-2
-------
• High—if the uncertainty associated with an assumption could influence the total
monetized benefits by more than 25%.
• Medium—if the uncertainty associated with an assumption could influence the total
monetized benefits by 5% to 25%.
• Low—if the uncertainty associated with an assumption could influence the total
monetized benefits by less than 5%.
For each uncertainty, we provide as much quantitative information as is available in the table to
support the classification.
Although many of the sources of uncertainty could affect both morbidity and mortality
endpoints, because mortality benefits comprise over 94% of the monetized benefits that we are able to
quantify in this analysis, uncertainties that affect the mortality estimate have the potential to have
larger impacts on the total monetized benefits than uncertainties affecting only morbidity endpoints.
One morbidity-related uncertainty that could have a significant impact on the benefits estimate is the
extent to which omitted morbidity endpoints are included in the benefits analysis. Including additional
morbidity endpoints that are currently not monetized would reduce the fraction of total benefits from
mortality. Ultimately, the magnitude classification is determined by professional judgment of EPA
staff based on the results of available information, including other U.S. EPA assessments of
uncertainty (U.S. EPA, 2010b, 2011)
Based on this assessment, the uncertainties that we classified as high or medium-high impact
are: the causal relationship between long-term and short-term ozone exposure and mortality, the shape
of the concentration-response function for both categories of ozone-related mortality, and the mortality
valuation, specifically for long-term exposure-related mortality. The classification of these
uncertainties as "high magnitude" is generally consistent with the results of EPA's Influence Analysis
(Mansfield et al, 2009), the Quantitative Health Risk Assessment for Paniculate Matter (U.S. EPA,
2010b), and the Benefits and Costs of the Clean Air Act 1990 to 2020 (U.S. EPA, 2011).
5A. 1.3 Confidence in Analytic Approach
The "confidence in analytic approach" column of Table 5A-1 is an assessment of the scientific
support for the analytic approach chosen (or the inherent assumption made) to account for the
relationship about which we are uncertain. In other words, based on the available evidence, how
EPA, 2010b), EPA applied classifications of "low" if the impact would not be expected to impact the interpretation of
risk estimates in the context of the PM NAAQS review, "medium" if the impact had the potential to change the
interpretation; and "high" if it was likely to influence the interpretation of risk in the context of the PM NAAQS review.
5A-3
-------
certain are we that EPA's selected approach is the most plausible of the potential alternatives. Similar
classifications have been included in previous risk and benefits analyses (U.S. EPA, 2010b, 2011).75
The three categories used to characterize the degree of confidence are:
• High—the current evidence is plentiful and strongly supports the selected approach;
• Medium—some evidence exists to support the selected approach, but data gaps are
present; and
• Low—limited data exists to support the selected approach.
Ultimately, the degree of confidence in the analytic approach is EPA staffs professional
judgment based on the volume and consistency of supporting evidence, much of which has been
evaluated in the O3 ISA (U.S. EPA, 2013a) and by EPA's independent SAB. The O3 ISA evaluated the
entire body of scientific literature on ozone science and was twice peer-reviewed by EPA's Clean Air
Scientific Advisory Committee (CASAC). In general, we regard a conclusion in the Os ISA or specific
advice from SAB as supporting a high degree of confidence in the selected approach.
Based on this assessment, we have low or low-medium confidence in the evidence available to
assess exposure error in epidemiology studies, morbidity valuation, baseline incidence projections for
morbidity, and omitted morbidity endpoints. However, because these uncertainties have been classified
as having a low or low-medium impact on the magnitude of the benefits, further investment in
improving the available evidence would not have a substantial impact on the total monetized benefits.
5A. 1.4 Uncertainty Quantification
The column of Table 5A-1 labeled "uncertainty quantification" is an assessment of the extent to
which we were able to use quantitative methods to characterize the residual uncertainty in the benefits
analysis, after addressing it to the extent feasible in the analytic approach for this RIA. We categorize
the level of quantification using the four tiers used in the WHO uncertainty framework (WHO, 2008).
The WHO uncertainty framework is a well-established approach to assess uncertainty in risk estimates
that systematically links the characterization of uncertainty to the sophistication of the health impact
assessment. The advantage of using this framework is that it clearly highlights the level of uncertainty
quantification applied in this assessment and the potential sources of uncertainty that require methods
75 We have applied the same classification as The Benefits and Costs of the Clean Air Act from 1990 to 2020 (U.S. EPA,
201 la) in this analysis. In the Quantitative Health Risk Assessment for Particulate Matter (U.S. EPA, 2010b), EPA
assessed the degree of uncertainty (low, medium, or high) associated with the knowledge-base (i.e., assessed how well
we understand each source of uncertainty), but did not provide specific criteria for the classification.
5A-4
-------
development in order to assess quantitatively. Specifically, EPA applied this framework in multiple
risk and exposure assessments (U.S. EPA, 201 Ob, 2014), and it has been recommended in EPA
guidance documents assessing air toxics-related risk and Superfund site risks (U.S. EPA, 2004 and
2001, respectively). Ultimately, the tier decision is the professional judgment of EPA staff based on the
availability of information for this assessment. The tiers used in this assessment are defined below.
• Tier 0—screening level, generic qualitative characterization.
• Tier 1—Scenario-specific qualitative characterization.
• Tier 2—Scenario-specific sensitivity analysis.
• Tier 3—Scenario-specific probabilistic assessment of individual and combined
uncertainty.
Within the limits of the data, we strive to use more sophisticated approaches (e.g., Tier 2 or 3)
for characterizing uncertainties that have the largest magnitudes and could not be completely addressed
through the analytic approach. The uncertainties for which we have conducted probabilistic (Tier 3)
assessments in this analysis are mortality causality, the shape of the concentration-response function,
and mortality and morbidity valuation. For lower magnitude uncertainties, we include qualitative
discussions of the potential impact of uncertainty on risk results (WHO Tier 0/1) and/or completed
sensitivity analyses assessing the potential impact of sources of uncertainty on risk results (WHO Tier
2).
5A.2 Organization of the Qualitative Uncertainty Table
Table 5A-1 is organized as follows. The uncertainties are grouped by category (i.e.,
concentration-response function, valuation, population and baseline incidence, omitted benefits
categories, and exposure changes). Within each category, the uncertainties are sorted by magnitude of
impact (i.e., high to low) then by confidence in our approach (i.e., low to high). In the table, red (bold)
text is used to indicate the uncertainties that likely have a high magnitude of impact on the total
benefits estimate. This organization highlights the uncertainty with the largest potential impact and the
lowest confidence at the top of each category.
5A-5
-------
Table 5A-1. Summary of Qualitative Uncertainty for Key Modeling Elements in Ozone Benefits
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Uncertainties Associated with Concentration-Response Functions
Overestimate, if short-term ozone
exposure does not have a causal
relationship with premature
mortality.
High
Causal relationship
between short-term
ozone exposure and
premature mortality
Mortality generally dominates
monetized benefits, so small
uncertainties could have large
impacts on the total monetized
benefits.
High
Our approach is consistent with the Os
Integrated Science Assessment (ISA),
which determined that premature mortality
has a likely causal relationship with short-
term ozone exposure based on the
collective body of evidence (p. 6-264). hi
addition, the NAS recommended that EPA
"should give little or no weight to the
assumption that there is no causal
association between estimated reductions
in premature mortality and reduced ozone
exposure" (NRC, 2008). hi 2010, the
Health Effects Subcommittee of the
Advisory Council on Clean Air
Compliance Analysis, while reviewing
EPA's The Benefits and Costs of the Clean
Air Act 1990 to 2020 (U.S. EPA, 2011),
also confirmed the NAS recommendation
to include ozone mortality benefits (U.S.
EPA-SAB,2010).
Tier 1 (qualitative)
Either
Medium-High
Medium
Tier 2 (sensitivity analysis)
5A-6
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Shape of the C-R
functions, particularly
at low concentrations
for short-term ozone
exposure-related
mortality
The direction of bias that
assuming linear-no threshold
model or alternative model
introduces depends upon the
"true" functional from of the
relationship and the specific
assumptions and data in a
particular analysis. For example, if
the true function identifies a
threshold below which health
effects do not occur, benefits may
be overestimated if a substantial
portion of those benefits were
estimated to occur below that
threshold. Alternately, if a
substantial portion of the benefits
occurred above that threshold, the
benefits may be underestimated
because an assumed linear no-
threshold function may not reflect
the steeper slope above that
threshold to account for all health
effects occurring above that
threshold.
The magnitude of this impact
depends on the fraction of
benefits occurring in areas with
lower concentrations. Mortality
generally dominates monetized
benefits, so small uncertainties
could have large impacts on total
monetized benefits.
The Os ISA did not find any evidence that
supports a threshold in the relationship
between short-term exposure to ozone and
mortality within the range of ozone
concentrations observed in the United
States, and recent evidence suggests that
the shape of the ozone-mortality C-R curve
remains linear across the full range of
ozone concentrations (p. 6-257). Consistent
with the Os ISA, we assume a log-linear
no-threshold model for the concentration-
response functions for short-term ozone
mortality. However, the ISA notes that
there is less certainty in the shape of the C-
R function below 20 ppb due to the low
density of data in this range (p. 6-254-255).
The comparison of short-term
mortality against the associated
distribution of (ozone season-
averaged) 8hr max ozone levels
(see Appendix 5C, section 5C.2)
suggests that the vast majority of
predicted reductions in mortality
are associated with days having
8hr max values that fall within the
range of increased confidence in
specifying the nature of the
mortality response (as identified
in the O3 ISA).
Causal relationship
between long-term
ozone exposure and
Overestimate, if short-term ozone
exposure does not have a causal
relationship with premature
mortality.
High
High
Tier 1 (qualitative)
5A-7
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
premature respiratory
mortality
The total dollar benefits of
reducing ozone levels are
comprised mostly of the value
placed on reducing the risk of
premature death, so small
uncertainties could have large
impacts on the total monetized
benefits.
While the Os ISA concludes that evidence
is suggestive of a causal association
between total mortality and long-term
ozone exposure (section 7.7.1), specifically
with regard to respiratory health effects
(including mortality), the ISA concludes
that there is likely to be a causal
association (section 7.2.8). Furthermore, in
their review of the HREA completed for
the Ozone NAAQS review, the CASAC
expressed their support for EPA's plan to
include a non-threshold-based modeling of
long-term exposure-related respiratory
mortality in the core estimate. The CASAC
also supported EPA's plan to consider the
potential for thresholds in the response (in
the range of 40-60 ppb) as a sensitivity
analysis in the HREA (see section 5.6.3.1).
This same approach (with regard to the
core approach and sensitivity analysis for
long-term exposure-related morality) has
been adopted for the RIA (see below).
Shape of the C-R
functions, particularly
at low concentrations
for long-term ozone
exposure-related
respiratory mortality
Either
Medium-High
Medium
Tier 2 (sensitivity analysis)
5A-8
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
The direction of bias that
assuming linear-no threshold
model or alternative model
introduces depends upon the
"true" functional form of the
relationship and the specific
assumptions and data in a
particular analysis. For example, if
the true function identifies a
threshold below which health
effects do not occur, benefits may
be overestimated if a substantial
portion of those benefits were
estimated to occur below that
threshold. Alternately, if a
substantial portion of the benefits
occurred above that threshold, the
benefits may be underestimated
because an assumed linear no-
threshold function may not reflect
the steeper slope above that
threshold to account for all health
effects occurring above that
threshold.
The magnitude of this impact
depends on the fraction of
benefits occurring in areas with
lower concentrations. Note, that
due to limitations in our ability to
predict the lag structure
associated with reductions in
long-term ozone exposure-related
mortality, we are not including
dollar benefits associated with
this endpoint in the core benefits
analysis (which significantly
reduces the role of this source of
uncertainty in impact core benefit
estimates).
hi their memo (see Sasser 2014) clarifying
the results of their study (Jerrett et al.,
2009) regarding long-term ozone exposure-
related respiratory mortality, the study
authors note that in terms of goodness of
fit, long-term health risk models including
ozone clearly performed better than models
without ozone, indicating the improved
predictions of respiratory mortality when
ozone is included, hi the article proper, the
authors state that, "There was limited
evidence that a threshold model
specification improved model fit as
compared with a non-threshold linear
mode...". Furthermore, in the memo
referenced above, the authors conclude that
considerable caution should be exercised in
using any specific threshold, particularly
when the more stringent statistical test
indicates there is no significantly improved
prediction. The CASAC was supportive of
the approach EPA used in the FfREA of
using a non-threshold C-R function based
on this study to generate core estimates and
consider the impact of potential thresholds
(ranging from 40-60 ppb) as a sensitivity
analysis.
As part of the sensitivity analyses
completed for the RIA, we did
examine the potential impact of
thresholds (from 40 to 60 ppb) in
the response function for long-
term exposure-related morality
(see Appendix 5B, section 5B. 1).
That analysis suggested that
thresholds between 55 and 60 ppb
would have a substantial impact
on overall modeled benefits,
while thresholds below 50 ppb
would have a minor impact on
predicted benefits (see Appendix
5B, Table 5B-1).
Exposure error in
epidemiology studies
Underestimate (generally)
The Os ISA states that exposure
measurement error can also be an
important contributor to
uncertainty in effect estimates
associated with both short-term
and long-term studies (p. Ixii).
Together with other factors (e.g.,
low data density), exposure error
can smooth the C-R functions and
obscure potential thresholds (p.
Ixix). hi addition, the O3 ISA
states that exposure error can bias
effect estimates toward or away
from the null and widen
confidence intervals (p. Ixii).
Medium
Recent analyses reported in
Krewski et al. (2009) demonstrate
the potentially significant effect
that this source of uncertainty can
have on effect estimates. These
analyses also illustrate the
complexity and site-specific
nature of this source of
uncertainty.
Low-Medium
Tier 1 (qualitative)
Although this underestimation is well
documented, including in the Os ISA, the
SAB has not suggested an approach to
adjust for this bias.
(No quantitative method
available)
Unknown
Medium
Medium
Tier 1 (qualitative)
5A-9
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Adjustment of risk
coefficients to 8-hour
maximum
Several of the mortality
epidemiological studies were
reported for a 24-hour average or
1-hour maximum ozone level.
These metrics are not the most
relevant to characterizing
population-level exposure. Thus,
we have converted ozone mortality
health impact functions that use
these metric to maximum 8-hour
average ozone concentration using
standard conversion functions.
The conversion ratios are based on
observed relationships between the
24-hour average and 8-hour
maximum in the underlying
studies.
This conversion also does not
affect the relative magnitude of
the health impact function.
However, we do note that the
pattern of 8hr max concentrations
for a particular location over an
ozone season could differ from
the pattern of Ihr max or 24 hour
average metrics for that same
location. Consequently, estimates
of incidence reductions (and
associated dollar benefits) could
differ for a particular location
depending on the metric used for
risk calculations.76
This practice is consistent both with the
available exposure modeling and with the
form of the current ozone standard.
However, in some cases, these conversions
were not specific to the ozone "warm"
season which was the period used in the
benefits analysis, which introduces
additional uncertainty due to the use of
effect estimates based on a mixture of
warm season and all year data in the
epidemiological studies.
(No quantitative method
available)
Confounding by
individual risk factors,
other than
socioeconomic status—
e.g., smoking, or
ecologic factors, which
represent the
neighborhood, such as
unemployment
Either, depending on the factor
and study
Individual, social, economic, and
demographic covariates can bias
the relationship between
particulate air pollution and
mortality, particularly in cohort
studies that rely on regional air
pollution levels.
Medium
Because mortality dominates
monetized benefits, even a small
amount of confounding could
have medium impacts on total
monetized benefits.
Medium
To minimize confounding effects, we use
risk coefficients that control for individual
risk factors to the extent practical.
Tier 2 (sensitivity analysis)
(Quantitative methods available
but not assessed in this analysis.)
Either, depending upon the
pollutant.
Medium
Medium
Tier 1 (qualitative)
76 We note that if metric adjustment ratios were derived for individual point locations (reflecting the relationship between metrics for each location) and those metrics were used to
generate adjusted C-R functions for each location, then some of this potential uncertainty could be reduced, however available data does not support the application of highly location-
specific adjustment ratios (with cross-city or national adjustment ratios being typically used instead).
5 A-10
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Confounding and effect
modification by co-
pollutants
Disentangling the health responses
of combustion-related pollutants
(i.e., PM, SOx, NOx, ozone, and
CO) is a challenge. The O3 ISA
defines a confounder as the true
cause of the association observed
between the exposure and the
outcome; in contrast, an effect
modifier changes the magnitude of
the association between the
exposure and the outcome (p. Ixii).
Because this uncertainty could
affect mortality and because
mortality generally dominates
monetized benefits, even small
uncertainties could have medium
impacts on total monetized
benefits.
The Os ISA states that there is high
confidence that unmeasured confounders
are not producing the findings when
multiple studies are conducted in various
settings using different subjects or
exposures, such as multi-city studies (p.
Ixi). When modeling effects of pollutants
jointly (e.g., PM and Os), we apply multi-
pollutant effect estimates when those
estimates are available to avoid double-
counting and satisfy other selection criteria.
hi addition, we apply multi-city effect
estimates when available.
(No quantitative method
available)
Application of C-R
relationships only to the
original study
population
Underestimate
Estimating health effects for only
the original study population may
underestimate the whole
population benefits of reductions
in pollutant exposures.
Low
Mortality generally dominates
monetized benefits, so further age
range expansions for morbidity
endpoints would have a small
impact on total monetized
benefits.
High
Following advice from the SAB (U.S.
EPA-SAB, 2004a, pg. 7) and NAS (NRC,
2002, pg. 114), we expanded the age range
for childhood asthma exacerbations beyond
the original study population to ages 6-18.
Tier 2 (sensitivity analysis)
(Quantitative methods available
but not assessed in this analysis.)
Uncertainties Associated with Economic Valuation
Mortality Risk
Valuation/Value-of-a-
Statistical-Life (VSL)
Unknown
Some studies suggest that EPA's
mortality valuation is too high,
while other studies suggest that it
is too low. Differences in age,
income, risk aversion, altruism,
nature of risk (e.g., cancer), and
study design could lead to higher
or lower estimates of mortality
valuation.
High
Mortality generally dominates
monetized benefits, so moderate
uncertainties could have a large
effect on total monetized benefits.
Medium
The VSL used by EPA is based on 26 labor
market and stated preference studies
published between 1974 and 1991. EPA is
in the process of reviewing this estimate
and will issue revised guidance based on
the most up-to-date literature and
recommendations from the SAB-EEAC in
the near future (U.S. EPA, 2010a, U.S.
EPA-SAB, 20lie).
Tier 3 (probabilistic)
Assessed uncertainty in mortality
valuation using a Weibull
distribution.
Underestimate
Medium-High
Low
Tier 2 (sensitivity analysis)
5 A-11
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Cessation lag structure
for long-term ozone
mortality
Jerrett et al. (2009) notes that,
"Allowing for a 10-year period of
exposure to ozone (5 years of
follow-up and 5 years before the
follow-up period) did not
appreciably alter the risk
estimates..." We acknowledge
substantial uncertainty associated
with specifying the lag for long-
term respiratory mortality.
Although the cessation lag does
not affect the number of
premature deaths attributable to
long-term ozone exposure, it
affects the timing of those deaths
and thus the discounted
monetized benefits. Mortality
generally dominates monetized
benefits, so moderate
uncertainties could have a large
effect on total monetized benefits.
As discussed in section 5.6.4.1, in
presenting dollar benefit estimates as part
of the sensitivity analysis (presenting dollar
benefits for long-term ozone-related
morality), we include both an assumption
of zero lag and a lag structure matching
that used for the core PM2.5 estimate (the
SAB 20 year segmented lag). Inclusion of
the zero lag reflects consideration for the
possibility that the long-term respiratory
mortality estimate captures primarily an
accumulation of short-term mortality
effects across the ozone season. The use of
the 20 year segmented lag reflects
consideration for advice provided by the
HES (USEPA-SAB, 2010).
As shown in sensitivity analysis
results presented in Appendix 5B,
the use of the 20 year segmented
lag results in a 10-20% reduction
in the total dollar benefit (using a
3% and 7% discount rate,
respectively) relative to the
alternative approach of applying
no lag (i.e., assuming all of the
mortality reductions occur in the
same year).
Income growth
adjustments
Either
Income growth increases
willingness-to-pay (WTP)
valuation estimates, including
mortality, over time. From 1997 to
2010, personal income and GDP
growth have begun to diverge. If
this trend continues, the
assumption that per capita GDP
growth is a reasonable proxy for
income growth may lead to an
overstatement of benefits, (ffic,
2012).
Medium
Income growth from 1990 to
2020 increases mortality
valuation by 20%. Alternate
estimates for this adjustment vary
by 20% (ffic, 2012). Because we
do not adjust for income growth
over the 20-year cessation lag,
this approach could also
underestimate the benefits for the
later years of the lag.
Medium
Consistent with SAB recommendations
(U.S. EPA,-SAB, 2000, pg. 16), we adjust
WTP for income growth. Difficult to
forecast future income growth. However, in
the absence of readily available income
data projections, per capita GDP is the best
available option.
Tier 2 (sensitivity analysis)
As shown in Appendix 5B, the
use of alternate income growth
adjustments would change the
monetized benefits by +33% to
-14%.
Morbidity valuation
Underestimate
Morbidity benefits such as
hospital admissions are calculated
using cost-of-illness (COI)
estimates, which are generally half
the WTP to avoid the illness
(Alberini and Krupnick, 2000). hi
addition, the morbidity costs do
not reflect physiological responses
or sequelae events, such as
increased susceptibility for future
morbidity.
Low
Even if we doubled the
monetized valuation of morbidity
endpoints using COI valuation
that are currently included in the
RIA, the change would still be
less than 5% of the monetized
benefits. It is unknown how much
including sequelae events could
increase morbidity valuation.
Low
Although the COI estimates for
hospitalizations reflect recent data, we have
not yet updated other COI estimates such
as for school loss days. The SAB
concluded that COI estimates could be
used as placeholders where WTP estimates
are unavailable, but it is reasonable to
presume that this strategy typically
understates WTP values (U.S. EPA-SAB,
2004b, pg. 3).
Tier 3 (probabilistic), where
available
Assessed uncertainty in morbidity
valuation using distributions
specified in the underlying
literature, where available (see
Table 5-10).
Uncertainties Associated with Baseline Incidence and Population Projections
Either
Low-Medium
Medium
Tier 1 (qualitative)
5 A-12
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Population estimates
and projections
The monetized benefits would
change in the same direction as the
over- or underestimate in
population projections in areas
where exposure changes.
Monetized benefits are
substantially affected by
population density. Comparisons
using historical census data show
that population projections are
±5% nationally, but projection
accuracy can vary by locality.
Historical error for Woods &
Poole's population projections
has been ±8.1% for county-level
projections and ±4.1% for states
(Woods and Poole, 2012). The
magnitude of impact on total
monetized benefits depends on
the specific location where PM is
reduced.
We use population projections for 5-year
increments for 304
race/ethnicity/gender/age groups (Woods
and Poole, 2012) at Census blocks.
Population forecasting is well-established
but projections of future migration due to
possible catastrophic events are not
considered. In addition, projections at the
small spatial scales used in this analysis are
inherently more uncertain than projections
at the county- or state-level.
(No quantitative method
available)
Uncertainty in
projecting baseline
incidence rates for
mortality
Unknown
Because the mortality rate
projections for future years reflect
changes in mortality patterns as
well as population growth, the
projections are unlikely to be
biased.
Low-Medium
Because mortality generally
dominates monetized benefits,
small uncertainties could have
medium impacts on total
monetized benefits.
Medium
The county-level baseline mortality rates
reflect recent databases (i.e., 2004-2006
data) and are projected for 5-year
increments for multiple age groups. This
database is generally considered to have
relatively low uncertainty (CDC Wonder,
2008). The projections account for both
spatial and temporal changes in the
population.
Tier 1 (qualitative)
(No quantitative method
available)
Uncertainty in
projecting baseline
incidence rates and
prevalence rates for
morbidity
Either, depending on the health
endpoint
Morbidity baseline incidence is
available for current year only
(i.e., no projections available).
Assuming current year levels can
bias the benefits for a specific
endpoint if the data has clear
trends over time. Specifically,
asthma prevalence rates have
increased substantially over the
past few years while hospital
admissions have decreased
substantially.
Low
The magnitude varies with the
health endpoint, but the overall
impact on the total benefits
estimate from these morbidity
endpoints is likely to be low.
Low-Medium
We do not have a method to project future
baseline morbidity rates, thus we assume
current year levels will continue. While we
try to update the baseline incidence and
prevalence rates as frequently as
practicable, this does not continue trends
into the future. Some endpoints such as
hospitalizations and ER visits have more
recent data (i.e., 2007) stratified by age and
geographic location. Other endpoints, such
as respiratory symptoms reflect a national
average. Asthma prevalence rates reflect
recent increases in baseline asthma rates
(i.e., 2008).
Tier 1 (qualitative)
(No quantitative method
available)
Uncertainties Associated with Omitted Benefits Categories
Underestimate
Medium-High
Low
Tier 2 (sensitivity analysis)
5 A-13
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Unqualified ozone
health benefit
categories, such as
worker productivity
and long-term mortality
EPA has not included monetized
estimates of these benefits
categories in the core benefits
estimate.
Although the potential magnitude
is unknown, including all of the
additional endpoints associated
with ozone exposure that are
currently not monetized could
increase the total benefits by a
large amount.
Current data and methods are insufficient
to value national quantitative estimates of
these health effects. The O3 ISA
determined that respiratory effects
(including mortality) are causally
associated with long-term ozone exposure
(p. 2-17). The O3 ISA also determined that
outdoor workers have an increased risk of
ozone-related health effects (p. 1-15), and
that studies on outdoor workers show
consistent evidence that short-term
increases in ambient ozone exposure can
decrease lung function in healthy adults
(p.6-38). Additional studies link short-term
ozone exposure to reduced productivity in
outdoor workers (Graf Zivin and Neidell,
2013; Crocker andHorst, 1981).
We include sensitivity analyses
reflecting long-term mortality
which shows that this endpoint
could add substantially to the total
core benefits range (see Tables 5-
20 and 5-27). We are still
considering options for updating
the worker productivity analysis
and including it as a sensitivity
analyses.
Uncertainties Associated with Estimated Exposure Changes
Spatial matching of air
quality estimates from
epidemiology studies to
air quality estimates
from air quality
modeling
Unknown
Epidemiology studies often
assume one air quality
concentration is representative of
an entire urban area when
calculating hazard ratios, while
benefits are calculated using air
quality modeling conducted at 12
km spatial resolution. This spatial
mismatch could introduce
uncertainty.
Unknown
Low
We have not controlled for this potential
bias, and the SAB has not suggested an
approach to adjust for this bias.
Tier 1 (qualitative)
(No quantitative method
available)
Uncertainties Associated with the Dollar-per-ton Approach Used in Modeling PM^s Co-benefits
Unknown
Unknown
Medium
Tier 1 (qualitative)
5 A-14
-------
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
Derivation of dollar-
per-ton estimates for
PM25
In the analysis used to generate the
dollar-per-ton values we assume
that all fine particles, regardless of
their chemical composition, are
equally potent in causing
premature mortality. However, the
scientific evidence is not yet
sufficient to allow differentiation
of effect estimates by particle type.
We also assume that the health
impact function for fine particles
is linear down to the lowest air
quality levels modeled in this
analysis. Thus, the estimates
include health benefits from
reducing fine particles in areas
with varied concentrations of
PM2.5, including regions that are in
attainment with the fine particle
standard.
While there are concerns regarding the
assumption of a uniform PM2.5 toxicity
across sectors and a linear C-R function all
the way down to zero, these sources of
uncertainty impact benefits modeling for
PM2.5 in general and have been discussed
as part of earlier RIAs (see section 5.7.2, of
the final PM RIA, U.S. EPA. 2012). We do
recognize that, as discussed below (and on
pp. 24-25 of US. EPA, 2013), there is
increased uncertainty when dollar-per-ton
benefits are applied outside of the specific
scenario used in their derivation.
(No quantitative method
available)
Unknown
Unknown
Medium
Application of dollar-
per-ton estimates in the
current Ozone NAAQS
review
As discussed in section 5.4.4, we
used a method to calculate the
regional benefit-per-ton estimates
that is a slightly modified version
of the national benefit-per-ton
estimates described in the TSD:
Estimating the Benefit per Ton of
Reducing PM2.5 Precursors from
17 Sectors (U.S. EPA, 2013b).
The national estimates were
derived using the approach
published in Farm et al. (2012c),
but they have since been updated
to reflect the epidemiology studies
and Census population data first
applied in the final PM NAAQS
RIA (U.S. EPA, 2012). These
dollar-per-ton estimates were
applied to sector-specific NOx
emissions reductions modeled as
part of the current ozone NAAQS
review.
While we acknowledge
uncertainty associated with
applying dollar-per-ton estimates
in the context of this ozone
NAAQS review (and outside of
the scenario in which they were
derived), we are not in a position
to characterize the magnitude or
direction of any bias that might
result from that application.
All benefit-per-ton estimates have inherent
limitations, including that the estimates
reflect the geographic distribution of the
modeled sector emissions, which may not
match the emissions reductions anticipated
by the proposed standards, and they may
not reflect local variability in population
density, meteorology, exposure, baseline
health incidence rates, or other local factors
for any specific locations reflected in
benefits modeling. However, the fact that
we are modeling regional/national benefits
rather than attempting a more spatially
refined application of the dollar-per-ton
values does reduce uncertainty in the
benefits estimates that are generated.
(No quantitative method
available)
5 A-15
-------
5A.3 References
Alberini, Anna and Alan Krupnick. 2000. "Cost-of-Illness and Willingness-to-Pay Estimates of the Benefits of
Improved Air Quality: Evidence from Taiwan." Land Economics
Vol. 76, No. 1 (Feb), pp. 37-53.
Centers for Disease Control and Prevention (CDC). 2008. National Center for Health Statistics. National Health
Interview Survey, 1999-2008.
Farm, N.; K.R. Baker and C.M. Fulcher. 2012. "Characterizing the PM2 s-related health benefits of emission
reductions for 17 industrial, area and mobile emission sectors across the U.S." Environment International 49
41-151.
Graff Zivin, J., Neidell, M. (2012). "The impact of pollution on worker productivity." American Economic Review,
102, 3652-3673.
Industrial Economics, Incorporated (ffic). 2012. Updating BenMAP Income Elasticity Estimates—Literature
Review. Memo to Neal Farm. March. Available on the Internet at
Jerrett M, Burnett RT, Pope CA, III, et al. 2009. "Long-Term Ozone Exposure and Mortality." New England
Journal of Medicine 360:1085-95.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi, Y, et al. 2009. "Extended follow-up and spatial analysis of
the American Cancer Society study linking paniculate air pollution and mortality." HEI Research Report, 140,
Health Effects Institute, Boston, MA.
Mansfield, Carol; Paramita Sinha; Max Henrion. 2009. Influence Analysis in Support of Characterizing Uncertainty
in Human Health Benefits Analysis: Final Report. Prepared for U.S. EPA, Office of Air Quality Planning and
Standards. November. Available on the internet at
.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press. Washington, DC.
National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. National Academies Press. Washington, DC.
Sasser, E. 2014. Response to Comments Regarding the Potential Use of a Threshold Model in Estimating the
Mortality Risks from Long-term Exposure to Ozone in the Health Risk and Exposure Assessment for Ozone,
Second External Review Draft. Memorandum to Holly Stallworth, Designated Federal Officer, Clean Air
Scientific Advisory Committee from EPA/OAQPS Health and Environmental Impacts Division.
Schwartz J, Coull B, Laden F. 2008. "The Effect of Dose and Timing of Dose on the Association between Airborne
Particles and Survival." Environmental Health Perspectives 116: 64-69.
U.S. Environmental Protection Agency (U.S. EPA). 2001. Risk assessment guidance for Superfund. Vol. Ill, Part A.
Process for conducting probabilistic risk assessment (RAGS 3 A). Washington, DC, United States
Environmental Protection Agency (EPA 540-R-02-002; OSWER 9285.7-45; PB2002 963302. Available on the
Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2004. EPA's risk assessment process for air toxics: History and
overview. In: Air toxics risk assessment reference library. Vol. 1. Technical resource manual. Washington, DC,
5 A-16
-------
United States Environmental Protection Agency, pp. 3-1-3-30 (EPA-453-K-04-001A. Available on the Internet
at.
U.S. Environmental Protection Agency (U.S. EPA). 2010a. Valuing Mortality Risk Reductions for Environmental
Policy: A White Paper (SAB Review Draft). December. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010b. Quantitative Health Risk Assessment for Paniculate
Matter—Final Report. EPA-452/R-10-005. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. September. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 201 la. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March.
Available on the Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Paniculate Matter. EPA-452/R-12-003. Office of Air Quality
Planning and Standards, Health and Environmental Impacts Division. December. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2013a. Integrated Science Assessment of Ozone and Related
Photochemical Oxidants (Final Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-
10/076F. Available at: .
U.S. Environmental Protection Agency (U.S. EPA). 2013b. Technical Support Document: Estimating the Benefit
per ton of Reducing PM2.5 Precursors from 17 sectors. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. February. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2014. Health Risk and Exposure Assessment for Ozone Final
Report. EPA-452/R-14-004a. Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division. August. Available at: < http://www.epa.gov/ttn/naaqs/standards/ozone/data/20140829healthrea.pdf>.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 1999. An SAB Advisory: The
Clean Air Act Section 812 Prospective Study Health and Ecological Initial Studies. EPA-SAB-COUNCIL-
ADV-99-005, Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2000. The Clean Air Act
Amendments (CAAA) Section 812 Prospective Study of Costs and Benefits (1999): Advisory by the Advisory
Council on Clean Air Compliance Analysis: Costs and Benefits of the CAAA. EPA-SAB-COUNCIL-ADV-00-
002. October. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2001. REVIEW OF THE
DRAFT ANALYTICAL PLAN FOR EPA'S SECOND PROSPECTIVE ANALYSIS - BENEFITS AND
COSTS OF THE CLEAN AIR ACT 1990- 2020. EPA-SAB-COUNCIL-ADV-01-004, Available on the
Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2004a. Advisory on Plans for
Health Effects Analysis in the Analytical Plan for EPA's Second Prospective Analysis—Benefits and Costs of
the Clean Air Act, 1990-2020. EPA-SAB-COUNCIL-ADV-04-002. March. Available on the Internet at
5 A-17
-------
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2004b. Review of the Draft
Analytical Plan for EPA's Second Prospective Analysis—Benefits and Costs of the Clean Air Act, 1990-2020:
An Advisory by the Advisory Council for Clean Air Compliance Analysis. EPA-SAB-COUNCIL-ADV-04-
004. May. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2004c. Advisory Council on
Clean Air Compliance Analysis Response to Agency Request on Cessation Lag. EPA-COUNCIL-LTR-05-001.
December. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2010. Review of EPA's
DRAFT Health Benefits of the Second Section 812 Prospective Study of the Clean Air Act. EPA-COUNCIL-
10-001. June. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 201 la. CASAC Review of
Quantitative Health Risk Assessment for Particulate Matter—Second External Review Draft (February 2010).
EPA-CASAC-10-008. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 201 Ib. Review of the Final
Integrated Report for the Second Section 812 Prospective Study of the Benefits and Costs of the Clean Air Act
(August 2010). EPA-COUNCIL-11-001. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 201 Ic. Review of Valuing
Mortality Risk Reductions for Environmental Policy: A White Paper (December 10, 2010). EPA-SAB-11-011.
July. Available on the Internet at
.
Woods & Poole Economics, Inc. 2012. Complete Economic and Demographic Data Source (CEDDS). CD-ROM.
Woods & Poole Economics, Inc. Washington, D.C.
World Health Organization (WHO). 2008. Part 1: Guidance Document on Characterizing and Communicating
Uncertainty in Exposure Assessment, Harmonization Project Document No. 6. Published under joint
sponsorship of the World Health Organization, the International Labour Organization and the United Nations
Environment Programme. WHO Press: Geneva, Switzerland. Available on the Internet at
.
5 A-18
-------
APPENDIX 5B: ADDITIONAL SENSITIVITY ANALYSES RELATED TO THE
OZONE HEALTH BENEFITS ANALYSIS
Overview
The benefits analysis presented in Chapter 5 of this RIA is based on our current
interpretation of the scientific and economic literature. That interpretation requires judgments
regarding the best available data, models, and analytical methodologies and the assumptions that
are most appropriate to adopt in the face of important uncertainties. The majority of the
analytical assumptions used to develop the main estimates of benefits have been reviewed and
supported by EPA's independent Science Advisory Board (SAB). Both EPA and the SAB
recognize that data and modeling limitations as well as simplifying assumptions can introduce
significant uncertainty into the estimates of benefits and that alternative choices exist for some
inputs to the analysis, such as the concentration-response functions for mortality.
This appendix assesses the sensitivity of the core benefits to: (a) the potential impact of
thresholds in long-term ozone exposure-related mortality in incidence and benefits estimates
(section 5B.2), (b) alternative response functions developed through expert elicitation (EE) for
long-term PIVh.s exposure-related mortality (section 5B.3), and (c) alternative assumptions
regarding income elasticity on benefits derived using willingness-to-pay (WTP) functions
(section 5B.3).
For the core analysis, we estimated incidence and dollar benefits for two scenarios: 2025
and post-2025 (see Chapter 5, Sections 5.2 and 5.4.3). However, in conducting these sensitivity
analyses, we used the 2025 scenario as the basis for making our calculations, since sensitivity
analysis findings for this scenario would generally hold for the post-2025 scenario.
In addition to the three sensitivity analyses covered in this appendix, we also included
two sensitivity analyses that are covered in detail in Chapter 5. The first of these are estimates of
dollar benefits associated with mortality resulting from long-term exposure to ozone. As
discussed in Section 5.2, while we felt that we had sufficient confidence to include incidence
estimates associated with long-term ozone exposure (based on effect estimates obtained from
Jerrett et al, 2009) in the core analysis, limitations in our ability to specify an appropriate lag for
5B-1
-------
reductions in this endpoint meant that we could not include dollar benefit estimates in the core
analysis. Instead, we have included those values as sensitivity analyses, including consideration
of alternative lag structures (they are presented in Tables 5-20 and 5-27, respectively for the
2025 and post-2025 scenarios).77 The second of the sensitivity analyses already covered in the
chapter addresses alternative models for short-term ozone exposure-related mortality. As
discussed in Section 5.6.3.1, in addition to the two epidemiological studies (Smith et al, 2009
and Zanobetti and Schwartz 2008) that provide effect estimates for the core benefits estimates,
we have also considered seven additional epidemiological studies as sensitivity analyses. These
additional studies include a mix of multi-city and meta-analysis designs. Results of this
sensitivity analysis are presented in Tables 5-19 and 5-20 (incidence and benefits, respectively
for the 2025 scenario) and Tables 5-26 and 5-27 (incidence and benefits, respectively for the
post-2025 scenario).
5B.1 Threshold Sensitivity Analysis for Premature Mortality Incidence and Benefits from
Long-term Exposure to Ozone
In estimating long-term ozone mortality, we employed a continuous non-threshold
concentration-response (C-R) function relating ozone exposure to premature death. However, as
discussed in Section 5.6.3.1, there is uncertainty regarding the potential existence and location of
a threshold in the C-R function relating mortality and long-term ozone concentrations. Thus, we
have included a sensitivity analysis exploring the impact of potential thresholds in the C-R
relationship on estimates of long-term exposure-related mortality that were evaluated in Jerrett et
al. (2009), consistent with advice from CASAC (Frey, 2014).
In their memo clarifying the results of their study (Sasser, 2014), the authors note that in
terms of goodness of fit, long-term health risk models including ozone clearly performed better
than models without ozone, indicating the improved predictions of respiratory mortality when
ozone is included. In exploring different functional forms, the authors report that the model
including a threshold at 56 ppb had the lowest log-likelihood value of all models evaluated (i.e.,
77 The sensitivity analysis-related benefits estimates presented in these tables include (a) benefits estimates reflecting
application of a zero lag model and (b) estimates reflecting application of the same 20-year segmented lag used in
modeling benefits for PM2s, together with application of a 3% and 7% discount rate. See Sections 5.6.3 and 5.6.4
for additional discussion of lags in relation to long-term ozone-related mortality.
5B-2
-------
linear models and models including thresholds ranging from 40-60 ppb), and thus provided the
best overall statistical fit to the data. However, they also note that it is not clear whether the 56
ppb threshold model is a better predictor of respiratory mortality than when using a linear (no-
threshold) model for this dataset. Using one statistical test, the model with a threshold at 56 ppb
was determined to be statistically superior to the linear model. Using another, more stringent test,
none of the threshold models considered were statistically superior to the linear model. Under the
less stringent test, although the threshold model produces a statistically superior prediction than
the linear model, there is uncertainty about the specific location of the threshold, if one exists.
This is because the confidence intervals on the model predictions indicate that a threshold could
exist anywhere from 0 to 60 ppb. The authors conclude that considerable caution should be
exercised in using any specific threshold, particularly when the more stringent statistical test
indicates there is no significantly improved prediction. Based on this additional information from
the authors, we have chosen to reflect the uncertainty about the existence and location of a
potential threshold by estimating mortality attributable to long-term ozone exposures using a
range of threshold-based effect estimates as sensitivity analyses. Specifically, we estimate long-
term ozone mortality benefits using unique risk coefficients that include a range of thresholds
from 40 ppb to 60 ppb in 5 ppb increments, while also including a model with a threshold equal
to 56 ppb, which had the lowest log-likelihood value for all models examined.78 Table 5B-1
provides the results of these sensitivity analyses (based on modeling incidence) for 60 ppb, 65
ppb and 70 ppb. We note that the same pattern in terms of relative reductions across thresholds
(relative to the core estimate) would hold for the dollar benefit estimates generated for this
endpoint.
! There is a separate effect estimate (and associated standard error) for each of the fitted threshold models estimated
in Jerrett et al. (2009). As a result, the sensitivity of estimated mortality attributable to long-term ozone
concentrations is affected by both the assumed threshold level (below which there is no effect of ozone) and the
effect estimate applied to ozone concentrations above the threshold.
5B-3
-------
Table 5B-1. Long-term Ozone Mortality Incidence at Various Assumed Thresholds
Threshold Concentration
No threshold (core
model)
40 ppb
45 ppb
50 ppb
55 ppb
56 ppb
60 ppb
70 ppb
680
520
410
120
5.6
3.3
<1
65 ppb
2,100
1,600
1,100
280
58
47
10
60 ppb
3,900
2,900
1,700
510
140
105
13
a All estimates rounded to two significant digits.
The results of the sensitivity analysis based on the suite of threshold-based risk
coefficients suggests that threshold models can result in substantially lower estimates of ozone-
attributable long-term mortality. For example, estimated incidence and dollar benefits for long-
term mortality using a model that includes a 55 ppb threshold are approximately 70% less than
long-term mortality benefits estimated using the core co-pollutant non-threshold model.
Generally, estimated long-term mortality benefits are progressively reduced when using models
with increasing thresholds, with the highest threshold considered (60 ppb) removing virtually all
of the estimated incidence reduction and associated benefits.
5B.2 Alternative Concentration-Response Functions for PMi.s-Related Mortality
In modeling PIVh.s cobenefits, we estimate that total dollar benefits are driven largely by
reductions in mortality (see Table 5-22 and 5-29). Therefore, it is particularly important to
attempt to characterize the uncertainties associated with reductions in premature mortality as
modeled in the PIVb.s cobenefits analysis. To better understand the concentration-response
relationship between PIVh.s exposure and premature mortality, the EPA conducted an expert
elicitation in 2006 (Roman et al, 2008; lEc, 2006).79 In general, the results of the expert
elicitation support the conclusion that the benefits of PIVh.s control are very likely to be
substantial.
Alternative concentration-response functions are useful for assessing uncertainty beyond
random statistical error, including uncertainty in the functional form of the model or alternative
study design. In this analysis, we present the results derived from the expert elicitation as
79 Expert elicitation is a formal, highly-structured and well-documented process whereby expert judgments, usually
of multiple experts, are obtained (Ayyub, 2002).
5B-4
-------
indicative of the uncertainty associated with a major component of the health impact functions,
and we provide the independent estimates derived from each of the twelve experts to better
characterize the degree of variability in the expert responses.
In previous RIAs, the EPA presented benefits estimates using concentration-response
functions derived from the PIVb.s Expert Elicitation (Roman et al, 2008) as a range from the
lowest expert value (Expert K) to the highest expert value (Expert E). However, this approach
did not indicate the Agency's judgment on what the best estimate of PIVb.s benefits may be, and
the EPA's independent SAB recommended refinements to the way EPA presented the results of
the elicitation (U.S. EPA-SAB, 2008). Therefore, we began to present the cohort-based studies
(Krewski et al., 2009; Laden et al., 2006)80 as our core estimates in the proposal RIA for the
Portland Cement NESHAP (U.S. EPA, 2009a). Using alternate relationships between PM2.5 and
premature mortality supplied by experts, higher and lower benefits estimates are plausible, but
most of the expert-based estimates of the mean PIVh.s effect on mortality fall between the two
epidemiology-based estimates (Roman et al., 2008). In addition to these studies, we have
included a discussion of other recent multi-state cohort studies conducted in North America, but
we have not estimated benefits using the effect coefficients from these studies (see Appendix
5D). Please note that the benefits estimates results presented are not the direct results from the
studies or expert elicitation; rather, the estimates are based in part on the effect coefficients
provided in those studies or by experts. In addition, because we are using a dollar-per-ton
approach in modeling PIVh.s cobenefits in this RIA, we cannot generate confidence intervals
reflecting the statistical fit characterized in the expert elicitation-based functions.
Even these multiple characterizations based on application of the range of expert
elicitation-based effect estimates omit the contribution to overall uncertainty from uncertainty in
air quality changes, baseline incidence rates, and populations exposed. Furthermore, the
approach presented here does not yet include methods for addressing correlation between input
80 We have since updated the Harvard Six Cities cohort study from Laden et al. (2006) to use the most recent follow-
up publication of this cohort (Lepeule et al, 2012). This study is reflected in the dollar-per-ton values used in this
RIA.
5B-5
-------
parameters and the identification of reasonable upper and lower bounds for input distributions
characterizing uncertainty in additional model elements.
The PM2.5 expert elicitation and the derivation of effect estimates from the expert
elicitation results (used in generating these alternative dollar-per-ton estimates) are described in
detail in the 2006 PM2.5 NAAQS RIA (U.S. EPA, 2006), the elicitation summary report (lEc,
2006) and Roman et al. (2008), and consequently, we do not present those effect estimates (and
associated functional forms) here.
Table 5B-2 presents the results of this sensitivity analysis as completed for the 2025
scenario (overall conclusions generated from this analysis are transferable to the post-2025
scenario). The alternative mortality estimates presented in Table 5B-2 were generated similar to
the core cobenefits PIVb.s mortality estimates, but applying dollar-per-ton values (for each expert
elicitation-based effect estimate) to the sector-level estimates of NOx reductions associated with
the 2025 scenario. We have also included the core cobenefits estimates for each alternative
standard to facilitate comparison against these alternative sensitivity analysis estimates. Because
application of these effect estimates (using a dollar-per-ton approach) represents a linear
calculation, results in terms of the relative ranking of the core estimates compared with expert
elicitation estimates will remain the same for all alternative standards evaluated. Therefore, we
only present estimates for the 70 ppb alternative standard, observing that observations drawn
from this sensitivity analysis would hold for other alternative standards considered (and for the
post-2025 scenario - for all alternative standards - as well).
5B-6
-------
Table 5B-2. Application of Alternative (Expert Elicitation-Based Effect Estimates) to the
Modeling of PMi.s Co-benefit Estimates for PMi.s (avoided incidence)
C-R Function (and effect estimate)13
Krewski et al., (2012) (core model)
Lepeule et al., (2012) (core model)
Expert K
Expert G
Expert L
Expert D
Expert H
Expert J
Expert F
Expert C
Expert I
Expert B
Expert A
Expert E
70 ppb
45
100
10
54
63
65
67
75
88
92
92
95
120
150
a All estimates rounded to two significant digits.
b Expert elicitation-based values ordered by magnitude of incidence reduction
The values presented in Table 5B-2 suggest that the two core incidence estimate fall
within the range of alternative C-R function based estimates obtained through expert elicitation.
This increases overall confidence in the core estimates with regard to the form of the functions
and magnitude of the effect estimates. We would expect the relationship between the core
estimates and the expert-derived estimates to remain constant for the remaining scenarios as
well.
5B.3 Income Elasticity of Willingness-to-Pay
As discussed in Chapter 5, our estimates of monetized benefits account for growth in real
GDP per capita by adjusting the WTP for individual endpoints based on the central estimate of
the adjustment factor for each of the categories (minor health effects, severe and chronic health
effects, premature mortality, and visibility). We examined how sensitive the estimate of total
benefits is to alternative estimates of the income elasticities. Income growth projections are only
currently available in BenMAP through 2024, so both the 2025 and post-2025 scenario estimates
use income growth only through 2024 and are therefore likely underestimates.
5B-7
-------
Table 5B-3 lists the ranges of elasticity values used to calculate the income adjustment
factors, while Table 5B-4 lists the ranges of corresponding adjustment factors. The results of this
sensitivity analysis, giving the monetized benefit subtotals for the four benefit categories, are
presented in Table 5B-5.
Table 5B-3. Ranges of Elasticity Values Used to Account for Projected Real Income
Growth a
Lower Sensitivity Upper Sensitivity
Benefit Category Bound Bound
Minor Health Effectb 0.04 0.30
Premature Mortality 0.08 1.00
a Derivation of these ranges can be found in Kleckner and Neumann (1999). COI estimates are assigned an
adjustment factor of 1.0.
b Minor health effects included in this RIA and valued using WTP-based functions include: upper and lower
respiratory symptoms, asthma exacerbations, minor restricted activity days, and acute bronchitis.
Table 5B-4. Ranges of Adjustment Factors Used to Account for Projected Real Income
Growth to 2024 a
Lower Sensitivity Upper Sensitivity
Benefit Category Bound Bound
Minor Health Effectb 1.021 1.170
Premature Mortality 1.043 1.705
a Based on elasticity values reported in Table C-4, U.S. Census population projections, and projections of real GDP
per capita.
b Minor health effects included in this RIA and valued using WTP-based functions include: upper and lower
respiratory symptoms, asthma exacerbations, minor restricted activity days, and acute bronchitis.
Table 5B-5. Sensitivity of Monetized Ozone Benefits to Alternative Income Elasticities in
2025 (Millions of 2011$)
Benefit Category
Minor Health Effect b
Premature Mortality °
No adjustment
70 ppb 65 ppb
$66
$2,000
$200
$6,400
Lower Sensitivity Bound
70 ppb 65 ppb
$67
$2,100
$210
$6,600
Upper Sensitivity Bound
70 ppb 65 ppb
$77
$3,500
$240
$11,000
a All estimates rounded to two significant digits. Only reflects income growth to 2024.
b For purposes of completing this sensitivity analysis, we have included minor restricted activity days (MRADS)
based resulting from short-term ozone exposure as the minor health effect evaluated here.
0 Using short-term mortality effect estimate from Smith et al. (2009) and 3% discount rate. Results using other short-
term mortality studies and a 7% discount rate would show the same proportional range.
Consistent with the impact of mortality on total benefits, the adjustment factor for
mortality has the largest impact on total benefits. The value of mortality in 2025 ranges from
86% to 133% of the main estimate for mortality based on the lower and upper sensitivity bounds
on the mortality income adjustment factor. The effect on the value of minor health effects is
5B-8
-------
much less pronounced, ranging from 96% to 108% of the main estimate for minor effects. These
observations (in terms of relative impact from alternative elasticities) hold for all three of the
alternative standard levels evaluated under both the 2025 and post-205 scenarios.
5B.4 References
Frey, C. 2014. CASAC Review of the EPA's Health Risk and Exposure Assessment for Ozone (Second External
Review Draft - February, 2014). U.S. Environmental Protection Agency Science Advisory Board. EPA-
CASAC-14-XXX.Gent, IF.; E.W. Triche; T.R. Holford; K. Belanger; M.B. Bracken; W.S. Beckett, et al. 2003.
Association of low-level ozone and fine particles with respiratory symptoms in children with asthma. Journal of
the American Medical Association. 290(14): 1859-1867.
Industrial Economics, Incorporated (ffic). 2006. Expanded Expert Judgment Assessment of the Concentration-
Response Relationship Between PM2s Exposure and Mortality. Prepared for: Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC. September. Available at
.
Jerrett M, Burnett RT, Pope CA, III, et al. 2009. "Long-Term Ozone Exposure and Mortality." New England
Journal of Medicine 360:1085-95.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi, Y, et al. 2009. Extended follow-up and spatial analysis of
the American Cancer Society study linking particulate air pollution and mortality. HEI Research Report, 140,
Health Effects Institute, Boston, MA.
Kleckner, N., and J. Neumann. 1999. "Recommended Approach to Adjusting WTP Estimates to Reflect Changes in
Real Income." Memorandum to Jim DeMocker, US EPA/OPAR. June 3, Available on the Internet at
.
Lepeule J, Laden F, Dockery D, Schwartz J. 2012. "Chronic Exposure to Fine Particles and Mortality: An Extended
Foliow-Up of the Harvard Six Cities Study from 1974 to 2009." Environmental Health Perspectives 120
(7):965-70.
Roman, Henry A., Katherine D. Walker, Tyra L. Walsh, Lisa Conner, Harvey M. Richmond, Bryan J. Hubbell, and
Patrick L. Kinney. 2008. "Expert Judgment Assessment of the Mortality Impact of Changes in Ambient Fine
Particulate Matter in the U.S." Environmental Science & Technology 42(7):2268-2274.
Smith, R.L.; B. Xu and P. Switzer. 2009. Reassessing the relationship between ozone and short- term mortality in
U.S. urban communities. Inhalation Toxicology. 21:37-61.
U.S. Environmental Protection Agency (U.S. EPA). 2009. Regulatory Impact Analysis: National Emission
Standards for Hazardous Air Pollutants from the Portland Cement Manufacturing Industry. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. April. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. EPA-452/R-12-003. Office of Air Quality
Planning and Standards, Health and Environmental Impacts Division. December. Available at:
.
U.S. Environmental Protection Agency—Science Advisory Board (EPA-SAB). 1999a. The Clean Air Act
Amendments (CAAA) Section 812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and
Ecological Effects Subcommittee on Initial Assessments of Health and Ecological Effects. Part 1. EPA-SAB-
5B-9
-------
COUNCIL-ADV-99-012. July. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (EPA-SAB). 1999b. The Clean Air Act
Amendments (CAAA) Section 812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and
Ecological Effects Subcommittee on Initial Assessments of Health and Ecological Effects. Part 2. EPA-SAB-
COUNCIL-ADV-00-001. October. Available on the Internet at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2004b. Advisory Council on
Clean Air Compliance Analysis Response to Agency Request on Cessation Lag. EPA-COUNCIL-LTR-05-001
December. Available on the Internet at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2008. Characterizing
Uncertainty in Particulate Matter Benefits Using Expert Elicitation. EPA-COUNCIL-08-002. July. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010. Review of EPA 's DRAFT
Health Benefits of the Second Section 812 Prospective Study of the Clean Air Act. EPA-COUNCIL-10-001. June.
Available on the Internet at
.
Zanobetti, A; J. Schwartz. 2008. Mortality displacement in the association of ozone with mortality: an analysis of 48
cities in the United States. American Journal of Respiratory and Critical Care Medicine. 177:184-189.
5B-10
-------
APPENDIX 5C: SUPPLEMENTAL ANALYSES RELATED TO THE OZONE HEALTH
BENEFITS ANALYSIS
Overview
A number of additional analyses have been completed to supplement the core estimates
generated for the RIA. These analyses give greater insight to the manner in which these impacts
are distributed among populations of different ages, the ambient levels of ozone at which the
avoided deaths are estimated to occur, and the benefits attributable to the known emissions
control measures. These supplemental analyses, which are presented in detail here (and
summarized in Section 5.7.3.2) include: (a) age group-differentiated aspects of short-term ozone
exposure-related mortality (including total avoided incidence, life years gained and percent
reduction in baseline mortality - Section 5C.1), (b) evaluation of mortality impacts relative to the
baseline pollutant concentrations (used in generating those mortality estimates) for both short-
term ozone exposure-related mortality and long-term PIVh.s exposure-related mortality (Section
5C.2)81 and (c) presentation of core incidence and benefits estimates reflecting application of
known controls for the 2025 scenario (Section 5C.3).
5C.1 Age Group-Differentiated Aspects of Short-Term Ozone Exposure-Related
Mortality
In their 2008 review of the EPA's approach to estimating ozone-related mortality
benefits, NRC indicated, "EPA should consider placing greater emphasis on reporting decreases
in age-specific death rates in the relevant population and develop models for consistent
calculation of changes in life expectancy and changes in number of deaths at all ages" (NRC,
2008). In addition, NRC noted in an earlier report that "[fjrom a public-health perspective, life-
years lost might be more relevant than annual number of mortality cases" (NRC, 2002). This
advice is consistent with that of the Health Effects Subcommittee of the Advisory Council on
Clean Air Compliance Analysis (SAB-HES), which agreed that ".. .the interpretation of mortality
risk results is enhanced if estimates of lost life-years can be made" (U.S. EPA-SAB, 2004a). To
address these recommendations, we use simplifying assumptions to estimate the number of life
81 The plot of long-term PM2 5 exposure-related mortality incidence versus PM2 5 levels is taken from previous RIAs
and reflects the benefits simulation used to generate the dollar-per-ton estimates used in deriving PM2 s-related
cobenefits for this analysis (i.e., these plots are not derived using new data specific to this PJA) (see Section 5C.3).
5C-1
-------
years that might be gained. We also estimate the reduction in the percentage of deaths attributed
to ozone resulting from the illustrative emissions reduction strategies to reach the proposed and
alternative primary standards. The EPA included similar estimates of life years gained in a
previous assessment of ozone and/or PIVh.s benefits (U.S. EPA, 2006, 2010c, 201 Ib), the latter of
which was peer reviewed by the SAB-HES (U.S. EPA-SAB, 2010a).
Changes in life years and changes in life expectancy at birth are frequently conflated,
thus it is important to distinguish these two very different metrics. Life expectancy varies by age.
The Centers for Disease Control and Prevention (CDC) defines life expectancy as the "average
number of years of life remaining for persons who have attained a given age" (CDC, 2011). In
other words, changes in life expectancy refer to an average change for the entire population, and
refer to the future. Over the past 50 years, average life expectancy at birth in the U.S. has
increased by 8.4 years (CDC, 2001). For example, life expectancy at birth was estimated in 2007
to be 77.9 years for an average person born in the U.S., but for people surviving to age 60,
estimated life expectancy is 82.5 years (i.e., 4.6 years more than life expectancy at birth) (CDC,
2011). Life years, on the other hand, measure the amount of time that an individual loses if they
die before the age of their life expectancy. Life years refer to individuals, and refer to the past,
e.g., when the individual has already died. If a 60-year old individual dies, we estimate that this
individual would lose about 22.5 years of life (i.e., the average population life expectancy for an
individual of this age minus this person's age at death).
Due to the use of benefit-per-ton estimates for the PIVb.s co-benefits, we are unable to
estimate the life years gained by reducing exposure to PIVh.s in this analysis. Instead, we refer the
reader to the 2012 PM NAAQs RIA for more information about the avoided life years lost from
PM2.5 exposure (U.S. EPA, 2012b). This analysis found that about half of the avoided PM-related
deaths occur in populations age 75 to 99, but half of the avoided life years lost would occur in
populations younger than 65 because the younger populations have the potential to lose more life
years per death than older populations. In addition, this analysis found that the average individual
who would otherwise have died prematurely from PM exposure would gain 16 additional years
of life.
5C-2
-------
Estimated Life Years Gained
For estimating the potential life years gained by reducing exposure to ozone in the U.S.
adult population, we use the same general approach as Hubbell (2006) and Fann et al. (2012a).
We have not estimated the change in average life expectancy at birth in this RIA. Because life
expectancy is an average of the entire population (including both those whose deaths would
likely be attributed to air pollution exposure as well as those whose deaths would not), average
life expectancy changes associated with air pollution exposure would be expected to always be
significantly smaller than the average number of life years lost by an individual who is projected
to die prematurely from air pollution exposure.
To estimate the potential distribution of life years gained for population subgroups
defined by the age range at which their reduction in air pollution exposure is modeled to occur,
we use standard life tables available from the CDC (2014) and the following formula:
Total Life Years = I?=1LŁ; x Mt (5.2)
where LE; is the average remaining life expectancy for age interval i, M; is the estimated change
in number of deaths in age interval i, and n is the number of age intervals.
To get M; (the estimated number of avoided premature deaths attributed to changes in
ozone exposure for the 2025 scenario), we use a health impact function that incorporates risk
coefficients estimated for the adult population in the U.S. and age-specific mortality rates. That
is, we use risk coefficients that do not vary by age, but use baseline mortality rates that do.
Because mortality rates for younger populations are much lower than mortality rates for older
populations, most but not all, of the avoided deaths tend to be in older populations. Table 5C-1
summarizes the number of avoided deaths (by age range) attributable to ozone for each
alternative standard for the 2025 scenario. Table 5C-2 summarizes the modeled number of life
years gained (for each age range) by reducing ozone for each alternative standard evaluated for
the 2025 scenario. We then calculated the average number of life years gained per avoided
premature mortality.
5C-3
-------
Table 5C-1. Potential Reduction in Premature Mortality by Age Range from Attaining
Alternate Ozone Standards (2025 scenario) a'b
Age Range b
0-4
5-9
10-14
15-19
20-24
25-29
30-34
35-44
45-54
55-64
65-74
75-84
85-99
Total ozone-attributable mortality
70 ppb
0.65
0.14
0.15
0.21
0.3
0.61
0.63
3.4
8.6
22
45
60
59
200
Standard Alternative
65 ppb
1.9
0.44
0.46
0.65
0.9
1.8
1.9
11
27
67
140
190
190
630
60 ppb
3.6
0.81
0.85
1.2
1.7
3.4
3.5
19
49
120
260
340
340
1,100
a Estimates rounded to two significant figures.
b Effects calculated using the core Smith et al. (2009) effect estimate for the 2025 scenario
Table 5C-2. Potential Years of Life Gained by Age Range from Attaining Alternate
Ozone Standards (2025 Scenario) a'b
Age Range b
0-4
5-9
10-14
15-19
20-24
25-29
30-34
35-44
45-54
55-64
65-74
75-84
85-99
Total life years gained
Average life years gained per individual
70 ppb
51
10
10
13
16
30
28
120
230
410
550
390
130
2,000
10.0
Standard Alternative
65 ppb
150
30
30
39
49
91
86
380
720
1,300
1,700
1,200
430
6,200
9.93
60 ppb
280
56
55
71
90
170
160
690
1,300
2,300
3,100
2,300
790
11,000
9.92
a Estimates rounded to two significant figures (except for average life years gained - presented to three significant
figures to allow differences across values to be evident.
b Effects calculated using the core Smith et al. (2009) effect estimate for the 2025 scenario
By comparing the projected age distribution of the avoided premature deaths with the age
distribution of life years gained, we observed that about half of the deaths occur in populations
age 75-99 (see Table 5C-1), but half of the life years would occur in populations younger than
5C-4
-------
65 (see Table 5C-2). This is because the younger populations have the potential to lose more life
years per death than older populations based on changes in ozone exposure for the 2025 scenario.
We estimate that the average individual who would otherwise have died prematurely from ozone
exposure would gain 10 additional years of life. However, this approach does not account for
whether or not people who are older are more likely to be susceptible to the health effects of air
pollution or whether that susceptibility was in and of itself caused by air pollution exposure (for
a more complete discussion of this issue, see Kunzli et al, 2001).
Percent of Ozone-related Mortality Reduced
To estimate the percentage reduction in all-cause mortality attributed to reduced ozone
exposure for the 2025 scenario as a result of the illustrative emissions reduction strategies, we
use M; from the equation above, dividing the number of excess deaths estimated for each
alternative standard by the total number of deaths in each county. Table 5C-3 shows the
reduction in all-cause mortality attributed to reducing ozone exposure to the proposed primary
standards for the 2025 scenario.
Table 5C-3. Estimated Percent Reduction in All-Cause Mortality Attributed to the
Proposed Primary Ozone Standards (2025 Scenario)
Standard Alternative
Age Range b
0-4
5-9
10-14
15-19
20-24
25-29
30-34
35-44
45-54
55-64
65-74
75-84
85-99
70 ppb
0.0183%
0.0176%
0.0178%
0.0180%
0.0177%
0.0177%
0.0176%
0.0174%
0.0174%
0.0179%
0.0179%
0.0176%
0.0164%
65 ppb
0.0547%
0.0538%
0.0542%
0.0546%
0.0536%
0.0535%
0.0535%
0.0535%
0.0536%
0.0554%
0.0557%
0.0549%
0.0522%
60 ppb
0.101%
0.100%
0.100%
0.101%
0.099%
0.099%
0.098%
0.099%
0.099%
0.101%
0.102%
0.101%
0.096%
a In order to illustrate the slight variations in percent reductions across age ranges (for a given alternative standard
level) we have presented results to three significant figures (rather than two as is typically done for other estimates
in this RIA).
5C-5
-------
Results presented in Table 5C-3 highlight that when reductions in ozone-attributable
mortality (in going from baseline to an alternative standard level) are considered as a percentage
of total all-cause baseline mortality, the estimates are relatively small and are fairly constant
across age ranges. However, it is important to point out that estimates of total ozone-attributable
mortality represent a substantially larger fraction of all-cause baseline mortality.
5C.2 Evaluation of Mortality Impacts Relative to the Baseline Pollutant Concentrations
(used in generating those mortality estimates) for both Short-Term Ozone
Exposure-Related Mortality and Long-Term PMi.s Exposure-Related Mortality
Analysis of baseline ozone levels used in modeling short-term ozone exposure-related mortality
Our review of the current body of scientific literature indicates that a log-linear no-
threshold model provides the best estimate of ozone-related short-term mortality (see section
2.5.4.4, in the O3 ISA, U.S. EPA, 2013a), which was reviewed by the EPA's Clean Air Scientific
Advisory Committee. Consistent with this finding, we estimate benefits associated with the full
range of ozone exposure. Our confidence in the estimated number of premature deaths avoided
(but not in the existence of a causal relationship between ozone and premature mortality)
diminishes as we estimate these impacts at successively lower concentrations. However, there
are uncertainties inherent in identifying any particular point at which our confidence in reported
associations becomes appreciably less, and the scientific evidence provides no clear dividing
line. The Os ISA noted that the studies indicate reduced certainty in specifying the shape of the
C-R function specifically for short-term ozone-attributable respiratory morbidity and mortality,
in the range generally below 20 ppb (for these reasons, the < 20 ppb range discussed in the Os
ISA should be viewed as a more generalized range to be considered qualitatively or semi-
quantitatively, along with many other factors, when interpreting the risk estimates rather than as
a fixed, bright-line).82
82 While clinical studies have suggested the presence of a threshold for respiratory effects, these should not be used
to support specification of population-level thresholds for use in the epidemiological-based risk assessment
focusing on short-term exposure-related endpoints. The clinical studies focus on relatively small and clearly
defined populations of healthy adults, which are not representative of the broader residential populations typically
associated with epidemiological studies, including older individuals and individuals with existing health
conditions that place them at greater risk for ozone-related effects. Therefore, the clinical studies are unlikely to
have the power to capture population thresholds in a broader and more diverse urban residential population,
should those thresholds exist.
5C-6
-------
Figures 5C-1 and 5C-2 compare the distribution of short-term ozone exposure-related
mortality to the underlying distribution of 8hr max baseline ozone levels used in generating those
estimates (these two plots present probability and cumulative probability plots, respectively).
Both figures are based on the core estimate of short-term mortality generated using effect
estimates obtained from Smith et al, 2009.83 In addition, each figure includes separate plots for
the three alternative standard levels being analyzed (with all being based on the 2025 scenario).
If we look at Figure 5C-1, we see that approximately 45% of the mortality benefit estimated for
the 65 ppb alternative standard is associated with days having baseline ozone level of between 40
and 45 ppb.84
01
c
o
OJ
00
ra
4->
c
0)
u
l_
OJ
Q_
Ozone season mean daily 8hr max baseline (ppb)
Figure 5C-1. Premature Ozone-related Deaths Avoided for the Alternative Standards
(2025 scenario) According to the Baseline Ozone Concentrations
83 The set of 12km-level mortality estimates (and associated 8hr max baseline values) generated using BenMAP
forms the basis for the plots.
84 As noted later in this section, this baseline range is actually for the mean across the ozone season of 8hr max
values within a given grid cell, so the actual distribution of baseline 8hr max values associated with this segment of
benefits reductions is likely wider than the 40-45 ppb range.
5C-7
-------
1.2
~ 0.8
0.6
O +J
d) JS
DO 0)
(D i-
= c
s s
i_ M
0) O
a E
0)
'43 +-"
JS
3
E
^
u
0.4
0.2
Ozone season mean daily 8hr max baseline (ppb)
Figure 5C-2. Cumulative Probability Plot of Premature Ozone-related Deaths Avoided for
the Alternative Standards (2025 scenario) According to the Baseline Ozone
Concentrations
When interpreting these results, it is important to understand that the avoided ozone-
related deaths are estimated to occur from ozone reductions in the baseline air quality simulation,
which assumes that 75 ppb is already met. When simulating attainment with proposed and
alternative standards, we adjust the design value at each monitor exceeding the standard
alternative to equal that standard and use an air quality interpolation technique to simulate the
change in ozone concentrations surrounding that monitor. This technique tends to simulate the
greatest air quality changes nearest the monitor. We estimate benefits using modeled air quality
data with 12 km grid cells, which is important because the grid cells are often substantially
smaller than counties and ozone concentrations vary spatially within a county. Therefore, there
may be a small number of grid cells with concentrations slightly greater than 75 ppb in the
gridded baseline even though all monitors could meet an annual standard of 75 ppb. In addition,
some grid cells in a county can be below the level of a standard even though the highest monitor
value is above that standard. Thus, emissions reductions can lead to benefits in grid cells that are
below a standard even within a county with a monitor that exceeds that standard. Furthermore,
our approach to simulating attainment can lead to benefits in counties that are below the
5C-8
-------
alternative standard being evaluated. Emissions reduction strategies designed to reduce ozone
concentrations at a given monitor will frequently improve air quality in neighboring counties. In
order to make a direct comparison between the benefits and costs of these emissions reduction
strategies, it is appropriate to include all the benefits occurring as a result of the emissions
reduction strategies applied, regardless of where they occur. Therefore, it is not appropriate to
estimate the fraction of benefits that occur only in counties that exceed the alternative standards
because it would omit benefits attributable to emissions reductions in exceeding counties.
One final caveat in interpreting the information presented in these figures is that in
modeling this mortality endpoint, rather than using a true distribution of daily 8hr max ozone
levels for each grid cell, due to resource limitations, we used a single mean value for the ozone
season within each grid cell. While this will generate the same total ozone benefit estimate for
each grid cell compared with application of a full distribution of daily 8hr max values, use of a
mean daily value means that an assessment such as this one that considers both the spatial and
temporal association between mortality benefit estimates and ozone levels, will be limited
somewhat in its treatment of the temporal dimension.
Consideration for the plots presented in Figures 5C-1 and 5C-2 results in a number of
observations. The vast majority of reductions in short-term exposure-related mortality for ozone
occur in grid cells with mean 8hr max baseline levels (across the ozone season) between 35 and
55ppb. Importantly, virtually all of the mortality reductions are associated with ozone levels
above the <20ppb range identified within the Ch ISA as being associated with less confidence in
specifying the nature of the C-R function for ozone mortality (Cb ISA, section 2.S.4.4).85 We also
note that as we compare patterns across the three alternative standard levels, we see that, as
expected, the upper end of the distribution is being shifted downwards as increasingly lower
standard levels are analyzed (see Figure 5C-2).
85 As noted earlier, care needs to be taken in interpreting these mortality vs. ozone air level distributions in the
context of the range of reduced confidence (<20ppb) identified in the Os ISA. The region of reduced confidence
identified by the ISA reflects the composite monitor daily time series values (including 8hr max values) used in
short-term mortality studies, while the ozone levels summarized in Figures 5C-1 and 5C-2 are the mean (across the
ozone season) of daily 8hr max values within each grid cell. The use of these mean values, while not impacting the
total mortality reductions estimated, could significantly reduce variability in the spread of values presented in these
two figures.
5C-9
-------
Concentration Benchmark Analysis for PIVh.s Benefit-per-ton Estimates
In general, we are more confident in the magnitude of the risks we estimate from simulated
PM2.5 concentrations that coincide with the bulk of the observed PM concentrations in the
epidemiological studies that are used to estimate the benefits. Likewise, we are less confident in
the risk we estimate from simulated PIVh.s concentrations that fall below the bulk of the observed
data in these studies. Concentration benchmark analyses (e.g., lowest measured level [LML], one
standard deviation below the mean of the air quality data in the study, etc.) allow readers to
determine the portion of population exposed to annual mean PIVh.s levels at or above different
concentrations, which provides some insight into the level of uncertainty in the estimated PIVh.s
mortality benefits. In this analysis, we apply two concentration benchmark approaches (LML and
one standard deviation below the mean) that have been incorporated into recent RIAs and EPA's
Policy Assessment for Particulate Matter (U.S. EPA, 201 Id). There are uncertainties inherent in
identifying any particular point at which our confidence in reported associations becomes
appreciably less, and the scientific evidence provides no clear dividing line. However, the EPA
does not view these concentration benchmarks as a concentration threshold below which we
would not quantify health co-benefits of air quality improvements.86 Rather, the co-benefits
estimates reported in this RIA are the best estimates because they reflect the full range of air
quality concentrations associated with the emissions reduction strategies. The PM ISA concluded
that the scientific evidence collectively is sufficient to conclude that the relationship between
long-term PM2.5 exposures and mortality is causal and that overall the studies support the use of a
no-threshold log-linear model to estimate PM-related long-term mortality (U.S. EPA, 2009b).
For this analysis, policy-specific air quality data is not available, and the compliance
strategies are illustrative of what states may choose to do. For this RIA, we are unable to
estimate the percentage of premature mortality associated with the emissions reductions at each
PM2.5 concentration, as we have done for previous rules with air quality modeling (e.g., U.S.
EPA, 201 Ib, 2012a). However, we believe that it is still important to characterize the distribution
of exposure to baseline concentrations. As a surrogate measure of mortality impacts, we provide
86 For a summary of the scientific review statements regarding the lack of a threshold in the PM2 5-mortality
relationship, see the TSD entitled Summary of Expert Opinions on the Existence of a Threshold in the
Concentration-Response Function for PM2 s-related Mortality (U.S. EPA, 2010b).
5C-10
-------
the percentage of the population exposed at each PIVh.s concentration in the baseline of the
source apportionment modeling used to calculate the benefit-per-ton estimates for this sector
using 12 km grid cells across the contiguous U.S.87 It is important to note that baseline exposure
is only one parameter in the health impact function, along with baseline incidence rates,
population and change in air quality. In other words, the percentage of the population exposed to
air pollution below the LML is not the same as the percentage of the population experiencing
health impacts as a result of a specific emissions reduction policy. The most important aspect,
which we are unable to quantify without rule-specific air quality modeling, is the shift in
exposure anticipated by implementing the proposed standards. Therefore, caution is warranted
when interpreting the LML assessment in this RIA because these results are not consistent with
results from RIAs that had air quality modeling.
Table 5C-4 provides the percentage of the population exposed above and below two
concentration benchmarks (i.e., LML and one standard deviation below the mean) in the
modeled baseline for the sector modeling. Figure 5C-3 shows a bar chart of the percentage of the
population exposed to various air quality levels in the baseline, and Figure 5C-4 shows a
cumulative distribution function of the same data. Both figures identify the LML for each of the
major cohort studies.
7
As noted above, the modeling used to generate the benefit-per-ton estimates does not reflect emissions reductions
anticipated from MATS rule. Therefore, the baseline PM2 5 concentrations in the LML assessment are higher than
would be expected if MATS was reflected.
5C-11
-------
Table 5C-4. Population Exposure in the Baseline Sector Modeling (used to generate the
benefit-per-ton estimates) Above and Below Various Concentrations
Benchmarks in the Underlying Epidemiology Studies a
Epidemiology Study
Below 1 Standard
Deviation.
Below AQ Mean
At or Above 1
Standard Deviation
Below AQ Mean
Below LML
At or Above LML
Krewski et al. (2009)
Lepeuleetal. (2012)
89%
N/A
11%
N/A
7%
23%
93%
67%
One standard deviation below the mean is equivalent to the middle of the range between the 10th and 25th
percentile. For Krewski, the LML is 5.8 ug/m3 and one standard deviation below the mean is 11.0 ug/m3. For
Lepeule et al., the LML is 8 ug/m3 and we do not have the data for one standard deviation below the mean. It is
important to emphasize that although we have lower levels of confidence in levels below the LML for each study,
the scientific evidence does not support the existence of a level below which health effects from exposure to PM2 5
do not occur.
25%
20%
15%
o
'•s
110%
LML of Krewski et
al. (2009) study
LML of Le|
(2012)stu
eule et a .
dy
1 ....
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 12-13 13-14 14-15 15-16 16-17 17-18 18-:
Baseline Annual Mean PM2.5 Level (ug/m3)
Among the populations exposed to PIVh.s in the baseline:
93% are exposed to PMi.s levels at or above the LML of the Krewski et al. (2009) study
67% are exposed to PM2 5 levels at or above the LML of the Lepeule et al. (2012) study
Figure 5C-3. Percentage of Adult Population (age 30+) by Annual Mean PMi.s Exposure in
the Baseline Sector Modeling (used to generate the benefit-per-ton estimates)*
* This graph shows the population exposure in the modeling baseline used to generate the benefit-per-ton estimates.
Similar graphs for analyses with air quality modeling show premature mortality impacts at each PM2 5 concentration.
Therefore, caution is warranted when interpreting this graph because it is not consistent with similar graphs from
PJAs that had air quality modeling (e.g., MATS).
5C-12
-------
100%
90%
80%
70%
I 60%
S 50%
5 40%
I
I 30%
I
20%
10%
LML of Krewski et
al. (2009) study
LMLof Lepeuleet
al. (2012) study
7 8 9 10 11 12 13 14
Baseline Annual Mean PM, < Level (ug/m3)
15
16
17
18
Among the populations exposed to PIVb.s in the baseline:
93% are exposed to PM2 5 levels at or above the LML of the Krewski et al. (2009) study
67% are exposed to PM2 5 levels at or above the LML of the Lepeule et al. (2012) study
Figure 5C-4. Cumulative Distribution of Adult Population (age 30+) by Annual Mean
PMi.s Exposure in the Baseline Sector Modeling (used to generate the benefit -
per-ton estimates)*
* This graph shows the population exposure in the modeling baseline used to generate the benefit-per-ton estimates.
Similar graphs for analyses with air quality modeling show premature mortality impacts at each PM2 5 concentration.
Therefore, caution is warranted when interpreting this graph because it is not consistent with similar graphs from
RIAs that had air quality modeling (e.g., MATS).
5C-13
-------
5C.3 Core Incidence and Dollar Benefits Estimates Reflecting Application of Known
Controls for the 2025 Scenario
This section presents a subset of the core incidence and dollar benefits estimates for the
2025 scenario reflecting only application of known controls in simulating each of the alternative
standard levels (i.e., partial-2025 scenario estimates). The presentation of these estimates
parallels results summarized for the 2025 and post-2025 scenario in Section 5.7 and the reader is
referred to that section for further explanation of the tables and types of estimates included in
those tables. However, before presenting detailed benefits summary tables for the partial-2025
scenario, we first (in Table 5C-5) present an overview of the percentage of benefits associated
with known controls for each of the alternative standards considered. Then, following that
overview table, we present a set of more detailed tables including: core incidence attributable to
reductions in ozone (Table 5C-6), core dollar benefits associated with ozone reductions (Table
5C-7), core incidence estimates associated with reductions in PIVh.s (Table 5C-8) and core dollar
benefits estimates associated with PIVh.s reductions (Table 5C-9). A summary of overall core
benefits associated with application of known controls is presented in Table 5C-10.
Table 5C-5. Fraction of Total Core Benefits Associated with Partial Attainment
(application of known controls) (2025 Scenario)
Percentage of Benefits Resulting from Application of Known
Controls
Category of Benefit
Ozone Benefits
PM2.5 Co-benefits
Total Benefits
70ppb
82%
76%
77%
65 ppb
59%
59%
59%
60 ppb
35%
33%
34%
5C-14
-------
Table 5C-6. Estimated Number of Avoided Ozone-Only Health Impacts for the
Alternative Annual Primary Ozone Standards (Incremental to the Analytical
Baseline) for the Partial Attainment of the 2025 Scenario (known controls) a'b
Health Effectb
Proposed and Alterative Standards
(95th percentile confidence intervals)
70ppb 65ppb 60ppb
Avoided Short-Term Mortality - Core Analysis
multi-city
studies
Smith et al. (2009) (all ages)
Zanobetti and Schwartz (2008) (all
ages)
160
(79 to 250)
270
(150 to 400)
370
(180 to 550)
620
(330 to 900)
400
(200 to 610)
680
(360 to 990)
Avoided Long-term Respiratory Mortality - Core Analysis
multi-city
study
Jerrett et al. (2009) (30-99yrs)
copollutants model (PM2.5)
550
(190 to 910)
1,200
(420 to 2,100)
1,400
(460 to 2,300)
Avoided Short-Term Mortality - Sensitivity Analysis
multi-city
studies
meta-
analyses
Smith et al. (2009) (all ages)
copollutants model (PM10)
Schwartz (2005) (all ages)
Huang et al. (2005)
(cardiopulmonary)
Bell etal. (2004) (all ages)
Bell etal. (2005) (all ages)
Ito etal. (2005) (all ages)
Levy etal. (2005) (all ages)
130
(-36 to 300)
200
(63 to 340)
190
(72 to 3 10)
130
(44 to 220)
430
(200 to 650)
590
(360 to 830)
600
(410 to 790)
290
(-80 to 660)
460
(140 to 770)
430
(160 to 710)
300
(99 to 490)
960
(460 to 1,500)
1,300
(800 to 1,900)
1,400
(930 to 1,800)
320
(-88 to 730)
500
(160 to 850)
480
(180 to 780)
330
(110 to 550)
1,100
(5 10 to 1,600)
1,500
(880 to 2,100)
1,500
(1,000 to 2,000)
Avoided Long-term Respiratory Mortality - Sensitivity Analysis c
multi-city
study
Jerrett et al. (2009) (age 30-99) (86
cities) (ozone-only)
Jerrett et al. (2009) (age 30-99) (96
cites) (ozone-only)
370
(110 to 640)
400
(140 to 660)
840
(240 to 1,400)
900
(320 to 1,500)
920
(260 to 1,600)
990
(350 to 1,600)
Avoided Morbidity - Core Analysis
Hospital admissions -respiratory
(age 65+)
Emergency department visits for
asthma (all ages)
Asthma exacerbation (age 6-18)
Minor restricted-activity days (age
18-65)
SchoolLoss Days (age 5-17)
300
(-79 to 670)
930
(87 to 2,900)
240,000
(-360,000 to 740,000)
760,000
(3 10,000 to 1,200,000)
270,000
(95,000 to 600,000)
680
(-180 to 1,500)
2,100
(200 to 6,600)
550,000
(-800,000 to 1,600,000)
1,700,000
(700,000 to 2,700,000)
600,000
(210,000 to 1,300,000)
740
(-200 to 1,700)
2,400
(230 to 7,500)
600,000
(-880,000 to 1,800,000)
1,900,000
(780,000 to 3,000,000)
660,000
(230,000 to 1,500,000)
a All incidence estimates are rounded to whole numbers with a maximum of two significant digits.
b All incidence estimates are based on ozone-only models unless otherwise noted.
0 The sensitivity analysis for long-term exposure-related mortality included an assessment of potential thresholds,
which was completed for the 2025 scenario (see Table 5-19). Care should be taken in applying the results of that
sensitivity analysis to the partial 2025 scenario, although general patterns of impact across the thresholds may apply.
5C-15
-------
Table 5C-7. Total Monetized Ozone-Only Benefits for the Alternative Annual Primary
Ozone Standards (Incremental to the Analytical Baseline) for the Partial
Attainment of the 2025 Scenario (using known controls) a'b
Health Effectb
Proposed and Alterative Standards
(95th percentile confidence intervals)
70ppb 65ppb 60ppb
Avoided Short-Term Mortality - Core Analysis
multi-city
studies
Smith et al. (2009) (all ages)
Zanobetti and Schwartz (2008) (all
ages)
$1,700
($150 to $4,700)
2,800
($250 to $7,900)
$3,700
($330 to $11,000)
6,300
($560 to $18,000)
$4,100
($360 to $12,000)
6,900
($610 to $20,000)
Avoided Short-Term Mortality - Sensitivity Analysis
multi-city
studies
meta-
analyses
Smith et al. (2009) (all ages)
copollutants model (PM10)
Schwartz (2005) (all ages)
Huang et al. (2005)
(cardiopulmonary)
Bell etal. (2004) (all ages)
Bell etal. (2005) (all ages)
Ito etal. (2005) (all ages)
Levy et al. (2005) (all ages)
$1,300
(-$3 10 to $4,800)
$2,100
($160 to $6,300)
$2,000
($160 to $5,800)
1,300
($110 to $4,000)
4,400
($380 to $12,000)
$6,000
($550 to $17,000)
$6,100
($570 to $17,000)
$3,000
(-$710 to $11,000)
$4,700
($370 to $14,000)
$4,400
($370 to $13,000)
3,000
($240 to $9,100)
9,800
($860 to $28,000)
$14,000
($1,200 to $38,000)
$14,000
($1,300 to $38,000)
$3,300
(-$780 to $12,000)
$5,100
($410 to $16,000)
$4,900
($400 to $14,000)
3,300
($270 to $10,000)
11,000
($950 to $3 1,000)
$15,000
($1,400 to $42,000)
$15,000
($1,400 to $41, 000)
Avoided Long-term Respiratory Mortality - Sensitivity Analysis
multi-city
study
Jerrett et al. (2009) (age 30-99)
copollutants model(PM2.5)no lag °
Jerrett et al. (2009) (age 30-99)
copollutants model (PM2.5) 20 yr
sesmented las d
$5,600
($460 to $17,000)
$4,600 to $5,100
($370 to $15,000)
$13,000
($1,000 to $38,000)
$10,000 to $11,000
($840 to $34,000)
$14,000
($1,100 to $41,000)
$11,000 to $13,000
($920 to $37,000)
a All benefits estimates are rounded to whole numbers with a maximum of two significant digits. The monetized
value of the ozone-related morbidity benefits are included in the estimates shown in this table for each mortality
study (and when combined account for from 4-6% of the total benefits, depending on the total mortality estimate
compared against. Note that asthma exacerbations accounts for «1% of the total).
b The sensitivity analysis for long-term exposure-related mortality included an assessment of potential thresholds
however this assessment was implemented using incidence estimates (see Table 5-19). Observations from that
analysis (in terms of fractional impacts on incidence can be directly applied to these benefit results) (see Appendix
5B, section 5B.1)
0 A single central-tendency value is provided in each cell, since the zero-lag model used here did not require
application of a 3% and 7% discount rates (see footnote d below) (note, however that as with all other entries in this
study, we do include a 95th percentile confidence interval range - these are the values within parentheses).
dThe range (outside of the parentheses) within each cell results from application of a 7% and 3% discount rates in
the context of applying the 20year segmented lag (with the 7% resulting in the lower estimate and the 3% the higher
estimate). The range presented within the parentheses reflects consideration for the 95th% confidence interval
generated for each of these estimates and ranges from a low value (2.5th% CI for the 7% discount-based dollar
benefit) to an upper value (97.5th% of the 3% discount-based dollar benefit).
5C-16
-------
Table 5C-8. Estimated Number of Avoided PMi.s-Related Health Impacts for the
Alternative Annual Primary Ozone Standards (Incremental to the Analytical
Baseline) for the Partial Attainment of the 2025 Scenario (using known controls)
a,b
Proposed and Alterative Standards
Health Effectb
60ppb
65ppb
70ppb
AvoidedPMj.s-related Mortality
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
Woodruff et al. (1997) (infant mortality)
390
870
1
860
1,900
1
880
2,000
1
Avoided PM2.5-related Morbidity
Non-fatal heart attacks
Peters et al. (2001) (age >18)
Pooled estimate of 4 studies (age >18)
Hospital admissions — respiratory (all ages)
Hospital admissions — cardiovascular (age > 18)
Emergency department visits for asthma (all ages)
Acute bronchitis (ages 8-12)
Lower respiratory symptoms (ages 7-14)
Upper respiratory symptoms (asthmatics ages 9-11)
Asthma exacerbation (asthmatics ages 6-18)
Lost work days (ages 18-65)
Minor restricted-activity days (ages 18-65)
450
49
120
140
210
600
7,700
11,000
11,000
49,000
290,000
1,000
110
260
310
470
1,300
17,000
24,000
25,000
110,000
640,000
1,000
110
260
320
480
1,400
17,000
25,000
26,000
110,000
660,000
a All incidence estimates are rounded to whole numbers with a maximum of two significant digits. Because these
estimates were generated using benefit-per-ton estimates, confidence intervals are not available. In general, the 95th
percentile confidence interval for the health impact function alone ranges from approximately ±30 percent for
mortality incidence based on Krewski et al. (2009) and ±46 percent based on Lepeule et al. (2012).
5C-17
-------
Table 5C-9. Monetized PM2.s-Related Health Co-Benefits for the Alternative Annual
Primary Ozone Standards (Incremental to Analytical Baseline) for the Partial
Attainment of the 2025 Scenario (using known controls) a'b'c
. , „ _ Proposed and Alterative Standards
Monetized Benefits _„ „ ™ ,
70 ppb 65 ppb 60 ppb
3% Discount Rate
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
7% Discount Rate
Krewski et al. (2009) (adult mortality age 30+)
Lepeule et al. (2012) (adult mortality age 25+)
$3,600
$8,200
$3,300
$7,400
$8,000
$18,000
$7,200
$16,000
$8,300
$19,000
$7,400
$17,000
a All estimates are rounded to two significant digits. Because these estimates were generated using benefit-per-ton
estimates, confidence intervals are not available. In general, the 95th percentile confidence interval for monetized
PM2 5 benefits ranges from approximately -90 percent to +180 percent of the central estimates based on Krewski et
al. (2009) and Lepeule et al. (2012). Estimates do not include unquantified health benefits noted in Table 5-2 or
Section 5.6.5 or welfare co-benefits noted in Chapter 6.
b The reduction in premature fatalities each year accounts for over 98% of total monetized benefits in this analysis.
Mortality risk valuation assumes discounting over the SAB-recommended 20-year segmented lag structure.
Table 5C-10. Combined Estimate of Monetized Ozone and PMi.s Benefits for the
Alternative Annual Primary Ozone Standards for the Partial Attainment of the
2025 Scenario (using known controls) (billions of 2011$) a'b
Total Benefits
Ozone-only Benefits (range reflects
Smith et al., 2009 and Zanobetti and
Schwartz, 2008)
PMi.s Co-benefits (range reflects
Krewski et al., 2009 and Lepeule et
al., 2012)
Discount
Rate
3%
7%
b
3%
70 ppb
$5.3to$ll+B
$5.0 to $10 +B
$1.7 to $2.8 +B
$3. 6 to $8.2 +B
65 ppb
$12 to $24 +B
$llto$23+B
$3.7 to $6.3 +B
$8.0 to $18 +B
60 ppb
$12 to $26 +B
$12 to $24 +B
$4.1to$6.9+B
$8.3 to $19 +B
a Rounded to two significant figures. The reduction in premature fatalities each year accounts for over 98% of total
monetized benefits in this analysis. Mortality risk valuation for PM2 5 assumes discounting over the SAB-
recommended 20-year segmented lag structure. Not all possible benefits are quantified and monetized in this
analysis. B is the sum of all unquantified health and welfare co-benefits. Data limitations prevented us from
quantifying these endpoints, and as such, these benefits are inherently more uncertain than those benefits that we
were able to quantify. These estimates reflect the economic value of avoided morbidities and premature deaths using
risk coefficients from the studies noted.
b Ozone-only benefits reflect short-term exposure impacts and as such are assumed to occur in the same year as
ambient ozone reductions. Consequently, social discounting is not applied to the benefits for this category.
5C-18
-------
5C.4 References
Bell, M. L., K. Ebisu, et al. 2008. "Seasonal and Regional Short-term Effects of Fine Particles on Hospital
Admissions in 202 US Counties, 1999-2005." American Journal of 'Epidemiology 168(11): 1301-1310.
Bell, M.L. and F. Dominici. 2008. Effect modification by community characteristics on the short-term effects of
ozone exposure and mortality in 98 U.S. communities. American Journal of Epidemiology. 167:986-997.
Bell, M.L.; A. McDermott; S.L. Zeger; J.M. Samet; F. Dominici. 2004. Ozone and short-term mortality in 95 U.S.
urban communities, 1987-2000. JAMA. 292:2372-2378.
Bell, ML; Dominici, F; Samet, JM. (2005). Ameta-analysis of time-series studies of ozone and mortality with
comparison to the national morbidity, mortality, and air pollution study. Epidemiology 16: 436-445.
http://dx.doi.org/10.1097/01.ede.0000165817.40152.85
Bell, ML; McDermott, A; Zeger, SL; Samet, JM; Dominici, F. (2004). Ozone and short-term mortality in 95 US
urban communities, 1987-2000. JAMA 292: 2372-2378. http://dx.doi.org/10.1001/jama.292.19.2372
Centers for Disease Control and Prevention (CDC). 2003. Health, United States. Table 30. Years of Potential Life
Lost Before Age 75 for Selected Causes of Death, According to Sex, Race and Hispanic Origin: United States,
Selected Years 1980- 2000. Available at .
Centers for Disease Control and Prevention (CDC). 2008. National Center for Health Statistics. National Health
Interview Survey, 1999-2008.
Centers for Disease Control and Prevention (CDC). 2010. Table Cl Adult Self-Reported Current Asthma Prevalence
Rate (Percent) and Prevalence (Number) by State or Territory. Centers for Disease Control and Prevention,
Behavioral Risk Factor Surveillance System (BRFSS). Available at:
.
Centers for Disease Control and Prevention (CDC). 2011. United States Life Tables, 2007 National Vital Statistics
Reports. Volume 59, Number 9. September. Available at
.
Farm N, Lamson A, Wesson K, Risley D, Anenberg SC, Hubbell BJ. 2012a. "Estimating the National Public Health
Burden Associated with Exposure to Ambient PM2s and ozone. Risk Analysis," Risk Analysis 32(1): 81-95.
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.
Hubbell BL. 2006. "Implementing QALYs in the analysis of air pollution regulations." Environmental and Resource
Economics 34:365-384.
Ito, K.; G.D. Thurston and R.A. Silverman. 2007. Characterization of PM2.5, gaseous pollutants, and meteorological
interactions in the context of time-series health effects models. Journal of Exposure Science and Environmental
Epidemiology. 17(S2):S45-60.
Jerrett M, Burnett RT, Pope CA, III, et al. 2009. "Long-Term Ozone Exposure and Mortality." New England
Journal of Medicine 360:1085-95.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi, Y, et al. 2009. Extended follow-up and spatial analysis of
the American Cancer Society study linking particulate air pollution and mortality. HEI Research Report, 140,
Health Effects Institute, Boston, MA.
5C-19
-------
Krupnick, A.J., and M.L. Cropper. 1992. "The Effect of Information on Health Risk Valuations." Journal of Risk
and Uncertainty 5(2):29-48.
Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Herry, F. Horak Jr., V. Puybonnieux-
Texier, P. Quenel, J. Schneider, R. Seethaler, J-C Vergnaud, and H. Sommer. 2000. "Public-Health Impact of
Outdoor and Traffic-Related Air Pollution: A European Assessment." The Lancet 356:795-801.
Lepeule J, Laden F, Dockery D, Schwartz J. 2012. "Chronic Exposure to Fine Particles and Mortality: An Extended
Foliow-Up of the Harvard Six Cities Study from 1974 to 2009." Environmental Health Perspectives 120
(7):965-70.
Levy JI, Baxter LK, Schwartz J. 2009. "Uncertainty and variability in health-related damages from coal-fired power
plants in the United States." Risk Analysis 29(7) 1000-1014.
National Center for Education Statistics (NCHS). 1996. The Condition of Education 1996, Indicator 42: Student
Absenteeism and Tardiness. U.S. Department of Education. Washington, DC.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press. Washington, DC.
National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. National Academies Press. Washington, DC.
Smith, R.L.; B. Xu and P. Switzer. 2009. Reassessing the relationship between ozone and short- term mortality in
U.S. urban communities. Inhalation Toxicology. 21:37-61.
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis, 2008 National Ambient Air
Quality Standards for Ground-level ozone, Chapter 6. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. March. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2010c. Final Regulatory Impact Analysis (RIA)for the SO2
National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning and Standards, Research
Triangle Park, NC. June. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2011b. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March.
Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2012b. Regulatory Impact Analysis for the Final Revisions to
the National Ambient Air Quality Standards for Particulate Matter. EPA-452/R-12-003. Office of Air Quality
Planning and Standards, Health and Environmental Impacts Division. December. Available at:
.
U.S. Environmental Protection Agency (U.S. EPA). 2013a. Integrated Science Assessment for Ozone and Related
Photochemical Oxidants: Final. Research Triangle Park, NC: U.S. Environmental Protection Agency. (EPA
document number EPA/600/R-10/076F).
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2004a. Advisory on Plans for
Health Effects Analysis in the Analytical Plan for EPA 's Second Prospective Analysis—Benefits and Costs of
the Clean Air Act, 1990-2020. EPA-SAB-COUNCIL-ADV-04-002. March. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010a. Review of EPA 's DRAFT
Health Benefits of the Second Section 812 Prospective Study of the Clean Air Act. EPA-COUNCIL-10-001.
5C-20
-------
June. Available at
.
Woodruff, T.J., J. Grille, and K.C. Schoendorf. 1997. "The Relationship Between Selected of postneonatal infant
mortality and paniculate air pollution in the United States." Environmental Health Perspectives 105(6): 608-
612.
Zanobetti, A; J. Schwartz. 2008. Mortality displacement in the association of ozone with mortality: an analysis of 48
cities in the United States. American Journal of Respiratory and Critical Care Medicine. 177:184-189.
5C-21
-------
APPENDIX 5D: DISCUSSION OF EFFECT ESTIMATES REFLECTED IN THE
DEVELOPMENT OF DOLLAR-PER-TON VALUES USED IN MODELING PM2.s
COBENEFITS
Overview
This section describes how we selected effect estimates to estimate the benefits of
reducing PIVh.s (see Section 5.4.4). This section mirrors that covering effect estimate selection
for ozone presented in Section 5.6.3 in terms of organization (in some cases, we have repeated
introductory/setup content here to facilitate review by the reader).
The first step in selecting effect coefficients is to identify the health endpoints to be
quantified. We base our selection of health endpoints on consistency with the EPA's Integrated
Science Assessments (which replace previous "Criteria Documents"), with input and advice from
the Health Effects Subcommittee (HES), a scientific review panel specifically established to
provide advice on the use of the scientific literature in developing benefits analyses for the
EPA 's Report to Congress on The Benefits and Costs of the Clean Air Act 1990 to 2020 (U.S.
EPA, 201 la). In addition, we have included more recent epidemiology studies from the PM ISA
(U.S. EPA, 2009b) and the PM Provisional Assessment (U.S. EPA, 2012b).88 In general, we
follow a weight of evidence approach, based on the biological plausibility of effects, availability
of concentration-response functions from well conducted peer-reviewed epidemiological studies,
cohesiveness of results across studies, and a focus on endpoints reflecting public health impacts
(like hospital admissions) rather than physiological responses (such as changes in clinical
measures like Forced Expiratory Volume [FEV1]).
There are several types of data that can support the determination of types and magnitude
of health effects associated with air pollution exposures. These sources of data include
toxicological studies (including animal and cellular studies), human clinical trials, and
observational epidemiology studies. All of these data sources provide important contributions to
the weight of evidence surrounding a particular health impact. However, only epidemiology
88 The peer-reviewed studies in the Provisional Assessment have not yet undergone external review by the Science
Advisory Board.
5D-1
-------
studies provide direct concentration-response relationships that can be used to evaluate
population-level impacts of reductions in ambient pollution levels in a health impact assessment.
For the data-derived estimates, we relied on the published scientific literature to ascertain
the relationship between PIVh.s and adverse human health effects. We evaluated epidemiological
studies using the selection criteria summarized in Table 5-6 (see section 5.6.3). These criteria
include consideration of whether the study was peer-reviewed, the match between the pollutant
studied and the pollutant of interest, the study design and location, and characteristics of the
study population, among other considerations. In general, the use of concentration-response
functions from more than a single study can provide a more representative distribution of the
effect estimate. However, there are often differences between studies examining the same
endpoint, making it difficult to pool the results in a consistent manner. For example, studies may
examine different pollutants or different age groups. For this reason, we consider very carefully
the set of studies available examining each endpoint and select a consistent subset that provides a
good balance of population coverage and match with the pollutant of interest. In many cases,
either because of a lack of multiple studies, consistency problems, or clear superiority in the
quality or comprehensiveness of one study over others, a single published study is selected as the
basis of the effect estimate.
When several effect estimates for a pollutant and a given health endpoint have been
selected, they are quantitatively combined or pooled to derive a more robust estimate of the
relationship. The BenMAP Manual Technical Appendices for an earlier version of the program
provide details of the procedures used to combine multiple impact functions (Abt Associates,
2012). In general, we used fixed or random effects models to pool estimates from different
single-city studies of the same endpoint. Fixed effect pooling simply weights each study's
estimate by the inverse variance, giving more weight to studies with greater statistical power
(lower variance). Random effects pooling accounts for both within-study variance and between-
study variability, due, for example, to differences in population susceptibility. We used the fixed
effect model as our null hypothesis and then determined whether the data suggest that we should
5D-2
-------
reject this null hypothesis, in which case we would use the random effects model.89 Pooled
impact functions are used to estimate hospital admissions and asthma exacerbations. When
combining evidence across multi-city studies (e.g., cardiovascular hospital admission studies),
we use equal weights pooling. The effect estimates drawn from each multi-city study are
themselves pooled across a large number of urban areas. For this reason, we elected to give each
study an equal weight rather than weighting by the inverse of the variance reported in each study.
For more details on methods used to pool incidence estimates, see the BenMAP Manual
Appendices (Abt Associates, 2012).
Effect estimates selected for a given health endpoint were applied consistently across all
locations nationwide. This applies to both impact functions defined by a single effect estimate
and those defined by a pooling of multiple effect estimates. Although the effect estimate may, in
fact, vary from one location to another (e.g., because of differences in population susceptibilities
or differences in the composition of PM), location-specific effect estimates are generally not
available.
The specific studies from which effect estimates were quantified for the core analysis for
PM2.5 (generated using the dollar-per-ton approach - see Section 5.4.4) are presented in Table 5-
8 and repeated below for ease of access in Table 5D-1. We highlight in red those studies that
have been added since the benefits analysis conducted for the ozone reconsideration (U.S. EPA,
2010d) or the Ozone NAAQS RIA (U.S. EPA, 2008). In all cases where effect estimates are
drawn directly from epidemiological studies, standard errors are used as a partial representation
of the uncertainty in the size of the effect estimate.
89 EPA recently changed the algorithm BenMAP uses to calculate study variance, which is used in the pooling
process. Prior versions of the model calculated population variance, while the version used here calculated sample
variance. This change did not affect the selection of random or fixed effects for the pooled incidence estimates
between the proposal and final PM RIA.
5D-2
-------
Table 5D-1. Health Endpoints and Epidemiological Studies Used to Quantify PMi.s-
related Health Impacts in the Core Analysis a
Endpoint
Study
Study Population
Relative Risk or Effect Estimate (P)
(with 95th Percentile Confidence
Interval orSE, respectively)
Premature Mortality
Premature mortality-
cohort study, all-cause
Premature mortality—
all-cause
Krewski et al. (2009)
Lepeule et al. (2012)
Woodruff etal. (1997)
> 29 years RR = 1.06 (1.04-1.06) per 10 u.g/m3
> 24 years RR = 1.14 (1.07-1.22) per 10 u.g/m3
Infant (< 1 OR = 1.04 (1.02-1.07) per 10 u.g/m3
year)
Chronic Illness
Nonfatal heart attacks
Peters et al. (2001)
Pooled estimate:
Pope et al. (2006)
Sullivan et al. (2005)
Zanobetti et al. (2009)
Zanobetti and Schwartz (2006)
Adults(>18
years)
OR = 1.62 (1.13-2.34) per 20 u.g/m3
P = 0.00481 (0.00199)
P = 0.00198 (0.00224)
P = 0.00225 (0.000591)
P = 0.0053 (0.00221)
Hospital Admissions
Respiratory
Cardiovascular
Asthma-related
emergency department
visits
Zanobetti et al. (2009)—ICD > 64 years
460-519 (All respiratory)
Kloog et al. (2012)-ICD 460-
519 (All Respiratory
Moolgavkar (2000)-ICD 490- 18-64 years
496 (Chronic lung disease)
Babinetal. (2007)-ICD493 < 19 years
(asthma)
Sheppard (2003)-ICD 493
(asthma)
Pooled estimate: > 64 years
Zanobetti et al. (2009)—ICD
390-459 (all cardiovascular)
Peng etal. (2009)-ICD 426-
427; 428; 430-438; 410-414;
429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)
Peng etal. (2008)-ICD 426-
427; 428; 430-438; 410-414;
429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)
Bell et al. (2008)-ICD 426-427;
428; 430-438; 410-414; 429;
440-449 (Cardio-, cerebro- and
peripheral vascular disease)
Moolgavkar (2000)-ICD 390- 20-64 years
429 (all cardiovascular)
Pooled estimate: All ages
Mar etal. (2010)
Slaughter etal. (2005)
Glad etal. (2012)
P=0.00207 (0.00446)
P=0.0007 (0.000961)
1.02 (1.01-1.03) per 36 |jg/m3
P=0.002 (0.004337)
RR = 1.04 (1.01-1.06) per 11.8 u.g/m3
P=0.00189 (0.000283)
P=0.00068
(0.000214)
P=0.00071
(0.00013)
P=0.0008
(0.000107)
RR=1.04 (t statistic: 4.1) per 10 u.g/m3
RR = 1.04 (1.01-1.07) per 7 u.g/m3
RR = 1.03 (0.98-1.09) per 10 u.g/m3
P=0.00392 (0.002843)
Other Health Endpoints
Acute bronchitis
Dockeryetal. (1996)
8-12 years OR = 1.50 (0.91-2.47) per 14.9 u.g/m3
5D-4
-------
Asthma exacerbations Pooled estimate: 6-18 yearsb
„ . . . . . OK — 1.U.D (U.yo—l.U/)
Ostro etal. (2001) (cough, OR = 1.06 1.01-1.11
wheeze and shortness of
b OR =1.08 (1.00-1.17) per 30 u.g/m3
breath)
Mar et al. (2004) (cough, RR = 1.21 (1-1.47) per
shortness of breath) RR = 1.13 (0.86-1.48) per 10 u.g/m3
Work loss days Ostro (1987) 18-65 years p=0.0046 (0.00036)
Acute respiratory Ostro and Rothschild (1989)
,.,n.rM ,.„• • -, • • -, > 18-65 years P=0.00220 0.000658
symptoms (MRAD) (Minor restricted activity days)
Upper respiratory „ , ,„„„„, Asthmatics. „ „_ ,„ „ „_. „„ . ,
KK K y Pope etal. 1991 „ „„ ' 1.003 1-1.006 per 10 u.g/m3
symptoms K v ' 9-11 years v /K ^&
Lower respiratory Schwartz and Neas (2000) 7-14 years OR = 1.33 (1.11-1.58) per 15 u.g/m3
symptoms
a Studies highlighted in red represent updates incorporated since the Ozone NAAQS RIA (U.S. EPA, 2008). These
updates were introduced in the PM NAAQS RIA (U.S. EPA, 2012).
b The original study populations were 8 to 13 for the Ostro et al. (2001) study and 7 to 12 for the Mar et al. (2004)
study. Based on advice from the SAB-HES, we extended the applied population to 6-18, reflecting the common
biological basis for the effect in children in the broader age group. See: U.S. EPA-SAB (2004) and NRC (2002).
5D.1 PM2.s Premature Mortality Effect Coefficients
Core Mortality Effect Coefficients for Adults. A substantial body of published scientific
literature documents the association between elevated PIVh.s concentrations and increased
premature mortality (U.S. EPA, 2009b). This body of literature reflects thousands of
epidemiology, toxicology, and clinical studies. The PM ISA completed as part of the most recent
review of the PM standards, which was twice reviewed by the SAB-CASAC (U.S. EPA-SAB,
2009, 2009b), concluded that there is a causal relationship between mortality and both long-term
and short-term exposure to PM2.5 based on the entire body of scientific evidence (U.S. EPA,
2009b). The size of the mortality effect estimates from epidemiological studies, the serious
nature of the effect itself, and the high monetary value ascribed to prolonging life make mortality
risk reduction the most significant health endpoint quantified in this analysis.
Researchers have found statistically significant associations between PM2.5 and
premature mortality using different types of study designs. Time-series methods have been used
to relate short-term (often day-to-day) changes in PM2.5 concentrations and changes in daily
mortality rates up to several days after a period of elevated PM2.5 concentrations. Cohort
methods have been used to examine the potential relationship between community-level PM2.5
exposures over multiple years (i.e., long-term exposures) and community-level annual mortality
rates that have been adjusted for individual level risk factors. When choosing between using
short-term studies or cohort studies for estimating mortality benefits, cohort analyses are thought
5D-5
-------
to capture more of the public health impact of exposure to air pollution over time because they
account for the effects of long-term exposures as well as some fraction of short-term exposures
(Kunzli et al., 2001; NRC, 2002). The NRC stated that "it is essential to use the cohort studies in
benefits analysis to capture all important effects from air pollution exposure" (NRC, 2002, p.
108). The NRC further notes that "the overall effect estimates may be a combination of effects
from long-term exposure plus some fraction from short-term exposure. The amount of overlap is
unknown" (NRC, 2002, p. 108-9). To avoid double counting, we focus on applying the risk
coefficients from the long-term cohort studies in estimating the mortality impacts of reductions
inPM2.5.
Over the last two decades, several studies using "prospective cohort" designs have been
published that are consistent with the earlier body of literature. Two prospective cohort studies,
often referred to as the Harvard "Six Cities Study" (Dockery et al., 1993; Laden et al., 2006;
Lepeule et al., 2012) and the "American Cancer Society" or "ACS study" (Pope et al., 1995;
Pope et al., 2002; Pope et al., 2004; Krewski et al., 2009), provide the most extensive analyses of
ambient PIVh.s concentrations and mortality. These studies have found consistent relationships
between fine particle indicators and premature mortality across multiple locations in the United
States. The credibility of these two studies is further enhanced by the fact that the initial
published studies (Pope et al., 1995; Dockery et al., 1993) were subject to extensive
reexamination and reanalysis by an independent team of scientific experts commissioned by the
Health Effects Institute (HEI) and by a Special Panel of the HEI Health Review Committee
(Krewski et al., 2000). Publication of studies confirming and extending the findings of the 1993
Six Cities Study and the 1995 ACS study using more recent air quality and a longer follow-up
period for the ACS cohort provides additional validation of the findings of these original studies
(Pope et al., 2002, 2004; Laden et al., 2006; Krewski et al., 2009; Lepeule et al., 2012). The
SAB-HES also supported using these two cohorts for analyses of the benefits of PM reductions,
and concluded, "the selection of these cohort studies as the underlying basis for PM mortality
benefit estimates to be a good choice. These are widely cited, well studied and extensively
reviewed data sets" (U.S. EPA-SAB, 2010a). As both the ACS and Six Cities studies have
inherent strengths and weaknesses, we present benefits estimates using relative risk estimates
from the most recent extended reanalysis of these cohorts (Krewski et al., 2009; Lepeule et al.,
2012). Presenting results using both ACS and Six Cities is consistent with other recent RIAs
5D-6
-------
(e.g., U.S. EPA, 2006, 2010c, 201 Ib, 201 Ic). The PM ISA concludes that the ACS and Six
Cities cohorts provide the strongest evidence of the association between long-term PIVh.s
exposure and premature mortality with support from a number of additional cohort studies
(described below).
The extended analyses of the ACS cohort data (Krewski et al, 2009) provides additional
refinements to the analysis of PM-related mortality by (a) extending the follow-up period by
2 years to the year 2000, for a total of 18 years; (b) incorporating almost double the number of
urban areas (c) addressing confounding by spatial autocorrelation by incorporating ecological, or
community-level, co-variates; and (d) performing an extensive spatial analysis using land use
regression modeling in two large urban areas. These enhancements make this analysis well-
suited for the assessment of mortality risk from long-term PIVh.s exposures for the EPA's
benefits analyses.
In 2009, the SAB-HES again reviewed the choice of mortality risk coefficients for
benefits analysis, concluding that "[t]he Krewski et al. (2009) findings, while informative, have
not yet undergone the same degree of peer review as have the aforementioned studies. Thus, the
SAB-HES recommends that EPA not use the Krewski et al. (2009) findings for generating the
Primary Estimate" (U.S. EPA-SAB, 2010a). Since this time, the Krewski et al. (2009) has
undergone additional peer review, which we believe strengthens the support for including this
study in this RIA. For example, the PM ISA (U.S. EPA, 2009b) included this study among the
key mortality studies. In addition, the risk assessment supporting the PM NAAQS (U.S. EPA,
2010a) utilized risk coefficients drawn from the Krewski et al. (2009) study, the most recent
reanalysis of the ACS cohort data. The risk assessment cited a number of advantages that
informed the selection of the Krewski et al. (2009) study as the source of the core effect
estimates, including the extended period of observation, the rigorous examination of model
forms and effect estimates, the coverage for ecological variables, and the large dataset with over
1.2 million individuals and 156 MSAs (U.S. EPA, 2010a). The CASAC also provided extensive
peer review of the risk assessment and supported the use of effect estimates from this study (U.S.
EPA-SAB, 2009, 201 Ob, c).
5D-7
-------
Consistent with the Quantitative Health Risk Assessment for Particulate Matter (U.S.
EPA, 2010a) which was reviewed by the CASAC (U.S. EPA-SAB, 2009), we use the all-cause
mortality risk estimate based on the random-effects Cox proportional hazard model that
incorporates 44 individual and 7 ecological covariates (RR=1.06, 95% confidence intervals 1.04-
1.08 per 10|ig/m3 increase in PIVh.s). The relative risk estimate (1.06 per 10|ig/m3 increase in
PM2.s) is identical to the risk estimate drawn from the earlier Pope et al. (2002) study, though the
confidence interval around the Krewski et al. (2009) risk estimate is tighter.
In the most recent Six Cities study, which was published after the last SAB-HES review,
Lepeule et al. (2012) evaluated the sensitivity of previous Six-Cities results to model
specifications, lower exposures, and averaging time using eleven additional years of cohort
follow-up that incorporated recent lower exposures. The authors found significant associations
between PIVh.s exposure and increased risk of all-cause, cardiovascular and lung cancer
mortality. The authors also concluded that the concentration-response relationship was linear
down to PM2.5 concentrations of 8 ug/m3, and that mortality rate ratios for PIVh.s fluctuated over
time, but without clear trends, despite a substantial drop in the sulfate fraction. We use the all-
cause mortality risk estimate based on a Cox proportional hazard model that incorporates 3
individual covariates. (RR=1.14, 95% confidence intervals 1.07-1.22 per 10 |ig/m3 increase in
PM2.s). The relative risk estimate is slightly smaller than the risk estimate drawn from Laden et
al. (2006), with relatively smaller confidence intervals.
Implicit in the calculation of PIVh.s-related premature mortality impacts are several key
assumptions, which are described in further detail later in this Appendix. First, we assume that
there is a "cessation" lag in time between the reduction in PM exposure and the full reduction in
mortality risk that affects the timing (and thus discounted monetary valuation) of the resulting
premature deaths (see Section 5.6.4.1). Second, following conclusions of the PM ISA, we
assume that all fine particles are equally potent in causing premature mortality (see Section
5.7.3). Third, following conclusions of the PM ISA, we assume that the health impact function
for fine particles is linear within the range of ambient concentrations affected by these standards
(see Section 5.7.3).
5D-8
-------
Alternate Mortality Effect Coefficients for Adults. In addition to the ACS and Six Cities
cohorts, several recent cohort studies conducted in North America provide evidence for the
relationship between long-term exposure to PIVh.s and the risk of premature death. Many of these
additional cohort studies are described in the PM ISA (U.S. EPA, 2009b) and the Provisional
Assessment (U.S. EPA, 2012b) (and thus not summarized here).90'91 Table 5D-2 provides the
effect estimates from each of these cohort studies for all-cause, cardiovascular, cardiopulmonary,
and ischemic heart disease (IHD) mortality, as well as the lowest measured air quality level
(LML) and mean concentration in the study.
We also draw upon the results of the 2006 expert elicitation sponsored by the EPA
(Roman et al, 2008; lEc, 2006) to demonstrate the sensitivity of the benefits estimates to 12
expert-defined concentration-response functions. The PIVb.s expert elicitation and the derivation
of effect estimates from the expert elicitation results are described in detail in the 2006 PIVh.s
NAAQS RIA (U.S. EPA, 2006), the elicitation summary report (lEc, 2006) and Roman et al.
(2008), and so we summarize the key attributes of this study relative to the interpretation of the
estimates of PM-related mortality reported here. We describe also how the epidemiological
literature has evolved since the expert elicitation was conducted in 2005 and 2006.
90 It is important to note that the newer studies in the Provisional Assessment are published in peer-reviewed
journals and meet our study selection criteria, but they have not been assessed in the context of an Integrated
Science Assessment nor gone through review by the SAB. In addition, only the ACS and Harvard Six Cities'
cohort studies have been recommended by the SAB as appropriate for benefits analysis of national rulemakings.
91 In this Appendix, we only describe multi-state cohort studies. There are additional cohort studies that we have not
included in this list, including cohort studies that focus on single cities (e.g., Gan et al., 2012) and cohort studies
focusing on methods development. In Appendix 5A, we provide additional information regarding cohort studies in
California, which is the only state for which we identified single state cohorts.
5D-9
-------
Table 5D-2. Summary of Effect Estimates from Associated with Change in Long-Term
Exposure to PMi.s in Recent Cohort Studies in North America
Study
Pope et al.
(2002)
Laden et al.
(2006)
Lipfert et al.
(2006)a
Miller et al.
(2007)b
Eftim et al.
(2008)
Zeger et al.
(2008)c
Krewski et
al. (2009)d
Puett et al.
(2009)b
Grouse et al.
(2011)^
Puett et al.
(2011)f
Lepeule et
al. (2012)d
Cohort (age)
ACS
(age >30)
Six Cities
(age > 25)
Veterans
(age 39-63)
WHI
(age 50-79)
Medicare (age >
65)
Medicare (age >
65)
ACS
(age >30)
NHS
(age 30-55)
Canadian
census
Health
Professionals
(age 40-75)
Six Cities
(age > 25)
LML
(jig/m3)
7.5
10
<14.1
3.4
6
<9.8
5.8
5.8
1.9
<14.4
8
Mean
(jig/m3)
18.2
16.4
14.3
13.5
13.6
13.2
14
13.9
8.7
17.8
15.9
Hazard Ratios per 10 jig/m3 Change in PMi.s
(95th percentile confidence intervals)
All Causes
1.06
(1.02-1.11)
1.16
(1.07-1.26)
1.15
(1.05-1.25)
N/A
1.21
(1.15-1.27)
1.068
(1.049-1.087)
1.06
(1.04-1.08)
1.26
(1.02-1.54)
1.06
(1.01-1.10)
0.86
(0.70-1.00)
1.14
(1.07-1.22)
Cardiovascular
1.12
(1.08-1.15)
1.28
(1.13-1.44)
N/A
1.76
(1.25-2.47)
N/A
N/A
N/A
N/A
N/A
1.02
(0.84-1.23)
1.26
(1.14-1.40)
Cardiopulmonary
1.09
(1.03-1.16)
N/A
N/A
N/A
N/A
N/A
1.13
(1.10-1.16)
N/A
N/A
N/A
N/A
IHD
N/A
N/A
N/A
2.21
(1.17^.16)
N/A
N/A
1.24
(1.19-1.29)
2.02
(1.07-3.78)
N/A
N/A
N/A
a Low socio-economic status (SES) men only. Used traffic proximity as a surrogate of exposure.
b Women only.
0 Reflects risks in the Eastern U.S. Risks in the Central U.S. were higher, but the authors found no association in the
Western U.S.
d Random effects Cox model with individual and ecologic covariates.
e Canadian population.
f Men with high socioeconomic status only.
The primary goal of the 2006 study was to elicit from a sample of health experts
probabilistic distributions describing uncertainty in estimates of the reduction in mortality among
the adult U.S. population resulting from reductions in ambient annual average PIVh.s levels.
These distributions were obtained through a formal interview protocol using methods designed to
elicit subjective expert judgments. These experts were selected through a peer-nomination
process and included experts in epidemiology, toxicology, and medicine. The elicitation
interview consisted of a protocol of carefully structured questions, both qualitative and
quantitative, about the nature of the PIVh.s-mortality relationship designed to build twelve
individual distributions for the coefficient (or slope) of the C-R function relating changes in
annual average PIVh.s exposures to annual, adult all-cause mortality. The elicitation also provided
useful information regarding uncertainty characterization in the PIVh.s-mortality relationship.
5D-10
-------
Specifically, during their interviews, the experts highlighted several uncertainties inherent within
the epidemiology literature, such as causality, concentration thresholds, effect modification, the
role of short- and long-term exposures, potential confounding, and exposure misclassification. In
Appendix 5C, we evaluate each of these uncertainties in the context of this health impact
assessment. For several of these uncertainties, such as causality, we are able to use the expert-
derived functions to quantify the impacts of applying different assumptions. The elicitation
received favorable peer review in 2006 (Mansfield and Patil, 2006).
Prior to providing a quantitative estimate of the risk of premature death associated with
long-term PIVh.s exposure, the experts answered a series of "conditioning questions." One such
question asked the experts to identify which epidemiological studies they found most
informative. The "ideal study attributes"92 according to the experts included:
• Geographic representation of the entire U.S. (e.g., monitoring sites across the country)
• Collection of information on individual risk factors and residential information both at
the beginning and throughout the follow-up period
• Large sample size that is representative of the general U.S. population
• Collection of genetic information from cohort members to identify and assess potential
effect modifiers
• Monitoring of individual exposures (e.g., with a personal monitor)
• Collection of data on levels of several co-pollutants (not only those that are monitored for
compliance purposes)
• Accurate characterization of outcome (i.e., cause of death)
• Follow-up for a long period of time, up to a lifetime
• Prospective study design
Although no single epidemiological study completely satisfies each of these criteria, the
experts determined that the ACS and Six Cities' cohort studies best satisfy a majority of these
ideal attributes. To varying degrees the studies examining these two cohorts are geographically
representative; have collected information on individual risk factors; include a large sample size;
have collected data on co-pollutants in the case of the ACS study; have accurately characterized
the health outcome; include a long (and growing) follow-up period; and, are prospective in
92
These criteria are substantively similar to EPA's study selection criteria identified in Table 5-5 of Chapter 5.
5D-11
-------
nature. The experts also noted a series of limitations in these two cohort studies. In the case of
the Six Cities study (Laden et al., 2006), the experts identified the "small sample size, limited
number of cities, and concerns about representativeness of the six cities for the U.S. as a whole"
as weaknesses. When considering the ACS study (Pope et al., 2002), the experts indicated that
the "method of recruitment for the study, which resulted in a group with higher income, more
education, and a greater proportion of whites than is representative of the general U.S.
population" represented a shortcoming. Several experts also argued that because the ACS study
relied upon ".. .whatever monitors were available to the study... a single monitor represented]
exposure for an entire metropolitan area.. .whereas [the Six Cities study] often had exposures
assigned at the county level." Despite these limitations, the experts considered the Pope et al.
(2002) extended analysis of the ACS cohort and the Laden et al. (2006) extended analysis of the
Six Cities cohort to be particularly influential in their opinions (see Exhibit 3-3 of the elicitation
summary report [lEc, 2006]).
Please note that the benefits estimates results presented are not the direct results from the
studies or expert elicitation; rather, the estimates are based in part on the effect coefficients
provided in those studies or by experts. In addition, the experts provided distributions around
their mean PIVh.s effect estimates, which provides more information regarding the overall range
of uncertainty, and this overall range is larger than the range of the mean effect estimates from
each of the experts.
Since the completion of the EPA's expert elicitation in 2006, additional epidemiology
literature has become available, including 9 new multi-state cohort studies shown in Table 5-12.
This newer literature addresses some of the weaknesses identified in the prior literature. For
example, in an attempt to improve its characterization of population exposure the most recent
extended analysis of the ACS cohort Krewski et al. (2009) incorporates two case studies that
employ more spatially resolved estimates of population exposure.
In light of the availability of this newer literature, we have updated the presentation of
results in the RIA. Specifically, we focus the core analysis on results derived from the two most
recent studies of the ACS and Six Cities cohorts (Krewski et al., 2009; Lepeule et al., 2012).
Because the other multi-state cohorts generally have limited geography and age/gender
5D-12
-------
representativeness, these limitations preclude us from using these studies in our core benefits
results, and we instead present the risk coefficients from these other multi-state cohorts in Table
5D-2. In addition, we now include as a sensitivity analysis, mortality estimates based on
application of the full set of expert-derived effect estimates (see Appendix 5B, Section 5B.2).
We do not combine the expert results in order to preserve the breadth and diversity of opinion on
the expert panel (Roman et. al, 2008). This presentation of the expert-derived results is generally
consistent with SAB advice (U.S. EPA-SAB, 2008), which recommended that the EPA
emphasize that "scientific differences existed only with respect to the magnitude of the effect of
PM2.5 on mortality, not whether such an effect existed" and that the expert elicitation "supports
the conclusion that the benefits of PIVh.s control are very likely to be substantial". Although it is
possible that the newer literature could revise the experts' quantitative responses if elicited again,
we believe that these general conclusions are unlikely to change.
Mortality Effect Coefficients for Infants. In addition to the adult mortality studies described
above, several studies show an association between PM exposure and premature mortality in
children under 5 years of age.93 The PM ISA states that less evidence is available regarding the
potential impact of PIVh.s exposure on infant mortality than on adult mortality and the results of
studies in several countries include a range of findings with some finding significant
associations. Specifically, the PM ISA concluded that evidence exists for a stronger effect at the
post-neonatal period and for respiratory-related mortality, although this trend is not consistent
across all studies. In addition, compared to avoided premature deaths estimated for adult
mortality, avoided premature deaths for infants are significantly smaller because the number of
infants in the population is much smaller than the number of adults and the epidemiology studies
on infant mortality provide smaller risk coefficients associated with exposure to PM2.5.
In 2004, the SAB-HES noted the release of the WHO Global Burden of Disease Study
focusing on ambient air, which cites several recently published time-series studies relating daily
PM exposure to mortality in children (U.S. EPA-SAB, 2004). The SAB-HES also cites the study
by Belanger et al. (2003) as corroborating findings linking PM exposure to increased respiratory
inflammation and infections in children. A study by Chay and Greenstone (2003) found that
93
For the purposes of this analysis, we only calculate benefits for infants age 0-1, not all children under 5 years old.
5D-13
-------
reductions in TSP caused by the recession of 1981-1982 were statistically associated with
reductions in infant mortality at the county level. With regard to the cohort study conducted by
Woodruff et al. (1997), the SAB-HES notes several strengths of the study, including the use of a
larger cohort drawn from a large number of metropolitan areas and efforts to control for a variety
of individual risk factors in infants (e.g., maternal educational level, maternal ethnicity, parental
marital status, and maternal smoking status). Based on these findings, the SAB-HES
recommended that the EPA incorporate infant mortality into the primary benefits estimate and
that infant mortality be evaluated using an impact function developed from the Woodruff et al.
(1997) study (U.S. EPA-SAB, 2004).
In 2010, the SAB-HES again noted the increasing body of literature relating infant
mortality and PM exposure and supported the inclusion of infant mortality in the monetized
benefits (U.S. EPA-SAB, 2010a). The SAB-HES generally supported the approach of estimating
infant mortality based on Woodruff et al. (1997) and noted that a more recent study by Woodruff
et al. (2006) continued to find associations between PIVh.s and infant mortality in California. The
SAB-HES also noted, "when PMio results are scaled to estimate PIVh.s impacts, the results yield
similar risk estimates." Consistent with the Costs and Benefits of the Clean Air Act (U.S. EPA,
201 la), we continue to rely on the earlier 1997 study in part due to the national-scale of the
earlier study.
5D.2 Hospital Admissions and Emergency Department Visits
Because of the availability of detailed hospital admission and discharge records, there is
an extensive body of literature examining the relationship between hospital admissions and air
pollution. For this reason, we pool together the incidence estimates using several different
studies for many of the hospital admission endpoints. In addition, some studies have examined
the relationship between air pollution and emergency department visits. Since most emergency
department visits do not result in an admission to the hospital (i.e., most people going to the
emergency department are treated and return home), we treat hospital admissions and emergency
department visits separately, taking account of the fraction of emergency department visits that
are admitted to the hospital. Specifically, within the baseline incidence rates, we parse out the
scheduled hospital visits from unscheduled ones as well as the hospital visits that originated in
the emergency department.
5D-14
-------
The two main groups of hospital admissions are estimated in this analysis for
respiratory admissions and cardiovascular admissions. There is not sufficient evidence linking
PM2.5 with other types of hospital admissions. Both asthma- and cardiovascular-related visits
have been linked to PIVh.s in the United States, though as we note below, we are able to assign an
economic value to asthma-related events only. To estimate the effects of PIVh.s air pollution
reductions on asthma-related ER visits, we use the effect estimate from a study of children 18
and under by Mar et al. (2010), Slaughter et al. (2005), and Glad et al. (2012). The first two
studies examined populations 0 to 99 in Washington State, while Glad et al. examined
populations 0-99 in Pittsburgh, PA. Mar and colleagues perform their study in Tacoma, while
Slaughter and colleagues base their study in Spokane. We apply random/fixed effects pooling to
combine evidence across these two studies.
To estimate avoided incidences of cardiovascular hospital admissions associated with
PM2.5, we used studies by Moolgavkar (2000), Zanobetti et al. (2009), Peng et al. (2008, 2009)
and Bell et al., (2008). Only Moolgavkar (2000) provided a separate effect estimate for adults 20
to 64, while the remainder estimate risk among adults over 64.94 Total cardiovascular hospital
admissions are thus the sum of the pooled estimate for adults over 65 and the single study
estimate for adults 20 to 64. Cardiovascular hospital admissions include admissions for
myocardial infarctions. To avoid double-counting benefits from reductions in myocardial
infarctions when applying the impact function for cardiovascular hospital admissions, we first
adjusted the baseline cardiovascular hospital admissions to remove admissions for myocardial
infarctions. We applied equal weights pooling to the multi-city studies assessing risk among
adults over 64 because these studies already incorporated pooling across the city-level estimates.
One potential limitation of our approach is that while the Zanobetti et al. (2009) study assesses
all cardiovascular risk, Bell et al. (2008), and Peng et al., (2008, 2009) studies estimate a subset
of cardiovascular hospitalizations as well as certain cerebro- and peripheral-vascular diseases. To
address the potential for the pooling of these four studies to produce a biased estimate, we match
94 Note that the Moolgavkar (2000) study has not been updated to reflect the more stringent GAM convergence
criteria. However, given that no other estimates are available for this age group, we chose to use the existing
study. Given the very small (<5%) difference in the effect estimates for people 65 and older with cardiovascular
hospital admissions between the original and reanalyzed results, we do not expect this choice to introduce much
bias. For a discussion of the GAM convergence criteria, and how it affected the size of effect coefficients reported
by time series epidemiological studies using NMMAPS data, see: http://www.healtheffects.org/Pubs/st-
timeseries.htm.
5D-15
-------
the pooled risk estimate with a baseline incidence rate that excludes cerebro- and peripheral-
vascular disease. An alternative approach would be to use the Zanobetti et al. (2009) study alone,
though this would prevent us from drawing upon the strengths of the three multi-city studies.
To estimate avoided incidences of respiratory hospital admissions associated with PIVh.s,
we used a number of studies examining total respiratory hospital admissions as well as asthma
and chronic lung disease. We estimated impacts among three age groups: adults over 65, adults
18 to 64 and children 0 to 17. For adults over 65, the multi-city studies by Zanobetti et al. (2009)
and Kloog et al. (2012) provide effect coefficients for total respiratory hospital admissions
(defined as ICD codes 460-519). We pool these two studies using equal weights. Moolgavkar et
al. (2003) examines PIVh.s and chronic lung disease hospital admissions (less asthma) in Los
Angeles, CA among adults 18 to 64. For children 0 to 18, we pool two studies using
random/fixed effects. The first is Babin et al. (2007) which assessed PIVh.s and asthma hospital
admissions in Washington, DC among children 1 to 18; we adjusted the age range for this study
to apply to children 0 to 18. The second is Sheppard et al. (2003) which assessed PIVh.s and
asthma hospitalizations in Seattle, Washington, among children 0 to 18.
5D.3 Acute Health Events and School/Work Loss Days
In addition to mortality, chronic illness, and hospital admissions, a number of acute
health effects not requiring hospitalization are associated with exposure to PIVh.s. The sources for
the effect estimates used to quantify these effects are described below.
Asthma exacerbations. For this RIA, we have followed the SAB-HES recommendations
regarding asthma exacerbations in developing the core estimate (U.S. EPA-SAB, 2004).
Although certain studies of acute respiratory events characterize these impacts among only
asthmatic populations, others consider the full population, including both asthmatics and non-
asthmatics. For this reason, incidence estimates derived from studies focused only on asthmatics
cannot be added to estimates from studies that consider the full population—to do so would
double-count impacts. To prevent such double-counting, we estimated the exacerbation of
asthma among children and excluded adults from the calculation. Asthma exacerbations
occurring in adults are assumed to be captured in the general population endpoints such as work
loss days and MRADs. Finally, we note the important distinction between the exacerbation of
5D-16
-------
asthma among asthmatic populations, and the onset of asthma among populations not previously
suffering from asthma; in this RIA, we quantify the exacerbation of asthma among asthmatic
populations and not the onset of new cases of asthma.
Based on advice from the SAB-HES (2004), regardless of the age ranges included in the
source epidemiology studies, we extend the applied population to ages 6 to 18, reflecting the
common biological basis for the effect in children in the broader age group. This age range
expansion is also supported by NRC (2002, pp. 8, 116).
To characterize asthma exacerbations in children from exposure to PIVh.s, we selected
two studies (Ostro et al, 2001; Mar et al, 2004) that followed panels of asthmatic children.
Ostro et al. (2001) followed a group of 138 African-American children in Los Angeles for 13
weeks, recording daily occurrences of respiratory symptoms associated with asthma
exacerbations (e.g., shortness of breath, wheeze, and cough). This study found a statistically
significant association between PIVh.s, measured as a 12-hour average, and the daily prevalence
of shortness of breath and wheeze endpoints. Although the association was not statistically
significant for cough, the results were still positive and close to significance; consequently, we
decided to include this endpoint, along with shortness of breath and wheeze, in generating
incidence estimates (see below).
Mar et al. (2004) studied the effects of various size fractions of particulate matter on
respiratory symptoms of adults and children with asthma, monitored over many months. The
study was conducted in Spokane, Washington, a semi-arid city with diverse sources of
particulate matter. Data on respiratory symptoms and medication use were recorded daily by the
study's subjects, while air pollution data was collected by the local air agency and Washington
State University. Subjects in the study consisted of 16 adults—the majority of whom participated
for over a year—and nine children, all of whom were studied for over eight months. Among the
children, the authors found a strong association between cough symptoms and several metrics of
particulate matter, including PM2.5. However, the authors found no association between
respiratory symptoms and PM of any metric in adults. Mar et al. therefore concluded that the
discrepancy in results between children and adults was due either to the way in which air quality
5D-17
-------
was monitored, or a greater sensitivity of children than adults to increased levels of PM air
pollution.
We employed the following pooling approach in combining estimates generated using
effect estimates from the two studies to produce a single estimate for PM-related asthma
exacerbation incidence. First, we used random/fixed effects pooling to combine the Ostro and
Mar estimates for shortness of breath and cough. Next, we pooled the Ostro estimate of wheeze
with the pooled cough and shortness of breath estimates to derive an overall estimate of asthma
exacerbation in children.
Acute Respiratory Symptoms. We estimate three types of acute respiratory symptoms related
to PM2.5 exposure: lower respiratory symptoms, upper respiratory symptoms, and minor
restricted activity days (MRAD).
Incidences of lower respiratory symptoms (e.g., wheezing, deep cough) in children aged
7 to 14 were estimated for PM2.5 using an effect estimate from Schwartz and Neas (2000).
Incidences of upper respiratory symptoms in asthmatic children aged 9 to 11 are estimated for
PM2.5 using an effect estimate developed from Pope et al. (1991). Because asthmatics have
greater sensitivity to stimuli (including air pollution), children with asthma can be more
susceptible to a variety of upper respiratory symptoms (e.g., runny or stuffy nose; wet cough; and
burning, aching, or red eyes). Research on the effects of air pollution on upper respiratory
symptoms has thus focused on effects in asthmatics.
MRADs result when individuals reduce most usual daily activities and replace them with
less strenuous activities or rest, yet not to the point of missing work or school. For example, a
mechanic who would usually be doing physical work most of the day will instead spend the day
at a desk doing paper work and phone work because of difficulty breathing or chest pain. The
effect of PM2.5 on MRAD was estimated using an effect estimate derived from Ostro and
Rothschild (1989).
More recently published literature examining the relationship between short-term PM2.5
exposure and acute respiratory symptoms was available in the PM ISA (U.S. EPA, 2009b), but
proved to be unsuitable for use in this benefits analysis. In particular, the best available study
5D-18
-------
(Patel et al., 2010) specified a population aged 13-20, which overlaps with the population in
which we asses asthma exacerbation. As we describe in detail below, to avoid the chance of
double-counting impacts, we do not estimate changes in acute respiratory symptoms and asthma
exacerbation among populations of the same age.
Acute Bronchitis. Approximately 4% of U.S. children between the ages of 5 and 17 experience
episodes of acute bronchitis annually (ALA, 2002). Acute bronchitis is characterized by
coughing, chest discomfort, slight fever, and extreme tiredness, lasting for a number of days.
According to the MedlinePlus medical encyclopedia,95 with the exception of cough, most acute
bronchitis symptoms abate within 7 to 10 days. Incidence of episodes of acute bronchitis in
children between the ages of 5 and 17 were estimated using an effect estimate developed from
Dockeryetal. (1996).
Work Loss Days. Health effects from air pollution can also result in missed days of work
(either from personal symptoms or from caring for a sick family member). Days of work lost due
to PM2.5 were estimated using an effect estimate developed from Ostro (1987). Children may
also be absent from school because of respiratory or other diseases caused by exposure to air
pollution, but we have not quantified these effects for this rule.
Work loss days. Health effects from air pollution can also result in missed days of work (either
from personal symptoms or from caring for a sick family member). Days of work lost due to
PlVLs were estimated using an effect estimate developed from Ostro (1987). Ostro (1987)
estimated the impact of PIVb.s on the incidence of work loss days in a national sample of the adult
working population, ages 18 to 65 living in metropolitan areas. Ostro reported that two-week
average PIVb.s levels were significantly linked to work loss days, but there was some year-to-year
variability in the results.
5D.4 Nonfatal Acute Myocardial Infarctions (AMI) (Heart Attacks)
Nonfatal heart attacks have been linked with short-term exposures to PIVb.s in the United
States (Mustafic et al., 2012; Peters et al., 2001; Sullivan et al., 2005; Pope et al., 2006;
Zanobetti and Schwartz, 2006; Zanobetti et al., 2009) and other countries (Poloniecki et al.,
95
See http://www.nlm.nih.gov/medlineplus/ency/article/001087.htm, accessed April 2012.
5D-19
-------
1997; Barnett et al, 2006; Peters et al, 2005). In previous health impact assessments, we have
relied upon a study by Peters et al. (2001) as the basis for the impact function estimating the
relationship between PIVh.s and nonfatal heart attacks. The Peters et al. (2001) study exhibits a
number of strengths. In particular, it includes a robust characterization of populations
experiencing acute myocardial infarctions (AMIs). The researchers interviewed patients within 4
days of their AMI events and, for inclusion in the study, patients were required to meet a series
of criteria including minimum kinase levels, an identifiable onset of pain or other symptoms and
the ability to indicate the time, place and other characteristics of their AMI pain in an interview.
Since the publication of Peters et al. (2001), a number of other single and multi-city
studies have appeared in the literature. These studies include Sullivan et al. (2005), which
considered the risk of PM2.s-related hospitalization for AMIs in King County, Washington; Pope
et al. (2006), based in Wasatch Range, Utah; Zanobetti and Schwartz (2006), based in Boston,
Massachusetts; and, Zanobetti et al. (2009), a multi-city study of 26 U.S. communities. Each of
these single and multi-city studies, with the exception of Pope et al. (2006), measure AMIs using
hospital discharge rates. Conversely, the Pope et al. (2006) study is based on a large registry with
angiographically characterized patients—arguably a more precise indicator of AMI. Because the
Pope et al. (2006) study reflected both myocardial infarctions and unstable angina, this produces
a more comprehensive estimate of acute ischemic heart disease events than the other studies.
However, unlike the Peters study (Peters et al., 2006), Pope and colleagues did not measure the
time of symptom onset, and PM2.5 data were not measured on an hourly basis.
As a means of recognizing the strengths of the Peters study while also incorporating the
newer evidence found in the four single and multi-city studies, we present a range of AMI
estimates. The upper end of the range is calculated using the Peters study, while the lower end of
the range is the result of an equal-weights pooling of these four newer studies. It is important to
note that when calculating the incidence of nonfatal AMI, the fraction of fatal heart attacks is
subtracted to ensure that there is no double-counting with premature mortality estimates.
Specifically, we apply an adjustment factor in the concentration-response function to reflect the
probability of surviving a heart attack. Based on recent data from the Agency for Healthcare
Research and Quality's Healthcare Utilization Project National Inpatient Sample database
(AHRQ, 2009), we identified death rates for adults hospitalized with acute myocardial infarction
5D-20
-------
stratified by age (e.g., 1.852% for ages 18-44, 2.8188% for ages 45-64, and 7.4339% for ages
65+). These rates show a clear downward trend over time between 1994 and 2009 for the
average adult and thus replace the 7% survival rate previously applied across all age groups from
Rosamond et al. (1999).
5D.5 References
Abt Associates, Inc. 2012. BenMAP User's Manual Appendices. Prepared for U.S. Environmental Protection
Agency Office of Air Quality Planning and Standards. Research Triangle Park, NC. September. Available at
.
Agency for Healthcare Research and Quality (AHRQ). 2009. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available at .
American Lung Association (ALA). 2002. Trends in Asthma Morbidity and Mortality. American Lung Association,
Best Practices and Program Services, Epidemiology and Statistics Unit. Available at
.
Babin, S. M, H. S. Burkom, et al. 2007. "Pediatric Patient Asthma-related Emergency Department Visits and
Admissions in Washington, DC, from 2001-2004, and Associations with Air Quality, Socio-economic Status
and Age Group." Environmental Health 6: 9.
Barnett, Adrian G. et al. 2006. "The Effects of Air Pollution on Hospitalizations for Cardiovascular Disease in
Elderly People in Australian and New Zealand Cities." Environmental Health Perspectives 114(7): 1018-1023.
Belanger, K., W. Beckett, E. Triche, M.B. Bracken, T. Holford, P. Ren, J.E. McSharry, D.R. Gold, T.A. Platts-Mills,
and B.P. Leaderer. 2003. "Symptoms of Wheeze and Persistent Cough in the First Year of Life: Associations
with Indoor Allergens, Air Contaminants, and Maternal History of Asthma." American Journal of Epidemiology
158:195-202.
Bell, M. L., K. Ebisu, et al. 2008. "Seasonal and Regional Short-term Effects of Fine Particles on Hospital
Admissions in 202 US Counties, 1999-2005." American Journal of Epidemiology 168(11): 1301-1310.
Chay, K.Y., and M. Greenstone. 2003. "The Impact of Air Pollution on Infant Mortality: Evidence from Geographic
Variation in Pollution Shocks Induced by a Recession." Quarterly Journal of Economics 118(3).
Crouse DL, Peters PA, van Donkelaar A, Goldberg MS, Villeneuve PJ, Brion O, et al. 2012. "Risk of Non-
accidental and Cardiovascular Mortality in Relation to Long-term Exposure to Low Concentrations of Fine
Particulate Matter: A Canadian National-level Cohort Study." Environmental Health Perspectives 120(5):708-
14.
Dockery, D.W., C.A. Pope, X.P. Xu, ID. Spengler, J.H. Ware, M.E. Fay, E.G. Ferris, andF.E. Speizer. 1993. "An
Association between Air Pollution and Mortality in Six U.S. Cities." New England Journal of Medicine
329(24): 1753-1759.
Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis, J.H. Ware, M. Raizenne,
and F.E. Speizer. 1996. "Health Effects of Acid Aerosols On North American Children-Respiratory
Symptoms." Environmental Health Perspectives 104(5):500-505.
Eftim SE, Samet JM, Janes H, McDermott A, Dominici F. 2008. "Fine paniculate matter and mortality: a
comparison of the Six Cities and American Cancer Society cohorts with a Medicare cohort." Epidemiology
19:209-216.
5D-21
-------
Gan, W. Q., H. W. Davies, et al. 2012. "Association of Long-term Exposure to Community Noise and Traffic-
related Air Pollution with Coronary Heart Disease Mortality." American Journal of Epidemiology. 1;175
(9):898-906.
Glad, J.A., L.L. Brink, E.G. Talbott, P.C. Lee, X. Xu, M. Saul, and J. Rager. 2012. "The Relationship of Ambient O3
and PM2 5 Levels and Asthma Emergency Department Visits: Possible Influence of Gender and Ethnicity."
Archives of Environmental & Occupational Health 62 (2): 103-108.
Industrial Economics, Incorporated (ffic). 2006. Expanded Expert Judgment Assessment of 'the Concentration-
Response Relationship Between PM2.s Exposure and Mortality. Prepared for: Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC. September. Available at
.
Kloog, I., B.A. Coull, A. Zanobetti, P. Koutrakis, J.D. Schwartz. 2012. "Acute and Chronic Effects of Particles on
Hospital Admissions in New England." PLoS ONE. Vol 7 (4): 1-8.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi, Y, et al. 2009. Extended follow-up and spatial analysis of
the American Cancer Society study linking particulate air pollution and mortality. HEI Research Report, 140,
Health Effects Institute, Boston, MA.
Krewski, D., R.T. Burnett, M.S. Goldbert, K. Hoover, J. Siemiatycki, M. Jerrett, M. Abrahamowicz, and W.H.
White. 2000. Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of Particulate
Air Pollution and Mortality. Special Report to the Health Effects Institute. Cambridge MA. July.
Kunzli, N., S. Medina, R. Kaiser, P. Quenel, F. Horak Jr, and M. Studnicka. 2001. "Assessment of Deaths
Attributable to Air Pollution: Should We Use Risk Estimates Based on Time Series or on Cohort Studies?"
American Journal of Epidemiology 153(11): 1050-55.
Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. "Reduction in Fine Particulate Air Pollution and
Mortality." American Journal of Respiratory and Critical Care Medicine 173:667-672.
Lepeule J, Laden F, Dockery D, Schwartz J. 2012. "Chronic Exposure to Fine Particles and Mortality: An Extended
Foliow-Up of the Harvard Six Cities Study from 1974 to 2009." Environmental Health Perspectives 120
(7):965-70.
Lipfert, F.W.; Wyzga, R.E.; Baty, J.D.; Miller, J.P. 2006. "Traffic Density as a Surrogate Measure of Environmental
Exposures in Studies of Air Pollution Health Effects: Long-Term Mortality in a Cohort of US Veterans."
Atmospheric Environment 40: 154-169.
Mansfield, Carol and Sumeet Patil, 2006. Peer Review of Expert Elicitation. Memo to Ron Evans, EPA. September
29, 2006. RTI International. Available at
.
Mar, T. F., J. Q. Koenig and J. Primomo. 2010. "Associations between asthma emergency visits and particulate
matter sources, including diesel emissions from stationary generators in Tacoma, Washington." Inhalation
Toxicology 22 (6): 445-8.
Mar, T. F., T. V. Larson, et al. 2004. "An analysis of the association between respiratory symptoms in subjects with
asthma and daily air pollution in Spokane, Washington." Inhalation Toxicology 16(13): 809-15.
Miller, Kristin A., David S. Siscovick, Lianne Sheppard, Kristen Shepherd, Jeffrey H. Sullivan, Garnet L. Anderson,
and Joel D. Kaufman. 2007. "Long-Term Exposure to Air Pollution and Incidence of Cardiovascular Events in
Women." New England Journal of Medicine 356 (5):447-458.
5D-22
-------
Moolgavkar, S.H. 2000. "Air Pollution and Hospital Admissions for Diseases of the Circulatory System in Three
U.S. Metropolitan Areas.'' Journal of the Air an d Waste Management Association 50:1199-1206.
Moolgavkar, S.H. 2003. "Air Pollution and Daily Deaths and Hospital Admissions in Los Angeles and Cook
Counties." In Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report. Boston,
MA: Health Effects Institute.
Mustafic H, Jabre P, Caussin C et al. 2012. "Main Air Pollutants and Myocardial Infarction: A Systematic Review
and Meta-analysis" Journal of the American Medical Association 2012;307(7):713-721.
National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press. Washington, DC.
Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens, andM. White. 2001. "Air Pollution and Exacerbation of Asthma
in African-American Children in Los Angeles." Epidemiology 12(2):200-208.
Ostro, B.D. 1987. "Air Pollution and Morbidity Revisited: A Specification Test." Journal of Environmental
Economics Management 14:87-98.
Ostro, B.D. and S. Rothschild. 1989. "Air Pollution and Acute Respiratory Morbidity: An Observational Study of
Multiple Pollutants." Environmental Research 50:238-247.
Patel, M. M., S. N. Chillrud, J. C. Correa, Y. Hazi, M. Feinberg, D. Kc, S. Prakash, J. M. Ross, D. Levy and P. L.
Kinney. 2010. "Traffic-Related Paniculate Matter and Acute Respiratory Symptoms among New York City
Area Adolescents." Environmental Health Perspectives 118(9): 1338-43.
Peng, R. D., H. H. Chang, et al. 2008. "Coarse paniculate matter air pollution and hospital admissions for
cardiovascular and respiratory diseases among Medicare patients." Journal of the American Medical
Association 299(18): 2172-9.
Peters A. 2005. "Paniculate matter and heart disease: evidence from epidemiological studies." Toxicology and
Applied Pharmacology Sept 1;207(2 Suppl):477-82.
Peters A. 2006. "When a myocardial infarction comes out of the not-so-blue air." Circulation 114(23):2430-1
Peters, A., D. W. Dockery, J. E. Muller and M. A. Mittleman. 2001. "Increased paniculate air pollution and the
triggering of myocardial infarction." Circulation 103 (23): 2810-5.
Poloniecki, J.D., R.W. Atkinson., A.P. de Leon., andH.R. Anderson. 1997. "Daily Time Series for Cardiovascular
Hospital Admissions and Previous Day's Air Pollution in London, UK." Occupational and Environmental
Medicine 54(8):535-540
Pope, C. A., Ill, J. B. Muhlestein, et al. 2006. "Ischemic heart disease events triggered by short-term exposure to
fine paniculate air pollution." Circulation 114(23): 2443-8.
Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. "Respiratory Health and PM10 Pollution:
A Daily Time Series Analysis." American Review of Respiratory Diseases 144:668-674.
Pope, C.A., III, M.J. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer, and C.W. Heath, Jr. 1995.
"Particulate Air Pollution as a Predictor of Mortality in a Prospective Study of U. S. Adults." American Journal
of Respiratory Critical Care Medicine 151:669-674.
Pope, C.A., III, R.T. Burnett, G.D. Thurston, M.J. Thun, E.E. Calle, D. Krewski, and J.J. Godleski. 2004.
"Cardiovascular Mortality and Long-term Exposure to Particulate Air Pollution." Circulation 109: 71-77.
5D-23
-------
Pope, C.A., III, 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 287:1132-1141.
Puett RC, Hart JE, Suh H, Mittleman M, Laden F. 2011. "Paniculate Matter Exposures, Mortality and
Cardiovascular Disease in the Health Professionals Follow-up Study." Environmental Health Perspectives
119(8): 1130-1135.
Puett RC, Hart JE, Yanosky JD, Paciorek C, Schwartz J, Suh H, et al. 2009. "Chronic fine and coarse paniculate
exposure, mortality and coronary heart disease in the Nurses' Health Study." Environmental Health
Perspectives 117:1697-1701.
Roman, Henry A., Katherine D. Walker, Tyra L. Walsh, Lisa Conner, Harvey M. Richmond, Bryan J. Hubbell, and
Patrick L. Kinney. 2008. "Expert Judgment Assessment of the Mortality Impact of Changes in Ambient Fine
Paniculate Matter in the U.S." Environmental Science & Technology 42(7):2268-2274.
Rosamond, W., G. Broda, E. Kawalec, S. Rywik, A. Pajak, L. Cooper, and L. Chambless. 1999. "Comparison of
Medical Care and Survival of Hospitalized Patients with Acute Myocardial Infarction in Poland and the United
States." American Journal of Cardiology 83:1180-1185
Schwartz, J., and L.M. Neas. 2000. "Fine Particles are More Strongly Associated than Coarse Particles with Acute
Respiratory Health Effects in Schoolchildren." Epidemiology 11:6-10.
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.
Slaughter, James C et al. 2005. "Association between paniculate matter and emergency room visits, hospital
admissions and mortality in Spokane, Washington." Journal of Exposure Analysis and Environmental
Epidemiology 15, 153-159.
Sullivan, J., L. Sheppard, et al. 2005. "Relation between short-term fine-paniculate matter exposure and onset of
myocardial infarction." Epidemiology 16(1): 41-8.
U.S. Environmental Protection Agency (U.S. EPA). 2006. Regulatory Impact Analysis, 2006 National Ambient Air
Quality Standards for Paniculate Matter, Chapter 5. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. October. Available on the Internet at
.
U. S. Environmental Protection Agency (U. S. EPA). 2009. Integrated Science Assessment for Particulate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division.
December. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2010a. Quantitative Health Risk Assessment for Particulate
Matter—Final Report. EPA-452/R-10-005. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. September. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010b. Final Regulatory Impact Analysis (RIA)for the SO2
National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning and Standards, Research
Triangle Park, NC. June. Available at .
5D-24
-------
U.S. Environmental Protection Agency (U.S. EPA). 2010c. Regulatory Impact Analysis (RIA)for the Proposed
Reconsideration of the O3 National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning
and Standards, Research Triangle Park, NC. January. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 201 la. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March.
Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2011b. Regulatory Impact Analysis for the Federal
Implementation Plans to Reduce Interstate Transport of Fine Particulate Matter and Os in 27 States;
Correction of SIP Approvals for 22 States. June. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. EPA-452/R-12-003. Office of Air Quality
Planning and Standards, Health and Environmental Impacts Division. December. Available at:
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2008. Characterizing
Uncertainty in Particulate Matter Benefits Using Expert Elicitation. EPA-COUNCIL-08-002. July. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2009a. Regulatory Impact Analysis: National Emission
Standards for Hazardous Air Pollutants from the Portland Cement Manufacturing Industry. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. April. Available at
.
U. S. Environmental Protection Agency (U. S. EPA). 2009b. Integrated Science Assessment for Particulate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division.
December. Available at .
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010a. Review of EPA 's DRAFT
Health Benefits of the Second Section 812 Prospective Study of the Clean Air Act. EPA-COUNCIL-10-001.
June. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010b. Review of Risk
Assessment to Support the Review of the Particulate Matter (PM) Primary National Ambient Air Quality
Standards—External Review Draft (September 2009). EPA-CASAC-10-003. Available at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010c. CASACReview of
Quantitative Health Risk Assessment for Particulate Matter—Second External Review Draft (February 2010).
EPA-CASAC-10-008. Available at
.
5D-25
-------
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2004. Advisory on Plans for
Health Effects Analysis in the Analytical Plan for EPA 's Second Prospective Analysis—Benefits and Costs of
the Clean Air Act, 1990-2020. EPA-SAB-COUNCIL-ADV-04-002. March. Available at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2009b. Review of Integrated
Science Assessment for P articulate Matter (Second External Review Draft, July 2009). EPA-CASAC-10-001.
November. Available at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2011. Review of Valuing
Mortality Risk Reductions for Environmental Policy: A White Paper (December 10, 2010). EPA-SAB-11-Oil
July. Available at
.
Woodruff TJ, Parker JD, Schoendorf KC. 2006. "Fine paniculate matter (PM2s) air pollution and selected causes of
postneonatal infant mortality in California." Environmental Health Perspectives 114(5):786-90.
Woodruff, T.J., J. Grille, and K.C. Schoendorf. 1997. "The relationship between Selected of postneonatal infant
mortality and paniculate air pollution in the United States." Environmental Health Perspectives 105 (6): 608-
612.
Zanobetti A. and Schwartz, J. 2006. "Air pollution and emergency admissions in Boston, MA." Journal of
Epidemiology and Community Health 60(10): 890-5.
Zanobetti, A., M. Franklin, et al. 2009. "Fine paniculate air pollution and its components in association with cause-
specific emergency admissions." Environmental Health 8: 58-60.
5D-26
-------
APPENDIX 5E: INPUTS TO PM2.5 COBENEFIT MODELING
Overview
This section presents inputs used in generating PIVh.s cobenefits estimates including (a)
benefit-per-ton estimates for each sector (dimensioned by mortality study and simulation year)96
(Table 5E-1), and (b) NOx emissions reductions by sector for both the 2025 and post-2025
scenarios (Table 5E-2).97 For additional detail on the approach used to generate PIVh.s cobenefits
estimates and the role played by these two types of inputs, see Chapter 5, Section 5.4.4.
96 Benefit-per-ton estimates were generated for each of the long-term exposure-related mortality studies used in
generating core benefits estimates for this RIA including Krewski et al, 2009 and Lepeule et al., 2012 (see
Appendix 5D, section 5D. 1). Estimates were available for 2025 and 2030, with those being used to model
cobenefits for the 2025 scenario and post-2025 scenario, respectively.
97 Sector-level NOx reductions (for each alternative standard level) were generated using methods described in
Chapter 4, section 4.2 and 4.3. As noted in section 5.4.4, NOx emissions reductions associated with alternative
standard levels considered for this NAAQS review involved seven of the 17 sectors for which we had benefit-per-
ton values and consequently, the cobenefits PM2 5 estimates are based on simulated benefits for those seven sectors.
5E-1
-------
Table 5E-1. Summary of Effect Estimates from Associated Sectors with Change in Long-Term Exposure to PMi.s in Recent
Cohort Studies in North America3
Long-term mortality
study
Emissions sector
air,
locamotive
and
marine
cement
kilns
coke
ovens
ECU
point
electric
arc
furnaces
farro
alloys
integrated
iron and
steel
iron and
steel
non-EGU
point
other
non-
point
other
nonroad
onroad
pulp and
paper
refineries
residenti
al wood
taconite
mining
ocean
going
vessels
Non-
specified
source b
2025 at 7% social discount
Krewskietal., 2009
Lepeuleetal., 2012
$7,221
$16,295
$5,692
$12,856
$10,438
$23,578
$5,245
$11,841
$9,610
$21,708
$4,358
$9,842
$13,424
$30,325
$16,944
$38,262
$6,341
$14,316
$7,859
$17,737
$6,979
$15,749
$7,620
$17,177
$3,746
$8,460
$6,924
$15,636
$13,598
$30,706
$5,998
$13,538
$2,033
$4,586
$6,433
$14,514
2025 at 3% social discount
Krewskietal., 2009
Lepeuleetal., 2012
$8,005
$18,071
6,311
14,257
$11,572
$26,149
$5,815
$13,132
$10,654
$24,074
$4,832
$10,916
$14,883
$33,630
$18,783
$42,433
$7,029
$15,876
$8,713
$19,671
$7,736
$17,466
$8,447
$19,049
$4,154
$9,383
$7,676
$17,340
$15,075
$34,054
$6,649
$15,014
$2,254
$5,086
$7,132
$16,096
2030 at 7% social discount
Krewskietal., 2009
Lepeuleetal., 2012
$7,829
$17,662
$6,125
$13,830
$11,056
$24,969
$5,591
$12,621
$10,219
$23,080
$4,637
$10,473
$14,264
$32,217
$18,373
$41,475
$6,814
$15,380
$8,469
$19,109
$7,587
$17,116
$8,214
$18,512
$4,017
$9,070
$7,531
$17,000
$14,695
$33,176
$6,403
$14,451
$2,258
$5,091
$6,941
$15,655
2030 at 3% social discount
Krewskietal., 2009
Lepeuleetal., 2012
$8,680
$19,587
$6,791
$15,338
$12,258
$27,691
$6,199
$13,996
$11,330
$25,596
$5,142
$11,615
$15,815
$35,729
$20,369
$45,997
$7,554
$17,057
$9,389
$21,192
$8,412
$18,983
$9,106
$20,530
$4,454
$10,059
$8,349
$18,854
$16,293
$36,794
$7,099
$16,026
$2,504
$5,646
$7,695
$17,362
a Benefit-per-ton estimates reflect application of the 20-year segmented lag used in the core analysis (see section 5.6.4.1) together with either a 3% or 7% social
discount rate as noted in the table. In addition, separate sets of benefit-per-ton estimates were generate for 2025 and 2030, reflecting application of appropriate
projected demographic and baseline incidence data (see section 5.4.4 for additional detail).
b Benefit-per-ton estimates for the non-specified source category were generated as a weighted average of values for the 17 source categories, with weighting
based on sector-specific NOx emissions for 2005 obtained from Farm et al., 2012.
5E-2
-------
Table 5E-2. Sector-Specific NOx Emissions Reductions for Each Alternative Standard Level"
Emissions Sector
Aircraft, locomotives and marine vessels
Area sources
Cement kilns
Electricity Generating Units
Industrial point sources
Non-road mobile sources
On-road mobile sources
Pulp and paper facilities
Refineries
Residential wood combustion
Unknown sector
TOTAL
Alternative Standard Level
70ppb
CANOx
-
-
-
-
-
-
-
-
-
-
53,289
53,289
nonCA NOx
-
45,708
20,135
32,315
382,743
4,984
-
357
8,243
-
154,343
648,828
65ppb
CANOx
-
-
-
-
-
-
-
-
-
-
104,708
104,708
nonCA NOx
-
94,340
43,867
217,881
741,588
12,863
-
617
12,384
-
752,162
1,875,702
eoppb
CANOx
-
-
-
-
-
-
-
-
-
-
143,916
143,916
nonCA NOx
-
95,332
43,867
244,079
750,620
12,863
-
617
12,411
-
2,234,709
3,394,497
a All values are tons of NOx reductions (75ppb vs alternative standard). Results are presented both for "CA NOx" (emissions in CA only - used in post-2025
scenario PM2 5 cobenefits modeling) and "nonCA NOx" (emissions reductions outside of CA - used in 2025 scenario PM2 5 cobenefits modeling)
5E-3
-------
CHAPTER 6: IMPACTS ON PUBLIC WELFARE OF ATTAINMENT STRATEGIES
TO MEET PRIMARY AND SECONDARY OZONE NAAQS
Overview
This chapter provides a discussion of the welfare-related benefits of meeting alternative
primary and secondary ozone standards. Welfare benefits of reductions in ambient ozone
include increased growth and/or biomass production in sensitive plant species, including forest
trees, increased crop yields, reductions in visible foliar injury, increased plant vigor (e.g.
decreased susceptibility to harsh weather, disease, insect pest infestation, and competition), and
changes in ecosystems and associated ecosystem services. We provide a limited quantitative
analysis for effects associated with changes in yields of commercial forests and agriculture, and
associated changes in carbon sequestration and storage.
The EPA is proposing to revise the level of the secondary standard to within the range
proposed for the primary standard of 65 parts per billion (ppb) to 70 ppb to provide increased
protection against vegetation-related effects on public welfare. As an initial matter, the EPA is
proposing that ambient ozone concentrations in terms of a three-year average W126 index value
within the range from 13 parts per million-hours (ppm-hours) to 17 ppm-hours, would provide
the requisite protection against known or anticipated adverse effects to the public welfare, which
data analyses indicate would provide air quality in terms of three-year average W126 index
values of a range at or below 13 ppm-hours to 17 ppm-hours. Data analyses also indicate that
actions taken to attain a standard in the range of 65 ppb to 70 ppb would also improve air quality
as measured by the W126 metric. The quantitative analysis in this chapter assesses the welfare
benefits of strategies to attain ozone standard levels of 65 to 70 ppb.
In addition to the direct welfare benefits of decreased levels of ambient ozone, the
emissions reduction strategies used to demonstrate attainment with alternative ozone standards
may result in additional benefits associated with reductions in nitrogen deposition and reductions
in ambient concentrations of PIVb.s and its components. These additional benefits include
reductions in nutrient enrichment and acidification impacts on sensitive aquatic and terrestrial
ecosystems and improvements in visibility in state and national parks, wilderness areas, and in
the areas where people live and work. We are not able to quantify or monetize these benefits in
this RIA.
6-1
-------
6.1 Welfare Benefits of Strategies to Attain Primary and Secondary Ozone Standards
The Clean Air Act defines welfare effects to include any non-health effects, including
direct economic damages in the form of lost productivity of crops and trees, indirect damages
through alteration of ecosystem functions, indirect economic damages through the loss in value
of recreational experiences or the existence value of important resources, and direct damages to
property, either through impacts on material structures or by soiling of surfaces (Section 302(h)
(42 U.S.C. § 7602(h)). For welfare effects associated with changes to ecosystem functions, we
use the concept of ecosystem services as a useful framework for analyzing the impact of
ecosystem changes on public welfare. Ecosystem services can be generally defined as the
benefits that individuals and organizations obtain from ecosystems. The EPA has defined
ecological goods and services as the "outputs of ecological functions or processes that directly or
indirectly contribute to social welfare or have the potential to do so in the future. Some outputs
may be bought and sold, but most are not marketed" (U.S. EPA, 2006). Changes in these
services can affect human well-being by affecting security, health, social relationships, and
access to basic material goods (MEA, 2005).
This RIA employs reductions in nitrogen oxides (NOx) and volatile organic compound
(VOC) emissions to demonstrate attainment with alternative levels of the NAAQS. Reductions
in these emissions will result in changes in ambient concentrations of ozone, as well as changes
in ambient concentrations of NOx, PIVh.s and its components, and deposition of nitrogen. It is
appropriate and reasonable to include all the benefits associated with these emissions reductions
to provide a comprehensive understanding of the likely public welfare impacts of attaining
alternative standards. Table 6-1 shows the welfare effects associated with emissions of NOx and
VOC. The following subsections discuss the direct benefits of reducing ambient ozone
concentrations and the additional welfare benefits associated with reduced emissions of NOx and
VOC.
6-2
-------
Table 6-1. Welfare Effects of NOx and VOC Emissions
. , , . „„,. , Atmospheric and
Atmospheric Effects _ .*. _ „.. ^
Deposition Ettects
Pollutant
Vegetation Visibility Materials „,.
T . 6 ,„ , T . J , _ Climate
Injury (Ozone) Impairment Damage
Deposition Effects
Ecosystem
Effects —
(Organics)
Acidification Nitrogen
(freshwater) Enrichment
NOX i/ i/ i/ i/ •/ •/
VOCs S S S S
6.2 Welfare Benefits of Reducing Ozone
Ozone can affect ecological systems, leading to changes in the ecological community and
influencing the diversity, health, and vigor of individual species (U.S. EPA, 2013). Ozone causes
discernible injury to a wide array of vegetation (U.S. EPA, 2013). In terms of forest productivity
and ecosystem diversity, ozone may be the pollutant with the greatest potential for region-scale
forest impacts (U.S. EPA, 2013). Studies have demonstrated repeatedly that ozone
concentrations observed in polluted areas can have substantial impacts on plant function (De
Steiguer et al,. 1990; Pye, 1988).
When ozone is present in ambient air, it can enter the leaves of plants, where it can cause
significant cellular damage. Like carbon dioxide and other gaseous substances, ozone enters
plant tissues primarily through the stomata in leaves in a process called "uptake" (Winner and
Atkinson, 1986). Once sufficient levels of ozone (a highly reactive substance), or its reaction
products, reaches the interior of plant cells, it can inhibit or damage essential cellular
components and functions, including enzyme activities, lipids, and cellular membranes,
disrupting the plant's osmotic (i.e., water) balance and energy utilization patterns (U.S. EPA,
2013; Tingey and Taylor, 1982). With fewer resources available, the plant reallocates existing
resources away from root growth and storage, above ground growth or yield, and reproductive
processes, and toward leaf repair and maintenance, leading to reduced growth and/or
reproduction. Studies have shown that plants stressed in these ways may exhibit a general loss
of vigor, which can lead to secondary impacts that modify plants' responses to other
environmental factors. Specifically, plants may become more sensitive to other air pollutants, or
-------
more susceptible to disease, pest infestation, harsh weather (e.g., drought, frost) and other
environmental stresses, which can all produce a loss in plant vigor in ozone-sensitive species that
over time may lead to premature plant death. Furthermore, there is evidence that ozone can
interfere with the formation of mycorrhizae, essential symbiotic fungi associated with the roots
of most terrestrial plants, by reducing the amount of carbon available for transfer from the host to
the symbiont (U.S. EPA, 2013).
This ozone damage may or may not be accompanied by visible injury on leaves, and
likewise, visible foliar injury may or may not be a symptom of the other types of plant damage
described above. Foliar injury is usually the first visible sign of injury to plants from ozone
exposure and indicates impaired physiological processes in the leaves (Grulke, 2003). When
visible injury is present, it is commonly manifested as chlorotic or necrotic spots, and/or
increased leaf senescence (accelerated leaf aging). Visible foliar injury reduces the aesthetic
value of ornamental vegetation and trees in urban landscapes and negatively affects scenic vistas
in protected natural areas.
Ozone can produce both acute and chronic injury in sensitive species depending on the
concentration level and the duration of the exposure. Ozone effects also tend to accumulate over
the growing season of the plant, so that even lower concentrations experienced for a longer
duration have the potential to create chronic stress on sensitive vegetation. Not all plants,
however, are equally sensitive to ozone. Much of the variation in sensitivity between individual
plants or whole species is related to the plant's ability to regulate the extent of gas exchange via
leaf stomata (e.g., avoidance of ozone uptake through closure of stomata) and the relative ability
of species to detoxify ozone-generated reactive oxygen free radicals (U.S. EPA, 2013; Winner,
1994). After injuries have occurred, plants may be capable of repairing the damage to a limited
extent (U.S. EPA, 2013). Because of the differing sensitivities among plants to ozone, ozone
pollution can also exert a selective pressure that leads to changes in plant community
composition. Given the range of plant sensitivities and the fact that numerous other
environmental factors modify plant uptake and response to ozone, it is not possible to identify
threshold values above which ozone is consistently toxic for all plants.
6-4
-------
Because plants are at the base of the food web in many ecosystems, changes to the plant
community can affect associated organisms and ecosystems (including the suitability of habitats
that support threatened or endangered species and below ground organisms living in the root
zone). Ozone impacts at the community and ecosystem level vary widely depending upon
numerous factors, including concentration and temporal variation of tropospheric ozone, species
composition, soil properties and climatic factors (U.S. EPA, 2013). In most instances, responses
to chronic or recurrent exposure in forested ecosystems are subtle and not observable for many
years. These injuries can cause stand-level forest decline in sensitive ecosystems (U.S. EPA,
2013, McBride et al, 1985; Miller et al, 1982). It is not yet possible to predict ecosystem
responses to ozone with certainty; however, considerable knowledge of potential ecosystem
responses is available through long-term observations in highly damaged forests in the U.S. (U.S
EPA, 2013). Biomass loss due to ozone exposure affects climate regulation by ecosystems by
reducing carbon sequestration. More carbon stays in the atmosphere because carbon uptake by
forests is reduced. The studies cited in the Ozone ISA demonstrate a consistent pattern of
reduced carbon uptake because of ozone damage, with some of the largest reductions projected
over North America (U.S. EPA, 2013).
Ozone also directly contributes to climate change because tropospheric ozone traps heat,
leading to increased surface temperatures. Projections of radiative forcing due to changing
ozone concentrations over the 21st century show wide variation, due in large part to the
uncertainty of future emissions of source gases (U.S. EPA 2014). However, reduction of
tropospheric ozone concentrations could provide an important means to slow climate change in
addition to the added benefit of improving surface air quality (U.S. EPA, 2014).
While it is clear that increases in tropospheric ozone lead to warming, the precursors of
ozone also have competing effects on methane, complicating emissions reduction strategies. A
decrease in carbon monoxide or VOC emissions would shorten the lifetime of methane, leading
to an overall cooling effect. A decrease in NOx emissions could lengthen the methane lifetime in
certain regions, leading to warming (U.S. EPA, 2014). Additionally, some strategies to reduce
ozone precursor emissions could also lead to the reduced formation of aerosols (e.g., nitrates and
sulfates) that currently have a cooling effect.
6-5
-------
In this RIA, we are able to quantify only a small portion of the welfare impacts associated
with reductions in ozone concentrations to meet alternative ozone standards. Using a model of
commercial agriculture and forest markets, we are able to analyze the effects on consumers and
producers of forest and agricultural products of changes in the W126 index resulting from
meeting alternative standards within the proposed range of 70 to 65 ppb, as well as a lower
standard level of 60 ppb. We also assess the effects of those changes in commercial agricultural
and forest yields on carbon sequestration and storage. This analysis provides limited quantitative
information on the welfare benefits of meeting these alternative standards, focused only on one
subset of ecosystem services. Commercial and non-commercial forests provide a number of
additional services, including medicinal uses, non-commercial food and fiber production, arts
and crafts uses, habitat, recreational uses, and cultural uses for Native American tribes. A more
complete discussion of these additional ecosystem services is provided in the final Welfare Risk
and Exposure Assessment for Ozone (WREA) (U.S. EPA, 2014).
6.3 Additional Welfare Benefits of Strategies to Meet the Ozone NAAQS
Reductions in emissions of NOx and VOC are associated with additional welfare benefits,
including reductions in nutrient enrichment and acidification impacts on sensitive aquatic and
terrestrial ecosystems and improvements in visibility in state and national parks, wilderness
areas, and in the areas where people live and work.
Excess nitrogen deposition can lead to eutrophication of estuarine waters, which is
associated with a range of adverse ecological effects. These include low dissolved oxygen
(DO), harmful algal blooms (HABs), loss of submerged aquatic vegetation (SAV), and low water
clarity. Low DO disrupts aquatic habitats, causing stress to fish and shellfish, which, in the
short-term, can lead to episodic fish kills and, in the long-term, can damage overall growth in
fish and shellfish populations. HAB are often toxic to fish and shellfish, lead to fish kills and
aesthetic impairments of estuaries, and can in some instances be harmful to human health. SAV
provides critical habitat for many aquatic species in estuaries and, in some instances, can also
protect shorelines by reducing wave strength. Low water clarity is in part the result of
accumulations of both algae and sediments in estuarine waters. In addition to contributing to
6-6
-------
declines in SAV, high levels of turbidity also degrade the aesthetic qualities of the estuarine
environment.
Nutrient enrichment from nitrogen deposition to terrestrial ecosystems is causally linked
to alteration of species richness, species composition, and biodiversity (U.S. EPA, 2008b).
Nitrogen enrichment occurs over a long time period; as a result, it may take as much as 50 years or
more to see changes in ecosystem conditions, indicators, and services.
Terrestrial acidification resulting from deposition of nitrogen can result in declines in
sensitive tree species, such as red spruce (Picea rubens) and sugar maple (Acer sacchamm), and
can also impact other plant communities including shrubs and lichen (U.S. EPA, 2008b).
Biological effects of acidification in terrestrial ecosystems are generally linked to aluminum
toxicity and decreased ability of plant roots to take up base cations (U.S. EPA, 2008b).
Terrestrial acidification affects several important ecosystem services, including declines in
habitat for threatened and endangered species, declines in forest aesthetics, declines in forest
productivity, and increases in forest soil erosion and reductions in water retention.
Aquatic acidification resulting from deposition of nitrogen can result in effects on health,
vigor, and reproductive success for aquatic species; and effects on biodiversity. Deposition of
nitrogen results in decreases in the acid neutralizing capacity and increases in inorganic aluminum
concentration, which contribute to declines in zooplankton, macro invertebrates, and fish species
richness in aquatic ecosystems (U.S. EPA, 2008b).
Reductions in NOx emissions will improve visibility in parks and wilderness areas and in
places where people live and work because of their impact on light extinction (U.S. EPA, 2009).
Good visibility increases quality of life where individuals live and work, and where they travel
for recreational activities, including sites of unique public value, such as the Great Smoky
Mountains National Park (U. S. EPA, 2009). Particulate nitrate is an important contributor to
light extinction in California and the upper Midwestern U.S., particularly during winter (U.S.
EPA, 2009). While EPA typically estimates the visibility benefits associated with reductions in NOx
(U.S. EPA, 2008a), we have not done so here because we do not have estimates of the changes in
particulate nitrate needed to calculate changes in light extinction and the resulting changes in
economic benefits.
6-7
-------
Strategies implemented by state and local governments to reduce emissions of ozone
precursors may also impact emissions of CC>2 or other long-lived climate gases. Our ability to
quantify the climate effects of the proposed standard levels is limited due to lack of available
information on the energy and associated climate gas implications of control technologies
assumed in the illustrative control strategy alternatives, remaining uncertainties regarding the
impact of ozone precursors on climate change, and lack of available information on the co-
controlled greenhouse gas (GHG) emission reductions. As a result, we do not attempt to
quantify the impacts of the illustrative attainment scenarios on GHG emissions and impacts.
6.4 Analysis of Commercial Agricultural and Forestry Related Benefits Using the
Forest and Agricultural Sector Optimization Model - Greenhouse Gas Version
(FASOMGHG)
To estimate the commercial timber effects of ozone induced biomass loss we used the
FASOMGHG (Adams et al, 2005) model for the forest and agricultural sectors to calculate the
market-based welfare benefits associated with the illustrative attainment strategies for the three
alternative ozone standard levels of 70, 65, and 60 ppb incremental to attainment of the current
standard level of 75 ppb. The air quality surfaces used are discussed in detail in Chapter 3; the
alternative primary standards modeled were recalculated to a W126 index appropriate for use
with the exposure-response functions available for trees and crops. This section provides a brief
summary of the analytical approach and results of the analysis. More details of the
FASOMGHG modeling conducted for this RIA are provided in Appendix 6A, while additional
details on the overall FASOMGHG methodology are provided in Appendix 6B of the WREA
(U.S. EPA, 2014).
6.4.1 Summary of the Analytical Approach
We used the ozone exposure-response functions evaluated in the ISA (U.S. EPA, 2013) for
tree seedlings to calculate relative yield loss (RYL), which is equivalent to relative biomass loss,
for trees over their entire life span. The RYL for species were aggregated into average RYL for
FASOMGHG forest types, based on mapping tree species to forest types using the Atlas of
United States Trees (Little, 1971, 1976, 1977, 1978).
6-8
-------
We used the NCLAN ozone exposure-response functions evaluated in the ISA to generate
RYL for commercial agricultural crops. For those crops that do not have E-R functions, we
assign them RYLs for each scenario based on the crop proxy mapping shown in Table 6 A-2 of
Appendix 6A.98 The RYL for each crop are aggregated to the regional level by computing a
weighted average RYL with weights determined by a county's share of production for the crop.
Additional details of the calculations of RYL are provided in Appendix 6 A.
Yield gains for both forest species and crops are calculated as the difference in RYL
between the 75 ppb standard baseline and the attainment scenarios for 70, 65, and 60 ppb
alternative standard levels. The FASOMGHG model requires estimates of yields over a time
horizon from 2010 to 2040. As such, we needed to specify yield gains over this range of years.
Yield gains are calculated for three periods, 2010 to 2025, 2025-2038, and post-2038. The pre-
2025 period has no changes in yields. The 2025-2038 period represents the effects on the W126
index of attainment of the alternative standards across the U.S. with the exception of California.
The post-2038 period includes the effects on the W126 index of attaining everywhere across the
U.S. including California. There is clearly uncertainty in the path of the yield changes
introduced by uncertainties about the specific time pattern of emissions reductions that will be
applied in California to attain alterative standards.
Changes in yield are associated with changes in consumer and producer/farmer surplus.
Consumer surplus is the difference between what a consumer would be willing to pay for a
product and the price they have to pay for the product. Producer surplus refers to the benefit, or
profit, a producer receives from providing a good or service at a market price when they would
have been willing to sell that good or service at a lower price. In general, increases in yields will
cause crop and timber prices to fall. These reductions in prices will have different impacts on
consumer and producer surplus. Overall effects on producer and consumer surplus depend on
the (1) ability of producers/farmers to substitute other crops that are less ozone sensitive, and (2)
responsiveness of demand and supply. The FASOMGHG model estimates changes in consumer
and producer surplus and net welfare (sum of consumer and producer surplus). The
98 For oranges, rice, and tomatoes, which have ozone E-R functions that are not W126-based (they are defined based
on alternative measures of ozone concentrations), we directly used the median RYG values under the "13 ppm-hrs"
ozone concentration reported in Table G-7 of Lehrer et al. (2007).
6-9
-------
FASOMGHG model also provides estimates of the changes in carbon sequestration for the
commercial forestry and agricultural sectors.
The model calculates market equilibria under each of the alternative standard scenarios
reflecting different forest and agricultural yields resulting from different ozone exposures. By
comparing the market equilibria under different scenarios, we can calculate the welfare and
carbon sequestration impacts of alternative ozone standards for the U.S. agricultural and forest
sector.
6.4.2 Summary of FASOMGHG Results
Tables 6-2 and 6-4 show the estimated changes in consumer and producer surplus
associated with attainment of the alternative ozone standards compared to attaining the current
standard for the forestry and agricultural sectors, respectively. Tables 6-3 and 6-5 show the
percent change in consumer and producer surplus associated with attainment of the alternative
ozone standards compared to attaining the current standard for the forestry and agricultural
sectors, respectively. Consumer and producer welfare are affected more in the forestry sector
than the agricultural sector. In general, consumer welfare increases in both the forestry and
agricultural sectors because higher yields lead to lower prices. Because the quantity demanded
for most forestry and agricultural commodities is not highly responsive to changes in price,
producer surplus often declines when lower prices reduce producer profits more than can be
offset by higher yields. In other words, consumers do not increase their demand in response to
the falling prices enough to offset the producer's loss of revenue. The increase in consumer
welfare is not as large as the loss of producer welfare resulting in net welfare losses in the
forestry sector nationally.
6-10
-------
Table 6-2. Change in Consumer and Producer Surplus in the Forestry Sector from
Attaining Alternative Ozone Standard Levels Compared to Attaining the
Current Ozone Standard (Million 2011$)
Consumer
Surplus
Producer
Surplus
Alternative Standard Level
70
65
60
ppb
ppb
ppb
70
65
60
ppb
ppb
ppb
2010
3
11
30
1,203
211
-233
2015
Change
24,592
24,596
24,742
Change
-22,160
-24,091
2020
with
-8
8
136
with
1,768
-480
-23,711 -1,731
2025
Respect to
111
305
719
Respect to
703
1,464
1,455
2030
2035
2040
Existing Standard
136
337
533
225
552
1,094
61
162
311
Existing Standard
-1,167
-3,044
-1,963
523
-39,691
-39,291
-52,496
-10,246
-8,082
Table 6-3. Percent Change in Consumer and Producer Surplus in the Forestry Sector from
Attaining Alternative Ozone Standard Levels Compared to Attaining the
Current Ozone Standard
Consumer
Surplus
Producer
Surplus
Alternative Standard Level
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
2010 2015 2020 2025
2030
2035 2040
Percent Change with Respect to Existing Standard
0.00 3.10 0.00 0.01
0.00 3.10 0.00 0.04
0.00 3.12 0.02 0.09
0.02
0.04
0.06
0.03 0.01
0.06 0.02
0.12 0.03
Percent Change with Respect to Existing Standard
0.15% -2.28 0.18 0.07
0.03% -2.48 -0.05 0.14
-0.03% -2.44 -0.18 0.14
-0.12
-0.32
-0.20
0.05 -5.04
-3.88 -0.98
-3.84 -0.78
Table 6-4. Change in Consumer and Producer Surplus in the Agricultural Sector from
Attaining Alternative Ozone Standard Levels Compared to Attaining the
Current Ozone Standard (Million 2011$)
Product
Consumer
Surplus
Producer
Surplus
Alternative
Standard
Level
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
2010
0
-1
0
1,202
216
-176
2015
0
1
47
2,454
530
1,015
2020 2025
Change with Respect to
0 113
4 262
9 462
Change with Respect to
2030
Existing Standard
66
148
100
Existing Standard
1,796 609 -1,144
-407 1,174 -2,933
-1,470 1,320 -1,492
2035
58
289
408
675
-39,605
-38,788
2040
-10
24
66
-52,504
-10,211
-8,119
6-11
-------
Table 6-5. Percent Change in Consumer and Producer Surplus in the Agricultural Sector
from Attaining Alternative Ozone Standard Levels Compared to Attaining the
Current Ozone Standard
Product
Consumer
Surplus
Producer
Surplus
Alternative
Standard
Level
70ppb
65 ppb
60ppb
70 ppb
65 ppb
60 ppb
2010
2015
2020
2025
Percent Change with Respect
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Percent Change with Respect
0.17
0.03
-0.02
0.30
0.06
0.12
0.22
-0.05
-0.18
0.07
0.14
0.15
2030
to Existing Standard
0.00
0.00
0.00
to Existing Standard
-0.14
-0.36
-0.18
2035
0.00
0.00
0.00
0.08
-4.53
-4.44
2040
0.00
0.00
0.00
-5.78
-1.12
-0.89
Since the forestry and agriculture sectors are interlinked and factors affecting one sector
can lead to changes in the other, it is important to consider the overall effect of ozone changes in
the context of producer and consumer welfare across both sectors. The impacts on consumer
surplus are positive for both sectors, with benefits increasing with lower alternative standards.
For producer surplus, however, impacts are negative for all alternative standards. Table 6-6 and
6-8 present the annualized surplus (over the period 2010 to 2040) for both sectors using 3 and 7
percent discount rates, while Table 6-7 and 6-9 present the percent change in surplus for both
sectors using 3 and 7 percent discount rates.
6-12
-------
Table 6-6. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry from Attaining Alternative Ozone Standard Levels Compared to
Attaining the Current Ozone Standard, 2010-2040, Million 2011$ (3% Discount
Rate)
Consumer surplus
Producer surplus
Total surplus
Alternative
Standard Level
70ppb
65 ppb
60ppb
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
Agriculture
Change
28
86
132
Change
-3,601
-5,035
-4,741
Change
-3,573
-4,949
-4,608
Forestry Total
with Respect to Existing Standard
4,552 4,580
4,592 4,678
4,744 4,877
with Respect to Existing Standard
-4,534 -8,135
-4,524 -9,559
-4,684 -9,425
with Respect to Existing Standard
18 -3,555
68 -4,882
60 -4,548
Table 6-7. Annualized Percent Changes in Consumer and Producer Surplus in Agriculture
and Forestry from Attaining Alternative Ozone Standard Levels Compared to
Attaining the Current Ozone Standard, 2010-2040, (3% Discount Rate)
Consumer surplus
Producer surplus
Total surplus
Alternative
Standard Level
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
Agriculture Forestry
Percent Change with Respect to
0.00 0.57
0.00 0.57
0.01 0.59
Percent Change with Respect to
-0.44 -3.35
-0.62 -3.34
-0.58 -3.46
Percent Change with Respect to
-0.13 0.00
-0.18 0.01
-0.17 0.01
Total
Existing Standard
0.17
0.17
0.18
Existing Standard
-0.85
-1.00
-0.99
Existing Standard
-1.10
-0.13
-0.12
Table 6-8. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry from Attaining Alternative Ozone Standard Levels Compared to
6-13
-------
Attaining the Current Ozone Standard, 2010-2040, Million 2011$ (7% Discount
Rate)
Consumer surplus
Producer surplus
Total surplus
Alternative
Standard Level
70ppb
65 ppb
60ppb
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
Agriculture
Change
21
61
100
Change
-945
-2,719
-2,635
Change
-923
-2,658
-2,535
Forestry Total
with Respect to Existing Standard
5,570 5,592
5,599 5,660
5,721 5,821
with Respect to Existing Standard
-5,561 -6,506
-5,555 -8,273
-5,694 -8,328
with Respect to Existing Standard
9 -914
45 -2,613
27 -2,508
Table 6-9. Annualized Percent Changes in Consumer and Producer Surplus in Agriculture
and Forestry from Attaining Alternative Ozone Standard Levels Compared to
Attaining the Current Ozone Standard, 2010-2040, (7% Discount Rate)
Consumer surplus
Producer surplus
Total surplus
Alternative
Standard Level
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
70 ppb
65 ppb
60 ppb
Agriculture Forestry
Percent Change with Respect to
0.00 0.71
0.00 0.72
0.01 0.73
Percent Change with Respect to
-0.12 -4.26
-0.34 -4.25
-0.33 -4.36
Percent Change with Respect to
-0.03 0.00
-0.10 0.00
-0.09 0.00
Total
Existing Standard
0.20
0.21
0.21
Existing Standard
-0.70
-0.89
-0.89
Existing Standard
-0.02
-0.07
-0.07
The impacts of the simulations of meeting the existing and alternative ozone standards on
carbon sequestration potential in U.S. forest and agricultural sectors are presented in Table 6-10,
in millions of metric tons of carbon dioxide equivalence, and the percent change in carbon
sequestration potential is presented in Table 6-11. As shown in the table, much greater
6-14
-------
sequestration changes are projected in the forest sector than in the agricultural sector. The
baseline stock of carbon storage decreases over time for agriculture because the agriculture
sector GHG emissions sources are released every year and soil carbon sequestration stabilizes
over the 30-year period. There are only small increases in net carbon sequestration compared to
the existing standard for each of the alternative scenarios modeled.
While we did not quantify the effects in this RIA, increases in growth for trees in urban
settings that results from reduced ozone concentrations can have additional benefits from
removal of air pollution. The WREA (U.S. EPA, 2014) provides a case study of the potential
impacts of ozone reductions on pollution removal in several eastern U.S. urban areas using the i-
Tree model. See appendix 6D of the WREA (U.S. EPA, 2014) for details and references for the
i-Tree model.
Table 6-10. Increase in Carbon Sequestration from Attaining Alternative Ozone
Standard Levels Compared to Attaining the Current Ozone Standard,
MMtCO2e
Agriculture
Forestry
Alternative Standard Level
70ppb
65 ppb
60ppb
70 ppb
65 ppb
60 ppb
2010
2020
2030 2040 2010-2040*
Change with Respect to Existing Standard
0
0
1
1
2
4
3 1 24
64 62
11 11 132
Change with Respect to Existing Standard
-1
-2
-12
-51
-53
-94
100 259 1,537
323 774 5,207
465 1,189 7,739
6-15
-------
Table 6-11. Percent Change in Carbon Sequestration from Attaining Alternative Ozone
Standard Levels Compared to Attaining the Current Ozone Standard
Agriculture
Forestry
Alternative Standard Level
70ppb
65 ppb
60ppb
70 ppb
65 ppb
60 ppb
2010 2020
2030 2040 2010-2040*
Percent Change with Respect to Existing Standard
0.00 0.00
0.00 0.01
0.00 0.03
0.03 0.01 0.01
0.05 0.06 0.02
0.10 0.13 0.05
Percent Change with Respect to Existing Standard
0.00 -0.07
0.00 -0.07
-0.02 -0.13
0.13 0.32 0.10
0.41 0.95 0.34
0.60 1.47 0.50
6.5 References
Adams, D.; Alig, R.; McCarl, B.A.; Murray, B.C. (2005). FASOMGHG Conceptual Structure, and Specification:
Documentation. Available at http://agecon2.tamu.edu/people/faculty/mccarl-bruce/FASOM.html
De Steiguer, I, Pye, I, Love, C. 1990. Air Pollution Damage to U.S. Forests. Journal of Forestry, 88(8), 17-22.
Grulke, N.E. 2003. The physiological basis of ozone injury assessment attributes in Sierran conifers. In A.
Bytnerowicz, M. J. Arbaugh, & R. Alonso (Eds.), Ozone air pollution in the Sierra Nevada: Distribution and
effects on forests, (pp. 55-81). New York, NY: Elsevier Science, Ltd.
McBride, J.R., Miller, P.R., Laven, R.D. 1985. Effects of oxidant air pollutants on forest succession in the mixed
conifer forest type of southern California. In: Air Pollutants Effects on Forest Ecosystems, Symposium
Proceedings, St. P, 1985, p. 157-167.
Millennium Ecosystem Assessment Board (MEA). 2005. Ecosystems and Human Well-being: Synthesis.
Washington, DC: World Resources Institute. Available on the Internet at
.
Miller, P.R., O.C. Taylor, R.G. Wilhour. 1982. Oxidant air pollution effects on a western coniferous forest
ecosystem. Corvallis, OR: U.S. Environmental Protection Agency, Environmental Research Laboratory
(EPA600-D-82-276).
Pye, J.M. 1988. Impact of ozone on the growth and yield of trees: A review. Journal of Environmental Quality, 17,
347-360.
Tingey, D.T., and Taylor, G.E. 1982. Variation in plant response to ozone: a conceptual model of physiological
events. In M.H. Unsworth & D.P. Omrod (Eds.), Effects of Gaseous Air Pollution in Agriculture and
Horticulture, (pp. 113-138). London, UK: Butterworth Scientific.
U.S. Environmental Protection Agency (U.S. EPA). 2006. Ecological Benefits Assessment Strategic Plan. EPA-240-
R-06-001. Office of the Administrator. Washington, DC. October. Available on the Internet at
.
6-16
-------
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis, 2008 National Ambient Air
Quality Standards for Ground-level Ozone, Chapter 6. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. March. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2008b. Integrated Science Assessment for Oxides of Nitrogen
and Sulfur—Ecological Criteria National (Final Report). National Center for Environmental Assessment,
Research Triangle Park, NC. EPA/600/R-08/139. December. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for Paniculate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division.
December. Available on the Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2013. Integrated Science Assessment of Ozone and Related
Photochemical Oxidants (Final Report). EPA-600/R-10/076F. February. Available on the Internet at
http://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=247492.
U.S. EPA. (U.S. Environmental Protection Agency). 2014 Welfare Risk and Exposure Assessment (Final). August.
Available on the internet at: http://www.epa.gov/ttn/naaqs/standards/ozone/data/20141021welfarerea.pdf.
Winner, W.E. 1994. Mechanistic analysis of plant responses to air pollution. Ecological Applications, 4(4), 651 -
661.
Winner, W.E., and C. J. Atkinson. 1986. Absorption of air pollution by plants, and consequences for growth.
Trends in Ecology and Evolution 1:15-18.
6-17
-------
APPENDIX 6A: METHODS AND DATA USED TO DEVELOP ESTIMATES OF
OZONE EFFECTS ON CROP AND FOREST PRODUCTIVITY
Incorporating the impacts of different ambient ozone concentration levels into
FASOMGHG requires determining crop yield and forest productivity impacts associated with
changes in concentrations. Productivity impacts are required for each crop/region and forest
type/region combination included within the model. In this section, we describe our methods for
calculating relative yield losses (RYLs) and relative yield gains (RYGs) of crops and tree species
under alternative ambient ozone concentration levels.
These data are essential for our market analysis because crop and forest yields play an
important role in determining the economic returns to agricultural and forest production
activities. Thus, they affect landowner decisions regarding land use, crop mix, forest rotation
lengths, production practices, and others. Alterations in ambient ozone concentration levels will
therefore change the supply curves of U.S. agricultural and forest commodities, resulting in new
market equilibriums. Because both the changes in ozone concentrations and the distribution of
ozone-sensitive crops and tree species vary spatially, there may be substantial differences in the
net impacts across regions. There may also be distributional impacts as commodity production
shifts between regions in response to changes in relative productivity.
6 A.I Methodology
There are several alternative metrics used for assessing ozone concentrations (see Lehrer et
al. [2007] for more information). For this assessment, we are using the W126 metric, which is a
weighted sum of all ozone concentrations observed from 8 a.m. to 8 p.m available in 2025 and
2038. More specifically, we are using W126 ozone concentration surfaces generated using
enhanced Voronoi Neighbor Averaging (eVNA). W126 concentration surfaces based on meeting
the current ozone standard" were provided by EPA in the previous iteration of this project
(2013) and served again as the baseline for this analysis. According to information provided by
EPA, the eVNA W126 ozone surface is built from monitor data fused with Community
"The current primary and secondary ozone standards are 75 parts per billion (ppb) based on the annual fourth-
highest daily maximum 8-hr concentration, averaged over 3 years. For the purposes of calculating impacts on crop
yields and forest growth rates, we used the W126 equivalent of the current standard.
6A-1
-------
Multiscale Air Quality (CMAQ) model-based gradient interpolations. The spatial resolution of
the ozone surface in ArcGIS Shapefile format is 12 km.
County-level values were extracted from the eVNA W126 ozone surface using ArcGIS.
Only the ozone concentrations for the cropland and forestland portions of the W126 ozone
surface are used to derive the county-level average crop and forest W126 ozone levels,
respectively. These weighting adjustments were made to better reflect the ozone concentration
that would affect the specific portions of each county containing forested land or cropland, rather
than basing county-level exposure on the ozone concentration across the whole county. Data
from the 2011 USGS National Land Cover Database (NLCD), updated from the previously used
2006 NLCD data are used to extract the cropland and forestland portions from the ozone surface
(Jin et al, 2013). The maps below demonstrate the change in forest and cropland area resulting
from updating the data from 2006 to 2011.
Table 6A-1. Comparison of Total Cropland and Forestland NLCD Area (sq m)
2006 NLCD
2011 NLCD
% Increase
Cropland Area
1,730,787,687,000 1,786,333,937,400
3.21
Forestland Area
1,938,825,202,500 1,963,044,107,100
1.25
Cropland Area by County 2006 (sq m)
0- 100.000.000
j j 100.000.001 - 250,000.000
250.000.001 - 500,000.000
| J 500,000.001 - 500.000,000
J 500.000,001 - 750,000,000
^H 750.000.001 - 1,000.000.000
H 1.000,000.001 -1.500,000,000
H 1.500,000.001 -2.000.000,000
H 2.000,000.001 -3.500,000,000
^H 3,500,000.001 -5.600,000,000
0 255 510
Figure 6A-1. Cropland Area (sq m) by County according to NLCD 2006 Data
6A-2
-------
Cropland Area by County 2011 (sq m)
0- 100,000,000
_ 100.000.001-250,000.000
1 250,000.001 - 500,000.000
| | 500.000,001 - 500,000,000
| 500,000,001 - 750,000,000
]__J 750,000,001 - 1.000,000.000
^B 1.000,000.001 -1,500,000,000
^H 1.500.000.001 -2.000.000.000
^B 2.000,000.001 -3.500.000.000
^H 3.500,000.001 -5.600.000.000
Figure 6A-2. Cropland Area (sq m) by County according to NLCD 2011 Data
Forest Area by County 2006 (sq m]
L 1 0-250,000,000
| | 250.000.001 - 500.000.000
500.000.001 - 750,000,000
^H 750.000.001 - 1.000.000.000
^B 1.000.000.001 - 1.500,000,000
^H 1,500,000.001 -2.000,000.000
IHi 2.000,000.001 - 3.000,000,000
^•f 3.000.000.001 -5.000.000.000
^B 5.000,000.001 - 10.000.000.000
^•j 10.000.000.001 - 16,500.000.000
Figure 6A-3. Forest Area (sq m) by County according to NLCD 2006 Data
6A-3
-------
Forest Area by County 2011 (sq
0 - 250.000.000
250.000.001 - 500.000.000
1 500.000,001 - 750.000.000
•J] 750.000.001 - 1-000.000.000
^B 1.000.000.001 -1.500,000,000
^B 1.500.000.001 -2.000.000.000
^H 2,000.000.001 - 3.000,000,000
^B 3.000,000.001 -5.000.000.000
^H 5.000,000.001 - 10.000.000.000
^H 10.000.000 001 - 16 500 000 000
Figure 6A-4. Forest Area (sq m) by County according to NLCD 2011 Data
6A.1.1 Calculation of Relative Yield Loss
The median W126 ozone concentration response (CR) functions for crops and tree
seedlings in the 2007 EPA technical report (Lehrer et al, 2007) are used to calculate the RYLs
for crops and tree species under each ambient ozone concentration scenario used in this analysis.
Table 6A-2 presents the a and ft parameters being used in the W126 ozone CR function for
different crops and tree species. The W126 ozone CR function is as follows: RYL = 1 —
6A-4
-------
Table 6A-2. Parameter Values Used for Crops and Tree Species
a P
Crops
Corn
Sorghum
Soybean
Winter wheat
Potato
Cotton
Tree Species
Ponderosa
Red alder
Black cherry
Tulip poplar
Sugar maple
Eastern white
Red maple
Douglas fir
Quaking aspen
Virginia pine
98.3
205.9
110.0
53.7
99.5
94.4
159.63
179.06
38.92
51.38
36.35
63.23
318.12
106.83
109.81
1,714.64
2.973
1.963
1.367
2.391
1.242
1.572
1.1900
1.2377
0.9921
2.0889
5.7785
1.6582
1.3756
5.9631
1.2198
1.0000
6A. 1.1.1 Relative Yield Loss for Crops
Specifically, for crops, we first calculate the FASOMGHG subregion RYLs for crops that
have W126 ozone CR functions using the subregion-level, cropland-based ozone concentration
values under each scenario. The FASOMGHG subregion-level ozone concentration values are
initially calculated for all crops as the simple averages of the county-level ozone concentration
values. For crops that do not have W126 ozone CR functions, we assign them W126 ozone CR
functions based on the crop proxy mapping shown in Table 6A-3. This crop mapping was based
on the authors' judgment and previous experience. 10° In addition, for oranges, rice, and tomatoes,
which have ozone CR functions that are not W126-based (they are defined based on alternative
measures of ozone levels), we directly used the median RYG values under the 13 ppm-hr ozone
level reported in Table G-7 of Lehrer et al. (2007). More details on RYG are presented in further
subsections.
100 Also, note that FASOMGHG defines short-rotation woody trees such as hybrid poplar and willow as crops.
Ozone impacts on short-rotation woody trees were based on ozone RYLs for aspen.
6A-5
-------
Table 6A-3. Mapping of Ozone Impacts on Crops to FASOMGHG Crops
Crops Used for
Estimating Ozone
Impacts
FASOMGHG Crops
W126 Crops
Corn
Cotton
Potatoes
Winter wheat
Sorghum
Soybeans
Aspen (tree)
Non-W126 Crops
Oranges
Rice
Tomatoes
Corn
Cotton
Potatoes
Soft white wheat, hard red winter wheat, soft red winter wheat, durum wheat,
hard red spring wheat, oats, barley, rye, sugar beet, grazing wheat, and
improved pasture
Sorghum, silage, hay, sugarcane, switchgrass, miscanthus, energy sorghum, and
sweet sorghum
Soybeans and canola
Hybrid poplar, willow (FASOMGHG places short-rotation woody biomass
production in the crop sector rather than in the forest sector)
Orange fresh/processed, grapefruit fresh/processed
Rice
Tomato fresh/processed
Moreover, for crops that have county-level production data and W126 ozone CR functions
(including functions based on proxy crops), we updated the RYLs with production-weighted
W126 values. The 2012 USDA Census of Agriculture (Ag Census) county-level production data
are used to derive the weighted FASOMGHG subregion RYLs, following Formula (6A. 1).
wRYLik = Ozone CR
iProd
ijk
where /' denotes FASOMGHG subregion, j indicates county, and k represents crop. Ozone
CR Function^ refers to the ozone concentration response function for crop k. Prod//* represents
the county-level production level of crop k, and W126// represents the cropland-based ozone
value for county y in subregion /'. Finally, wRYL^ stands for the weighted FASOMGHG
subregion RYL for crop k. RYLs are calculated for each ozone concentration level being
considered.
6A-6
-------
6A. 1.1.2 Relative Yield Loss for Trees
The ozone CR functions for tree seedlings were used to calculate RYLs for FASOMGHG
trees over their whole life span. To derive the FASOMGHG region-level RYLs for trees under
each ozone concentration scenario, we used FASOMGHG region ozone values and the mapping
in Table 6A-4.
Table 6A-4. Mapping of Ozone Impacts on Forests to FASOMGHG Forest Types
FASOMGHG Forest Type FASOMGHG Region(s)
Tree Species Used for Estimating
Ozone Impacts
Black cherry, tulip poplar
Douglas fir
Eastern white pine
Ponderosa pine
Quaking aspen
Quaking aspen, black cherry, red maple,
sugar maple, tulip poplar
Red alder
Red maple
Virginia pine
Virginia pine, eastern white pine
Virginia pine, eastern white pine
Upland hardwood
Douglas fir
Softwood
Softwood
Hardwood
Hardwood
Hardwood
Bottomland hardwood
Natural pine, oak-pine, planted pine
Natural pine, oak-pine, planted pine
Softwood
SC, SE
PNWW
CB, LS
PNWE, PNWW, PSW,
RM
RM
CB, LS, NE
PNWE, PNWW, PSW
SC, SE
SC
SE
NE
Note: CB = Corn Belt; LS = Lake States; NE = Northeast; PNWE = Pacific Northwest—East side; PNWW = Pacific
Northwest—West side; PSW = Pacific Southwest; RM = Rocky Mountains; SC = South Central; SE = Southeast.
Specifically, the FASOMGHG region-level RYLs are first calculated for each tree species
listed in first column of Table 6A-4. Then, a simple average of RYLs for each tree species
mapped to a FASOMGHG forest type in a given region is calculated. The mapping of tree
species to FASOMGHG forest types is based on Elbert L. Little, Jr.'s Atlas of United States
Trees (1971, 1976, 1977, 1978). Note that crop RYLs are generated at the FASOMGHG
subregion level, whereas forest RYLs are calculated at the FASOMGHG region level, consistent
with the greatest level of regional disaggregation available for these sectors within
FASOMGHG.
6A. 1.1.3 Calculation of Relative Yield Gain
As described by Lehrer et al. (2007), the RYL is the relative yield loss compared with the
baseline yield under a "clean air" environment. For implementation within FASOMGHG, we
calculate the RYG for crops and trees from moving between ambient ozone concentrations (i.e.,
RYG is calculated as a change in RYL when moving between scenarios).
6A-7
-------
Thus, to obtain the RYG for crops and trees under alternative ozone concentrations, we
need the RYLs under each scenario. For example, to derive RYG under the current standard
75 ppb scenario relative to current conditions "currcond," we use Formula (6A.2):
r>\r *" 75pp,
KY (j — -- 1 —
The FASOMGHG subregion-level crop RYGs and the FASOMGHG region-level tree
RYGs for changes associated with moving from one scenario to another were calculated for
additional comparisons in the same way.
6A. 1.2 Conducting Model Scenarios in FASOMGHG
The current crop/forest budgets included in FASOMGHG are assumed to reflect
input/output relationships under current ambient ozone concentrations as these budgets are based
on historical data. To model the effects of changing ozone concentrations on the agricultural and
forest sectors, the following five scenarios were constructed and run through the model:
1 . "Current Conditions" scenario, where no RYGs of crops and trees are considered
(assumed to be consistent with current ambient ozone concentration levels);
2. 75 ppb scenario, where crop and forest yields are assumed to increase by the
percentages calculated in RYG75PP&, calculated relative to the current scenario;
3. 70 ppb scenario, using RYG/oppfc, calculated relative to both current and 75 ppb
scenarios
4. 65 ppb scenario, using RYG^fe, calculated relative to both current and 75 ppb
scenarios
5. 60 ppb scenario, using RYG<50p/>6, calculated relative to both current and 75 ppb
scenarios
The time scope of the FASOMGHG model scenarios used for these analyses is 2000-
2050, solved in 5-year time steps.101 The crop and tree RYGs are introduced into the model
starting in 2025. During the time-steps of 2025 and 2030, crop and tree RYGs are assumed to be
same as the RYGs in 2025. After 2040, RYGs remain constant at the RYGs obtained in 2038.
The special case is in 2035, for which the weighted average of RYGs in 2025 and 2038 (60% for
RYGs in 2025 and 40% for RYGs in 2038) are assigned to RYGs in 2035. Figure 6A-5 presents
101 Because of terminal period effects, the model is run out to the 2050 time period, but only results through the 2040
model time period (representative of 2040-2044) are used in the analyses.
6A-8
-------
the modeling process of simulating ozone scenarios using F ASOMGHG. The changes in crop
and tree yield growth potentially lead to new market equilibriums for agricultural and forestry
commodities, as well as land use changes between agricultural and forestry uses and
consequently GHG emissions and sequestration changes.
By comparing the market equilibriums under different scenarios, we can calculate the
welfare, land use, and GHG impacts of alternative ozone standards on the U.S. agricultural and
forest sector, including changes in consumer and producer welfare, land use allocation, and GHG
mitigation potential over time.
Figure 6A-5. FASOMGHG Modeling Flowchart
6A.1.3 Data Inputs
In this section, we summarize the input data used in the FASOMGHG scenarios specified
for this assessment. Following the methods described above, we calculated W126 ozone
concentration levels by region and crop. Effects on crop yields and forest productivity were
calculated for each FASOMGHG region. We present the values used as model inputs in tabular
6A-9
-------
and map format, with a primary focus here on comparison of the more stringent scenarios to the
current standard.
6A. 1.3.1
Ambient Ozone Concentration Data
The county-level forested and cropland W126 ozone values were aggregated at regional
and sub-regional levels, respectively.
Table 6A-5. Forestland W126 Ozone Values under Alternative Scenarios
FASOMGHG Region
CB
GPC
LS
NE
PNWE
PNWW
PSW
RM
SC
SE
SWC
c.c.a
11.84
8.95
6.33
8.48
5.55
3.79
17.28
13.36
11.84
13.10
10.03
2025
75 ppbb
5.24
5.15
2.83
3.65
2.24
1.49
6.65
7.32
4.12
3.03
6.08
70ppb
4.58
4.47
2.71
3.19
2.24
1.49
6.65
7.29
3.46
2.81
4.58
65 ppb
3.11
3.73
2.23
2.29
2.22
1.49
6.64
6.83
2.48
2.26
3.43
60 ppb
1.99
2.87
1.83
1.64
2.19
1.49
6.54
5.75
1.75
1.82
2.47
2038
75
ppbb
5.24
5.06
2.83
3.65
2.16
1.49
4.52
6.69
4.11
3.03
6.04
70
ppb
4.58
4.37
2.71
3.19
2.14
1.48
3.61
6.57
3.46
2.81
4.55
65
ppb
3.10
3.63
2.23
2.29
2.10
1.48
2.91
6.08
2.48
2.26
3.40
60
ppb
1.99
2.79
1.83
1.64
2.06
1.48
2.45
5.07
1.75
1.82
2.44
a Current Conditions
b Current Standard
CGP and SW are modeled as agriculture-only regions in FASOMGHG
Note: CB = Corn Belt; GP = Great Plains; LS = Lake States; NE = Northeast; PNWE = Pacific Northwest—East side;
PNWW = Pacific Northwest—West side; PSW = Pacific Southwest; RM = Rocky Mountains; SC = South Central; SE
= Southeast; SW = Southwest.
Table 6A-5 displays the agricultural W126 ozone values at the sub-region level. Similar to
the forest ozone values, the Pacific Southwest, South Central, Southeast, and Rocky Mountains
regions experience the greatest agricultural ozone reductions under the 75 ppb scenario
compared with current conditions. The Corn Belt, Southwest, and Northwest regions also see
noteworthy agricultural ozone reductions.
6 A-10
-------
Table 6A-6. Cropland W126 Ozone Values under Modeled Scenarios
FASOMGHG Region
CB
GPC
LS
Sub-region
Illinois, Northern
Illinois, Southern
Indiana, Northern
Indiana, Southern
Iowa, Central
Iowa, Northeast
Iowa, Southern
Iowa, Western
Missouri
Ohio, Northeast
Ohio, Northwest
Ohio, Southern
Kansas
Nebraska
North Dakota
South Dakota
Michigan
Minnesota
Wisconsin
C.C.a
7.84
11.46
10.38
13.13
5.71
6.23
6.65
5.74
11.50
12.71
12.07
13.49
10.93
8.87
4.46
5.66
9.50
4.97
7.02
2025
75 ppbb
3.90
6.54
4.65
6.66
2.79
2.69
3.31
3.33
5.17
5.34
5.89
5.93
8.17
5.81
3.03
3.73
5.20
2.63
3.04
70 ppb 65
3.51
5.79
4.43
6.40
2.46
2.42
2.68
2.90
4.01
5.08
5.65
5.73
6.88
5.27
3.00
3.50
4.96
2.49
2.85
ppb
2.47
3.86
3.03
4.22
2.06
2.02
2.07
2.51
2.76
3.51
3.88
3.64
5.54
4.41
2.86
3.12
3.57
2.31
2.25
60 ppb
1.65
2.38
1.99
2.54
1.65
1.67
1.53
2.07
1.81
2.35
2.54
2.12
4.12
3.24
2.72
2.68
2.50
2.12
1.78
2038
75 ppbb
3.90
6.53
4.64
6.65
2.79
2.69
3.31
3.31
5.17
5.34
5.88
5.93
8.02
5.65
3.03
3.67
5.20
2.63
3.04
70 ppb 65
3.51
5.79
4.43
6.40
2.46
2.42
2.68
2.90
4.01
5.08
5.65
5.73
6.74
5.12
2.96
3.43
4.96
2.49
2.85
ppb
2.46
3.86
3.02
4.22
2.06
2.02
2.06
2.49
2.76
3.51
3.87
3.64
5.39
4.26
2.86
3.09
3.57
2.31
2.25
60 ppb
1.65
2.38
1.99
2.54
1.65
1.67
1.53
2.07
1.81
2.35
2.54
2.12
4.03
3.13
2.69
2.62
2.50
2.11
1.78
(continued)
6 A-11
-------
FASOMGHG Region
NE
PNWE
Sub-region
Connecticut
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Pennsylvania
Rhode Island
Vermont
West Virginia
Oregon
Washington
PNWW
PSW
California, Northern
California, Southern
C.C.a
11.90
17.45
3.70
17.20
10.23
5.84
16.70
8.36
12.14
11.72
5.51
10.74
6.66
4.96
3.59
21.86
21.46
2025
75ppb
4.53
7.88
1.62
7.37
2.98
1.99
6.95
3.93
5.37
4.47
2.01
4.63
2.64
1.79
1.35
9.61
12.49
70 ppb 65
3.69
6.29
1.47
5.88
2.39
1.73
5.42
3.53
4.42
3.71
1.84
4.24
2.64
1.79
1.35
9.61
12.48
ppb
2.71
4.15
1.30
3.77
1.69
1.32
3.49
2.78
2.98
2.69
1.55
2.73
2.62
1.79
1.36
9.58
12.43
60 ppb
1.85
2.58
1.20
2.29
1.18
1.06
2.19
2.18
1.97
1.85
1.35
1.68
2.55
1.78
1.36
9.49
12.31
2038
75 ppb
4.53
7.88
1.62
7.37
2.98
1.99
6.93
3.93
5.37
4.47
2.01
4.63
2.53
1.78
1.34
6.20
6.20
70 ppb 65
3.69
6.29
1.47
5.87
2.39
1.73
5.42
3.53
4.42
3.71
1.84
4.24
2.48
1.78
1.34
4.54
5.52
ppb
2.71
4.15
1.30
3.77
1.69
1.32
3.49
2.78
2.98
2.69
1.55
2.73
2.43
1.77
1.35
3.21
4.82
60 ppb
1.85
2.58
1.20
2.29
1.18
1.06
2.19
2.18
1.97
1.85
1.34
1.68
2.35
1.77
1.35
2.35
4.23
(continued)
6 A-12
-------
FASOMGHG Region
RM
SC
SE
Sub-region
Arizona
Colorado
Idaho
Montana
Nevada
New Mexico
Utah
Wyoming
Alabama
Arkansas
Kentucky
Louisiana
Mississippi
Tennessee
Texas, East
Florida
Georgia
North Carolina
South Carolina
Virginia
C.C.a
13.53
16.33
12.75
6.61
15.48
12.58
18.06
14.28
13.16
12.30
13.79
9.59
11.21
15.55
9.92
9.16
12.79
14.31
12.82
12.62
2025
75ppb
9.63
10.71
5.05
3.25
6.57
8.54
8.62
7.56
2.89
4.81
5.59
4.62
3.32
4.57
5.67
2.95
2.56
3.35
2.18
3.95
70 ppb 65
9.63
10.41
5.04
3.25
6.56
8.15
8.61
7.50
2.70
3.76
5.35
3.58
2.75
4.22
3.88
2.86
2.45
3.08
2.07
3.47
ppb
9.17
8.98
4.79
3.15
6.29
7.40
7.81
6.79
2.29
2.57
3.52
2.66
2.11
2.85
2.79
2.76
2.21
2.49
1.77
2.35
60 ppb
7.91
6.39
4.15
3.00
5.58
6.20
5.85
5.20
1.91
1.71
2.16
1.95
1.55
1.83
1.99
2.60
1.95
2.00
1.52
1.59
2038
75 ppb
8.24
10.23
4.62
3.16
5.58
8.31
7.53
7.02
2.89
4.81
5.59
4.62
3.32
4.57
5.64
2.95
2.56
3.35
2.18
3.95
70 ppb 65
8.11
9.88
4.57
3.16
5.25
7.89
7.40
6.91
2.70
3.76
5.34
3.58
2.75
4.22
3.87
2.86
2.45
3.08
2.07
3.47
ppb
7.51
8.46
4.30
3.09
4.75
7.11
6.58
6.22
2.29
2.57
3.52
2.66
2.11
2.85
2.77
2.76
2.21
2.49
1.77
2.35
60 ppb
6.24
6.06
3.75
2.89
4.03
5.91
4.95
4.76
1.91
1.70
2.16
1.95
1.54
1.83
1.98
2.60
1.95
2.00
1.51
1.59
(continued)
6 A-13
-------
FASOMGHG
SWC
Region Sub-
region
Oklahoma
Texas, Central
Blacklands
Texas, Coastal
Bend
Texas, Edwards
Plateau
Texas, High Plains
Texas, Rolling
Plains
Texas, South
Texas, Trans Pecos
C.C.a
11.52
9.17
7.24
9.02
11.74
10.64
4.33
11.83
2025
75ppb
7.44
5.45
5.23
6.38
8.49
7.13
3.45
7.49
70ppb
5.72
4.06
4.13
5.24
7.50
5.71
2.91
7.09
65 ppb
4.31
3.05
3.24
4.28
6.44
4.56
2.48
6.61
60 ppb
3.07
2.18
2.44
3.38
5.18
3.45
2.09
5.86
2038
75 ppb
7.38
5.43
5.23
6.32
8.36
7.07
3.44
7.38
70 ppb
5.67
4.05
4.12
5.17
7.38
5.66
2.90
6.96
65 ppb
4.25
3.03
3.23
4.20
6.27
4.48
2.47
6.44
60 ppb
3.04
2.16
2.42
3.31
5.03
3.38
2.05
5.68
a Current Conditions
b Current Standard
CGP and SW are modeled as agriculture-only regions in FASOMGHG
6 A-14
-------
Figure 6A-6 presents the incremental ozone reductions under alternative 2038 ozone
standards with respect to the current 75 ppb standard. As the standard is tightened from the 70
ppb scenario to the 60 ppb scenario, the greatest ozone reductions are observed in the Pacific
Southwest region. Central parts of the Rocky Mountains region extending into northern areas of
the Southwest and South central regions also see substantial ozone reductions. These ozone
reductions would affect the production of crops and timber that are susceptible to ground-level
ozone in these regions.
Legend
_| FASOM Region
FASOM Subregion
J-1.0--0.6
-0.5--0.3
-0.2 - 0.0
0.1 -0.5
0.6-0.9
1.0- 1.3
1.4-1.6
1.7-2.0
2.1 -2.4
I2.5 - 2.8
2.9-3.2
3.3-3.5
3.6-3.9
4.0 - 4.3
4.4-4.7
4.8-5.1
5.2 - 5.4
5.5-5.8
5.9-6.2
6.3-6.6
6.7-6.9
70-7.3
7.4 -7.7
7.8-8.1
8.2-8.5
8.6-8.8
8.9-9.2
9.3-9.6
9.7- 10.0
10.1 - 10.4
10.5- 10.7
10.8-11.1
0 150300 600 900 1.200
I Mile
ro
o
CO
00
CT>
O
O
CO
oo
ro
0>
o
—*1
0>
01
Tl
Tl
ro
w
oo
05
o
Tl
Tl
ro
Figure 6A-6. Ozone Reductions with Respect to 75 ppb under Alternative Scenarios
6-15
-------
6A. 1.3.2 Changes in Crop and Forest Yields with Respect to 75 ppb Scenario
Figures 6A-7 through 6A-12 display major crops' RYGs under alternative ozone
standards scenarios at the FASOMGHG subregion level, with respect to the current 75 ppb
standard. These are the values that were directly incorporated into FASOMGHG to define the
scenarios modeled. Figures 6A-13 and 6A-14 display changes in forest RYGs. As discussed
previously, the Rocky Mountains, Corn Belt, and parts of the southern regions of the United
States (e.g., within the Pacific Southwest, South Central, and Southeast regions) are shown to
experience the most significant further ozone reductions under the alternative policy scenarios.
Hence, one would expect to see the most sizable increases in RYGs for crops and tree species
grown in those regions. This finding is consistent with our calculations.
6-16
-------
Legend
| | FASOM Region
FASOM Subregions
0.00%
0.01% -4.00%
4.01% -4.55%
i - 5.56%
i - 6.67%
, - 8.70%
.71% -9.62%
.63% - 11.76%
11.77%- 12.50%
12.51%-20.00%
0.01%-25.71%
5.72%- 100.00%
0 150300 600
900
1,200
iMiles
Figure 6A-7. Percentage Changes in Corn RYGs with Respect to the 75 ppb Scenario
6-17
-------
Legend
| | FASOM Region
FASOM Subregions
0.00%
0.01% -2.85%
2.86% - 3.72%
3.73% - 4.84%
O
4.85% - 10.27%
10.28%- 11.59%
11.60%-13.31%
13.32%-20.33%
20.34%-21.99%
22.00% - 25.27%
.28% - 28.73%
28.74% - 32.54%
0 150300 600 900 1,200
i Mites
Figure 6A-8. Percentage Changes in Cotton RYGs with Respect to the 75 ppb Scenario
6-18
-------
Legend
| | FASOM Region
FASOM Subregions
0.00%
0.01% -9.23%
9.24% -13.53%
n 13.54%-17.59%
17.60% -20.30%
20. 31% -21.82%
21. 83% -29.88%
o •29.89%- 35.11%
Q.
.12%-44.22%
4.23% - 46.86%
6.87% - 54.67%
4.68% - 70.70%
0 150300 600 900 1,200
i Mites
Figure 6A-9. Percentage Changes in Potato RYGs with Respect to the 75 ppb Scenario
6-19
-------
Legend
| | FASOM Region
FASOM Subregions
0.00%
0.01% -3.03%
3.04% - 4.76%
4.77% - 7.69%
7.70% - 14.81%
14.82%-18.18%
18.19%-18.75%
18.76%-20.00%
20.01%-21.67%
21.68%-27.03%
.04% - 35.29%
35.30% - 69.23%
0 150300 600 900 1,200
i Mites
Figure 6A-10. Percentage Changes in Sorghum RYGs with Respect to the 75 ppb Scenario
6-20
-------
Legend
| | FASOM Region
FASOM Subregions
g. 0.00%
m
r-
oo
.2 4.48%-8.13%
«
&
8.14%-13.97%
I
Ł
13.98%- 18.08%
o 18.09% - 20.80%
v>
c
•
27.86%-31.38%
31.39%-37.36%
37.37%-41.39%
40% - 57.34%
57.35% - 79.00%
0 150300 600 900 1,200
i Mites
Figure 6A-11. Percentage Changes in Soybean RYGs with Respect to the 75 ppb Scenario
6-21
-------
a
Legend
| | FASOM Region
FASOM Subregions
0.00%
0.01% - 1.96%
1.97% -3.57%
3.58% - 4.49%
a> 4.50% - 7.26%
7.27% - 10.14%
10.15%-12.25%
12.26%-15.53%
15.54%-16.92%
16.93% -21.56%
21.57% -30.80%
30.81%-37.16%
0 150300 600 900 1,200
i Mites
Figure 6A-12. Percentage Changes in Winter Wheat RYGs with Respect to the 75 ppb
Scenario
6-22
-------
Legend
| | FASOM Region
0.00% - 2.59%
n 2.60%-5.19%
o
fM
O
§ 5.20% - 7.78%
I
c 7.79%-10.37%
'™
(D
2
.2 10.38%-12.96%
I
12.97% -15.56%
2 115.57% - 18.15%
18.16%- 20.74%
0.75% - 23.33%
3.34% - 25.93%
5.94% - 28.52%
8.53%-31.11%
0 150300 600 900 1,200
i Mites
ro
o
oo
oo
--J
o
T)
-D
DO
O
OJ
00
CD
cn
Tl
T)
CD
ro
o
co
oo
05
o
Tl
T)
CO
Figure 6A-13. Percentage Changes in Softwood RYGs with Respect to the 75 ppb Scenario
6-23
-------
Legend
| | FASOM Region
0.00%
0.01%- 1.06%
1.07%- 1.72%
1.73% - 7.44%
7.45% - 8.52%
8.53%- 10.02%
10.03%- 17.09%
17.10%-20.41%
20.42% - 27.25%
27.26% - 30.98%
0 150300 600 900 1,200
i Mites
ro
o
oo
oo
--J
o
T)
T)
DO
O
CO
00
CD
cn
Tl
T)
CO
ro
o
co
oo
05
o
Tl
T)
CO
Figure 6A-14. Percentage Changes in Hardwood RYGs with Respect to the 75 ppb Scenario
6-24
-------
6A.2 Model Results
FASOMGHG was used to estimate the projected effects of alternative ozone
concentration standards on the U.S. agricultural and forestry sectors. As introduced earlier, the
comparisons considered for this report focus on the differences between a scenario assuming
compliance with the existing 2008 standards (75 ppb) and scenarios in which three more
stringent ozone standards are met. Those three scenarios are 70 ppb, 65 ppb, and 60 ppb W126
values. Our analysis included changes to production, prices, forest inventory, land use, welfare,
and GHG mitigation potential associated with achieving each of the more stringent standards.
6A. 2.1 Agricultural Sector
Ozone negatively affects growth in many plants, leading to lower crop yields. In addition,
some crops are more sensitive to ozone than others, so the percentage changes in yield will vary
by crop and region. However, reducing ambient ozone concentrations would generally increase
agricultural yields and total production, though the reductions in ozone concentrations that would
be achieved under a given standard vary across regions. Our analysis began by determining the
extent to which current yield losses caused by ozone could be reversed by reducing ozone levels.
Increased crop yields lead to a greater available supply of most agricultural crops, which in turn
tends to reduce market prices. There is also an overall tendency toward acreage shifting away
from ozone-sensitive crops. In general, impacts in the agricultural sector are relatively limited,
especially when compared with the forestry sector. By and large, more stringent standards led to
increased incremental impacts, but the additional impact in moving to increasingly stringent
ozone standards was relatively small.
6A.2.1.1 Production and Prices
Changes in U.S. agricultural production and prices were measured using Fisher indices
(sees Tables 6A-7 and 6A-7a).102 Both primary and secondary commodity production levels are
projected to increase by 2040 as a result of heightened productivity. Agricultural production
changes were generally relatively small across products, rarely exceeding an increase of 0.50%
with respect to the current standard and often changing by 0.02% or less.
102 The Fisher price index is known as the "ideal" price index. It is calculated as the geometric mean of an index of
current prices and an index of past prices.
6-25
-------
Table 6A-7. Agricultural Production Fisher Indices (Current conditions =100)
Sector
Policy
2010
2020
2030
2040
Primary
Commodities
Crops
Livestock
Farm products3
75ppb
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
99.8
99.8
99.8
99.8
99.9
99.9
99.9
99.9
99.8
99.8
99.8
99.9
100.7
100.7
100.8
100.9
100.2
100.2
100.2
100.2
100.1
100.4
100.5
100.5
101.3
101.3
101.4
101.4
100.3
100.3
100.4
100.5
100.3
100.3
100.4
100.5
Secondary
Commodities
Processed
Meats
Mixed feeds
a Farm Products is the
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
99.9
99.9
100.0
100.0
100.0
100.0
100.0
99.9
99.9
99.9
99.9
100.1
100.1
100.1
100.1
100.1
100.1
100.1
100.2
100.1
100.1
100.1
100.2
100.1
100.1
100.1
100.1
100.3
100.2
100.3
100.3
100.5
100.5
100.5
100.5
composite of Crops and Livestock.
Table 6A-7a. Agricultural Price
Sector
Policy
Fisher Indices
2010
(Current
2020
Conditions = 100)
2030
2040
Primary Commodities
Crops
75 ppb
70 ppb
100.0
100.0
100.2
100.2
98.5
98.4
97.2
97.0
6-26
-------
Sector
Livestock
Farm products3
Policy
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010
100.0
100.0
100.1
100.6
100.1
100.1
100.0
100.0
100.0
100.0
2020
100.2
100.2
101.4
102.2
102.2
101.8
100.2
100.2
100.2
100.2
2030
98.3
98.3
99.2
100.3
100.6
100.4
100.1
98.4
98.3
98.3
2040
97.0
97.0
99.9
97.0
97.0
97.0
99.9
97.0
97.0
97.0
Secondary Commodities
Processed
Meats
Mixed feeds
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
100.1
100.1
100.1
100.1
100.1
100.1
100.1
100.1
100.1
100.1
100.1
100.1
100.3
100.3
100.3
100.3
99.8
99.8
99.8
99.8
99.7
99.7
99.7
99.7
98.8
98.6
98.2
98.2
100.1
100.1
100.1
100.1
98.2
98.1
97.4
97.5
98.3
98.2
98.0
97.9
99.9
99.9
100.0
100.0
99.3
99.5
99.3
99.5
Increased production led to a general decline in market prices because the equilibrium
price adjusts to higher levels of supply. This result is consistent with expectations because higher
productivity leads to greater supply, which tends to decrease market prices. Changes in price
were generally more pronounced than changes in production, with the largest decreases in the
Farm Products, Livestock, and Processed categories. Agricultural prices tend to decline by a
greater percentage than production increases because the demand for most agricultural
6-27
-------
commodities is inelastic.103 However, almost all declines in price were less than 2.0% of prices at
the current standards, and most were less than 0.5%.
6A. 2.1.2 Crop Acreage
Crop acreage was projected to decline with the introduction of the ozone standards because
additional productivity per acre reduces the demand for crop acreage. In aggregate, farmers will
be able to meet the demand for agricultural commodities using less land under scenarios with
lower ozone concentrations. Consistent with these expectations, the total cropped area is slightly
smaller for each model year in the alternative standard cases. However, land allocation also
depends on relative returns across various uses and is influenced by forest harvest timing.
Changes in land allocation between the agricultural and forestry sectors are discussed later in this
section.
Table 6A-8 provides projections of acreage in each of the major U.S. crops, as well as
composites of all remaining crops and total cropland. The absolute change relative to the current
standard is presented for each alternative standard. Larger changes occurred in sorghum acreage,
whereas only minor changes occurred in all other crops, leading to almost no net change in crop
acreage across all crops. This shift occurred largely because of differential crop sensitivity to
ozone concentrations. Note that the sum of the crop-specific changes will not necessarily equal
the total changes shown in Table 6A-8 because some double-cropping is reflected in the model
(e.g., soybeans and winter barley).
103 Demand elasticities are measures of the responsiveness of the quantity demanded to a change in price.
Commodities with inelastic demands are those where consumers change the quantity of a good they purchase by a
smaller percentage than the change in market price. Many food products fall into this category because they are
relatively low-priced necessities.
6-28
-------
Table 6A-8.
Crop
Corn
Soybeans
Hay
Hard Red
Winter Wheat
Cotton
Hard Red
Spring Wheat
Sorghum
Switch Grass
Major Crop Acreage, Million Acres
Policy
75ppb
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
2010 2020 2030
91.2 85.8 77.4
Change with Respect to Current Standard
0.00 0.02 0.00
0.00 0.01 0.06
0.00 0.01 0.04
4.9 5.2 5.8
Change with Respect to Current Standard
0.00 0.00 0.00
0.00 0.00 0.00
0.00 0.00 0.00
43.7 41.1 41.7
Change with Respect to Current Standard
0.00 -0.02 0.03
0.00 -0.01 0.05
0.00 0.01 0.06
25.5 23.7 23.2
Change with Respect to Current Standard
0.00 0.00 0.00
0.00 0.00 -0.06
0.00 0.00 -0.03
13.6 13.6 14.2
Change with Respect to Current Standard
0.00 0.00 -0.04
0.00 0.00 -0.05
0.00 0.00 -0.04
13.5 12.9 13.4
Change with Respect to Current Standard
0.00 0.00 0.01
0.00 0.01 -0.03
0.00 0.01 -0.08
73.1 71.9 72.0
Change with Respect to Current Standard
0.00 0.20 0.02
0.00 0.24 0.02
0.00 0.26 0.09
0.1 0.1 0.1
Change with Respect to Current Standard
2040
70.9
-0.05
-0.07
-0.08
6.9
0.00
-0.03
-0.03
42.0
0.00
-0.01
-0.07
22.4
-0.06
-0.11
-0.10
14.0
-0.01
-0.01
0.00
13.0
0.01
0.00
-0.01
72.2
-4.88
-4.90
-4.89
0.1
6-29
-------
Crop
All Others3
Total
Policy
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010
0.00
0.00
0.00
352.2
Change with
0.00
0.00
0.00
5.6
Change with
0.00
0.00
0.00
2020
0.00
0.00
0.00
366.7
Respect to Current
0.10
0.13
0.19
5.4
Respect to Current
0.00
-0.01
-0.01
2030
0.00
0.00
0.00
355.8
Standard
0.01
-0.12
-0.25
4.6
Standard
0.01
0.02
0.01
2040
0.00
0.00
0.00
340.0
-0.17
-0.32
-0.45
4.0
-0.01
0.00
0.00
a Canola, durum wheat, fresh grapefruit, fresh orange, fresh tomato, grazing wheat, hybrid poplar, oats, potato,
processed grapefruit, processed orange, processed tomato, rice, rye, silage, soft red winter wheat, soft white
wheat, spring barley, sugar beet, sugarcane, sweet sorghum, winter barley.
6A.2.2 Forestry Sector
As with agricultural crops, ozone diminishes growth in most tree species, and our analysis
began by estimating how much of this diminished growth would be reversed under the more
stringent ozone standards. Impacts are significantly higher in the forestry sector, especially in
hardwood species and species more prevalent in the southern regions. Impacts are more
significant for southern regions because of the higher baseline ozone concentrations in the South
Central and Southeast regions. Higher initial concentrations resulted in higher reductions to meet
the alternative standards and, thus, higher impacts to tree growth. This relationship also
contributes to the larger changes in the forestry sector as a whole.
6A.2.2.1 Production and Prices
Reducing ozone concentrations led to increased forest growth, which was reflected in
increased production in FASOMGHG. Some of the most substantial ozone standard impacts
occurred in saw log and pulp log harvest quantities and prices. Compared with the current
standard, alternative standard cases had consistently higher production except for hardwood pulp
logs in 2030 and softwood pulp logs in 2040, where production increased only marginally and at
times fell below the baseline estimates, especially in 2040. The most significant impacts
6-30
-------
occurred in hardwood pulp logs, where harvests were projected to be more than 1% higher than
under the current standard level by 2040 for the 60 ppb concentration. There are some cases
where production of pulp logs and saw logs moved in opposite directions. There are two primary
explanations for these trends. The first is that as softwood saw log production expands, the price
for softwood saw logs drops, allowing processers to substitute saw logs for cases of production
in which pulp logs are traditionally used. The second is that even when the primary log size
being harvested is pulp logs, saw logs will generally also be present because of natural variation
in tree growth rates (and vice versa for harvest of saw logs). With higher growth rates, there
would tend to be more saw logs in stands harvested primarily for pulp logs over time.
The largest changes in production occurred at the 60 ppb level. Changes from the current
standard to 70 ppb were fairly small, but increased in 65 ppb, and changes were even larger in 60
ppb. Table 6A-9 presents these changes by major product.
The impact of policy intervention on timber market prices was more substantial than the
change in production in terms of percentage changes compared with the current standard.
Although increases in production of forest products did not exceed 1.5% compared with the
current standard, changes in price were as large as 12.8%. As with agricultural products, many
forest products have relatively inelastic demand so prices tend to change by a larger percentage
than quantities. Table 6A-10 lists absolute changes with respect to the current standard, whereas
Table 6A-11 lists the percentage change in forest product prices for each year, alternative
standard, and forest product.
Table 6A-9. Forest Products Production, Million Cubic Feet
Product Policy
Hardwood Saw logs 75 pp^
70 ppb
65 ppb
60 ppb
Hardwood Pulp logs 75 pp^
70 ppb
65 ppb
2010 2020 2030
3,588 3,394 3,708
Change with Respect to Existing Standard
1 1 38
1 0 33
3 0 54
2,448 2,162 2,510
Change with Respect to Existing Standard
2 1 -29
3 -2 -29
2040
4,246
3
-6
-13
2,220
o
5
14
6-31
-------
Softwood saw logs
Softwood pulp logs
Table 6A-10. Forest
Product
Hardwood saw logs
Hardwood pulp logs
Softwood saw logs
Softwood pulp logs
60ppb
75 ppb
70ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
Product Prices,
Policy
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
10 -1 -40
4,568 5,114 5,449
Change with Respect to Existing Standard
-1 10 -2
0 22 1
5 35 21
3,437 3,878 4,350
Change with Respect to Existing Standard
1 -2 3
0 1 4
2 1 8
U.S. Dollars per Cubic Foot
2010 2020 2030
0.80 0.89 0.58
Change with Respect to Existing Standard
0.00 -0.01 -0.01
0.00 -0.01 -0.02
0.00 -0.01 -0.03
0.30 0.64 0.45
Change with Respect to Current Standard
0.00 -0.01 -0.01
0.00 -0.01 -0.03
0.00 -0.02 -0.04
2.46 2.08 1.78
Change with Respect to Current Standard
0.00 0.00 0.00
0.00 -0.01 -0.01
-0.01 -0.02 -0.02
1.49 1.33 1.47
Change with Respect to Current Standard
0.00 0.00 0.00
0.00 0.00 -0.01
-0.01 0.00 -0.03
25
6,542
7
53
61
4,298
-17
-23
-8
2040
0.33
-0.02
-0.03
-0.04
0.24
-0.01
-0.02
-0.03
1.50
0.00
-0.02
-0.04
1.14
0.00
-0.02
-0.05
6-32
-------
Table 6A-11. Forest Product Prices and Percentage Change, U.S. Dollars per Cubic Foot
Product
Hardwood saw logs
Hardwood pulp logs
Softwood saw logs
Softwood pulp logs
Policy
75ppb
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010 2020 2030
0.80 0.89 0.58
% Change with Respect to Existing Standard
-0.41% -0.75% -2.16%
-0.24% -1.00% -3.89%
-0.25% -1.22% -5.60%
0.30 0.64 0.45
% Change with Respect to Existing Standard
0.07% -0.96% -3.19%
0.12% -1.92% -5.57%
-0.06% -2.87% -8.98%
2.46 2.08 1.78
% Change with Respect to Existing Standard
0.01% -0.17% -0.16%
0.03% -0.27% -0.36%
-0.28% -0.97% -1.20%
1.49 1.33 1.47
% Change with Respect to Existing Standard
-0.05% 0.28% -0.28%
-0.18% 0.10% -0.76%
-0.57% -0.27% -1.87%
2040
0.33
-5.07%
-9.11%
-11.36%
0.24
-5.15%
-10.53%
-12.75%
1.50
-0.13%
-1.08%
-2.65%
1.14
-0.13%
-1.90%
-4.11%
6A.2.2.2
Forest Acres Harvested
Harvested acres are projected to decline in hardwoods as a result of higher productivity in
the policy cases. Conversely softwood acres harvested increases over time. The difference
between the hardwood harvested acres in the current standard case and in the alternative
standards widens from 2010 to 2040, increasing to a difference of more than 4% under the 60
ppb case. The impact to total acres of softwood harvested shows a less uniform pattern, with the
largest impacts under the 65 ppb scenario, and an increase of up to 6% of harvested acres by
2040 compared to 2010. Table 6A-12 presents the model results for forest acres harvested.
6-33
-------
Table 6A-12. Forest Acres Harvested, Thousand Acres
Product
Total hardwood
Total softwood
Policy
75ppb
70ppb
65 ppb
60ppb
75 ppb
2010 2020
14,923 11,701
Change with Respect to
4 50
8 37
37 50
17,539 16,158
2030
12,277
Current Standard
-139
-265
-377
14,911
2040
13,138
-151
-352
-513
19,031
Change with Respect to Current Standard
70 ppb
65 ppb
60 ppb
-11 35
-13 107
-2 142
47
223
328
106
361
314
6A.2.2.3 Forest Inventory
Under FASOMGHG definitions, existing inventory includes only trees that have been
standing since the initial model year of 2000. All trees planted since then, including both
reforestation and afforestation, are included in new inventory. The model projected significant
increases in existing inventory for hardwood species under the current standard, and consistent
with the increase in the acres harvested of softwood, the existing inventory of softwoods declines
through 2040 under the current standard. The difference in responses is partially explained by
differential sensitivity to ozone between species. Hardwood species show a much higher
sensitivity to ozone levels and are thus modeled to respond more dramatically to reductions in
ozone concentration.
Some relatively large differences between ozone standards occurred in the forest inventory
projections. For example, existing hardwood inventory was projected to be 4.0% small under the
60 ppb case than the 70 ppb case by 2040. New hardwood inventory is similarly sensitive, with
the model projecting a 2% decrease for this same comparison. For new and existing inventory of
both hardwoods and softwoods, the largest impacts occurred at the 70 ppb standard. This type of
nonlinear response can occur because of differences in the relative impacts on alternative forest
and agricultural products that lead to land reallocation. Table 6A-13 presents the model results
for forest inventory.
6-34
-------
Table 6A-13. Existing and New Forest Inventory, Million Cubic Feet
Product
Existing Hardwood
Existing Softwood
New Hardwood
New Softwood
Policy
75ppb
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010 2020 2030
281,924 273,055 292,685
Change with Respect to Existing Standard
0 -64 1,983
0 -138 5,974
0 -278 8,896
184,828 153,771 135,137
Change with Respect to Existing Standard
0 -19 185
0 -38 683
-6 -77 1,059
1,932 9,437 18,872
Change with Respect to Existing Standard
0 -152 -107
0 -150 -63
1 -156 0
8,837 64,254 114,454
Change with Respect to Existing Standard
0 -42 -17
4 6 -128
7 -18 -312
2040
306,296
4,998
14,937
22,548
133,794
494
1,583
2,553
29,732
77
577
931
128,581
-497
-1,164
-1,798
6A.2.3 Cross-Sectoral Policy Impacts
One of the advantages of a model such as FASOMGHG for analysis of impacts on major
land-using activities is the ability to account for shifts in land use. Differentiated impacts on
productivity across products will lead to changes in market prices and in the relative profitability
of alternative land uses. In response, landowners will change their allocation of land across
different productive activities, which will contribute to market impacts. In addition, these
changes in land use have implications for GHG emissions and other environmental impacts. In
this section, we discuss changes in land use, net GHG emissions, and producer and consumer
welfare across the agricultural and forest sectors.
6-35
-------
6A.2.3.J
Land Use
FASOMGHG projected changes in eight land use categories: existing forest,
reforestation, afforestation, cropland, pasture, cropland pasture,104 and lands enrolled in the
Conservation Research Program (CRP).105 The largest impacts under the current standard were
projected in afforestation. The general projected pattern under the alternative standards within
these categories was a decline in reforested area, afforested area, and cropland in 2030 and 2040,
coupled with increases in the pasture and cropland pastured areas. There was no change in
acreage retained in CRP or rangeland.
The incremental impact of more stringent ozone standards appears to be non-linear in
some cases, especially in existing and reforested areas. Existing forest exhibits a general decline
in the area under the 70 ppb scenario; in the 60 ppb scenario, however, there are small declines
through 2020, followed by large increases in 2030 and 2040. This is due to the different trends in
the hardwood and softwood species. Table 6A-14 presents the model results by major land use
type.
Table 6A-14. Land Use by Major Category, Thousand Acres
Product
Existing forest
Reforested
Afforested
Policy
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
2010 2020 2030
257,565 201,587 161,426
Change with Respect to Current Standard
-7 82 -113
000
-2 -84 159
72,201 117,974 148,837
Change with Respect to Current Standard
-10 -7 -56
-16 52 -212
28 107 -267
14,086 10,886 6,404
Change with Respect to Current Standard
2040
131,210
-269
0
691
172,267
-189
-878
-1,346
11,474
104 Cropland pasture is managed land suitable for crop production (i.e., relatively high productivity) that is being
used as pasture.
105 Rangeland estimates are also included, but rangeland is held fixed in FASOMGHG by assumption because it
cannot be allocated to any other use.
6-36
-------
Product
Cropland
Pasture
Cropland pasture
Rangeland
CRP
Policy
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010 2020 2030
0 0 -134
0 0 -309
0 0 -427
311,713 313,325 304,227
Change with Respect to Current Standard
0 96 -38
0 124 -102
0 180 -157
84,280 85,049 86,133
Change with Respect to Current Standard
3 -7 86
10 -11 189
10 24 257
45,381 44,634 55,061
Change with Respect to Current Standard
0 1 29
0 1 161
0 6 235
302,210 301,104 300,049
Change with Respect to Current Standard
000
000
000
36,879 36,659 36,659
Change with Respect to Current Standard
000
000
000
2040
-134
-309
-427
292,678
-60
-157
-200
82,571
64
168
263
61,042
51
216
277
299,039
0
0
0
36,659
0
0
0
6A.2.3.2
Welfare
Welfare impacts resulting from the implementation of alternative standard levels
followed the same pattern between the agriculture and forestry sectors, although it was more
pronounced in forestry. Consumer surplus typically increased in both cases as higher
6-37
-------
productivity under reduced ozone conditions tended to increase total production and reduce
market prices. Because demand for most forestry and agricultural commodities is inelastic, there
are more instances in which producer surplus declines. In some year/ozone concentration
combinations, the effect of falling prices on producer profits more than outweighs the effects of
higher production levels.
Percentage changes in agricultural sector consumer and producer surplus between the
current standard and the alternative standards were relatively small in many cases, with the
largest percentage change being a 5.8% decline in producer surplus in the 2040 model period.
However, the agricultural sector is a very large market, and even small percentage changes in
welfare can result in annualized values of tens or even hundreds of millions of dollars. Table 6A-
15 provides consumer and producer surplus for the agricultural sectors under the current
standard, along with the change in surplus for each alternative standard. There is considerable
variability in the magnitude of consumer and producer impacts from year to year, which is not
surprising given the dynamic nature of the model and numerous adjustments taking place over
time in response to changes in net returns associated with alternative land uses.
Table 6A-15. Consumer and Producer Surplus in Agriculture, Million 2010 U.S. Dollars
Policy
2010
2015
2020
2025
2030
2035
2040
Consumer
Surplus
75ppb
1,907,219
1,928,147
1,955,266 1,983,375 2,013,518 2
Change with Respect
Producer
Surplus
70ppb
65 ppb
60ppb
75 ppb
0
-1
0
718,105
0
1
47
824,802
0
4
9
to Existing
113
262
462
Standard
66
148
100
815,996 865,123 817,707
Change with Respect
70 ppb
65 ppb
60 ppb
1,202
216
-176
2,454
530
1,015
1,796
-407
-1,470
to Existing
609
1,174
1,320
Standard
-1,144
-2,933
-1,492
,042,566
58
289
408
874,549
675
-39,605
-38,788
2,071,338
-10
24
66
908,352
-52,504
-10,211
-8,119
The impacts of the scenarios with more stringent ozone standards were larger in the
forestry sector, with bigger increases in consumer surplus and greater declines in producer
surplus. Table 6 A-16 presents the model results of the welfare analysis in the forestry sector.
6-38
-------
Table 6A-16. Consumer and Producer Surplus in Forestry, Million 2010 U.S. Dollars
Consumer
surplus
Producer
surplus
Policy
75ppb
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010
715,634
3
12
30
93,795
1
-5
-57
2015 2020
760,957 801,653
Change with Respect to
24,592 -9
24,595 4
24,695 127
147,719 154,137
Change with Respect to
-24,614 -28
-24,621 -73
-24,727 -261
2025 2030
819,407 867,332
Current Standard
-2 70
43 189
257 433
144,888 147,039
Current Standard
94 -24
291 -111
135 -471
2035
885,687
167
263
686
147,797
-152
-86
-503
2040
926,837
72
138
245
132,850
8
-35
37
Because of the complex dynamics of the agriculture and forestry sectors and variability in
welfare impacts over time, it is often helpful to summarize the impacts in terms of annualized
values. Table 6 A-17 and Table 6 A-18 summarize the annualized impacts of alternative ozone
standards on consumer and producer surplus in the agricultural and forestry sectors for 2010-
2044 at a discount rate of 3% and 7% respectively.106 The impacts of alternative standards on
consumer surplus are positive for each of the tighter standards for both agricultural and forestry
sectors, with the benefits increasing with more stringent requirements. For producer surplus, on
the other hand, annualized impacts are negative for all of the tighter standards for both
agriculture and forestry sector, becoming more negative as stringency is increased. Overall, total
surplus declines for the agriculture sector while rises slightly for the forest sector across
alternative standard levels, leading to a net decrease for both sectors.
Table 6A-17. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry, 2010-2044, Million 2010 U.S. Dollars (3% Discount Rate)
Policy
Agriculture
Forestry
Total
Consumer surplus
75 ppb 1,969,838 805,563
Change with Respect to Current Standard
2,775,401
106 Each model period in FASOMGHG is representative of the 5-year period starting with that year, so results
reported for 2040 are representative of 2040-2044. Thus, we use values through 2044 in the annualization
calculations.
6-39
-------
Producer surplus
Total surplus
70ppb
65 ppb
60ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
28
86
132
817,744
Change with
-3,601
-5,035
-4,741
2,787,582
-3,573
-4,949
-4,608
4,552
4,592
4,744
135,478
Respect to Current Standard
-4,534
-4,524
-4,684
941,041
18
68
60
4,580
4,678
4,877
953,222
-8,135
-9,559
-9,425
3,728,623
-3,555
-4,882
-4,548
Table 6A-18. Annualized Changes in Consumer and Producer Surplus in Agriculture and
Forestry, 2010-2044, Million 2010 U.S. Dollars (7% Discount Rate)
Product
Consumer surplus
Producer surplus
Total surplus
Policy
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
Agriculture Forestry Total
1,951,843 782,748 2,734,591
Change with Respect to Existing Standard
21 5,570 5,592
61 5,599 5,660
100 5,721 5,821
800,022 130,681 930,703
Change with Respect to Existing Standard
-945 -5,561 -6,506
-2,719 -5,555 -8,273
-2,635 -5,694 -8,328
2,751,865 913,429 3,665,294
Change with Respect to Existing Standard
-923 9 -914
-2,658 45 -2,613
-2,535 27 -2,508
6A.2.3.3
Greenhouse Mitigation Potential
The capacity for both the agricultural and forest sectors to sequester carbon is enhanced
in each of the alternative standard cases, with increasing magnitude as policy stringency is
6-40
-------
increased. Although FASOMGHG projects fewer acres of forestland and total cropland, the
accelerated storage of carbon in trees and forestland and cropland soils outweighs any decline
from reductions in covered area. Carbon storage in both sectors is consistently higher in the
alternative standard cases, with the gap widening over time (see Figure 6A-15 for change in
forest carbon stock). By 2040, the agricultural sector sequestered 0.01%~0.05%% more carbon
under the alternative standard cases and the forestry sector up to 0.5% more, resulting in gains of
more than 1,500 million metric tons CCh equivalent (MMtCChe). Table 6A-19 presents carbon
sequestration projections under the current standard and changes under each alternative
standard.107 Note that negative values in the row for the current standard indicate sequestration or
carbon storage. Negative values in the change rows indicate that the alternative standard stores
more carbon than the current standard (and vice versa for positive changes).
Notice that for the agricultural sector, the overall stock of net GHG would decrease over
time in the baseline because cropping activities involve fertilizer and chemical usage, fossil
fuels, running machinery, livestock emissions from enteric fermentation and manure
management, and so forth—all these GHG emissions are being released each year, while soil
carbon sequestration moves toward equilibrium within 25 years of a change in tillage. As soil
carbon reaches equilibrium, little additional sequestration is taking place each year but annual
emissions from other sources continue. Thus, over time, the annual emissions tend to outweigh
the increase in carbon stocked in agricultural soils, and net stock of GHG tends to become less
negative and eventually positive relative to the starting point.108
107 These are total stocks of net GHG emissions over time, not annual emissions. If the total stock of GHG is
becoming more negative over time, more net sequestration is taking place than emissions. If the total stock of GHG
is becoming less negative or positive over time, emissions are greater than the increase in sequestration.
108 This change is consistent with the fact that U.S. agriculture is a net source of emissions on an annual basis. The
value of the total GHG stock associated with agriculture is starting at a negative value because of the FASOMGHG
convention of accounting for total carbon sequestration present in agricultural soils in the first year of the model run.
A large stock of carbon is sequestered, but it does not increase by much over time.
6-41
-------
(78,000)
(80,000)
(82,000)
(84,000)
Base
75 ppb
70 ppb
65 ppb
Figure 6A-15. Carbon Storage in Forestry Sector, MMtCChe
•60 ppb
Table 6A-19. Carbon Storage, MMtCChe
Product
Agriculture
Forestry
Policy
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010
-18,621
0
0
-1
-73,321
1
2
12
2020 2030 2040
-15,240 -11,772 -8,009
Change with Respect to Existing Standard
-1 -3 -1
-2 -6 -4
-4 -11 -11
-74,338 -77,963 -81,063
Change with Respect to Existing Standard
51 -100 -259
53 -323 -774
94 -465 -1,189
2010-2044
-268,210
-24
-62
-132
-1,533,424
-1,537
-5,207
-7,739
Changes in forestry sector carbon sequestration are largely driven by changes in forest
management, which include the increases in tree yield in the lower ozone environments. The
increased sequestration in this category outweighs losses in sequestration in the other major
6-42
-------
forestry categories: afforestation and forest soil. Table 6A-20 presents the detailed changes in
forestry carbon sequestration.
Table 6A-20. Forestry Carbon Sequestration, MMtCChe
Product
Afforestation,
Trees
Afforestion, Soils
Forest
Management
Forest Soils
Policy
75 ppb
70ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
75 ppb
70 ppb
65 ppb
60 ppb
2010
-730
Change
0
0
0
-767
Change
0
0
0
-39,827
Change
1
2
11
-28,320
Change
0
1
1
2020
-1,594
with Respect to
0
0
0
-598
with Respect to
0
0
0
-37,995
with Respect to
43
44
77
-27,698
with Respect to
10
13
21
2030
-963
Existing Standard
22
48
66
-550
Existing Standard
11
27
37
-40,023
Existing Standard
-125
-387
-553
-27,243
Existing Standard
5
20
27
2040
-1,502
29
64
88
-1,018
11
27
37
-39,556
-291
-852
-1,298
-27,474
4
19
28
6A.3 Summary
Impacts to both sectors generally mirror one another, although they are more prominent in
the forestry sector. Not only are tree species more responsive to changes in ozone, but the largest
reductions to meet the alternative standards will occur in regions with large forestry sectors:
South Central, Southeast, and Rocky Mountains. Reductions in agricultural regions are
comparatively moderate. Productivity of both crops and forests is projected to increase at each of
the alternative standard levels. This increase in supply resulted in decreased prices for forest
products and agricultural commodities, which benefits consumer welfare while reducing
producer welfare. Unless there are significant changes to wood products markets in particular,
producers will be forced to sell at reduced prices to absorb the increased supply. Nonetheless,
6-43
-------
gains to consumers become increasingly large with more stringent ozone standards. Gains to
agricultural producers in the most stringent case are associated with a decline in forestry returns
that results in a net shift in land use toward agriculture.
Increased productivity is also projected to affect land use both within and between the
agricultural and forest sectors. Within sectors, acreage is projected to shift from crops and tree
species that are more sensitive to ozone to those that are less sensitive because productivity in the
former will be more substantially affected by reductions in ozone concentrations. For ozone-
sensitive crops and species, producers are projected to require less land to produce at the same or
higher levels. Forest acreage in particular is projected to decline sharply, driven by declines in
both reforestation and afforestation.
Despite reductions in crop and forest area, carbon sequestration is expected to increase
over time, led almost entirely by increased forest sequestration. Although there is less
reforestation and afforestation and lower sequestration in new inventory, the change is a result of
existing inventories becoming so much larger as trees grow faster. Lower sequestration in new
inventory is outweighed by increased inventory in standing forests, represented in the model as a
change in forest management.
Increased stringency in the ozone standard generally produces larger impacts on all of the
model outputs. However, the additional impact of moving from the current standard to 65 ppb, or
to 60 ppb was sometimes marginal compared with changes occurring between the current
standard to 70 ppb. In particular, the impacts to the forestry sector, most notably in forest
inventories and the forest sector welfare analysis, tended to increase at a decreasing rate after
meeting the 70 ppb standard.
The model results are subject to several limitations: First, the ozone concentration response
functions applied to crops and trees were using "median" parameters in Lehrer et al. (2007)—the
RYLs and RYGs calculated are thus "median" ones; second, the use of crop proxy mapping and
the forest-type mapping due to incomplete data specified in Section 6.1 adds to the uncertainty of
these model results; third, the potential changes in tree species mixes within forest types due to
ground ozone-level changes were not considered; and last, the international trade component in
FASOMGHG that assumes USDA-based future projections under current conditions may present
6-44
-------
another uncertainty for the model results, especially when soybeans and wheat are among the
major crop commodities for U.S. exports and have relatively large responses to changed ozone
environments.
6A.4 References
Jin, S., Yang, L., Danielson, P., Homer, C., Fry, I, and Xian, G. 2013. A comprehensive change detection method
for updating the National Land Cover Database to circa 2011. Remote Sensing of Environment, 132: 159 -
175.
Lehrer, J.A., M. Bacou, B. Blankespoor, D. McCubbin, J. Sacks, C.R. Taylor, and D.A. Weinstein. 2007. Technical
Report on Ozone Exposure, Risk, and Impact Assessments for Vegetation. EPA 452/R-07-002.
Little, E.L., Jr. 1971. Atlas of United States Trees, Volume 1, Conifers and Important Hardwoods. U.S. Department
of Agriculture Miscellaneous Publication 1146, 9 p., 200 maps.
Little, E.L., Jr. 1976. Atlas of United States Trees, Volume 3, Minor Western Hardwoods. U.S. Department of
Agriculture Miscellaneous Publication 1314, 13 p., 290 maps.
Little, E.L., Jr. 1977'. Atlas of United States Trees, Volume 4, Minor Eastern Hardwoods. U.S. Department of
Agriculture Miscellaneous Publication 1342, 17 p., 230 maps.
Little, E.L., Jr. 1978. Atlas of United States Trees, Volume 5, Florida. U.S. Department of Agriculture
Miscellaneous Publication 1361, 262 maps.
U.S. Department of Agriculture, National Agricultural Statistics Service (USDA NASS). Various years. USDA
Agricultural Statistics (1990-2012). Available at: http://www.nass.usda.gov/Publications/Ag_Statistics/.
U.S. Department of Agriculture, Natural Resource Conservation Service (USDA NRCS). 2003. Annual NRI—Land
Use. Available at: http://www.nrcs.usda.gov/technical/NRI.
6-45
-------
CHAPTER 7: ENGINEERING COST ANALYSIS AND ECONOMIC IMPACTS
Overview
This chapter summarizes the data sources and methodologies used to estimate
engineering costs of attaining the alternative, more stringent levels for the ozone primary
standards analyzed in this regulatory impact analysis (RIA). The chapter also provides estimates
of the engineering costs of control strategies presented in Chapter 4 for the alternative standards
of 70, 65, and 60 ppb. The discussion is presented as follows: Section 7.1 presents the costs
associated with the application of known controls and is followed by discussions about the
challenges of estimating costs for unknown controls (Section 7.2); costs associated with
emissions reductions from unknown controls that are needed to demonstrate full attainment of
the alternative standards analyzed (Section 7.3); total compliance cost estimates (Section 7.4);
updated methodology (Section 7.5); economic impacts (Section 7.6); and the uncertainties and
limitations associated with these components of the RIA (Section 7.7).
The engineering costs described in this chapter generally include the costs of purchasing,
installing, operating, and maintaining the referenced technologies. The costs associated with
monitoring, testing, reporting, and record keeping for affected sources are not included in the
annualized cost estimates. For a variety of reasons, actual control costs may vary from the
estimates the EPA presents. As discussed throughout this document, the technologies and control
strategies selected for analysis are illustrative of one way in which nonattainment areas could
meet a revised standard. There are numerous ways to construct and evaluate potential control
programs that would bring areas into attainment with alternative standards, and the EPA
anticipates that state and local governments will consider programs that are best suited for local
conditions. Also, the EPA recognizes the unknown emissions control portion of the engineering
cost estimates (Section 7.3) reflects substantial uncertainty about the sectors and technologies
that might become available for cost-effective application in the future.
The engineering cost estimates are limited in their scope. This analysis focuses on the
emissions reductions needed for attainment of a range of alternative revised standards. The EPA
understands that some states will incur costs both designing State Implementation Plans (SIPs)
for and implementing new control strategies to meet final revised standards. However, the EPA
7-1
-------
does not know what specific actions states will take to design their SIPs to meet final revised
standards. Therefore, we do not present estimated costs that government agencies may incur for
managing the requirement, implementing these (or other) control strategies, or for offering
incentives that may be necessary to encourage the implementation of specific technologies,
especially for technologies that are not necessarily market driven. This analysis does not assume
specific control measures that would be required in order to implement these technologies on a
regional or local level.
7.1 Estimating Engineering Compliance Costs
7.1.1 Methods and Data
After designing the hypothetical control strategy using the methodology discussed in
Chapter 4, the EPA used the Control Strategy Tool (CoST) (U.S. EPA, 2014a) to estimate
engineering control costs for non-electric generating unit (non-EGU point) point, nonpoint and
mobile nonroad sources. CoST calculates engineering costs using one of two different methods:
(1) an equation that incorporates key operating unit information, such as unit design capacity or
stack flow rate, or (2) an average annualized cost-per-ton factor multiplied by the total tons of
reduction of a pollutant. Most control cost information within CoST was developed based on the
cost-per-ton approach because estimating engineering costs using an equation requires more
detailed data, and parameters used in these equations may not be readily available or broadly
representative across sources within the emissions inventory. The cost equations used in CoST
estimate annual, capital and/or operating and maintenance (O&M) costs and are used primarily
for some larger sources such as industrial/commercial/institutional (ICI) boilers and petroleum
refinery process heaters. Information on CoST control measures information, including cost-per-
ton factors and cost equations can be found at www.epa.gov/ttnecasl/cost.htm. Costs for
selective reduction catalysts (SCR) applied as part of the analysis for reducing NOX at coal-fired
electric generating units (EGUs) were estimated using documentation for the Integrated Planning
Model (IPM) (Sargent & Lundy, 2013).
7-2
-------
Capital costs are converted to annual costs using the capital recovery factor (CRF).109
Where possible, calculations are used to calculate total annual control cost (TACC), which is a
function of capital costs (CC) and O&M costs. The CRF incorporates the interest rate and
equipment life (in years) of the control equipment. Operating costs are calculated as a function of
annual O&M and other variable costs. The resulting TACC equation is TACC = (CRF * CC) +
O&M. For more information on this cost methodology, refer to the EPA Air Pollution Control
Cost Manual (U.S. EPA, 2003) and EPA's Guidelines for Preparing Economic Analyses,
Chapter 6 (US. EPA, 2014b).
Engineering costs will differ depending on the quantity of emissions reduced, emissions
unit capacity, or stack flow, which can vary over time. Engineering costs will also differ in
nominal terms by the year for which the costs are calculated (e.g., 2011$ versus 2008$).uo For
capital investment, in order to attain standards in 2025 we assume capital investment occurs at
the beginning of 2025. We make this simplifying assumption because we do not know what all
firms making capital investments for control measures will do and when they will do it. Our
estimates of annualized costs include annualized capital and annual O&M costs for those
controls included in our known control strategy analysis. Our engineering cost analysis uses the
equivalent uniform annual costs (EUAC) method, in which annualized costs are calculated based
on the equipment life for the control measure and the interest rate incorporated into the CRF.
Annualized costs represent an equal stream of yearly costs over the period the control technology
is expected to operate. We make no presumption of additional capital investment in years beyond
2025. The EUAC method is discussed in detail in the EPA Air Pollution Control Cost Manual
(U.S. EPA, 2003). The controls applied and their respective engineering costs are described in
the Chapter 7 Appendix.
109 The capital recovery factor formula is expressed as r*(l+r)An/[(l+r)An -1]. Where r is the real rate of interest and
n is the number of time periods.
11 ° The engineering costs will not be any different in real (inflation-adjusted) terms if calculated in 2011 versus other
year dollars, if the other-year dollars are properly adjusted. For this analysis, all costs are reported in real 2011
dollars.
7-3
-------
7.1.2 Compliance Cost Estimates for Known Controls
In this section, we provide engineering cost estimates for the known controls identified in
Chapter 4 that include control technologies for EGUs, non-EGU point, nonpoint and mobile
nonroad sources. Onroad mobile source controls were not applied because they are largely
addressed in existing rules such as the recent Tier 3 rule. Engineering costs generally refer to the
equipment installation expense, the site preparation costs for the application, and annual
operating and maintenance costs. Note that in many cases the application of these control
strategies does not result in areas reaching attainment for the alternative ozone standards of 70,
65, and 60 ppb and additional emission reductions beyond known controls are needed.
The EPA evaluated the costs of all known NOX controls contained in the CoST control
measures database for this RIA and found that all nonpoint and nonroad sector controls were
below $14,000 per ton of NOX emission reduction, and that the bulk of the non-EGU point source
controls were below this cost. Overall, for all NOX controls prior to application of any cost cap,
controls costing less than $14,000 per ton account for 96 percent of known emission reductions.
Figure 7-1 represents the marginal cost curve for all the NOx control measures contained in the
CoST database. This is an incomplete representation of the marginal abatement cost curve for all
NOx abatement, because we do not have information on the control measures and costs for the
remaining uncontrolled NOx emissions (see discussion in section 7.2). The controls above
$14,000 were investigated and we determined that the higher cost controls were primarily due to
errors in the cost equations for a few types of sources (mainly ICI Boilers and Process Heaters)
for certain source sizes. We are taking steps to correct these equations for the final ozone RIA.
As a result of the error mentioned above, a cost cap of $14,000 per ton of NOX emissions
reduction for the non-EGU point source controls was applied, to remove the incorrectly applied
controls from the analysis. A small number of NOX controls were applied for non-EGU point
sources above $14,000 per ton but these had little impact on the overall emission reductions or
control cost. A significant portion of the EGU SCR controls were above this level; no cost cap
was applied to the EGU SCR controls. Note that control costs for California were lower than for
other areas because there were no EGU SCR controls applied in California because there were
no coal-fired utility boilers without SCR already in place.
7-4
-------
We intended to apply a similar cost cap but inadvertently applied a slightly higher cap of
$15,000 per ton of VOC emission reduction. At the beginning of the analysis we were
anticipating VOC reductions would play a relatively minor role in the analysis so we did not do a
separate marginal cost analysis for VOC. However, for the final ozone RIA we will conduct a
separate analysis for VOC and will make costing decisions accordingly.
238,000
224,000
210,000
196,000
_ 182,000
.0 168,000
^ 154,000
"g 140,000
C 126,000
;S 112,000
^ 98,000
g 84,000
u 70,000
56,000
42,000
28,000
14,000
0 4
Marginal Costs for NOx Controls
MA
•^ ^
«•
:-
•
•
•
1
I
1
1
^r
l» tf^^*^^*^^^
0 200,000 400,000 600,000 800,000 1,000,000 1,200,000 1,400,000
Cumulative NOx Emission Reductions (tons/yr)
Figure 7-1. Marginal Costs for Known NOX Controls for All Source Sectors (EGU, non-
EGU Point, Nonpoint, and Nonroad)
See Tables 7-1 through 7-3 for summaries of control costs from the application of known
controls for alternative standards of 70, 65, and 60 ppb. Costs are listed by sector for both
eastern and western U.S., except California and presented at 3 and 7 percent discount rates. Note
that any incremental costs for known controls for California (post-2025) for alternative standards
7-5
-------
of 70, 65, and 60 ppb are zero because all known controls for California were applied in the
demonstration of attainment for the baseline standard of 75 ppb.
These numbers reflect the engineering costs annualized at discount rates of 3 percent and
7 percent, which is to the extent possible consistent with the guidance provided in the Office of
Management and Budget's (OMB) (2003) Circular A-4. Discount rates refer to the rate at which
capital costs are annualized.111 A higher discount, or interest, rate results in a larger annualized
cost of capital estimate. It is important to note that it is not possible to estimate both 3 percent
and 7 percent discount rates for a number of the controls included in this analysis. Because we
obtain control cost data from many sources, we are not always able to obtain consistent data
across original data sources.112 If disaggregated control cost data is unavailable (i.e., where
capital, equipment life value, and O&M costs are not separated out), EPA typically assumes that
the estimated control costs are annualized using a 7 percent discount rate. When disaggregated
control cost data is available (i.e., where capital, equipment life value, and O&M costs are
explicit) we can recalculate costs using a 3 percent discount rate.
In addition, the EGU control costs were not estimated for either 3 or 7 percent. The
interest rate used for this analysis reflects an internal rate of return of 11.51 percent for retrofit
controls as described in the IPM v5.13 documentation (U.S. EPA, 2013).
For non-EGU point source controls, some disaggregated data is available, and we were
able to calculate those costs at both 3 and 7 percent discount rates for those controls. For the
alternative standards analyzed in this RIA, approximately 23 percent of known control costs are
disaggregated at a level that could be discounted at 3 percent. Because we do not have
disaggregated control cost data for any EGU, nonpoint, or nonroad source controls, total
annualized costs for these sectors are assumed to be calculated using a 7 percent discount rate.
Because we do not have a full set of costs at the 3 percent discount rate, the 3 percent columns in
111 In this analysis, the discount rate refers to the interest rate used in the discounted cash flow analysis to determine
the present value of future cash flows. A social discount rate is a discount rate used in computing the value of
monies spent on social projects or investments, such as environmental protection. The social discount rate is directly
analogous to the discount rate we use in the engineering cost analysis, as well as certain rates used in corporate
finance (e.g., hurdle rate or a project appropriate discount rate), so the mathematics are identical.
112 Data sources can include states and technical studies, which do not always include the original data source.
7-6
-------
Tables 7-1 to 7-3 reflect the sum of some non-EGU point source controls at a 3 percent discount
rate, some non-EGU point source controls at a 7 percent discount rate, and the other sectors at a
7 percent discount rate. With the exception of the 3 percent Total Annualized Cost estimates in
Tables 7-1 to 7-3, engineering cost estimates presented throughout this chapter and elsewhere in
this document are based on a 7 percent discount rate.
The total annualized engineering costs associated with the application of known controls,
incremental to the baseline and using a 7 percent discount rate, are approximately $1.6 billion for
an alternative annual standard of 70 ppb, $4.2 billion for a 65 ppb alternative standard, and $4.4
billion for a 60 ppb alternative standard. Costs of NOX controls in terms of dollars per ton of
NOX reduction for the alternative standards analyses were approximately $12,000/ton on average
for the EGU sector with a range of $2,000/ton to $38,000/ton; $3,000/ton for the non-EGU point
sector on average, with a range of $17 to $90,000/ton; $l,100/ton for the nonpoint sector on
average ,with a range of $520 to $2,200/ton; and $4,600/ton for the nonroad sector on average
with a range of $3,300/ton to $5,300/ton. The overall average cost range was $2,300 to
$3,000/ton for all sectors combined.
The cost trend in terms of dollars per ton of emissions reduction is increasing for some
sectors and decreasing for others. The variation is small for most sectors and depends on the
sources that happen to be in the geographic areas of control. In general, the same set of controls
is being applied at each level of the alternative standards analyzed. The primary difference in the
control strategies for the alternative standards is the greater size of the geographic area of control
as the stringency of the alternative standards increases. Overall for all sectors as a whole, the
average cost per ton is actually increasing slightly with increasing stringency of the alternative
standard analyzed. The reason for slight increase in cost per ton relative to the increase in
stringency is that newly affected areas are estimated to obtain emissions reductions using slightly
more expensive known technologies as the alternative standards become more stringent.
7-7
-------
Table 7-1. Summary of Known Annualized Control Costs by Sector for 70 ppb for 2025
- U.S., except California (millions of 2011$)a
Geographic Area Emissions Sector Known Control Costs
7 Percent 3 Percent
Discount Rate Discount Rate
East
West
ECU
Non-EGU Point
Nonpoint
Nonroad
Total
ECU
Non-EGU Point
Nonpoint
Nonroad
Total
Total Known Control Costs
310b
640
610
22
1,600
-
-
-
-
-
1,600
310b
620C
610d
22d
l,600e
-
-
-
-
-
l,600e
a All values are rounded to two significant figures.
b EGU control cost data is calculated using an 11.51 percent for retrofit controls.
0 Non-EGU control cost data is calculated at a 3% interest where control cost equations are utilized.
d Nonpoint and nonroad control costs are calculated using a 7% interest rate because no 3% data exists.
e Because we obtain control cost data from many sources, we are not always able to obtain consistent data across
original data sources. Where disaggregated control cost data is available (i.e., where capital, equipment life value,
and O&M costs are explicit) we can calculate costs using a 3 percent discount rate. Therefore the cost estimate
provided here is a summation of costs at 3 percent and 7 percent discount rates.
-------
Table 7-2. Summary of Known Annualized Control Costs by Sector for 65 ppb for 2025
- U.S., except California (millions of 2011$)a
Geographic Area Emissions Sector
ECU
Non-EGU Point
East Nonpoint
Nonroad
Total
ECU
Non-EGU Point
West Nonpoint
Nonroad
Total
Total Known Control Costs
Known
7 Percent
Discount Rate
l,500b
1,100
1,100
53
3,800
230b
86
87
5.7
410
4,200
Control Costs
3 Percent
Discount Rate
l,500b
1,100C
l,100d
53d
3,800e
230b
86C
87d
5.7d
410e
4,200e
a All values are rounded to two significant figures.
b EGU control cost data is calculated using an 11.51 percent for retrofit controls.
0 Non-EGU control cost data is calculated at a 3% interest where control cost equations are utilized.
d Nonpoint and nonroad control costs are calculated using a 7% interest rate because no 3% data exists.
e Because we obtain control cost data from many sources, we are not always able to obtain consistent data across
original data sources. Where disaggregated control cost data is available (i.e., where capital, equipment life value,
and O&M costs are explicit) we can calculate costs using a 3 percent discount rate. Therefore the cost estimate
provided here is a summation of costs at 3 percent and 7 percent discount rates.
Table 7-3. Summary of Known Annualized Control Costs by Sector for 60 ppb for 2025
- U.S., except California (millions of 2011$)a
Geographic Area Emissions Sector Known Control Costs
__ , _.. .-„. 3 Percent Discount
7 Percent Discount Rate „ ,
Rate
East
West
EGU
Non-EGU Point
Nonpoint
Nonroad
Total
EGU
Non-EGU Point
Nonpoint
Nonroad
Total
Total Known Control Costs
l,500b
1,200
1,100
53
3,800
400
98
88
5.7
590
4,400
l,500b
1,100C
l,100d
53d
3,700e
400b
970
88d
5.7d
590e
4,400e
a All values are rounded to two significant figures.
b EGU control cost data is calculated using an 11.51 percent for retrofit controls.
0 Non-EGU control cost data is calculated at a 3% interest where control cost equations are utilized.
d Nonpoint and nonroad control costs are calculated using a 7% interest rate because no 3% data exists.
e Because we obtain control cost data from many sources, we are not always able to obtain consistent data across
original data sources. Where disaggregated control cost data is available (i.e., where capital, equipment life value,
and O&M costs are explicit) we can calculate costs using a 3 percent discount rate. Therefore the cost estimate
provided here is a summation of costs at 3 percent and 7 percent discount rates.
7-9
-------
7.2 The Challenge of Estimating Costs for Unknown Controls
As described in Chapter 4, the known control measures were applied to EGU, non-EGU
point, nonpoint (area), and nonroad mobile sources for demonstration of attainment with the
current and alternative standards. Table 4-7 lists the specific control technologies applied in the
known control analysis. There were several areas where known controls did not achieve enough
emissions reductions to attain the alternative standards of 70, 65, and 60 ppb. To complete the
analysis, the EPA then estimated the additional emissions reductions beyond known controls
needed to reach attainment, also referred to as unknown controls. For information on the
methodology used to develop the emissions reductions estimates, see Chapter 3.
The estimation of engineering costs for unspecified emission reductions needed to reach
attainment many years in the future is inherently a difficult task. This is because it is likely that
the abatement supply function will shift out or change shape over time due to a variety of
economic, technical, and regulatory influences. Our experience with Clean Air Act
implementation shows that numerous factors, such as technical change and development of
innovative strategies, can lead to emissions reductions that may not seem possible today, while
potentially reducing costs over time. For example, facility-level data collected through the U.S.
Census Bureau's Pollution Abatement Costs and Expenditures (PACE) survey suggests that this
may have happened in the manufacturing sector in recent decades. Based on surveys of
approximately 20,000 plants classified in manufacturing industries, the PACE data show during
the 1994-2005 time period, a period of increasing regulatory stringency, spending on air
pollution abatement as a percentage of revenues decreased for the manufacturing
sector.113 Although exogenous factors such as changes in economic conditions may have
113 The Pollution Abatement Costs and Expenditures (PACE) survey collects facility-level data on pollution
abatement capital expenditures and operating costs for compliance with local, state, and federal regulations and
voluntary or market-driven pollution abatement activities. In 2005, the most recent year PACE data were collected,
the U.S. manufacturing sector spent $3.9 billion dollars on air capital expenditures and incurred $8.6 billion dollars
in operating costs for air pollution prevention and treatment. These figures represent less than 3% of total new
capital expenditures and less than 0.18% of total revenue for the manufacturing sector, respectively. These
percentages have declined since 1994, when air capital expenditures were less than 4% of total new capital
expenditures and air pollution abatement operating costs were less than 0.2% of total revenue. Levinson (2009) finds
that most of the pollution reductions in the U.S. come from changes in technology as opposed to changes in imports
or changes in the types of domestically produced goods. He finds that even though manufacturing output increased
by 24% from 1987 to 2001, emissions of four common air pollutants from the sector declined 25% over that time
period and the most important factor contributing to the decrease in pollution is technical change or innovation.
7-10
-------
contributed to the relative share in costs of pollution abatement, it is also possible that
technological change and innovation may have contributed to this relative decline.
In addition to considering the potential for technological innovation, it is also important
to understand that EPA's control strategy tools largely focus on a limited set of emissions
inventory sectors, whereas abatement opportunities exist in other sectors. EPA's control strategy
tools undergo continuous improvement, and as the need for additional abatement opportunities
grows, more evaluation of uncontrolled emissions takes place. During these evaluations,
additional abatement opportunities from applying identified controls typically are found.
This section discusses various factors that must be considered in developing a
methodology for estimating the costs of unknown controls. First, we explain why the abatement
supply curve from known controls presented in the previous section provides an incomplete
picture of all currently available abatement opportunities. Second, we show how, as time passes
and the EPA reviews NAAQS standards, relevant information is revealed in the current RIA
development process that was not available to analysts developing RIAs for previous reviews,
such as unforeseen regulatory programs or other exogenous factors that account for significant
emissions reductions. Third, we discuss a related issue, that technical change may affect the
marginal abatement cost curve. Additionally, we present evidence from the literature that
regulatory action can act as a forcing function for technical change. Fourth, we discuss how
regulatory costs can decrease over time as regulated entities gain experience reducing emissions
and reduce per unit costs, commonly referred to as "learning by doing". Fifth, we discuss how
NOx offset prices could serve as reasonable proxies for the costs associated with emissions
reductions from unknown controls. Finally, we describe how we use this information to help
inform the unknown control cost methodology applied in section 7.3.
7.2.1 Incomplete Characterization of NOx Marginal Abatement Cost Curves
Underlying the selection of controls as described in Appendix 4 A is the concept of the
marginal abatement cost curve (MACC). The marginal abatement cost curve (MACC) is a
representation of how the marginal cost of additional emissions abatement changes with
increasing levels of abatement. Adding new technologies, or changing either the abatement
amount or cost of the technology, will change the shape of the overall MACC.
7-11
-------
In developing engineering cost estimates in section 7.1, the focus was largely on end-of-
pipe controls and only includes limited process-oriented control measures, such as switching to
lower-emitting fuel or energy sources and the installation of energy efficiency measures. These
measures can result in significant emissions abatement, but are not reflected in the marginal
abatement cost curve based on traditional control measures. As a result, the MACC derived in
the previous section from known controls represents an incomplete supply curve that partially
captures the "true" abatement supply. An illustrative, hypothetical depiction of an "observed but
incomplete" MACC and the "true" underlying MACC is presented in Figure 7-2.
$/ton
* Adding abatement 'unobserved1
by current tools shifts curve to the
right, implying a greater supply of
abatement than observed
Emissions Reductions
Figure 7-2. Observed but incomplete MACC (solid line) based on known controls
identified by current tools and complete MACC (dashed line) where gaps
indicate abatement not identified by current tools
In the figure, the solid line traces out a hypothetical observed MACC, while the dashed line
characterizes the combination of observed and unobserved abatement possibilities. The
inclusion of the unobserved abatement pushes the abatement supply out.
Due to the incomplete characterization of the full range of the MACC, it is important to
understand the composition of the cost information that is available to construct the partial
7-12
-------
MACC. The nature of available information on the cost of NOx abatement measures is
somewhat complex. The highest cost per ton estimates are often associated with controls that
achieve very small total reductions in NOx or are special cases. For example, in some cases,
controls have been developed primarily to address other pollutant emissions, such as SCh, but
achieve NOx reductions as a co-benefit. These controls are well characterized in the CoST
database because they have been applied for SO2 control, but the degree to which sources would
adopt those controls for NOx is uncertain, and it is unlikely that those very high per ton cost
measures would be the efficient marginal abatement cost (MAC) level for the targeted level of
NOx emissions reductions needed for attainment. In addition, there are some controls that have
been identified as applicable for some sources even though they have very high marginal costs
per ton of NOx, because those controls are required to meet other provisions of the CAA. For
example, LNB+SCR for process heaters is relatively common at refineries, mainly due to a New
Source Review enforcement initiative that has been underway since 2001. These plants are
being required to install the controls regardless of cost as part of the enforcement action, and thus
the costs may not represent the actual marginal cost at the level of abatement needed to reach the
NAAQS attainment targets.
Lack of information about the MAC for emissions reductions not characterized in CoST
is not an indication that controlling those tons is necessarily more difficult than controlling NOx
from other sources that are in the database, or that the MAC for those tons is necessarily higher
than all of the costs of controls already in the database. Some sectors are controlled at a higher
rate than others, and in those cases, getting additional NOx reductions may indeed require higher
cost controls. For example, EGU NOx has been heavily controlled, and the additional SCR units
applied in the analysis are relatively more expensive per ton than the typical SCRs that have been
applied in past analyses. However, other sectors may not be as well-controlled, and lower cost
controls may be available.
7.2.2 Comparison of Baseline Emissions and Controls across Ozone NAA QS RIAs from 1997
to 2014
While each ozone NAAQS analysis since 1997 has required at least some emissions
reductions from controls that were unknown at the time of the analysis, evidence suggests that
over time new information on exogenous factors affecting baseline emissions and emissions
7-13
-------
controls becomes available that can shift emissions reductions from the unknown to the known
category. Some exogenous factors that might affect baseline emissions include changes in
economic conditions that may affect production levels, as well as plant closures and openings.
Baseline emissions may also be affected by EPA or state regulations that require specific
controls that may not be fully characterized in the set of known controls applied in an earlier
RIA, or by EPA or state regulations targeting other pollutants, e.g., air toxics, that may result in
reductions in NOx or VOC as a co-benefit. For example, in the 1997 ozone NAAQS RIA, the
NOx emissions reductions from the mobile source Tier 2 standards were not included as known
controls, even though the RIA acknowledged the potential for these standards to provide
substantial cost effective controls. Likewise, the 2008 ozone NAAQS RIA did not include
controls on EGUs reflecting the Mercury and Air Toxics Standards or the Clean Power Plan. As
a result, emissions reductions from unknown controls were much higher in some regions of the
U.S. than they are in this RIA.
Furthermore, many of these emission reductions may be achieved at a cost less than was
originally applied to emissions reductions from unknown tons. Several of the large NOx-
reducing regulations issued between the 1997 and 2008 RIAs had ex ante estimates of costs per
ton of NOx reduced well below the $10,000 per ton value applied to emissions reductions from
unknown controls in the 1997 Ozone NAAQS RIA. Table 7-4 provides information on the cost
per ton for NOx reductions from five major NOx-reducing regulations. To the extent possible,
we determined from the RIA for each regulation the projected emissions reductions in 2010 (or
the closest year) and the estimated costs in 2010. Costs were adjusted from the year reported in
the RIAs to constant 2010 dollars. In some cases, only an annualized cost was reported. In those
cases we divided the year specific emissions projection by the annualized cost, recognizing that
this may under or overstate the actual year specific cost per ton. Where possible, we report the
separate costs for NOx emissions reductions. However, for most programs, total costs which
include reductions in multiple pollutants are reported. As such, the cost per ton of NOx alone is
likely overstated.
Table 7-4. Emissions and Cost Information for Major NOx Rules Issued Between 1997 and
2008
Projected NOx Total Annual Cost Average Cost/Tons
Emissions Reductions in in 2010 NOx Reduced
Regulation 2010 (thousands) (Million 2010$) (2010$)
7-14
-------
(A) (B) (C = B/A)
NOx SIP Call" 1,141 2,515 $2,204
Clean Air Interstate Ruleb 1,200 $3,034 $2,528
Tier 2 Standards0 1,236 $5,256 $4,253
Heavy Duty Diesel Engines'1 403 $4,570 $11,340
Nonroad Diesel6 203 (in 2015) $660 $3,251
a Costs and emissions reductions obtained from Table ES-2 in the Regulatory Impact Analysis for the NOx SIP Call,
FIP, and Section 126 Petitions, Volume 2: Health and Welfare Benefits (U.S. EPA, 1998).
b Emissions reductions obtained from Table 7-2 and Costs obtained from Table 7-3 in the Regulatory Impact
Analysis for the Final Clean Air Interstate Rule (U.S. EPA, 2005).
0 Costs and emissions reductions obtained from Appendix VI-C in the Regulatory Impact Analysis - Control of Air
Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control
Requirements (U.S. EPA, 1999).
d Emission reductions obtained from Table II.B-4, and costs obtained from Table V.D-1 in the Regulatory Impact
Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements (U.S.
EPA, 2000)
e .Emissions reductions obtained from Table 8.6-1 (NOx+NMHC), and costs obtained from Table 8.5-2
(NOx+NMHC) in the Final Regulatory Impact Analysis: Control of Emissions from Nonroad Diesel Engines (U.S.
EPA, 2004). In this RIA, no information was provided for NOx emissions alone, in all cases, NOx+non-methane
hydrocarbons (NMHC) was provided. Calculated cost per ton is thus the average for NOx and NMHC emissions
reductions, rather than just for NOx emissions reductions.
The NOx State Implementation Plan (the "NOx SIP call") was partially represented in the
1997 RIA. The other rules in Table 7-4 were not included in 1997 RIA, although the RIA notes
that the Tier 2 standards may be a way to get additional reductions to meet the ozone NAAQS.
Table 7-4 shows that for these five major regulations, which account for 4.1 million tons of NOx
reductions, the expected average cost per ton ranged from about $2,200 to $11,300 per ton of
NOx reduced. Only in the single case of the heavy duty diesel engine rule was the cost per ton
NOx reduced expected to be greater than the $10,000/ton value used for unknown controls in the
1997 RIA. This suggests that unknown controls may be implemented at lower cost than the
highest point of the MACC for known controls anticipated in the RIA.
Comparing the MACC over time and across analyses is complicated because of (1)
differences in the regions or areas affected by the proposed changes to the ozone standards, (2)
the information available at the time of the analyses, and (3) analytical assumptions made about
the universe of sources that can be controlled. Table 7-5 provides information about
assumptions used in generating the costs of control measures for the 1997, 2008, and 2014 RIAs.
7-15
-------
Table 7-5. Comparison of Key Assumptions Used in Developing Estimates of NOx Emissions Controls and Costs across Past
and Current Ozone NAAQS Regulatory Impact Analyses a
1997
2008
2014
Sectors Excluded
No additional utility NOx controls beyond
Title IV and OTAG recommendations.
Some mobile source control measures
excluded due to mismatch between
attainment dates and implementation
timelines for the control measures.
Utility NOx controls beyond
baseline only applied in the East
(including East TX)
On-road mobile sources.
Exceeding county plus CMSA for county
exceeding the standard. If no CMSA, just
use the exceeding county
(12 areas in East, 7 areas in West)
Geographic
Definition of
Control Areas
Non-EGU point and area controls
applied to counties exceeding the
standard plus surrounding counties
out to 200km. Some additional
controls were placed on large point
sources in counties touching the
buffer.
For mobile controls, both local
(within 200 km buffer) and
statewide controls were applied.
Counties outside the state in which
the exceeding county resides are
excluded. Controls were applied
to all states in the OTC excepting
VT.
NOx controls for 70 ppb were applied to counties
exceeding the standard, then surrounding counties
within 200 km that were also within state
boundaries (Texas, California) and within OTC
boundaries including counties closest to exceeding
counties first (Northeast). Where more controls
were needed they were applied to counties in other
states within the region, nearest the county
exceeding the standard (Oklahoma, Arkansas,
Louisiana and Kansas for Houston). VOC controls
for 70 ppb were applied to special areas and
counties within 100 km from them (California
North and South, Dallas, Houston, Baltimore, and
New York City).
NOx controls for 65 and 60 ppb were applied in
regions (California, Southwest, Central, Midwest,
Northeast). VOC controls were applied in special
areas plus counties within 100 km from these
special areas (Denver, Dallas, Houston, Chicago,
New York City)
Cost Cap
$10,000/ton (1990 dollars) -justification is
that states generally have not chosen to
require existing sources to apply control
measures with incremental costs above the
threshold, even in severe areas like the
South Coast of CA. Sensitivity on
$7,000/ton to $20,000/ton.
$23,000/ton (2006$) for non-EGU
point and area sources - 98% of
possible reductions are achieved at
82% of total costs. Based on
evaluation of marginal cost curves
for all counties in control areas
(1,300 counties).
For NOx: $14,000/ton for point non-EGU sources,
although an error resulted in a small number of
point non-EGU controls above $14,000/ton. No
nonpoint (area) source NOx controls were above
$14,000/ton so cost caps had no effect on nonpoint
source controls. For ICI boilers, costs were
calculated through equations without cost
constraints.
7-16
-------
1997
2008
2014
For VOC: $15,000/ton for point non-EGU and
nonpoint (area) sources
Minimum
Emission
Reduction
Minimum
Emissions
Assumed
Not clearly defined
Not clearly defined
95 to 100 percent
For NOx: 5 tons for point non-
EGU and non-point sources
For VOC: 1 ton for point non-
EGU and non-point
50 tons for NOx
Variable
For NOx: 5 tons for point non-EGU and non-point
For VOC: 1 ton for point non-EGU and non-point
25 tons for NOx
Variable
Control
Effectiveness (%
of intended
effect)
Rule
Penetration13
Obtained from published reports from state
and local agencies
75% for onroad and nonroad SCR
and diesel particulate filters
75% penetration in CA and 25% in TX and
Northeast for nonroad
a These analyses also included controls for VOC emissions, however, those controls were generally very local in nature. The current analysis is focused
primarily on controlling NOx to meet the alternative ozone standards, with relatively modest VOC controls where they are expected to help reach attainment.
Also, for the 1997 and 2008 RIAs, attainment of the PM standards in place at the time of the analysis was assumed, and as a result, some additional NOx
measures were already in place.
b Percent of county-level mobile or area source inventory affected by control measure.
7-17
-------
7.2.3 Impact of Technological Innovation and Diffusion
In general, the MACC at any particular point in time for a defined set of emitting sectors
will be an increasing function of the level of abatement. That is, marginal costs are increasing as
the amount of emissions are reduced. However, in regulatory analyses of NAAQS, we are
typically assessing costs of abatement in a future year or years selected to represent
implementation of the standards. As such, a MACC constructed based on currently available
information on abatement opportunities will not be the best representation of a future MACC.
MACC in the future can differ from MACC in the present due to technological innovation and
diffusion, such as the introduction of new technologies or improvements in effectiveness or
applicability of existing technologies. Additionally, environmental policy can create incentives
and constraints that influence the rate and direction of technical change (Jaffe et al. 2002) as well
as the rate of diffusion and adoption of the innovations (Sterner and Turnheim 2009) .
In the context of emissions controls examined in this RIA, technological innovation and
diffusion can affect the MACC in several ways. The following bullets present some examples of
the potential effects of technical change:
Case 1: New control technologies can be developed that cost less than existing technologies.
Case 2: A new control technology is developed to address an uncontrolled emissions source, at
a higher marginal cost than existing technologies, but still lower than the cost threshold value.
Case 3: The efficiency of an existing control measure increases. In some cases, the control
efficiency of a measure can be improved through technological advances.
Case 4: The cost of an existing control measure decreases.
Case 5: The applicability of an existing control measure to other emissions sources increases.
Overall, these five cases describe ways that technological change can reduce both the
amount of unidentified abatement needed, decrease the MAC, decrease average costs, and
decrease total costs relative to the case where it is assumed that technological change does not
occur in response to increased demand for abatement. It is also possible in cases where there is a
strictly binding emissions reduction target that new technologies can be introduced and adopted
with much higher marginal costs. However, if there are cost off-ramps, such as those provided by
Section 185 of the CAA, those higher cost technologies will not be adopted.
7-18
-------
Regulatory policies can also help induce technological change when a standard cannot be
met either (1) with existing technology or (2) with existing technology at an acceptable cost, but
over time market demand will provide incentives for industry to invest in research and
development of appropriate technologies. These incentives are discussed in Gerard and Lave
(2005), who demonstrate that the 1970 Clean Air Act induced significant technical change that
reduced emissions for 1975 and 1976 automobiles. Those mandated improvements went beyond
the capabilities of existing technologies by using regulatory pressure to incentivize the
development of catalytic converting technology in 1975. Induced technological change can
correspond to Cases 1 through 3 above.
There are many other examples of low-emission technologies developed and/or
commercialized over the past 15 or 20 years, such as:
• Selective catalytic reduction (SCR) and ultra-low NOx burners for NOx emissions
• Scrubbers that achieve 95 percent or greater 862 control on boilers
• Sophisticated new valve seals and leak detection equipment for refineries and chemical
plants
• Low or zero VOC paints, consumer products and cleaning processes
• Chlorofluorocarbon (CFC) free air conditioners, refrigerators, and solvents
• Water and powder-based coatings to replace petroleum-based formulations
• Vehicles with lower emissions than believed possible in the late 1980s due to
improvements in evaporative controls, catalyst design and fuel control systems for light-
duty vehicles; and treatment devices and retrofit technologies for heavy-duty engines
• Idle-reduction technologies for engines, including truck stop electrification efforts
• Increasing market penetration of gas-electric hybrid vehicles and cleaner fuels
These technologies were not commercially available two decades ago, and some were not
even in existence. Yet today, all of these technologies are on the market, and many are widely
employed. Several are key components of major pollution regulatory programs.
As Brunnermeier and Cohen (2003) demonstrate, there is a positive correlation, other
things held constant, between environmental innovations (measured as the number of relevant
environmental patent applications) and specific regulations imposed on an industry (measured in
terms of the frequency of government compliance inspections). Lanjouw and Mody (1996)
7-19
-------
show empirically a positive relationship between responses to environmental regulations (i.e.,
increases in pollution abatement expenditure) and new technology (i.e., relevant patent
applications) in the United States, Japan, and Germany. They show that in each of these
countries, even though on different timelines, the share of environmental patents increased
considerably in response to stricter environmental regulations. Similarly, Popp (2004) studied
the relationship between environmental regulation and new technology focusing on SC>2 and
NOX. The study was performed using patent data from the United States, Japan, and Germany.
Popp found that more stringent regulation enhanced domestic patenting by domestic inventors.
While regulation may influence the direction and intensity of emissions-related research
and development activities, "crowding out" of investment resources may occur as resources are
directed away from other opportunities, potentially leading to opportunity costs that offset
savings resulting from research and development successes (Popp and Newell 2012). In a study
that links energy-related patent activity and firm financial data, Popp and Newell (2012) find that
while increases in alternative energy patents result in fewer patents for other energy
technologies, this result is due to firm-level profit-maximizing behavior rather than constraints
on the magnitude of research and development resources. Alternatively, Kneller and Manderson
(2012) find evidence in the United Kingdom that environment-related research and development
resulting from more stringent regulation may crowd out other research and development
activities but that environment-related capital does not crowd out non-environmental capital.
Another factor to consider is the degree to which a particular sector is likely to be close to fully
controlled, e.g., in comparing existing emissions with uncontrolled emissions levels, is the
percent of control close to 100 percent? In those cases, achieving additional reductions through
technological change is likely to be more difficult and costly, because the benefits of investment
in those technologies is smaller, due to smaller remaining potential for abatement.
7.2.4 Learning by Doing
What is known as "learning by doing" or "learning curve impacts" has also made it
possible to achieve greater emissions reductions than had been feasible earlier, or reduce the
costs of emissions control relative to original estimates. Learning curve impacts can be defined
generally as the extent to which variable costs (of production and/or pollution control) decline as
firms gain experience with a specific technology. This type of change corresponds to case 4 in
7-20
-------
the discussion of the ways technological change can affect the MACC that appeared earlier. Such
impacts have been identified to occur in a number of studies conducted for various production
processes. These impacts would manifest themselves as a lowering of expected costs for
operation of technologies in the future below what they may otherwise have been. For example,
Rubin et al. 2004 show that capital costs of Flue gas desulphurization (FGD) and selective
catalytic reduction (SCR) systems have decreased over time as a result of research and
development activities and learning by doing, among other factors, and that failing to account for
these technological dynamics can lead to incorrect estimates of future regulatory costs.
Rubin et al. (2012) also note that when technologies succeed, costs tend to fall over time.
They offer the example of post-combustion SO2 and NOX combustion systems. After an increase
in costs during an initial commercialization period, costs decreased by at least 50 percent over
the course of two decades. Table 9.5 in the 1997 Ozone NAAQS RIA summarizes historical and
projected "progress ratios" for existing technologies. These ratios show declining costs over
time, due to learning by doing, economies of scale, reductions in O&M costs, and technological
improvements in manufacturing processes. There are other discrete examples, for example prices
of the catalyst used in operating SCR have dropped dramatically over time. From 1980 to 2005,
catalyst prices dropped by roughly 85 percent (Cichanowicz, 2010). This follows the "learning
curve," which finds that production and implementation costs decrease as learning and repetitive
use occurs. In addition, Table 3a.7 in Appendix 3a of the 2008 Ozone NAAQS RIA lists controls
applied to new source types (Case 3). For example, SCR is now applicable to the cement
manufacturing sector, and SNCR is now applicable to a large number of additional boiler source
categories. In some cases, more effective controls were determined to be applicable where in
past cases, less effective controls were applied. For example, for industrial and manufacturing
incinerators, where previously SNCR was the NOx control technology, SCR was applied in
2008, increasing the control efficiency from 45 percent to 90 percent.
The magnitude of learning curve impacts on pollution control costs has been estimated for
a variety of sectors as part of the cost analyses done for the Direct Cost Estimates Report for the
7-21
-------
Second EPA Section 812 Prospective Analysis of the Clean Air Act Amendments of 1990.114 In
the Report, learning curve adjustments were included for those sectors and technologies for
which learning curve data was available. For all technologies and industries, a default learning
rate of 10 percent was adopted based on SAB advice. No adjustments were used for on-road and
non-road controls. The 10 percent adjustment is a 10 percent cost reduction per doubling of
emission reductions. The literature supports a rate of up to 20 percent for many technologies
(Button and Thomas, 1984). The impact of this on costs in the Report was to reduce costs of
local controls in nonattainment areas by 9.9 percent in 2020.
A typical learning curve adjustment is to reduce either capital or operation and
maintenance costs by a certain percentage given a doubling of output from that sector or for that
technology. In other words, capital or operation and maintenance costs will be reduced by some
percentage for every doubling of output for the given sector or technology. In addition, learning
by doing may also lead to instances where existing control technologies are found to be
applicable to additional sources. This corresponds to Case 5 in the discussion in section 7.2.1.
For example, scrubber technologies applied to electric utilities have been adapted to apply to
industrial boilers. As a result of this increased applicability due to learning, potential abatement
has increased at a cost less than the cost threshold. For this RIA, however, we do not have the
necessary data to properly generate control costs that reflect learning curve impacts.
7.2.5 Using Regional NOx Offset Prices to Estimate Costs of Unknown Emissions Controls
In ozone nonattainment areas, new sources interested in locating in that area and existing
sources interested in expanding are required to offset any emissions increases. If those emissions
increases are NOX emissions, the source typically purchases NOX emission reduction credits
(ERCs), or offsets, from within that particular nonattainment area. Within nonattainment areas,
offset prices fluctuate because of changes in the available supply of offsets and changes in
demand for offsets. Offset supply increases when facilities shut down or when they make
114 Industrial Economics, Incorporated and E.H. Pechan and Associates, Direct Cost Estimates for the Clean Air Act
Second Section 812 Prospective Analysis: Final Report, prepared for U.S. EPA, Office of Air and Radiation,
February 2011. Available at http://www.epa.gov/cleanairactbenefits/febll/costfullreport.pdf.
7-22
-------
process or other changes that reduce emissions permanently. Offset demand depends on the
industrial base in a given area and fluctuates with changes in economic growth. For example, in
the San Joaquin Valley, in recent years offset prices have increased because of increased oil and
gas industry development.
We identified historical NOX offset prices in several nonattainment areas, including the
San Joaquin Valley and the South Coast in California, Houston, TX, and New York region. For
the San Joaquin Valley Air Pollution Control District, we collected information on NOX offset
prices using the California Air Resources Board's Emission Reduction Offset Transaction Cost
Summary Reports for 2002 through 2013115. For the South Coast Air Quality Management
District, we collected information on prices for perpetual NOX RECLAIM Trading Credit (RTC)
for 2003 through 2012 from the Listing of Trade Registrations116 Lastly, we collected
information on NOx offset prices in the Houston-Galveston nonattainment area for 2010 through
2013 from the Trade Report117 and the New York-New Jersey-Connecticut nonattainment area
from 2000 through 2013 from industry representatives.
Table 7-6 presents the price data we were able to collect for these four regions, adjusted
to 2011 dollars using the Gross Domestic Product Implicit Price Deflator. The offset prices in
this table are denominated in units of perpetual tons, or tons per year. The prices constitute
average of the trades in the regions for the year given.
115 http://www.arb.ca.gov/nsr/erco/erco.htm
116http://www.aqmd.gov/home/programs^usiness/about-reclaim/reclaim-trading-credits
117 http://www.tceq.state.tx.us/airquality/banking/mass_ect_prog.html
7-23
-------
Table 7-6. Average NOX Offset Prices for Four Areas (2011$);
Annualized NOx Offset Prices ($/ton)
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Average
Maximum
San Joaquin
Valley
N/A
N/A
36,000
28,000
25,000
25,000
21,000
21,000
48,000
58,000
62,000
64,000
47,000
42,000
40,000
64,000
California South
Coast
N/A
N/A
N/A
N/A
12,000
31,000
163,000
206,000
210,000
128,000
98,000
56,000
47,000
N/A
106,000
210,000
Houston TX
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
36,000
N/A
N/A
97,000
66,000
97,000
New York Region
25,000
12,000
12,000
12,000
12,000
11,000
11,000
N/A
N/A
N/A
N/A
N/A
N/A
4,000
12,000
25,000
a All values are rounded to two significant figures.
The data series for the California regions are more complete than those for Houston and the New
York region. We are working to obtain more complete data series for future analysis.
To more directly compare offset prices to potential annual costs for unidentified
emissions controls, we annualized the tons per year prices using the same engineering cost
equations as used in the main analysis to estimate annualized control cost. We converted the
offset cost to an annual costs by using the capital recovery factor (CRF) discussed in Section
7.1.1. In a capital cost context, the CRF incorporates the interest rate and lifetime of the
purchased capital. In this instance, although the offsets are perpetual in nature, we assumed a
lifetime of 20 years in order to make the cost basis more comparable to the control cost
estimates. Also, we used 7 percent for the interest rate. Table 7-7 presents the average and
maximum annualized NOX offset prices in 2011 dollars.
7-24
-------
Table 7-7. Annualized NOx Offset Prices for Four Areas (2011$)a
Annualized NOx Offset Prices ($/ton)
Average
Maximum
San Joaquin
Valley
$ 4,000
$ 6,000
California South
Coast
$ 10,000
$ 20,000
Houston TX
$ 6,000
$ 9,000
New York Region
$ 1,000
$ 2,000
a All values are rounded to two significant figures.
From an economic perspective, these offset prices may represent the shadow value of a
ton of emissions. It is possible that these offset prices could serve as reasonable proxies for the
costs associated with emissions reductions from unknown controls. The cost information
informing the known control strategy traces out an incomplete marginal abatement cost curve in
that, as discussed in Chapter 4, the controls used in the known control analysis are primarily end-
of-pipe technologies. The known control estimates for NOX do not account for other forms of
abatement, switching to lower emitting fuels or increasing energy efficiency, for example. The
estimates also do not account for institutional or market arrangements that allow firms to buy or
sell emissions offsets in nonattainment regions with emissions constraints. These voluntary
exchanges may enable abatement at lower costs than may otherwise be available. The benefit of
these market transaction data is that the prices are revealed by the interaction of offset supply and
demand in regions with differentiated characteristics and air quality profiles.
7.2.6 Conclusion
The preceding sections have discussed the ways in which various factors might affect the
observed marginal abatement costs and the resulting total abatement costs estimated in this RIA.
Based on past experience with Clean Air Act implementation, the EPA believes that it is
reasonable to anticipate that the marginal cost of emissions reductions will decline over time due
to technological improvements and more widespread adoption of previously considered niche
control technologies as well as the development of innovative strategies.118 As the EPA
continuously improves its data and tools, we expect to better characterize the currently
unobserved pieces of the MACC. As a result of our consideration of these complexities, we are
118 See Chapter 4, Section 4.5 for additional discussion of uncertainties associated with predicting technological
advancements that may occur between now and 2025.
7-25
-------
currently unable to quantitatively predict future shifts in the abatement supply curve because
many factors are intertwined and data are incomplete or highly uncertain.
7.3 Compliance Cost Estimates for Unknown Emissions Controls
This section presents the methodology and results for the costs of emissions reductions
from unidentified controls needed for attainment of the alternative ozone standards. As
discussed in Chapter 4, the application of the modeled control strategy was not successful in
reaching full nationwide attainment of the alternate ozone standards. Many areas remained in
nonattainment under all four alternate standard scenarios. Therefore, the engineering costs
detailed in Section 7.1 represent only the costs of partial attainment.
7.3.7 Methods
Prior to presenting the methodology for estimating costs for unspecified emission
reductions in this RIA, it is important to provide information from EPA's Science Advisory
Board Council Advisory,119 dated June 8, 2007, on the issue of estimating costs of unidentified
control measures:
572 Council Advisory, Direct Cost Report, Unidentified Measures (charge question 2.a)
"The Project Team has been unable to identify measures that yield sufficient emission
reductions to comply with the National Ambient Air Quality Standards (NAAQS) and
relies on unidentified pollution control measures to make up the difference. Emission
reductions attributed to unidentified measures appear to account for a large share of
emission reductions required for a few large metropolitan areas but a relatively small
share of emission reductions in other locations and nationwide.
"The Council agrees with the Project Team that there is little credibility and hence
limited value to assigning costs to these unidentified measures. It suggests taking great
care in reporting cost estimates in cases where unidentified measures account for a
significant share of emission reductions. At a minimum, the components of the total cost
associated with identified and unidentified measures should be clearly distinguished. In
some cases, it may be preferable to not quantify the costs of unidentified measures and to
119 U.S. Environmental Protection Agency. June 2007. Advisory Council on Clean Air Compliance
Analysis (COUNCIL), Council Advisory on OAR's Direct Cost Report and Uncertainty
Analysis Plan. Washington, DC.
7-26
-------
simply report the quantity and share of emissions reductions attributed to these
measures.
"When assigning costs to unidentified measures, the Council suggests that a simple,
transparent method that is sensitive to the degree of uncertainty about these costs is best.
Of the three approaches outlined, assuming a fixed cost/ton appears to be the simplest
and most straightforward. Uncertainty might be represented using alternative fixed costs
per ton of emissions avoided. "
While we have considered alternative methodologies to predict future abatement supply
curves, we are currently unable to quantitatively predict future shifts in the supply curve with
sufficient confidence to use in this RIA. For most NAAQS RIAs prepared during the past five
years, EPA estimated the costs for unidentified controls using a pair of methodologies: what we
termed a "fixed cost" approach, following the SAB advice, and a "hybrid" approach that has not
yet been reviewed by the SAB. We now refer to the fixed-cost approach as the "average cost"
approach because it more accurately characterizes the concepts underlying the approach. The
average cost methodology uses an assumed national average cost per ton for unidentified
controls needed for attainment, as well as two alternative assumed values employed for
sensitivity analysis. The range of estimates reflects different assumptions about the cost of
additional emissions reductions beyond those in the modeled control strategy. While we use a
constant cost per ton of emissions reduction to estimate the costs of the emissions reductions
beyond known controls, this does not imply that marginal costs are not increasing in needed
emissions abatement. Rather, the average cost per ton is intended to capture what might be the
total costs associated with the abatement of the emissions reductions from unknown controls.
The alternative estimates implicitly reflect different assumptions about the amount of
technological progress and innovation in emission reduction strategies. The average cost
methodology reflects a view that because no cost data exists for unspecified future strategies, it is
unclear whether approaches using hypothetical cost curves will be more accurate or less accurate
in forecasting total national costs of unspecified controls than an average-cost approach that uses
a range of national cost per ton values.
The hybrid cost methodology assumed increasing marginal costs of control along an
upward-sloping marginal cost curve. The hybrid cost methodology assumed the rate of increase
7-27
-------
in the marginal costs of abatement is proportional to the weighted ratio of the amount of
abatement using identified controls to the remaining needed abatement using unidentified
controls.120 Under this approach, the relative costs of unspecified controls in different
geographic areas reflected the expectation that average per-ton control costs are likely to be
higher in areas needing a higher ratio of emissions reductions from unspecified and known
controls. However, the weight, which reflected the anticipated degree of difficulty of achieving
needed emissions reductions, and the ratios that informed the slope of the marginal abatement
cost curve in previous NAAQS analyses were strong assumptions that have not been empirically
tested.
When used to estimate costs for end-of-pipe technologies, the hybrid methodology
assumed all emissions reductions come from the highest cost margin of the abatement supply
curve which, as explained in the previous section, is unlikely for much of the unobserved
abatement capacity in the present and future. For example, EPA's control strategy tools largely
focus on a limited set of emissions inventory sectors, whereas abatement opportunities exist in
other sectors. When new abatement opportunities are identified in other sectors, they typically
are not at the higher end of the cost curve.
For areas needing significant additional emission reductions, much pollution abatement is
likely needed from sources within regulated sectors that historically have not been intensively
regulated. However, if national standards become more stringent, new regions or firms will be
added to the regulated domain. These new entrants, with their relatively untapped abatement
supply, will contribute to an outward shift in abatement supply. The newly regulated regions and
firms will also face new incentives for technical change and innovation that may lower costs
over the long run by developing new, more efficient compliance strategies. Because the point of
departure for the hybrid approach cost curve is based on our current database, which includes
only existing controls, it will systematically overstate future costs if any cost-reducing
technological change occurs.
120 See, for example, Section 7.2 and Appendix &.A.2 is the December 2012 RIA for the final PM NAAQS,
available at http://www.epa.gov/ttn/ecas/regdata/RIAs/fmalria.pdf.
7-28
-------
As noted in previous NAAQS analyses, the EPA continues to explore other sources of
information to inform the estimates of extrapolated costs. For this RIA we examined the full set
of known controls, examined evidence that suggests that over time new information and data
emerges that shifts emissions reductions from the unknown to the known category, as well as
explored whether NOx offset prices can serve as reasonable proxies for the costs of emissions
reductions not identified by current tools.
Based upon deliberations informing this discussion, the EPA Council's advice, and the
requirements of E.O. 12866 and OMB circular A-4, which provides guidance on the estimation
of benefits and costs of regulations, in this RIA, we follow the Council recommendations by
using an average cost per ton as a central estimate and conduct sensitivity analysis using
alternative average costs to explore how sensitive total costs are to these assumptions. While the
average cost methodology has limitations, we agree with the Council that the approach is both
transparent and strikes a balance between the likelihood that some unidentified abatement would
arise at lower segments of the identified cost curve while other sources of abatement may come
at the higher cost margin.
While the known control analysis limited the application of controls with costs above
$14,000 per ton, we examined the full set of controls available for application in regions needing
emissions reductions. The MAC curves from this analysis are presented above in Section 7.1.
For NOx controls, a total of 1.22 million tons of reductions are available, and about 1.18 million
of these tons are available for less than $15,000 per ton. The known reductions available for less
than $15,000 per ton represent about 96 percent of the total known reductions in areas needing
emissions reductions to meet an alternative standard level. In addition, the average cost per ton
across all of these abatement opportunities is about $3,400 per ton. As a result, we decided to
use $15,000 per ton NOx as the main estimate for the extrapolated cost analysis. This assumed
cost is representative of higher cost controls available in the analysis. If, for example, the true
costs of the unidentified controls are distributed at the upper end of the identified control costs
depicted in Figure 7-1, $15,000 per ton may under-estimate the average value of the unidentified
abatement. Alternatively, the assumed value might overestimate the '"unobserved" abatement
discussed in Section 7.2. The results shown in Table 7-4 (per ton cost of NOx reductions for five
major NOx-reducing rules) and Table 7-7 (annualized per ton cost of NOs offsets in four
7-29
-------
regions) may suggest that the $15,000 per ton assumed average cost may be over-estimating the
average value of the unidentified abatement.
Because of this uncertainty, we use alternative assumptions of the average cost in the
Appendix, a first sensitivity analysis using an assumed cost of $10,000 per ton and a second
sensitivity analysis using an assumed $20,000 per ton.121 This range is inclusive of the
annualized NOx offset prices observed in recent years in the areas likely to need unknown
controls to achieve the proposed standard, and if anything, suggests the central estimate of
$15,000/ton is conservative. EPA requests comments on the methods presented to estimate
emission reductions needed beyond known controls including the parameter estimate of
$15,000/ton.
Because cost changes due to technological change will be available on a national-level, it
makes sense to use national-level average cost per ton in the primary analysis. However, as
indicated by the variation in NOx offset prices across regions shown in Table 7-6, regional
factors may play a significant role in the estimation of control costs. As a result, the EPA will
continue to explore alternative methodologies and sources of regional information that may make
the average cost methodology more regionally specific for the RIA for the final rule.
7.3.2 Unknown Compliance Cost Estimates
Table 7-8 presents the extrapolated control cost estimates for the East and West in 2025,
except for California for the alternative standards using an assumed average cost of $15,000/ton.
Values of $10,000/ton and $20,000/ton are used for the sensitivity analyses found in Appendix
7.2.
121 As shown in Section 7.1, we also performed a similar analysis for VOC controls, which indicated that about 52
percent of VOC controls available in the analysis for less than $15,000 per ton, with an average of about $12,000
per ton. While a limited amount of extrapolated VOC emissions reductions were needed for the 60 ppb alternative
level for the East (41,000 tons), we decided to use the same $15,000 per ton (with $10,000 and $20,000 per ton for
the sensitivity analysis) for VOC controls for simplicity.
7-30
-------
Table 7-8. Extrapolated Control Costs in 2025 by Alternative Standard for 2025 - U.S.,
except California (millions of 2011$)
Alternative Level
70 ppb
65 ppb
60 ppb
Geographic Area
East
West
Total
East
West
Total
East
West
Total
Extrapolated Cost
2,300
-
2,300
11,000
-
11,000
28,000
5,200
34,000
' All values are rounded to two significant figures. Extrapolated costs are based on the average-cost methodology
using a $15,000/ton assumed average cost.
Table 7-9 presents the extrapolated control cost estimates for post-2025 for California
across the alternative standards using an assumed average cost of $15,000/ton.
Table 7-9. Extrapolated Control Costs in 2025 by Alternative Standard for Post-2025 ~
California (millions of 2011$)
Alternative Level Geographic Area Extrapolated Costs
70 ppb California 800
65 ppb California 1,600
60 ppb California 2,200
a All values are rounded to two significant figures. Extrapolated costs are based on the average-cost methodology
using a $15,000/ton assumed average cost.
7.4 Total Compliance Cost Estimates
As discussed throughout this RIA, we present the primary costs and benefits estimates for
2025. We assume that potential nonattainment areas everywhere in the U.S., excluding
California, will be designated such that they are required to reach attainment by 2025, and we
developed our projected baselines for emissions, air quality, and populations for 2025.
In estimating the incremental costs and benefits of potential alternative standards, we
recognize that there are several areas that are not required to meet the existing ozone standard by
2025. The Clean Air Act allows areas with more significant air quality problems to take
additional time to reach the existing standard. Several areas in California are not required to
7-31
-------
meet the existing standard by 2025 and may not be required to meet a revised standard until
sometime between 2032 and December 31, 2037.122 We were not able to project emissions and
air quality beyond 2025 for California, however, we adjusted baseline air quality to reflect
mobile source emissions reductions for California that would occur between 2025 and 2030;
these emissions reductions were the result of mobile source regulations expected to be fully
implemented by 2030. While there is uncertainty about the precise timing of emissions
reductions and related costs for California, we assume costs occur through the end of 2037 and
beginning of 2038. In addition, we model benefits for California using projected population
demographics for 2038.
Because of the different timing for incurring costs and accruing benefits and for ease of
discussion throughout the analyses, we refer to the different time periods for potential attainment
as 2025 and post-2025 to reflect that (1) we did not project emissions and air quality for any year
other than 2025; (2) for California, emissions controls and associated costs are assumed to occur
through the end of 2037 and beginning of 2038; and (3) for California benefits are modeled using
population demographics in 2038. It is not straightforward to discount the post-2025 results for
California to compare with or add to the 2025 results for the rest of the U.S. While we estimate
benefits using 2038 information, we do not have good information on precisely when the costs of
controls will be incurred. Because of these differences in timing related to California attaining a
revised standard, the separate costs and benefits estimates for post-2025 should not be added to
the primary estimates for 2025.
Tables 7-9 and 7-10 present summaries of the total national annual costs (known and
extrapolated) of attaining the alternative standards of 70, 65, and 60 ppb. To calculate total cost
estimates at a 3 percent discount rate and to include the extrapolated costs in those totals, we
added the known control estimates at a 3 percent discount rate, where available, to the known
control estimates at a 7 percent discount rate where the costs could not be determined for the 3
percent rate; we added these to the extrapolated costs at a 7 percent discount rate. Table 7-10
122 The EPA will likely finalize designations for a revised ozone NAAQS in late 2017. Depending on the precise
timing of the effective date of those designations, nonattainment areas classified as Severe 15 will likely have to
attain sometime between late 2032 and early 2033 and nonattainment areas classified as Extreme will likely have
to attain by December 31, 203.
7-32
-------
presents the total national annual costs by alternative standard for 2025 for all of the U.S., except
California. Table 7-11 presents the total national annual costs by alternative standard for post-
2025 for California.
Table 7-10. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S., except California (millions of 2011$, 7% Discount Rate)"
~ , . . Total Control Costs
... T , Geographic Area ,T, ,
Alternative Level (Known and
Extrapolated)
70 ppb
65 ppb
60 ppb
East
West
Total
East
West
Total
East
West
Total
3,900
-
$3,900
15,000
400
$15,000
33,000
5,800
$39,000
' All values are rounded to two significant figures. Extrapolated costs are based on the fixed-cost methodology.
Table 7-11. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for post-2025 - California (millions of 2011$, 7% Discount Rate)"
Total Control Costs
Alternative Level Geographic Area (Known and
Extrapolated)
70 ppb California 800
65 ppb California 1,600
60 ppb California 2,200
a All values are rounded to two significant figures. Extrapolated costs are based on the fixed-cost methodology.
7.5 Updated Methodology Presented in this RIA
The cost analysis presented in this chapter incorporates an array of methodological and
technical updates that the EPA has adopted since the previous review of the ozone standards in
2008 and proposed reconsideration in 2010. The updates to models, methods, and data are too
numerous to be able to quantitatively estimate the impact of any of the updates individually.
Therefore, we present the major updates below qualitatively. Many of these changes reflect
updates to inputs to the cost analysis, but are discussed here for completeness. Below we note the
aspects of this analysis that differ from the 2008 RIA as well as the 2010 reconsideration RIA
7-33
-------
(U.S. EPA, 2010). A few overarching changes that are worth mentioning: the incremental costs
and benefits for this analysis are measured from a baseline of the current 75 ppb ozone standard;
in the previous analysis the baseline was the 84 ppb ozone standard. Also, the currency year was
updated from 2006$ to 2011$.
Emissions and Air Quality Updates
The base year emissions for this analysis are 2011, and the future analysis year is 2025
(previously the base year was 2002 and the future analysis year was 2020). Key changes in the
emission estimates include: increased accuracy of stationary source emissions estimates, updates
in models and Annual Energy Outlook (AEO) projections for mobile sources and EGUs,
inclusion of oil and gas sector emissions and numerous updates to the nonpoint emissions. In
addition, from 2002 to 2011 there have been a number of changes in the energy system, cleaner
mobile sources, and economic changes that have resulted in reduced emissions since the last
ozone analysis. There are additional federal control programs included in the emissions
projections to 2025, including Tier 3, MATS, and the Clean Power Plan.
Air quality monitor design values were updated to reflect 2009 through 2013 air quality
in contrast the 2008 and 2010 analyses which used 2000 through 2004 air quality. As shown in
Chapter 2 Figure 2-2, ozone design values are generally decreasing over time. For example, the
figure shows that for the period of years from 2000 through 2004 the 90th percentile
concentrations ranged from 85.9 to 102.8 ppb; the 75th percentile values for the same period
ranged from 78.8 to 87.3 ppb. In contrast, the design values for the period of 2009 through 2013
the 90th percentile concentrations ranged from 77 to 86.2 ppb and the 75th percentile
concentrations ranged from 70.8 to 81 ppb. These design values are the basis for future year air
quality projections. A quick comparison across analyses reveals that in the 2008 analysis 89
counties were projected to exceed the higher end of the proposed range of 70ppb in the future
year (2020), while only 9 counties are projected to exceed 70 ppb in this analysis for 2025. The
same is true for the lower end of the proposed range, in 2008 231 counties were projected to
exceed 65 ppb and that number has decreased to 68 counties in this analysis.
As emissions decrease in the base year and air quality design values decrease, the result is
a smaller incremental change in air quality needed to meet the revised standards as compared to
7-34
-------
the 2008 and 2010 analyses. The smaller increment needed to achieve attainment also means
that fewer controls are needed across a reduced number of geographic areas
Control Strategy Updates
Many improvements were made in the non-EGU point control measures used in this
analysis that make the data more accurate and defensible. These changes include: removal of
incorrect links between control measures and SCCs; updates to cost equations where more
recent data was available to improve their accuracy; inclusion of information that has recently
become available concerning known and emerging technologies for reducing NOx emissions;
and revising costs and control efficiencies from control measures in the dataset based on recently
obtained information from industry and multi-jurisdictional organizations (e.g., Ozone Transport
Commission and Lake Michigan Air Directors Consortium). In addition, a different mix of
known control measures across EGU, mobile, nonpoint and non-EGU point were applied due to
the recently promulgated rules mentioned above.
Costing methodology updates have occurred since previous ozone analyses, the 'hybrid'
approach was not utilized to estimate costs of emission reductions needed beyond known
controls. Holding cost methodology constant (average/fixed cost approach), fewer emissions
reductions were needed beyond known controls in this analysis due to the increased application
of known control measures mentioned above.
The above mentioned technical and methodological changes resulted in the lower cost
estimates in this analysis. The most influential factors on the cost analysis were the lower
number of exceeding counties and increased data accuracy of the known control measures. The
effect of fewer exceeding counties is that fewer emissions reductions are needed, and therefore
the costs are lower. The effect of additional and lower-cost known control measures being
applied is a lowering of the emissions reductions needed beyond known controls, which are
typically the higher cost emissions reductions.
7-35
-------
7.6 Economic Impacts
7.6.1 Introduction
This section addresses the potential economic impacts of the illustrative control strategies
for the potential alternative ozone standards. The control costs are uncertain for several reasons.
The controls that the states ultimately choose to implement will likely differ from the illustrative
control strategies for which costs are estimated in earlier sections of this chapter. The flexibility
afforded to states by the Clean Air Act also allows them to adopt programs that include design
elements that may mitigate or promote particular economic impacts based on their individual
priorities. The cost estimates become more uncertain because of the length of time before they
will be implemented. By the 2025 and post-2025 time frames, changes in technology, changes in
implemented regulations, and changes in relative prices will all add to the uncertainty in the cost
analysis. Finally, the portion of costs that is extrapolated is not allocated to particular sectors.
Economic impacts focus on the behavioral response to the costs imposed by a policy being
analyzed. The responses typically analyzed are market changes in prices, quantities produced
and purchased, changes in international trade, changes in profitability, facility closures, and
employment. Often, these behavioral changes are used to estimate social costs if there is
indication that the social costs differ from the estimate of control costs because behavioral
change results in other ways of meeting the requirements (e.g., facilities choosing to reduce
emissions by producing less rather than adding pollution control devices).
The potential alternative ozone standards are anticipated to impact multiple markets in
many times and places. Computable General Equilibrium (CGE) models are designed to address
such problems. To support the Final Ozone NAAQS of March 2008 (Final Ozone NAAQS
Regulatory Impact Analysis), among other rulemakings, the EPA used the Economic Model for
Policy Analysis (EMPAX) to estimate the market impacts of the portion of the cost that was
associated with the application of known controls (excluding the extrapolated costs). EMPAX is
a dynamic computable general equilibrium (CGE) model that forecasts a new equilibrium for the
entire economy after a policy intervention. While the external Council on Clean Air Compliance
Analysis (Council) peer review of The Benefits and Costs of the Clean Air Act from 1990 to
2020(Hammitt 2010) stated that inclusion of benefits in an economy-wide model, specifically
7-36
-------
adapted for use in that study, "represented] a significant step forward in benefit-cost analysis,"
EPA recognizes that serious technical challenges remain when attempting to evaluate the
benefits and costs of potential regulatory actions using economy-wide models. Consistent with
the Council's advice regarding the importance of including benefit-side effects demonstrated by
the Benefits and Costs of the Clean Air Act from 1990 to 2020, and the lack of available multi-
year air quality projections needed to include these benefit-side effects, EPA has not conducted
CGE modeling for this analysis.
However, the EPA recognizes that serious technical challenges remain when attempting to
evaluate the impacts of potential regulatory actions using economy-wide models. The EPA is
therefore establishing a new Science Advisory Board (SAB) panel on economy-wide modeling
to consider the technical merits and challenges of using this analytical tool to evaluate costs,
benefits, and economic impacts in regulatory development. The EPA will use the
recommendations and advice of this SAB panel as an input into its process for improving
benefit-cost and economic impact analyses that are used to inform decision-making at the
Agency. The panel will also be asked to identify potential paths forward for improvements that
could address the challenges posed when economy-wide models are used to evaluate the effects
of regulations.
The advice from the SAB panel formed specifically to address the subject of economy-
wide modeling will not be available in time for this analysis. Given the ongoing SAB panel on
economy-wide modeling, and the uncertain nature of costs, this section proceeds with a
qualitative discussion of market impacts.
7.6.2 Summary of Market Impacts
Consider an added cost to produce a good associated with the pollution control required to
reach the alternative ozone standards. Such a good is either one developed for the consumer
(called a consumption good), or one used in the production of other goods for consumption
(called an intermediate good). Some goods are both consumption and intermediate goods. First,
consider the direct impact on the market facing the increased cost. In this case for the market
facing the increased cost, the price will go up and the amount sold will go down. The magnitude
of these shifts depends on a number of factors. The greater the unit cost increase relative to the
7-37
-------
price of the good the greater will be the changes. The more responsive a consumer is to a change
in the price of a consumption good or the more responsive a purchase of an intermediate good is
the greater will be the changes. For the alternative ozone standards, many goods will have direct
changes in costs of production. This makes the assumption of isolated markets too simple. With
multiple intermediate goods affected, then the intermediate goods and consumption goods they
are used to produce are affected. As fewer intermediate goods and consumption goods are
purchased at a higher price, other intermediate goods and consumption goods that serve as
substitutes become more attractive and more are sold at a higher price. All of these market
changes lead to changes in income, which can lead to changes in purchases of consumption
goods. Quantities of intermediate goods used to reduce emissions would also change.
Considering all of these changes, it is not possible to qualitatively conclude the direction of price
and quantity changes for any single market. Any conclusions about changes in international
trade, profits, closures, or social cost is impossible in a qualitative analysis.
7.7 Uncertainties and Limitations
The EPA acknowledges several important limitations of this analysis, which include the
following:
Boundary of the cost analysis: In this engineering cost analysis we include only the impacts to
the regulated industry, such as the costs for purchase, installation, operation, and maintenance of
control equipment over the lifetime of the equipment. As mentioned above, recordkeeping,
reporting, testing and monitoring costs are not included. In some cases, costs are estimated for
changes to a process such as switching from one fuel to another less polluting fuel. Additional
profit or income may be generated by industries supplying the regulated industry, especially for
control equipment manufacturers, distributors, or service providers. These types of secondary
impacts are not included in this engineering cost analysis.
Cost and effectiveness of control measures: Our application of control measures reflect
average retrofit factors and equipment lives that are applied on a national scale. We do not
account for regional or local variation in capital and annual cost items such as energy, labor,
materials, and others. Our estimates of control measure costs may over- or under-estimate the
costs depending on how the difficulty of actual retrofitting and equipment life compares with our
7-38
-------
control assumptions. In addition, our estimates of control efficiencies for the known controls
assume that the control devices are properly installed and maintained. There is also variability in
scale of application that is difficult to reflect for small area sources of emissions.
Discount rate: Because we obtain control cost data from many sources, we are not always able
to obtain consistent data across original data sources. If disaggregated control cost data are
unavailable (i.e., where capital, equipment life value, and operation and maintenance [O&M]
costs are not separated out), the EPA typically assumes that the estimated control costs are
annualized using a 7 percent discount rate. When disaggregated control cost data are available
(i.e., where capital, equipment life value, and O&M costs are explicit), we can recalculate costs
using a 3 percent discount rate. In general, we have some disaggregated data available for non-
EGU point source controls, and we do not have any disaggregated control cost data for area
source controls. In addition, these discount rates are consistent with OMB guidance, but the
actual real discount rates may vary regionally or locally.
Known control costs: We estimate that there is an accuracy range of+/- 30 percent for non-
EGU point source control costs. This level of accuracy is described in the EPA Air Pollution
Control Cost Manual, which is a basis for the estimation of non-EGU control cost estimates
included in this RIA. This level of accuracy is consistent with either the budget or bid/tender-
level of cost estimation as defined by the AACE International.123 The accuracy for nonpoint
control costs estimates has not been determined, but it is likely no more accurate than those for
non-EGU point source control costs.
Differences between ex ante and ex post compliance cost estimates: In comparing regulatory
cost estimates before and after regulation, ex ante cost estimate predictions often overestimate or
underestimate costs. Harrington et al. (2000) surveyed the predicted and actual costs of 28
federal and state rules, including 21 issued by the U.S. Environmental Protection Agency and the
Occupational Safety and Health Administration (OSHA). In 14 of the 28 rules, predicted total
costs were overestimated, while analysts underestimated costs in three of the remaining rules. In
123 AACE International. Recommended Practice No. 18R-97. Cost Estimate Classification System - As Applied in
Engineering, Procurement, and Construction For the Process Industries. Revised on November 29, 2011.
Available at http://www.aacei.org/non/rps/18R-97.pdf.
7-39
-------
EPA rules where per-unit costs were specifically evaluated, costs of regulations were
overestimated in five cases, underestimated in four cases, and accurately estimated in four cases
(Harrington et al. 2000). The collection of literature regarding the accuracy of cost estimates
seems to reflect these splits. The "Retrospective Study of the Costs of EPA Regulations" found
that several of the case studies124 suggested that cost estimates were over-estimated. However,
the EPA stated in the report that the small number of regulatory actions covered and data and
analytical challenges associated with the case studies limited the certainty of this conclusion.
Costs of unknown controls (extrapolated costs): In addition to the application of known
controls, the EPA assumes the application of unidentified future controls that make possible the
additional emissions reductions needed beyond known controls for attainment in the projection
year for this analysis.
7.8 References
Brunnermeier, S.B., Cohen, M.A., 2003. Determinants of environmental innovation in US manufacturing industries.
Journal of Environmental Economics and Management 45, 278-293.
Button, J. M, Thomas, A., 1984. Treating Progress Functions as a Managerial Opportunity. Academy of
Management Review 9(2), 235-247.
Gerard, D., Lave, L.B., 2005. Implementing technology-forcing policies: The 1970 Clean Air Act Amendments and
the introduction of advanced automotive emissions controls in the United States. Technological Forecasting and
Social Change 72, 761-778.
Hammitt, J.K.,2010. Review of the final integrated report for the second section 812 prospective study of the
benefits and costs of the clean air act. Available at
http://yosemite.epa.gov/sab/sabproduct.nsf/9288428b8eeea4c885257242006935a3/lE6218DE3BFF682E852577F
B005D46F l/$File/EPA-COUNCIL-11 -001 -unsigned.pdf.
Harrington, W., Morgenstern, R.D., Nelson, P., 2000. On the accuracy of regulatory cost estimates. Journal of
Policy Analysis and Management 19, 297-322.
Jaffe, A., Newell, R., Stavins, R., 2002. Environmental Policy and Technological Change. Environmental and Resource
Economics 22, 41-70.
Kneller, R., Manderson, E., 2012. Environmental regulations and innovation activity in UK manufacturing
industries. Resource and Energy Economics 34, 211-235.
124 The four case studies in the 2014 Retrospective Study of the Costs of EPA Regulations examine five EPA
regulations: the 2001/2004 National Emission Standards for Hazardous Air Pollutants and Effluent Limitations
Guidelines, Pretreatment Standards, and New Source Performance Standards on the Pulp and Paper Industry;
Critical Use Exemptions for Use of Methyl Bromide for Growing Open Field Fresh Strawberries in California for
the 2004-2008 Seasons; the 2001 National Primary Drinking Water Regulations for Arsenic; and the 1998
Locomotive Emission Standards.
7-40
-------
Levinson, A. Technology, 2009. International Trade, and Pollution from US Manufacturing. American Economic
Review 99, 2177-2192.
Popp, D., 2006. International innovation and diffusion of air pollution control technologies: the effects of NOX and
SCh regulation in the US, Japan, and Germany. Journal of Environmental Economics and Management 51, 46-71.
Popp, D., Newell, R., 2012. Where does energy R&D come from? Examining crowding out from energy
R&D. Energy Economics 34, 980-991.
Rubin, E.S., Yeh, S., Hounshell, D.A., Taylor, M.R., 2004. Experience curves for power plant emission control
technologies. International Journal of Energy Technology and Policy 2, 52-69.
Sargent & Lundy, L.L.C. 2013. IPM Model - Updates to Cost and Performance for APC Technologies, SCR Cost
Development Methodology. Chicago, IL. Available at http://www.epa.gov/airmarkets/progsregs/epa-
ipm/docs/v513/attachment5_3.pdf.U.S. Environmental Protection Agency (U.S. EPA). 1997. Regulatory Impact
Analyses for the Paniculate Matter and Ozone National Ambient Air Quality Standards and Proposed Regional
Haze Rule. Office of Air Quality Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/ttn/oarpg/naaqsfin/ria.html.
Sterner, T., Turnheim, B., 2009. Innovation and diffusion of environmental technology: Industrial NOx abatement in
Sweden under refunded emission payments. Ecological Economics 68, 2996-3006.
U.S. Environmental Protection Agency (U.S. EPA). 2003. EPA Air Pollution Control Cost Manual. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. Available at
http://epa.gov/ttn/catc/products.html#cccinfo.
U.S. Environmental Protection Agency (U.S. EPA). 2008. Final Ozone NAAQS Regulatory Impact Analysis. Office
of Air Quality Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/ttn/ecas/regdata/RIAs/452_R_08_003.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2010. Regulatory Impact Analysis (RIA) for the Proposed
Reconsideration of the ozone National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning
and Standards, Research Triangle Park, NC. January. Available at .
U.S. Environmental Protection Agency (U.S. EPA). 2013. Documentation for EPA Base Case v.5.13 Using the
Integrated Planning Model. Office of Air Atmospheric Programs, Washington, DC. Available at
http://www. epa. gov/powersectormodeling/BaseCasev513. html.
U.S. Environmental Protection Agency (U.S. EPA). 2014a. Control Strategy Tool (CoST) Documentation Report.
Office of Air Quality Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/ttnecasl/cost.htm.
U.S. Environmental Protection Agency (U.S. EPA). 2014b. Guidelines for Preparing Economic Analyses, Chapter 6.
Available at http://yosemite.epa.gov/ee/epa/eed.nsf/webpages/Guidelines.htmMdownload.
U.S. Environmental Protection Agency (U.S. EPA). 2014c. Retrospective Study of the Costs of EPA Regulations: A
Report of Four Case Studies. Available at http://yosemite.epa.gov/ee/epa/eerm.nsf/vwAN/EE-0575.pdf/$file/EE-
0575.pdf.
7-41
-------
APPENDIX 7A: ENGINEERING COST ANALYSIS
Overview
Chapter 7 describes the engineering cost analysis approach that EPA used in applying to
demonstrate attainment of alternative ozone standard levels of 70, 65, and 60 ppb. This
Appendix contains more detailed information about the control costs of the known control
strategy analyses by control measure as well as sensitivity analyses for the fixed cost approach
used to estimate costs for the unknown emissions controls.
7A.I Cost of Known Controls in Alternative Standards Analyses
This section presents costs of known controls for the alternative standards analyses. Costs
are in terms of 2011 dollars and include values for all portions of the U.S. that were part of the
analyses. However, because all available known controls for California were applied as part of
the baseline analysis, no known controls were available for the alternative standards analyses in
California so these costs do not include any known control costs for California. Costs for the
alternative standard analyses are incremental to the baseline attainment demonstration. Tables
7A-1 and 7A-2 present the Costs for known controls by measure for the 70ppb alternative
standard analysis for NOx and VOC respectively. Tables 7A-3 and 7A-4 present the Costs for
known controls by measure for the 65ppb alternative standard analysis, and Tables 7A-5 and 7A-
6 present the costs for known controls by measure for the 60 ppb alternative standard analysis.
Table 7A-1. Costs for Known NOx Controls in the 70 ppb Analysis (millions of 2011$)a
NOx Control Measure Cost
Adjust Air to Fuel Ratio and Ignition Retard - Gas Fired 1C Engines 0.35
Biosolid Injection Technology - Cement Kilns 0.42
Ignition Retard - 1C Engines 0.037
Low Emission Combustion - Gas Fired Lean Burn 1C Engines 1
Low NOx Burner - Commercial/Institutional Boilers & 1C Engines 6.6
Low NOx Burner - Industr/Commercial/Institutional (ICI) Boilers 1.4
Low NOx Burner - Industrial Combustion 0.37
Low NOx Burner - Lime Kilns 0.28
Low NOx Burner - Natural Gas-Fired Turbines 8.4
Low NOx Burner - Residential Furnaces 5.2
Low NOx Burner - Residential Water Heaters & Space Heaters 13
Low NOx Burner and Flue Gas Recirculation - Coke Oven/Blast Furnace 0.13
Low NOx Burner and Flue Gas Recirculation - Iron & Steel Mills - Reheating 0.026
Low NOx Burner and SCR - Coal-Fired ICI Boilers 4.1
Low NOx Burner and SCR - Industr/Commercial/Institutional Boilers 27
7A-1
-------
NOx Control Measure Cost
Natural Gas Reburn - Natural Gas-Fired EGU Boilers 0
Nonroad Diesel Retrofits & Engine Rebuilds - e.g., Construction Equipment 29
Non-Selective Catalytic Reduction (NSCR) - 4 Cycle Rich Burn 1C Engines 10
OXY-Firing - Glass Manufacturing 20
Selective Catalytic Reduction (SCR) - Cement Kilns 16
Selective Catalytic Reduction (SCR) - Fluid Catalytic Cracking Units 4
Selective Catalytic Reduction (SCR) - 1C Engines, Diesel 5.5
Selective Catalytic Reduction (SCR) - ICI Boilers 8.5
Selective Catalytic Reduction (SCR) - Industrial Incinerators 4
Selective Catalytic Reduction (SCR) - Petroleum Refinery Gas-Fired Process Heaters 3.1
Selective Catalytic Reduction (SCR) - Sludge Incineration 1.4
Selective Non-Catalytic Reduction (SNCR) - Comm/Inst. Incinerators 0.029
Selective Non-Catalytic Reduction (SNCR) - Industrial Incinerators 0.24
Selective Non-Catalytic Reduction (SNCR) - Utility Boilers 0.43
a All values are rounded to two significant figures.
Table 7A-2. Costs for Known VOC Controls in the 70 ppb Analysis (millions of 2011$) a
VOC Control Measure Cost
Control of Fugitive Releases - Oil & Natural Gas Production 0.044
Control Technology Guidelines - Wood Furniture Surface Coating 1.9
Flare - Petroleum Flare 0.36
Gas Recovery - Municipal Solid Waste Landfill 0.27
Improved Work Practices, Material Substitution, Add-On Controls - Printing 0.4
Incineration - Other 0.84
Incineration - Surface Coating 200
Low-VOC Coatings and Add-On Controls - Surface Coating 0.04
LP V Relief Valve - Underground Tanks 7.1
Permanent Total Enclosure (PTE) - Surface Coating 3.5
RACT - Graphic Arts 20
Reduced Solvent Utilization - Surface Coating 2.7
Reformulation - Architectural Coatings 140
Reformulation - Industrial Adhesives 6.4
Reformulation-Process Modification - Automobile Refinishing 38
Reformulation-Process Modification - Cold Cleaning 3.3
Reformulation-Process Modification - Cutback Asphalt 0.02
Reformulation-Process Modification - Surface Coating 13
Solvent Recovery System - Printing/Publishing 0.038
Wastewater Treatment Controls- POTWs 0.73
a All values are rounded to two significant figures.
7A-2
-------
Table 7A-3. Costs for Known NOX Controls in the 65 ppb Analysis (millions of 2011$) a
NOx Control Measure Cost
Adjust Air to Fuel Ratio and Ignition Retard - Gas Fired 1C Engines 12
Biosolid Injection Technology - Cement Kilns 2.7
Ignition Retard - 1C Engines 0.96
Low Emission Combustion - Gas Fired Lean Burn 1C Engines 110
Low NOx Burner - Coal Cleaning 0.77
Low NOx Burner - Commercial/Institutional Boilers & 1C Engines 40
Low NOx Burner - Industr/Commercial/Institutional (ICI) Boilers 41
Low NOx Burner - Industrial Combustion 3.3
Low NOx Burner - Lime Kilns 4.8
Low NOx Burner - Miscellaneous Sources 0.058
Low NOx Burner - Natural Gas-Fired Turbines 44
Low NOx Burner - Residential Furnaces 16
Low NOx Burner - Residential Water Heaters & Space Heaters 110
Low NOx Burner - Steel Foundry Furnaces 0.27
Low NOx Burner - Surface Coating Ovens 0.092
Low NOx Burner and Flue Gas Recirculation - Coke Oven/Blast Furnace 2.2
Low NOx Burner and Flue Gas Recirculation - Iron & Steel Mills - Reheating 0.55
Low NOx Burner and Over Fire Air - Utility Boilers 1.6
Low NOx Burner and SCR - Coal-Fired ICI Boilers 87
Low NOx Burner and SCR - Industr/Commercial/Institutional Boilers 210
Low NOx Burner and SNCR - Industr/Commercial/Institutional Boilers 3.6
Natural Gas Reburn - Natural Gas-Fired ECU Boilers 1.2
Nonroad Diesel Retrofits & Engine Rebuilds - e.g., Construction Equipment 58
Non-Selective Catalytic Reduction (NSCR) - 4 Cycle Rich Burn 1C Engines 170
Non-Selective Catalytic Reduction (NSCR) - Nitric Acid Mfg 1.4
OXY-Firing - Glass Manufacturing 140
SCR and Flue Gas Recirculation - Fluid Catalytic Cracking Units 0.91
SCR and Flue Gas Recirculation - ICI Boilers 4.7
SCR and Flue Gas Recirculation - Process Heaters 2.9
Selective Catalytic Reduction (SCR) - Ammonia Mfg 15
Selective Catalytic Reduction (SCR) - Cement Kilns 230
Selective Catalytic Reduction (SCR) - Fluid Catalytic Cracking Units 19
Selective Catalytic Reduction (SCR) - 1C Engines, Diesel 24
Selective Catalytic Reduction (SCR) - ICI Boilers 76
Selective Catalytic Reduction (SCR) - Industrial Combustion 24
Selective Catalytic Reduction (SCR) - Industrial Incinerators 5.6
Selective Catalytic Reduction (SCR) - Iron Ore Processing 1.3
Selective Catalytic Reduction (SCR) - Petroleum Refinery Gas-Fired Process Heaters 61
Selective Catalytic Reduction (SCR) - Process Heaters 0.26
Selective Catalytic Reduction (SCR) - Sludge Incineration 39
Selective Catalytic Reduction (SCR) - Space Heaters 1.3
Selective Catalytic Reduction (SCR) - Utility Boilers 1,700
Selective Non-Catalytic Reduction (SNCR) - Coke Mfg 6.4
Selective Non-Catalytic Reduction (SNCR) - Comm/Inst. Incinerators 2.9
Selective Non-Catalytic Reduction (SNCR) - ICI Boilers 0.75
Selective Non-Catalytic Reduction (SNCR) - Industrial Combustion 0.065
Selective Non-Catalytic Reduction (SNCR) - Industrial Incinerators 2.8
7A-3
-------
NOx Control Measure Cost
Selective Non-Catalytic Reduction (SNCR) - Miscellaneous 0.25
Selective Non-Catalytic Reduction (SNCR) - Municipal Waste Combustors 2.5
Selective Non-Catalytic Reduction (SNCR) - Utility Boilers 0.46
Ultra-Low NOx Burner - Process Heaters 1.8
a All values are rounded to two significant figures.
Table 7A-4. Costs for Known VOC Controls in the 65 ppb Analysis (millions of 2011$) a
VOC Control Measure Cost
Control of Fugitive Releases - Oil & Natural Gas Production 0.083
Control Technology Guidelines - Wood Furniture Surface Coating 3
Flare - Petroleum Flare 0.36
Gas Recovery - Municipal Solid Waste Landfill 0.37
Improved Work Practices, Material Substitution, Add-On Controls - Printing 1.4
Incineration - Other 0.91
Incineration - Surface Coating 370
Low VOC Adhesives and Improved Application Methods - Industrial Adhesives 0.06
Low-VOC Coatings and Add-On Controls - Surface Coating 0.57
LP V Relief Valve - Underground Tanks 13
Permanent Total Enclosure (PTE) - Surface Coating 17
Petroleum and Solvent Evaporation - Surface Coating Operations 0.037
RACT - Graphic Arts 38
Reduced Solvent Utilization - Surface Coating 3.1
Reformulation - Architectural Coatings 300
Reformulation - Industrial Adhesives 6.1
Reformulation-Process Modification - Automobile Refinishing 62
Reformulation-Process Modification - Cutback Asphalt 0.075
Reformulation-Process Modification - Surface Coating 24
Solvent Recovery System - Printing/Publishing 1.3
Solvent Substitution and Improved Application Methods - Fiberglass Boat Mfg 0.059
Wastewater Treatment Controls- POTWs 0.82
a All values are rounded to two significant figures.
7A-4
-------
Table 7A-5. Costs for Known NOx Controls in the 60 ppb Analysis (millions of 2011$) a
NOx Control Measure Cost
Adjust Air to Fuel Ratio and Ignition Retard - Gas Fired 1C Engines 12
Biosolid Injection Technology - Cement Kilns 2.7
Ignition Retard - 1C Engines 1
Low Emission Combustion - Gas Fired Lean Burn 1C Engines 120
Low NOx Burner - Coal Cleaning 0.77
Low NOx Burner - Commercial/Institutional Boilers & 1C Engines 40
Low NOx Burner - Industr/Commercial/Institutional (ICI) Boilers 42
Low NOx Burner - Industrial Combustion 3.3
Low NOx Burner - Lime Kilns 4.8
Low NOx Burner - Miscellaneous Sources 0.077
Low NOx Burner - Natural Gas-Fired Turbines 47
Low NOx Burner - Residential Furnaces 17
Low NOx Burner - Residential Water Heaters & Space Heaters 120
Low NOx Burner - Steel Foundry Furnaces 0.27
Low NOx Burner - Surface Coating Ovens 0.092
Low NOx Burner and Flue Gas Recirculation - Coke Oven/Blast Furnace 2.2
Low NOx Burner and Flue Gas Recirculation - Iron & Steel Mills - Reheating 0.55
Low NOx Burner and Over Fire Air - Utility Boilers 1.6
Low NOx Burner and SCR - Coal-Fired ICI Boilers 87
Low NOx Burner and SCR - Industr/Commercial/Institutional Boilers 210
Low NOx Burner and SNCR - Industr/Commercial/Institutional Boilers 3.6
Natural Gas Reburn - Natural Gas-Fired EGU Boilers 1.3
Nonroad Diesel Retrofits & Engine Rebuilds - e.g., Construction Equipment 58
Non-Selective Catalytic Reduction (NSCR) - 4 Cycle Rich Burn 1C Engines 170
Non-Selective Catalytic Reduction (NSCR) - Nitric Acid Mfg 1.4
OXY-Firing - Glass Manufacturing 140
SCR and Flue Gas Recirculation - Fluid Catalytic Cracking Units 0.91
SCR and Flue Gas Recirculation - ICI Boilers 4.7
SCR and Flue Gas Recirculation - Process Heaters 2.9
Selective Catalytic Reduction (SCR) - Ammonia Mfg 15
Selective Catalytic Reduction (SCR) - Cement Kilns 230
Selective Catalytic Reduction (SCR) - Fluid Catalytic Cracking Units 19
Selective Catalytic Reduction (SCR) - 1C Engines, Diesel 26
Selective Catalytic Reduction (SCR) - ICI Boilers 76
Selective Catalytic Reduction (SCR) - Industrial Combustion 24
Selective Catalytic Reduction (SCR) - Industrial Incinerators 5.6
Selective Catalytic Reduction (SCR) - Iron Ore Processing 1.3
Selective Catalytic Reduction (SCR) - Petroleum Refinery Gas-Fired Process Heaters 61
Selective Catalytic Reduction (SCR) - Process Heaters 0.26
Selective Catalytic Reduction (SCR) - Sludge Incineration 39
Selective Catalytic Reduction (SCR) - Space Heaters 1.5
Selective Catalytic Reduction (SCR) - Utility Boilers 1,900
Selective Non-Catalytic Reduction (SNCR) - Coke Mfg 6.4
Selective Non-Catalytic Reduction (SNCR) - Comm/Inst. Incinerators 2.9
Selective Non-Catalytic Reduction (SNCR) - ICI Boilers 0.75
Selective Non-Catalytic Reduction (SNCR) - Industrial Combustion 0.087
Selective Non-Catalytic Reduction (SNCR) - Industrial Incinerators 2.8
7A-5
-------
NOx Control Measure Cost
Selective Non-Catalytic Reduction (SNCR) - Miscellaneous 0.25
Selective Non-Catalytic Reduction (SNCR) - Municipal Waste Combustors 2.5
Selective Non-Catalytic Reduction (SNCR) - Utility Boilers 0.46
Ultra-Low NOx Burner - Process Heaters 1.8
a All values are rounded to two significant figures.
Table 7A-6. Costs for Known VOC Controls in the 60 ppb Analysis (millions of 2011$) a
VOC Control Measure Cost
Control of Fugitive Releases - Oil & Natural Gas Production 0.089
Control Technology Guidelines - Wood Furniture Surface Coating 3.3
Flare - Petroleum Flare 0.36
Gas Recovery - Municipal Solid Waste Landfill 0.41
Improved Work Practices, Material Substitution, Add-On Controls - Printing 1.4
Incineration - Other 0.91
Incineration - Surface Coating 370
Low VOC Adhesives and Improved Application Methods - Industrial Adhesives 0.064
Low-VOC Coatings and Add-On Controls - Surface Coating 0.97
LP V Relief Valve - Underground Tanks 13
Permanent Total Enclosure (PTE) - Surface Coating 28
Petroleum and Solvent Evaporation - Surface Coating Operations 0.071
RACT - Graphic Arts 40
Reduced Solvent Utilization - Surface Coating 3.2
Reformulation - Architectural Coatings 310
Reformulation - Industrial Adhesives 6.1
Reformulation-Process Modification - Automobile Refinishing 66
Reformulation-Process Modification - Cutback Asphalt 0.087
Reformulation-Process Modification - Surface Coating 25
Solvent Recovery System - Printing/Publishing 1.3
Solvent Substitution and Improved Application Methods - Fiberglass Boat Mfg 0.059
Wastewater Treatment Controls- POTWs 0.82
a All values are rounded to two significant figures.
7A.2 Alternative Estimates of Costs Associated with Emissions Reductions from
Unknown Controls
This section presents alternative estimates of the extrapolated control costs using
alternative average cost per ton of emission reductions from unknown controls. The alternative
values used are $10,000/ton and $20,000, as well as the $15,000/ton value used as the primary
estimate in the RIA. Table 7A-7 presents the alternative estimates for the East and West Regions
in 2025, without California, while Table 7A-8 presents the estimates for post-2025 California.
7A-6
-------
Table 7A-7. Extrapolated Control Costs in 2025 by Alternative Standard for 2025 U.S.,
except California, using Alternative Average Cost Assumptions (millions of
2011$)
Alternative Level
70ppb
65 ppb
60ppb
Geographic Area
East
West
East
West
East
West
$10,000/ton
1,500
-
7,500
19,000
3,500
Extrapolated Cost
$15,000/ton
2,300
-
11,000
-
28,000
5,200
$20,000/ton
3,100
-
15,000
38,000
7,000
a All values are rounded to two significant figures.
Table 7A-8. Extrapolated Control Costs in 2025 by Alternative Standard for Post-2025
California, using Alternative Average Cost Assumptions (millions of 2011$)
Alternative Level
70 ppb
65 ppb
60 ppb
Geographic Area
California
California
California
Extrapolated Cost
$10,000/ton
530
1,000
1,400
$15,000/ton
800
1,600
2,200
$20,000/ton
1,100
2,100
2,900
a All values are rounded to two significant figures.
Note, by definition, the lower per ton estimate is 50% less than the upper end estimate, and the
range between the extrapolated cost bounds becomes larger as the alternative levels become
more stringent and rely more heavily on unknown controls.
Tables 7A-9 presents the estimates of the total control costs for 2025 East and West,
without California, when using the alternative per ton cost assumptions for emissions reductions
from unknown controls. Tables 7 A-10 presents the estimates of the total control costs for post-
2025 California when using the alternative per ton cost assumptions for emissions reductions
from unknown controls.
Table 7A-9. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S. using Alternative Cost Assumption for Extrapolated Costs,
except California (millions of 2011$)"
Total Control Costs (Known and Extrapolated)
Alternative Level
70 ppb
65 ppb
Geographic Area
East
West
East
Extrapolated
Cost =
$10,000/ton
3,100
-
11,000
Extrapolated
Cost =
$15,000/ton
3,900
-
15,000
Extrapolated
Cost =
$20,000/ton
4,700
-
19,000
7A-7
-------
Total Control Costs (Known and Extrapolated)
Alternative Level
60ppb
Geographic Area
West
East
West
Extrapolated
Cost =
$10,000/ton
400
23,000
4,100
Extrapolated
Cost =
$15,000/ton
400
33,000
5,800
Extrapolated
Cost =
$20,000/ton
400
42,000
7,600
a All values are rounded to two significant figures.
Table 7A-10. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for Post-2025 California - U.S. using Alternative Cost Assumption for
Extrapolated Costs (millions of 2011$)"
Total Control Cost
Alternative Level
70ppb
Geographic Area
California
Extrapolated
Cost =
$10,000/ton
530
Extrapolated
Cost =
$15,000/ton
800
Extrapolated
Cost =
$20,000/ton
1,100
65 ppb
California
1,000
1,600
2,100
60 ppb
California
1,400
2,200
2,900
a All values are rounded to two significant figures.
7A-8
-------
CHAPTER 8: COMPARISON OF COSTS AND BENEFITS
Overview
The EPA has performed an illustrative analysis to estimate the costs and human health
benefits of nationally attaining alternative ozone standards. The EPA Administrator is proposing
to revise the level of the primary ozone standard to within a range of 65 to 70 ppb and is
soliciting comment on alternative standard levels below 65 ppb, as low as 60 ppb. Per Executive
Order 12866 and the guidelines of OMB Circular A-4, this Regulatory Impact Analysis (RIA)
presents the analyses of the following alternative standard levels — 60 ppb, 65 ppb, and 70 ppb.
This chapter summarizes these results and discusses the implications of the analysis. The cost
and benefit estimates below are calculated incremental to a 2025 baseline assuming attainment of
the existing ozone standard of 75 ppb and incorporating air quality improvements achieved
through the projected implementation of existing regulations.
8.1 Results
In this RIA we present the primary costs and benefits estimates for full attainment in
2025. For analytical purposes, we assume that almost all areas of the country will meet each
alternative standard level in 2025 through the adoption of technologies at least as effective as the
control strategies used in this illustration. It is expected that some costs and benefits will begin
occurring earlier, as states begin implementing control measures to attain earlier or to show
progress towards attainment. For California, we provide estimates of the costs and benefits of
attaining the standard in a post-2025 time frame.
In estimating the incremental costs and benefits of potential alternative standard levels,
we recognize that there are several areas that are not required to meet the existing ozone standard
by 2025. The Clean Air Act allows areas with more significant air quality problems to take
additional time to reach the existing standard. Several areas in California are not required to
meet the existing standard by 2025, and depending on how areas are ultimately designated for a
revised standard, many areas may not be required to meet a revised standard until sometime
between 2032 and December 31, 2037. We were not able to project emissions and air quality
beyond 2025 for California; however, we adjusted baseline air quality to reflect mobile source
emissions reductions for California that would occur between 2025 and 2030: these emissions
3-1
-------
reductions were the result of mobile source regulations expected to be fully implemented by
2030. While there is uncertainty about the precise timing of emissions reductions and related
costs for California, we assume costs occur through the end of 2037 and beginning of 2038. In
addition, we model benefits for California using projected population demographics for 2038.
Because of the different timing for incurring costs and accruing benefits and for ease of
discussion throughout the analyses, we refer to the different time periods for potential attainment
as 2025 and post-2025 to reflect that (1) we did not project emissions and air quality for any year
other than 2025; (2) for California, emissions controls and associated costs are assumed to occur
through the end of 2037 and beginning of 2038; and (3) for California benefits are modeled using
population demographics in 2038. It is not straightforward to discount the post-2025 results for
California to compare with or add to the 2025 results for the rest of the U.S. While we estimate
benefits using 2038 information, we do not have good information on precisely when the costs of
controls will be incurred. Because of these differences in timing related to California attaining a
revised standard, the separate costs and benefits estimates for post-2025 should not be added to
the primary estimates for 2025.
By the 2030s, various mobile source rules, such as the onroad and nonroad diesel rules
are expected to be fully implemented. Because California will likely not have all of its areas in
attainment with a revised standard until sometime after its attainment date for the existing
standard, it is important to reflect the impact these mobile source rules might have on the
emissions that affect ozone nonattainment. To reflect the emissions reductions that are expected
from these rules, we subtract those from the estimates of the emissions reductions that might be
needed for California to fully attain in 2025, making our analysis more consistent with full
attainment later than 2025. The EPA did the analysis this way to be consistent with the
requirements in the Clean Air Act and because forcing full attainment in California in an earlier
year would likely lead to overstating costs due to (1) benefits those areas might enjoy from
existing federal or state programs implemented between 2025 and the future potential attainment
year, (2) the likelihood that energy efficiency and cleaner technologies will be further
implemented, and/or (3) the potential decline in costs of existing technologies due to economies
of scale or improvements in the efficiency of installing and operating controls ('learning by
doing').
8-2
-------
Tables 8-1 and 8-2 summarize the costs and benefits of the three potential alternative
standard levels analyzed and shows the net benefits for each of the levels across a range of
modeling assumptions related to the calculation of costs and benefits. Tables 8-3 and 8-4
provide information on the costs by geographic region for the U.S., except California in 2025
and on the costs for California for post-2025. Tables 8-5 and 8-6 provide a regional breakdown
of benefits for 2025 and a regional breakdown of benefits for post-2025.
The estimates for benefits reflect the variability in the functions available for estimating
the largest source of benefits - avoided premature mortality associated with simulated reductions
in ozone and PIVb.s (as a co-benefit). The low end of the range of net benefits is constructed by
subtracting the cost from the lowest benefit, while the high end of the range is constructed by
subtracting the cost from the highest benefit. Following these tables is a discussion of the
implications of these estimates, as well as the uncertainties and limitations that should be
considered in interpreting the estimates.
In the RIA we provide estimates of costs of emissions reductions to attain the proposed
standards in three regions — California, the rest of the western U.S., and the eastern U.S. In
addition, we provide estimates of the benefits that accrue to each of these three regions resulting
from (i) control strategies applied within the region, (ii) reductions in transport of ozone
associated with emissions reductions in other regions, and (iii) the control strategies for which
the regional cost estimates are generated. These benefits are not directly comparable to the costs
of control strategies in a region because the benefits include benefits not associated with those
control strategies.
The net benefits of emissions reductions strategies in a specific region would be the
benefits of the emissions reductions occurring both within and outside of the region minus the
costs of the emissions reductions. Because the air quality modeling is done the national level, we
do not estimate separately the nationwide benefits associated with the emissions reductions
occurring in any specific region.125 As a result, we are only able to provide net benefits
estimates at the national level. The difference between the costs for a specific region and the
125 For California, we provide separate estimates of the costs and nationwide estimates of benefits, so it is
appropriate to calculate net benefits. As such, we provide net benefits for the post-2025 California analysis.
-------
benefits accruing to that region is not an estimate of net benefits of the emissions reductions in
that region.
Table 8-1. Total Costs, Total Monetized Benefits, and Net Benefits in 2025 for U.S., except
California (billions of 2011$)a
Proposed Alternative Standard
70
65
Total Costs
7% Discount
Rate
Levels
$3.9
$15
Monetized Benefits
7% Discount
Rate
$6.4 to $13
$19 to $38
Net Benefits
7% Discount
Rate
$2.5 to $9.1
$4 to $23
Alternative Standard Level
60
$39
$34 to $70
($5) to $31
a EPA believes that providing comparisons of social costs and social benefits at 3 and 7 percent is
appropriate. Estimating multiple years of costs and benefits is not possible for this RIA due to data and resource
limitations. As a result, we provide a snapshot of costs and benefits in 2025, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
Table 8-2. Total Costs, Total Monetized Benefits, and Net Benefits of Control Strategies
Applied in California, Post-2025 (billions of 2011$)a
Proposed Alternative Standard
70
65
Total Costs
7% Discount
Rate
Levels
$0.80
$1.6
Monetized Benefits
7% Discount
Rate
$1.1 to $2
$2.2 to $4.1
Net Benefits
7% Discount
Rate
$0.3 to $1.2
$0.6 to $2.5
Alternative Standard Level
60
$2.2
$3.2 to $5.9
$1 to $3.7
a EPA believes that providing comparisons of social costs and social benefits at 3 and 7 percent is appropriate.
Estimating multiple years of costs and benefits is not possible for this RIA due to data and resource limitations. As
a result, we provide a snapshot of costs and benefits in 2025, using the best available information to approximate
social costs and social benefits recognizing uncertainties and limitations in those estimates.
EPA believes that providing comparisons of social costs and social benefits at 3 and 7
percent is appropriate. Ideally, streams of social costs and social benefits over time would be
estimated and the net present values of each would be compared to determine net benefits of the
illustrative attainment strategies. The three different uses of discounting in the RIA - (i)
construction of annualized engineering costs, (ii) adjusting the value of mortality risk for lags in
mortality risk decreases, and (iii) adjusting the cost of illness for non-fatal heart attacks to adjust
for lags in follow up costs — are all appropriate. Our estimates of net benefits are the
approximations of the net value (in 2025) of benefits attributable to emissions reductions needed
to attain just for the year 2025.
8-4
-------
Table 8-3. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for 2025 - U.S., except California (billions of 2011$, 7% Discount Rate)"
„ , . . Total Control Costs
.,, T , Geographic Area ,T, ,
Alternative Level (Known and
Extrapolated)
East 3.9
Total $3.9
East 15
65ppb w^I oT
Total $15
East 33
6°Ppb
Total $39
a All values are rounded to two significant figures. Extrapolated costs are based on the average-cost methodology.
Table 8-4. Summary of Total Control Costs (Known and Extrapolated) by Alternative
Level for post-2025 - California (billions of 2011$, 7% Discount Rate)3
Total Control Costs
Alternative Level Geographic Area (Known and
Extrapolated)
70ppb California $0.8
65 ppb California $1.6
60 ppb California $2.2
a All values are rounded to two significant figures. Extrapolated costs are based on the average-cost methodology.
Table 8-5. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the 2025
Scenario (nationwide benefits of attaining each alternative standard everywhere
in the U.S. except California) - Full Attainment
Eastb
California
Rest of West
70 ppb
99%
0%
1%
Proposed and Alterative Standards
65 ppb
96%
0%
4%
60 ppb
92%
0%
7%
a Because we use benefit-per-ton estimates to calculate the PM2 5 co-benefits, a regional breakdown for the co-
benefits is not available. Therefore, this table only reflects the ozone benefits.
b Includes Texas and those states to the north and east. Several recent rules such as Tier 3 will have substantially
reduced ozone concentrations by 2025 in the East, thus few additional controls would be needed to reach 70 ppb.
3-5
-------
Table 8-6. Regional Breakdown of Monetized Ozone-Specific Benefits Results for the post-
2025 Scenario (nationwide benefits of attaining each alternative standard just in
California) - Full Attainment a
Region
Proposed and Alterative Standards
70ppb
65 ppb
60ppb
East
California
Rest of West
0%
93%
6%
0%
94%
6%
0%
94%
6%
a Because we use benefit-per-ton estimates to calculate the PM2 5 co-benefits, a regional breakdown for the co-
benefits is not available. Therefore, this table only reflects the ozone benefits.
In this RI A, we quantify an array of adverse health impacts attributable to ozone and
PM2.5. The Integrated Science Assessment for Ozone and Related Photochemical Oxidants
("Ozone ISA") (U.S. EPA, 2013a) identifies the human health effects associated with ozone
exposure, which include premature death and a variety of illnesses associated with acute (days-
long) and chronic (months to years-long) exposures. Similarly, the Integrated Science
Assessment for Paniculate Matter ("PM ISA") (U.S. EPA, 2009) identifies the human health
effects associated with ambient particles, which include premature death and a variety of
illnesses associated with acute and chronic exposures. Air pollution can affect human health in a
variety of ways, and in Table 8-7 we summarize the "categories" of effects and describe those
that we could quantify in our "core" benefits estimates and those we were unable to quantify
because we lacked the data, time or techniques.
Table 8-7. Human Health Effects of Pollutants Potentially Affected by Strategies to Attain
the Primary Ozone Standards
Benefits Category
Effect Has
Specific Effect Been
Quantified
Effect Has
_ More
Been _ „
,„ ,. , Information
Monetized
Improved Human Health
Reduced incidence
of premature
mortality from
exposure to ozone
Reduced incidence
of morbidity from
exposure to ozone
Premature mortality based on short-term S
exposure (all ages)
Premature respiratory mortality based on S
long-term exposure (age 30-99)
Hospital admissions — respiratory causes S
(age > 65)
Emergency department visits for asthma S
(all ages)
Asthma exacerbation (age 6-18) -S
Minor restricted-activity days (age 18-65) -S
School absence days (age 5-17) -S
Decreased outdoor worker productivity a
(age 18-65)
S Section 5.6
a Section 5.6
•S Section 5.6
•S Section 5.6
•S Section 5.6
•S Section 5.6
a Section 5.6
8-6
-------
Benefits Category
Specific Effect
Effect Has
Been
Quantified
Effect Has
_
Been
,.„ ,. ,
Monetized
,.
More
T „
Information
Other respiratory effects (e.g., mediation
use, pulmonary inflammation, decrements
in lung functioning)
Cardiovascular (e.g., hospital admissions,
emergency department visits)
Reproductive and developmental effects
(e.g., reduced birthweight, restricted fetal
growth)
ages)
Other respiratory effects (e.g., pulmonary
function, non-asthma ER visits, non-
bronchitis chronic diseases, other ages
and populations)
Reproductive and developmental effects
(e.g., low birth weight, pre-term births,
etc.)
Cancer, mutagenicity, and genotoxicity
effects
ozone ISA'
ozone ISA'
ozone ISA'
Reduced incidence
of premature
mortality from
exposure to PM2 5
Reduced incidence
of morbidity from
exposure to PM2 5
Adult premature mortality based on S
cohort study estimates and expert
elicitation estimates (age >25 or age >30)
Infant mortality (age <1) S
Non-fatal heart attacks (age > 18) S
Hospital admissions — respiratory (all S
ages)
Hospital admissions — cardiovascular (age S
>20)
Emergency department visits for asthma S
(all ages)
Acute bronchitis (age 8-12) S
Lower respiratory symptoms (age 7-14) S
Upper respiratory symptoms (asthmatics S
age 9-11)
Asthma exacerbation (asthmatics age 6- S
18)
Lost work days (age 18-65) S
Minor restricted-activity days (age 18-65) S
Chronic Bronchitis (age >26) —
Emergency department visits for —
cardiovascular effects (all ages)
Strokes and cerebrovascular disease (age —
50-79)
Other cardiovascular effects (e.g., other —
•S Section 5.6 of
PMRIA
•S Section 5.6 of
PMRIA
•S Section 5.6 of
PMRIA
•S Section 5.6 of
PMRIA
•S Section 5.6 of
PMRIA
S Section 5.6 of
PMRIA
S Section 5.6 of
PMRIA
S Section 5.6 of
PMRIA
S Section 5.6 of
PMRIA
S Section 5.6 of
PMRIA
S Section 5.6 of
PMRIA
•S Section 5.6 of
PMRIA
— Section 5.6 of
PMRIA
— Section 5.6 of
PMRIA
— Section 5.6 of
PMRIA
— PM ISA b
— PM ISA b
— PM ISA b>c
— PM ISA b>c
Asthma hospital admissions (all ages)
— NO2 ISAl
3-7
-------
Benefits Category
Reduced incidence
of morbidity from
exposure to NO2
Effect Has
Specific Effect Been
Quantified
Chronic lung disease hospital admissions —
(age > 65)
Respiratory emergency department visits —
(all ages)
Asthma exacerbation (asthmatics age 4- —
18)
Acute respiratory symptoms (age 7-14) —
Premature mortality —
Other respiratory effects (e.g., airway —
hyperresponsiveness and inflammation,
lung function, other ages and populations)
Effect Has ,.
_ More
Been T „
,.„ ,. , Information
Monetized
— NO2 ISA d
— NO2 ISA d
— NO2 ISA d
— NO2 ISA d
— NO2ISAb'c
— NO2ISAb>c
a We are in the process of considering an update to the worker productivity analysis for ozone based on more recent
literature.
b We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
0 We assess these benefits qualitatively because current evidence is only suggestive of causality or there are other
significant concerns over the strength of the association.
d We assess these benefits qualitatively due to time and resource limitations for this analysis.
8.2 Discussion of Results
The costs and benefits presented in this RIA incorporate an array of methodological and
technical changes that the EPA has adopted since the previous review of the ozone standards in
2008 (U.S. EPA, 2008) and proposed reconsideration in 2010 (U.S. EPA 2010).
Several factors contributed to lower cost estimates, including shifting the baseline year
from 2020 to 2025, allowing for more time to attain and for Federal measures to work. The
baseline starting point has changed from 84 ppb to 75 ppb, substantially reducing the amount of
emissions reductions needed and the associated costs of attainment. Also, we have identified
additional known controls, which are less expensive per ton than unknown controls. Lastly,
there are fewer counties exceeding the alternative standards analyzed, therefore fewer emissions
reductions are needed for attainment, resulting in lower cost estimates.
While the costs presented in this analysis decreased compared to the prior analyses, the
benefits estimates remained about the same despite the baseline differences. The main factors
that affected the benefits estimates included the updated analysis year of 2025, which affects
population projections, baseline mortality rates, and income growth adjustment. The Value of a
Statistical Life was revised, and we removed thresholds and the assumption of no causality for
ozone mortality (which as assumed in 2008). The differences resulted in a tighter benefits range,
but the total benefits are about the same as the previous ozone RIA analyses.
-------
5.2.7 Relative Contribution of PM Benefits to Total Benefits
Because of the relatively strong relationship between PIVh.s concentrations and premature
mortality, PM co-benefits resulting from reductions in NOx emissions can make up a large
fraction of total monetized benefits, depending on the specific PM mortality impact function
used, and on the relative magnitude of ozone benefits, which is dependent on the specific ozone
mortality function assumed. PM co-benefits based on daily average concentrations are
calculated over the entire year, while ozone related benefits are calculated only during the
summer ozone season. Because the control strategies evaluated in this RIA are assumed to
operate year round rather than only during the ozone season, this means that PM benefits will
accumulate during both the ozone season and the rest of the year.
For primary benefits estimates in 2025, PM2.5 co-benefits account for between 70 and 75
percent of co-benefits, depending on the standard analyzed and on the choice of ozone and PM
mortality functions used.126 The estimate with the lowest fraction from PM co-benefits occurs
when we model benefits for the lowest alternative standard level analyzed (60ppb) and add the
lower bound core estimates for both ozone-related and PM2.5 (co-benefit) related mortality. The
estimate with the highest fraction from PM co-benefits results from modeling the highest
alternative standard level analyzed (70ppb) based on combining the high-bound core ozone and
PM2.5 (co-benefit) related mortality estimates.
8.2.2 Developing Future Control Strategies with Limited Data
Because of relatively higher ozone levels in several large urban areas (Southern
California, Houston, and the Northeastern urban corridor, including New York and Philadelphia)
and because of limitations associated with the data on currently known emissions control
technologies, the EPA recognized that known and reasonably anticipated emissions controls
would likely not be sufficient to bring some areas into attainment with either the existing or
alternative, more stringent ozone standard levels. Therefore, we designed this analysis in two
stages: the first stage focused on analyzing the air quality improvements that could be achieved
through application of documented, well-characterized emissions controls, and the costs and
126 For the separate results for post-2025, PM2 5 co-benefits account for between 30 and 45 percent of total benefits,
again depending on the standard level analyzed and the choice of ozone and PM mortality functions.
8-9
-------
benefits associated with those controls. The second stage took the emissions reductions beyond
known controls and used an extrapolation method to estimate the costs and benefits of these
additional emissions reductions needed to bring all areas into full attainment with the alternative
standard levels analyzed.
The structure of the RIA reflects this two-stage analytical approach. Separate chapters are
provided for the emissions, air quality, and cost impacts of modeled controls. We used the
information currently available to develop reasonable approximations of the costs and benefits of
the extrapolated portion of the emissions reductions necessary to reach attainment. However,
because of the uncertainty associated with the extrapolation of costs, we judged it appropriate to
provide separate estimates of the costs and benefits for partial attainment (based on known
controls) and full attainment (based on known controls and extrapolation), as well as an overall
estimate for reaching full attainment. There is a single chapter on benefits, because the
methodology for estimating benefits does not change between stages. However, in that chapter,
we again provide separate estimates of the benefits associated with the partial attainment and full
attainment portions of the analysis.
In both stages of the analysis, it should be recognized that all estimates of future costs and
benefits are not intended to be forecasts of the actual costs and benefits of implementing
potentially revised standards. Ultimately, states and local areas will be responsible for
developing and implementing emissions control programs to reach attainment with the ozone
NAAQS, with the timing of attainment being determined by future decisions by states and the
EPA. Our estimates are intended to provide information on the general magnitude of the costs
and benefits of alternative standard levels rather than on precise predictions of control measures,
costs, or benefits. With these caveats, we expect that this analysis can provide a reasonable
picture of the types of emissions controls that are currently available, the direct costs of those
controls, the levels of emissions reductions that may be achieved with these controls, the air
quality impact that can be expected to result from reducing emissions, and the public health
benefits of reductions in ambient ozone levels. This analysis identifies those areas of the U.S.
where our existing knowledge of control strategies is not sufficient to allow us to model
attainment, and where additional data or research may be needed to develop strategies for
attainment.
8-10
-------
In many ways, RIAs for proposed actions are learning processes that can yield valuable
information about the technical and policy issues that are associated with a particular regulatory
action. This is especially true for RIAs for proposed NAAQS, where we are required to stretch
our understanding of both science and technology to develop scenarios that illustrate how certain
we are about how economically feasible the attainment of these standards might be regionally.
The proposed ozone NAAQS RIA provided great challenges when compared to previous RIAs
primarily because as we tighten standards across multiple pollutants with overlapping precursors
(e.g., the recent tightening of the PIVb.s standards), we move further down the list of cost-
effective known and available controls in our database. With the more stringent NAAQS, more
areas will need to find new ways of reducing emissions. While we can speculate on what some
of these technologies might look like based on new developments in energy efficiency and clean
technology, the specific technological path in different nonattainment areas is not clear.
Because of the uncertainty regarding the development of future emissions reduction
strategies, a significant portion of the analysis is based on extrapolating from available data on
known control technologies to generate the emissions reductions necessary to reach full
attainment of an alternative ozone NAAQS and the resulting costs and benefits. Studies indicate
that it is not uncommon for pre-regulatory cost estimates to be higher than later estimates, in part
because of difficulty in predicting technological changes. Over longer time horizons, such as the
time allowed for areas with high levels of ozone pollution to meet the ozone NAAQS, the
opportunity for technical change is greater (See Chapter 7, Section 7.2 for additional discussion).
Also, because of the nature of the extrapolation method for benefits (which focuses on reductions
in ozone only at monitors that exceed the NAAQS), we generally understate the total benefits
that would result from implementing additional emissions controls to fully attain the ozone
NAAQS (i.e., assuming that the application of control strategies would result in ozone reductions
both at nonattainment and attainment monitors). On the other hand, the possibility also exists
that benefits are overestimated, because it is possible that new technical changes might not meet
the specifications, development time lines, or cost estimates provided in this analysis.
8.3 Framing Uncertainty
This section includes a qualitative presentation of key factors that (1) could impact how
air quality changes over time; (2) could impact the timing for meeting an alternative standard; (3)
8-11
-------
are difficult to predict and quantify; and (4) introduce additional uncertainty into this analysis.127
These factors, summarized in Table 8.8 below, include energy development, distribution, and use
trends; land use development patterns; economic factors; energy and research and development
policies; climate signal changes; and the influence of technological change. Additional factors
that could have an impact on how air quality changes over time include environmental indicators
other than climate change and societal preferences and attitudes toward the environment and
conservation; the potential direction and magnitude of these additional factors is less clear.
These key factors can impact the estimated baseline air quality used in the analysis, and
as a result the types of control measures and associated costs needed to meet an alternative
standard. In addition, some combinations of the key factors could have significant effects
beyond the effects of any individual factor. We cannot estimate the probability that any one
factor or combination of factors will occur, but we do believe that they introduce additional,
broader uncertainties about future trends that provide important context for the costs and benefits
presented in this analysis.
Table 8-8. Relevant Factors and Their Potential Implications for Attainment
Individual Factors
Potential Implications for NAAQS
Attainment
Information on Trends
Recently there has been an increase in domestic
production of oil and a relative decrease in
imported oil, in addition to policies and
investments geared toward the development of
alternative fuels and energy efficiency.128 This
is likely a result of all of these activities, which
have led to a reduction of U.S. dependence on
imports of foreign oil.
Several trends have emerged within the last ten
years and are expected to continue, including
increased natural gas production and
consumption, renewable energy installations,
Energy —
Extraction,
conversion,
distribution and
storage, efficiency,
international energy
trends
Geopolitics, reserves, international
and domestic demand, and
technological breakthroughs in
energy technologies can drive fuel
prices up or down.
If more renewable sources of energy
are employed and use of natural gas
increases, then emissions may be
lower, potentially lowering
attainment costs.
127 OMB Circular A-4 indicates that qualitative discussions should be included in analyses whenever there is
insufficient data to quantify uncertainty.
128 http://energy.gov/articles/us-domestic-oil-production-exceeds-imports-first-time-18-years
8-12
-------
Individual Factors
Potential Implications for NAAQS
Attainment
Information on Trends
and energy efficiency technology
installations.129
Land Use
Development
Patterns -
Design of urban areas,
vehicle-miles travelled
A move toward denser urban
settlements, slowing of growth in
vehicle miles travelled (VMT) and
increased use of public transit could
decrease emissions, potentially
lowering attainment costs.130
Recent trends in VMT illustrate some of the
uncertainty around future emissions from
mobile sources131. In 2006, projections of
VMT showed a sustained increase,132 yet VMT
growth slowed in recent years and actually
declined in 2008 and 2009.133 Between 2000
and 2010 average growth in VMT was 0.8%, as
compared to 2.9% from the previous decade.
The Economy
An increase in economic growth,
investment in technologies that have
high energy use, and a return of U.S.
manufacturing could lead to higher
emissions making attainment
potentially more costly. A slowing
of the economy, investments in
energy efficient technologies, and a
continuation of a service-based
economy could lead to lower
emissions making attainment
potentially less costly.134
Affluence leads to increased consumption and
energy use. However, this increase may not be
proportional. Energy and materials is not
directly proportional to economic growth,
decreasing or stabilizing over time in spite of
continued economic growth.135
Policies3 -
Energy efficiency,
energy security,
direction of research
and development,
renewable energy
A move toward energy security and
independence would mean an
increased use of domestic energy
sources. If this results in a fuel mix
where emissions decrease, then
attainment could likely be less costly.
State and local policies related to energy
efficiency, cleaner energy, energy security136,
as well as the direction of research and
development of technology can have a direct or
indirect effect on emissions. Policies that result
in energy efficiency, renewable energy, the use
129 http://www.eia.gov/forecasts/aeo/er/pdf/0383er%282014%29.pdf;
http://energy.gov/sites/prod/files/2014/08/fl8/2013%20Wind%20Technologies%20Market%20Report%20Present
ation.pdf; http://www.eia.gov/electricity/monthly/update/archive/april2014/;
http://www.ercot.com/content/news/presentations/2014/GCPA%20%2002%20Oct%202013%20FINAL.pdf;
http://energy.gov/eere/sunshot/photovoltaics.
130 For example, see Cervero (1998), the Center for Clean Air Policy's Transportation Emissions Guidebook
(http://www.trb.org/Main/Blurbs/156164.aspx). For ongoing research see
http://apps.trb.org/cmsfeed/TRBNetProjectDisplay.asp?ProjectID=3092.
131 For example, see the Transportation Research Board's National Cooperative Highway Research Program
(NCHRP) 2014.
132https://www.fhwa.dot.gov/policy/2006cpr/chap9.htm#body
133 https://www.fhwa.dot.gov/policyinformation/travel_monitoring/13jantvt/page2.cfm
134 For example, Bo (2011), and http://www.epa.gov/regionl/airquality/nox.html for manufactures contributions to
NOx emissions.
135 UNEP 2011, http://www.unep.org/resourcepanel/decoupling/files/pdf/decoupling_report_english.pdf
136 For example, http://www2.epa.gov/laws-regulations/summary-energy-independence-and-security-act.
8-13
-------
Individual Factors
Potential Implications for NAAQS
Attainment
Information on Trends
If not, attainment could likely be
more costly. A move toward
investments in fuel efficiency and
low emissions fuels could decrease
emissions and likely lower
attainment costs.
of cleaner fuels and conservation measures
would likely result in decreased emissions and
likely decrease attainment costs.137 Growth in
energy demand has stayed well below growth
in gross domestic product, likely as a result of
technological advances, federal, state and local
energy efficiency standards and policies, and
othermacroeconomic factors.138 U.S.
productivity per energy expended relative to
other countries suggests that additional
efficiency gains are possible.139
Intensity, Location
and Outcome of the
Climate Change
Signal
Strong climate signals that bring high
temperatures could increase ozone,
likely making attainment more
costly.
Uncertainty exists regarding how the climate
signal will interact with air quality, as well as
with other factors. However, research
demonstrates that in areas where there are both
high levels of emissions and high temperatures,
attaining an ozone standard will likely be much
harder. The magnitudes of these impacts will
depend on atmospheric chemical and physical
processes, as well as anthropogenic activities
that increase or decrease NOx and/or VOC
emissions.140
Technological
Change — Including
emissions reductions
technologies and other
technological
developments
Innovation in production and
emissions control technologies,
learning that lower costs, and
breakthroughs in battery/energy
storage technologies for use with
renewable energy could improve air
quality, reducing emissions and
likely lowering attainment costs.
Examples of emerging technologies include
carbon capture and sequestration (CCS), battery
technologies, emerging advanced biofuels,
which could all have breakthroughs that could
impact fuel use. Similarly, shifts in industrial
production processes, such as a move from
using primary metals to more recycling could
impact energy use141 .
aPolicies refer to any policies or regulations that are not environmental regulations set by U.S. EPA, states, tribes, or
local authorities.
8.4 Key Observations from the Analysis
The following are key observations about the RIA results.
• Tightening the ozone standards can incur significant, but uncertain, costs. Our
estimates of costs for a set of modeled NOx and VOC controls comprise only a small part
of the estimated costs of full attainment. These estimated costs for the modeled set of
137
7 For example, see http://www.dsireusa.org/solar/solarpolicyguide/.
138http:^ipartisanpolicy.org/sites/default/files/BPC%20SEPI%20Energy%20Report%202013_0.pdf, p. 5.
139http:^ipartisanpolicy.org/sites/default/files/BPC%20SEPI%20Energy%20Report%202013_0.pdf, p. 69.
140 See Jacobs (2009).
141 http ://www. eia. gov/todayinenergy/detail. cfm?id= 16211
8-14
-------
controls are still uncertain, but they are based on the best available information on control
technologies, and have their basis in real, tested technologies. Estimating costs of full
attainment was based on a generalized relationship between emissions and ozone levels.
This introduces significant uncertainty into the calculation of the emissions reductions that
might be needed to reach full attainment.
• Tightening the ozone standards can also result in significant benefits. Estimates of
benefits are driven largely by projected reductions in ozone-related short-term mortality
and co-benefits associated with reductions in PIVh.s-related long-term mortality. Although
using a benefit-per-ton approach in modeling PIVh.s-related cobenefits (rather than direct
modeling) has increased uncertainty, this approach is peer-reviewed and robust. We also
modeled reductions in ozone-related long-term respiratory mortality, however due to
concerns over potential double counting of benefits and limitations in our ability to project
the lag-structure of reductions in this mortality endpoint, we did not include these
estimates as part of the core benefit estimate. In addition to these mortality endpoints, we
did quantify a wide-range of morbidity endpoints for both ozone and PIVh.s, although these
contribute only minimally to total monetized benefits.
• Air quality modeling approach can introduce uncertainty. Based on air quality
modeling sensitivity analyses, there is significant spatial variability in the relationship
between local and regional NOx emission reductions and ozone levels across urban areas.
We performed a national scale air quality modeling analysis to estimate ozone
concentrations for the future base case year of 2025. To accomplish this, we modeled
multiple emissions cases for 2025, including the 2025 base case and twelve (12) 2025
emissions sensitivity simulations. The 12 emissions sensitivity simulations were used to
develop ozone sensitivity factors (ppb/ton) from the modeled response of ozone to changes
in NOx and VOC emissions from various sources and locations. These ozone sensitivity
factors were then used to determine the amount of emissions reductions needed to reach
the 2025 baseline and evaluate potential alternative standard levels of 70, 65, and 60 ppb
incremental to the baseline. We used the estimated emissions reductions needed to reach
each of these standard levels to analyze the costs and benefits of alternative standard
levels.
• Available technologies that might achieve NOx and VOC reductions to attain
alternative ozone NAAQS are not sufficient. In some areas of the U.S., the information
we have about existing controls does not result in sufficient emissions reductions needed to
meet the existing standard. After applying existing rules and the illustrative known
controls across the nation (excluding California), in order to reach 70 ppb we were able to
identify controls that reduce overall NOx emissions by 490,000 tons and VOC emissions
by 55,000 tons. In order to reach 65 ppb we were able to identify controls that reduce
overall NOx emissions by 1,100,000 tons and VOC emissions by 110,000 tons. After these
reductions, in order to reach 70 ppb over 150,000 tons of NOx emissions remained, and in
order to reach 65 ppb over 750,000 tons of NOx emissions remained.
• California costs and benefits are highly uncertain. California faces large challenges in
meeting any alternative standard, but their largest challenges may be in attaining the
existing standard. Because our analysis suggested that all available controls would be
8-15
-------
exhausted in attempting to meet the current 75 ppb standard, all of the benefits and costs of
lower standards in California are based on the application of unknown controls. Both the
benefits and the costs associated with the assumed NOx and VOC reductions in California
are particularly uncertain.
Some EPA existing mobile source programs will help areas reach attainment. These
programs promise to continue to help areas reduce ozone concentrations beyond 2025. In
California, continued implementation of mobile source rules, including the onroad and
nonroad diesel rules and the locomotive and marine engines rule, are projected to reduce
NOx emissions by an additional 14,000 tons and VOC emissions by an additional 6,300
tons between 2025 and 2030. These additional reductions will likely reduce the overall
emissions reductions needed for attainment relative to what California might have needed
to reduce from other sectors if attainment were to be required in 2025.
The economic impacts (i.e., social costs) of the cost of these modeled controls were not
included in this analysis. Incorporating the economic impact of the extrapolated portion
of the costs was too uncertain to be included as part of these estimates, and it was
determined best to keep the modeled and extrapolated costs on the same basis.
Costs and benefits will depend on implementation timeframes. States will ultimately
select the specific timelines for implementation as part of their State Implementation Plans.
To the extent that states seek classification as extreme nonattainment areas, the timeline for
implementation may be extended beyond 2025, meaning that the amount of emissions
reductions that will be required in 2025 will be less, and costs and benefits in 2025 will be
lower.
8.5 References
Cervero, R. (1998) The Transit Metropolis, A Global Inquiry. Island Press, Washington, D.C.
Jacob, D. I, & Winner, D. (2009). Effect of climate change on air quality. Atmospheric Environment, 43, 51-63.
NCHRP (2014). The Effects of Socio-Demographics on Future Travel Demand, Transportation Research Board of
the National Academies, Washington, D.C. Strategic Issues Facing Transportation, Report 750, Volume 6.
U.S. Environmental Protection Agency (U.S. EPA). 2008. Final Ozone NAAQS Regulatory Impact Analysis. Office
of Air Quality Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/ttn/ecas/regdata/RIAs/452_R_08_003.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2010. Regulatory Impact Analysis (RIA) for the Proposed
Reconsideration of the ozone National Ambient Air Quality Standards (NAAQS). Office of Air Quality Planning
and Standards, Research Triangle Park, NC. January. Available at .
U.S. EPA. 2009. Integrated Science Assessment for Paniculate Matter: Final. Research Triangle Park, NC: U.S.
Environmental Protection Agency. (EPA document number EPA/600/R-08/139F).
U.S. EPA. 2013a. Integrated Science Assessment for Ozone and Related Photochemical Oxidants: Final. Research
Triangle Park, NC: U.S. Environmental Protection Agency. (EPA document number EPA/600/R-10/076F).
8-16
-------
CHAPTER 9: STATUTORY AND EXECUTIVE ORDER IMPACT ANALYSIS
Overview
This section explains the statutory and executive orders applicable to EPA rules, and
discusses EPA's actions taken pursuant to these orders.
9.1 National Technology Transfer and Advancement Act
Section 12(d) of the National Technology Transfer and Advancement Act of 1995
(NTTAA), Public Law No. 104-113, §12(d) (15 U.S.C. 272 note) directs the EPA to use
voluntary consensus standards in its regulatory activities unless to do so would be inconsistent
with applicable law or otherwise impractical. Voluntary consensus standards are technical
standards (e.g., materials specifications, test methods, sampling procedures, and business
practices) that are developed or adopted by voluntary consensus standards bodies. The NTTAA
directs the EPA to provide Congress, through OMB, explanations when the Agency decides not
to use available and applicable voluntary consensus standards.
Today's proposed rulemaking does not involve technical standards. Therefore, the EPA is
not considering the use of any voluntary consensus standards.
9.2 Paperwork Reduction Act
This action does not impose an information collection burden under the provisions of the
Paperwork Reduction Act, 44 U.S.C. 3501 et seq. There are no information collection
requirements directly associated with the establishment of a NAAQS under section 109 of the
CAA.
Burden means the total time, effort, or financial resources expended by persons to
generate, maintain, retain, or disclose or provide information to or for a federal agency. This
includes the time needed to review instructions; develop, acquire, install, and utilize technology
and systems for the purposes of collecting, validating, and verifying information, processing and
maintaining information, and disclosing and providing information; adjust the existing ways to
comply with any previously applicable instructions and requirements; train personnel to be able
to respond to a collection of information; search data sources; complete and review the collection
3-1
-------
of information; and transmit or otherwise disclose the information. Burden is defined at 5 CFR
1320.3(b).
An agency may not conduct or sponsor, and a person is not required to respond to a
collection of information unless it displays a currently valid OMB control number. The OMB
control numbers for the EPA's regulations in 40 CFR are listed in 40 CFR part 9. Per the
Implementation of the 2008 National Ambient Air Quality Standards for Ozone: State
Implementation Plan Requirements, the annual burden for this information collection averaged
over the first 3 years is estimated to be a total of 40,000 labor hours per year at an annual labor
cost of $2.4 million (present value) over the 3-year period or approximately $91,000 per state for
the 26 state respondents, including the District of Columbia. The average annual reporting
burden is 690 hours per response, with approximately 2 responses per state for 58 state
respondents. There are no capital or operating and maintenance costs associated with the
proposed rule requirements.
In addition, per the draft Supporting Statement for Revisions to Ambient Air Monitoring
Regulations for Ozone (Proposed Rule) (EPA ICR Number 0940), as part of the ozone NAAQS
proposed rulemaking EPA is proposing revisions to the length of the required ozone monitoring
season (see Section VI of the preamble). The draft Information Collection Request (ICR) is
estimated to involve 158 respondents for a total average cost of approximately $24 million (total
capital, and labor and operation and maintenance costs) plus a total burden of 339,930 hours for
the support of all operational aspects of the entire ozone monitoring network. The labor costs
associated with these hours are approximately $19.8 million. Also included in the total costs are
other costs of operation and maintenance of $2.2 million and equipment and contract costs of
$2.1 million. EPA typically funds 60 percent of the approximately $24 million through grants to
the State/local agencies. In addition to the costs at the State, local, and Tribal air quality
management agencies, there is a burden to EPA of 41,418 hours and $2.6 million.
9.3 Regulatory Flexibility Act
The Regulatory Flexibility Act (RFA) generally requires an agency to prepare a
regulatory flexibility analysis of any rule subject to notice and comment rulemaking
requirements under the Administrative Procedure Act or any other statute unless the agency
A.9-2
-------
certifies that the rule will not have a significant economic impact on a substantial number of
small entities. Small entities include small businesses, small organizations, and small
governmental jurisdictions.
For purposes of assessing the impacts of today's proposed rule on small entities, small
entity is defined as: (1) a small business that is a small industrial entity as defined by the Small
Business Administration's (SBA) regulations at 13 CFR 121.201; (2) a small governmental
jurisdiction that is a government of a city, county, town, school district or special district with a
population of less than 50,000; and (3) a small organization that is any not-for-profit enterprise
which is independently owned and operated and is not dominant in its field.
After considering the economic impacts of today's proposed rule on small entities, I
certify that this action will not have a significant economic impact on a substantial number of
small entities. This proposed rule will not impose any requirements on small entities. Rather, this
rule establishes national standards for allowable concentrations of ozone in ambient air as
required by section 109 of the CAA. See also American Trucking Associations v. EPA. 175 F.
3d at 1044-45 (NAAQS do not have significant impacts upon small entities because NAAQS
themselves impose no regulations upon small entities). We continue to be interested in the
potential impacts of the proposed rule on small entities and welcome comments on issues related
to such impacts.
9.4 Unfunded Mandates Reform Act
Title II of the Unfunded Mandates Reform Act of 1995 (UMRA), Public Law 104-4,
establishes requirements for federal agencies to assess the effects of their regulatory actions on
state, local, and tribal governments and the private sector. Under section 202 of the UMRA, the
EPA generally must prepare a written statement, including a cost-benefit analysis, for proposed
and final rules with "federal mandates" that may result in expenditures to state, local, and tribal
governments, in the aggregate, or to the private sector, of $100 million or more in any 1 year.
Before promulgating an EPA rule for which a written statement is needed, section 205 of the
UMRA generally requires the EPA to identify and consider a reasonable number of regulatory
alternatives and to adopt the least costly, most cost-effective or least burdensome alternative that
achieves the objectives of the rule. The provisions of section 205 do not apply when they are
A.9-2
-------
inconsistent with applicable law. Moreover, section 205 allows the EPA to adopt an alternative
other than the least costly, most cost-effective or least burdensome alternative if the
Administrator publishes with the final rule an explanation for why that alternative was not
adopted. Before the EPA establishes any regulatory requirements that may significantly or
uniquely affect small governments, including tribal governments, it must have developed under
section 203 of the UMRA a small government agency plan. The plan must provide for notifying
potentially affected small governments, enabling officials of affected small governments to have
meaningful and timely input in the development of the EPA regulatory proposals with significant
federal intergovernmental mandates, and informing, educating, and advising small governments
on compliance with the regulatory requirements.
Today's proposed rule contains no federal mandates (under the regulatory provisions of
Title II of the UMRA) for state, local, or tribal governments or the private sector. The proposed
rule imposes no new expenditure or enforceable duty on any state, local or tribal governments or
the private sector, and the EPA has determined that this proposed rule contains no regulatory
requirements that might significantly or uniquely affect small governments. Furthermore, as
indicated previously, in setting aNAAQS the EPA cannot consider the economic or
technological feasibility of attaining ambient air quality standards, although such factors may be
considered to a degree in the development of state plans to implement the standards. See also
American Trucking Associations v. EPA. 175 F. 3d at 1043 (noting that because the EPA is
precluded from considering costs of implementation in establishing NAAQS, preparation of a
Regulatory Impact Analysis (RIA) pursuant to the Unfunded Mandates Reform Act would not
furnish any information that the court could consider in reviewing the NAAQS). Accordingly,
the EPA has determined that the provisions of sections 202, 203, and 205 of the UMRA do not
apply to this proposed decision. The EPA acknowledges, however, that any corresponding
revisions to associated SIP requirements and air quality surveillance requirements, 40 CFR part
51 and 40 CFR part 58, respectively, might result in such effects. Accordingly, EPA will address,
as appropriate, unfunded mandates if and when it proposes any revisions to 40 CFR parts 51 or
58.
A.9-4
-------
9.5 Executive Order 12866: Regulatory Planning and Review
Under section 3(f)(l) of Executive Order (EO) 12866 (58 FR 51735, October 4, 1993),
the ozone NAAQS action is an "economically significant regulatory action" because it is likely
to have an annual effect on the economy of $100 million or more. Accordingly, the EPA
submitted this action to the Office of Management and Budget (OMB) for review under EO
12866 and any changes made in response to OMB recommendations have been documented in
the docket for this action. In addition, the EPA prepared this RIA of the potential costs and
benefits associated with this action. A copy of the analysis is available in the RIA docket (EPA-
HQ-OAR-2013-0169) and the analysis is briefly summarized here. The RIA estimates the costs
and monetized human health and welfare benefits of attaining four alternative ozone NAAQS
nationwide. Specifically, the RIA examines the alternatives of 75 ppb, 70 ppb, 65 ppb, and 60
ppb. The RIA contains illustrative analyses that consider a limited number of emissions control
scenarios that states and Regional Planning Organizations might implement to achieve these
alternative ozone NAAQS. However, the Clean Air Act (CAA) and judicial decisions make clear
that the economic and technical feasibility of attaining ambient standards are not to be
considered in setting or revising NAAQS, although such factors may be considered in the
development of state plans to implement the standards. Accordingly, although this RIA has been
prepared, the results of the RIA have not been considered in issuing today's proposed rule.
9.6 Executive Order 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations
Executive Order 12898 (59 FR 7629 (Feb. 16, 1994)) establishes federal executive policy
on environmental justice. Its main provision directs federal agencies, to the greatest extent
practicable and permitted by law, to make environmental justice part of their mission by
identifying and addressing, as appropriate, disproportionately high and adverse human health or
environmental effects of their programs, policies, and activities on minority populations and low-
income populations in the United States.
To gain a better understanding of the populations within the areas potentially impacted by
a revised ozone NAAQS, the EPA conducted a proximity analysis that examined socio-
demographic attributes of populations in these areas. The areas of interest for these analyses
were defined as all counties contained within any Core-Based Statistical Area (CBSA) with at
A.9-5
-------
least one monitor with a current (2011-2013) design value above the proposed range of standard
levels (65 to 70 ppb), as well as counties not in a CBSA with a current (2011-2013) design value
above the proposed range of standard levels. The tabulation of results from this analysis are
presented below. Table 9-1 shows the percent of minority populations within CBS As with
monitors above the levels of the proposed ozone standard levels, as well as the national
percentages for comparison. Table 9-2 shows the percent of populations of different ages,
education, and income level for those same areas. Additional details can be found in Appendix
9A.
Table 9-1. Summary of Population Totals and Demographic Categories for Areas of
Interest and National Perspective
Demographic
Summary
Area of Interest
65 ppb
70 ppb
Population
Total
221,431,286
193,316,836
White
153,706,027
132,112,738
African
American
30,429,108
27,193,155
Native
American
1,726,110
1,488,364
Other or
Multiracial1
35,570,041
32,522,579
Minority/Non-
White
Hispanic"
67,725,259
61,204,098
% of Area of Interest Total
65 ppb
70 ppb
National
Total
%of
National
Total
312,861,256
69%
68%
226,405,205
72%
14%
14%
39,475,216
13%
1%
1%
2,952,087
1%
16%
17%
44,028,748
14%
31%
32%
86,456,051
28%
a The race Minority'/Non-White Hispanic field is computed by subtracting the white population from the total
population.
Table 9-2.
Summary of Population Totals and Demographic Categories for Areas of
Interest and National Perspective
Demographic Linguistically Without a
Summary Isolated Age 0-4 Age 0-17 Age 65+ HS Diploma
Low
Income2
Area of Interest Total
65 ppb
70 ppb
% of Area
65 ppb
13,072,109 14,695,948 54,008,810 27,163,990 20,914,891
12,179,896 12,849,637 47,296,147 23,518,071 18,495,474
of Interest Total
6% 7% 24% 12% 14%
67,027,700
58,296,224
30%
1 Appendix 9A clarifies that "other or multiracial" is derived from individual reporting on Census forms and
includes citations to the specific 2010 Census data used in this analysis.
2 Appendix 9A clarifies that "low income" in this analysis is defined as income two times the poverty line or less.
A.9-6
-------
70ppb 6% 7% 24% 12% 15% 30%
National Total 19,196,507 20,465,065 75,217,176 40,830,262 30,952,789 101,429,436
% of National
Total 6% 7% 24% 13% 15% 32%
The proposed rule will establish uniform national standards for ozone air pollution. This
analysis identifies, on a limited basis, the subpopulations that may be exposed to elevated ozone
concentrations, who thus are expected to benefit most from this regulation. This analysis does
not identify the demographic characteristics of the most highly affected individuals or
communities; nor does it quantify the level of risk faced by those individuals or communities. To
the extent that any minority, low-income or indigenous subpopulation is disproportionately
impacted by ozone levels because it resides in an area of interest, that subpopulation also stands
to see increased environmental and health benefits from the emission reductions called for by
this proposed rule. Available data suggest that the counties most likely to experience risk
reductions from the proposed rule are approximately 14% African American, 1% Native
American, 16-17% Other and multi-racial, and 31-32% Minority/Hispanic, which are
approximately equal to the proportions of these populations represented in the U.S.
9.7 Executive Order 13045: Protection of Children from Environmental Health &
Safety Risks
Executive Order 13045, "Protection of Children from Environmental Health Risks and
Safety Risks" (62 FR 19885, April 23, 1997) applies to any rule that: (1) is determined to be
"economically significant" as defined under Executive Order 12866, and (2) concerns an
environmental health or safety risk that the EPA has reason to believe may have a
disproportionate effect on children. If the regulatory action meets both criteria, the Agency must
evaluate the environmental health or safety effects of the planned rule on children, and explain
why the planned regulation is preferable to other potentially effective and reasonably feasible
alternatives considered by the Agency.
Today's proposed rule is subject to Executive Order 13045 because it is an economically
significant regulatory action as defined by Executive Order 12866, and we believe that the
environmental health risk addressed by this action may have a disproportionate effect on
children. The proposed rule will establish uniform national ambient air quality standards for
A.9-7
-------
ozone; these standards are designed to protect public health with an adequate margin of safety, as
required by CAA section 109. However, the protection offered by these standards may be
especially important for children because children, especially children with asthma, along with
other sensitive population subgroups such as all people with lung disease and people active
outdoors, are potentially susceptible to health effects resulting from ozone exposure. Because
children are considered a potentially susceptible population, we have carefully evaluated the
environmental health effects of exposure to ozone pollution among children. Discussions of the
results of the evaluation of the scientific evidence, policy considerations, and the exposure and
risk assessments pertaining to children occurs throughout the preamble. Table 9-2 above
includes a summary of available data that indicates that the counties most likely to experience
risk reductions from the proposed rule are comprised of approximately 24% of the population
between the ages of zero and 17, which is approximately equal to the proportions of this segment
of the population represented in the U.S.
9.8 Executive Order 13132: Federalism
Executive Order 13132, entitled "Federalism" (64 FR 43255, August 10, 1999), requires
EPA to develop an accountable process to ensure "meaningful and timely input by state and local
officials in the development of regulatory policies that have federalism implications." "Policies
that have federalism implications" are defined in the Executive Order to include regulations that
have "substantial direct effects on the states, on the relationship between the national government
and the states, or on the distribution of power and responsibilities among the various levels of
government."
Today's proposed rule does not have federalism implications. It will not have substantial
direct effects on the states, on the relationship between the national government and the states, or
on the distribution of power and responsibilities among the various levels of government, as
specified in Executive Order 13132. The rule does not alter the relationship between the federal
government and the states regarding the establishment and implementation of air quality
improvement programs as codified in the CAA. Under section 109 of the CAA, the EPA is
mandated to establish NAAQS; however, CAA section 116 preserves the rights of states to
establish more stringent requirements if deemed necessary by a state. Furthermore, this proposed
rule does not impact CAA section 107 which establishes that the states have primary
A.9-8
-------
responsibility for implementation of the NAAQS. Finally, as noted in section E (above) on
UMRA, this rule does not impose significant costs on state, local, or tribal governments or the
private sector. Thus, Executive Order 13132 does not apply to this rule.
However, as also noted in section D (above) on UMRA, EPA recognizes that states will
have a substantial interest in this rule and any corresponding revisions to associated SIP
requirements and air quality surveillance requirements, 40 CFR part 51 and 40 CFR part 58,
respectively. Therefore, in the spirit of Executive Order 13132, and consistent with the EPA
policy to promote communications between the EPA and state and local governments, the EPA
has specifically solicited comment on today's proposed rule from state and local officials.
9.9 Executive Order 13175: Consultation and Coordination with Indian Tribal
Governments
Executive Order 13175, entitled "Consultation and Coordination with Indian Tribal
Governments" (65 FR 67249, November 9, 2000), requires the EPA to develop an accountable
process to ensure "meaningful and timely input by tribal officials in the development of
regulatory policies that have tribal implications." This rule concerns the establishment of ozone
NAAQS. The Tribal Authority Rule gives tribes the opportunity to develop and implement CAA
programs such as the ozone NAAQS, but it leaves to the discretion of the tribe whether to
develop these programs and which programs, or appropriate elements of a program, they will
adopt.
Today's proposed rule does not have tribal implications, as specified in Executive Order
13175. It does not have a substantial direct effect on one or more Indian Tribes, since tribes are
not obligated to adopt or implement any NAAQS. In addition, tribes are not obligated to conduct
ambient monitoring for ozone or to adopt the ambient monitoring requirements of 40 CFR part
58. Thus, Executive Order 13175 does not apply to this rule.
The EPA specifically solicits comment on this rule from tribal officials. Prior to
finalization of this proposal, the EPA intends to conduct outreach consistent with the EPA Policy
on Consultation and Coordination with Indian Tribes. Outreach to tribal environmental
professionals will be conducted through participation in Tribal Air call, which is sponsored by
the National Tribal Air Association. In addition, the EPA intends to offer formal consultation to
A.9-9
-------
the tribes during the public comment period. If consultation is requested, a summary of the result
of that consultation will be presented in the notice of final rulemaking and will be available in
the docket.
9.10 Executive Order 13211: Actions that Significantly Affect Energy Supply,
Distribution, or Use
Today's proposed rule is not a "significant energy action" as defined in Executive Order
13211, "Actions Concerning Regulations That Significantly Affect Energy Supply, Distribution,
or Use" (66 FR 28355 (May 22, 2001)) because in the Agency's judgment it is not likely to have
a significant adverse effect on the supply, distribution, or use of energy. The purpose of this rule
is to establish revised NAAQS for ozone. The rule does not prescribe specific pollution control
strategies by which these ambient standards will be met. Such strategies will be developed by
states on a case-by-case basis, and the EPA cannot predict whether the control options selected
by states will include regulations on energy suppliers, distributors, or users. Thus, the EPA
concludes that today's proposed rule is not likely to have any adverse energy effects and does not
constitute a significant energy action as defined in Executive Order 13211.
Application of the modeled illustrative control strategy containing known controls for
power plants, shown in Chapter 7, means that 3 percent of the total projected coal-fired EGU
capacity nationwide in 2025 could be affected by controls for the alternative standard level of 70
ppb. Similarly, 21 percent of total projected coal-fired EGU capacity in 2025 could be affected
for the alternative standard level of 65 ppb, and 22 percent for the alternative standard level of 60
ppb. In addition, some fuel switching might occur that could alter these percentages, though we
are unable to estimate the effect on energy impacts from fuel switching. Controls on EGUs
powered by fuels other than coal were not part of this illustrative analysis, and thus, would not be
affected. In addition, we are unable to estimate energy impacts resulting from application of
controls to non-EGUs or mobile sources. It is important to note that the estimates presented
above are just one illustrative strategy and states may choose to apply control on sources other
than EGUs for the purposes of attaining a more stringent ozone standard.
A.9-10
-------
APPENDIX 9A: SOCIO-DEMOGRAPHIC CHARACTERISTICS OF POPULATIONS
IN CORE BASED STATISTICAL AREAS WITH OZONE MONITORS EXCEEDING
PROPOSED OZONE STANDARDS
Overview
This appendix describes a limited screening-level analysis of the socio-demographic
characteristics of populations living in areas with an ozone monitor with a current (2011-2013)
design value exceeding the proposed range of ozone standard levels, 65 to 70 parts per billion
(ppb). This analysis does not include a quantitative assessment of exposure and/or risk for
specific populations of potential interest from an environmental justice (EJ) perspective, and
therefore it cannot be used to draw any conclusions regarding potential disparities in exposure or
risk across populations of interest from an EJ perspective. This appendix describes the technical
approach used in the analysis, discusses uncertainties and limitations associated with the
analysis, and presents results.
Executive Order 12898, Federal Actions to Address Environmental Justice in Minority
Populations and Low-Income Populations (59 FR 7629; Feb. 16, 1994), directs federal agencies,
to the greatest extent practicable and permitted by law, to make environmental justice part of
their mission by identifying and addressing, as appropriate, disproportionately high and adverse
human health or environmental effects of their programs, policies, and activities on minority
populations and low-income populations in the United States. . In addition, Executive Order
13045, Protection of Children from Environmental Health Risks and Safety Risks (62 FR 19885,
April 23, 1997) applies to any rule that: (1) is determined to be "economically significant" as
defined under Executive Order 12866, and (2) concerns an environmental health or safety risk
that the EPA has reason to believe may have a disproportionate effect on children. Accordingly,
the Environmental Protection Agency's (EPA) Office of Air Quality Planning and Standards
(OAQPS) has conducted a limited analysis of population demographics in some areas that may
be affected by the proposed revisions to the National Ambient Air Quality Standards (NAAQS)
for ozone.
The EPA Administrator is proposing to revise the NAAQS for ozone from the current
level of 75 ppb to within the range of 65 to 70 ppb, while also soliciting comment on retaining
A.9-1
-------
the current standard and levels down to 60 ppb. This proposed rule will establish uniform
national standards for ozone in ambient air. The proposed revisions would improve public health
protection for at-risk groups, especially children. The Agency has elected to conduct a limited
analysis of key socio-demographic characteristics of populations living in areas of interest,
defined for this analysis as any Core Based Statistical Area (CBSA) with at least one county
having an ozone monitor with a current (2011-2013) design value exceeding the proposed range
of ozone standard levels (65 to 70 ppb), and also including individual counties not included in a
CBS A with a design value exceeding the proposed range of standards.
9A.I Design of Analysis
To gain a better understanding of the populations within the areas of interest, the EPA
conducted an analysis at the county level for this ozone NAAQS review. The areas of interest for
these analyses were defined as all counties contained within any CBSA with at least one monitor
with a current (2011-2013) design value above the proposed range of standard levels (65 to 70
ppb) as well as counties not in a CBSA with a current (2011-2013) design value above the
proposed range of standard levels (65 ppb to 70 ppb). The areas of interest were designed to try
to capture population and communities most likely to benefit from improved air quality resulting
from implementation of the proposed ozone NAAQS revisions.
At the lower end of the range of proposed standard levels of 65 ppb, this definition of the
areas of interest resulted in 953 counties being analyzed, 888 in 265 CBSAs and 65 outside
CBSAs. At the upper end of the range of proposed standard levels of 70 ppb, this resulted in 707
counties being analyzed, 680 in 182 CBSAs and 27 outside CBSAs. The demographic variables
used in this analysis include race, ethnicity, age, economic and education data. Details on these
demographic groups are provided in the following section (9A. 1.1).
To compare the demographic data in the areas of interest to the national data, the data
from the identified counties were aggregated to represent the areas of interest identified by the
lower and upper end of the proposed range of standard levels. This analysis identifies, on a
limited basis, the subpopulations that are most likely to experience reductions in ozone
concentrations as a result of actions taken to meet the proposed range of standard levels and thus
are expected to benefit most from this regulation. This analysis does not identify the
9A-2
-------
demographic characteristics of the most highly affected individuals or communities nor does it
quantify the level of risk faced by those individuals or communities. To the extent that any
minority, low-income or indigenous subpopulation is disproportionately impacted by ozone
levels because they reside in an area of interest, that subpopulation also stands to see increased
environmental and health benefit from meeting the more protective proposed standard levels.
The aggregated demographic sub-population values across the areas of interest are
compared to the national data and are listed in Table 9A-2 in Section 9A-3.
9A. 1.1 Demographic Variables Included in Analysis
This analysis includes race, ethnicity, and age data derived from the 2010 Census SF1
dataset1 and economic and education data from the Census Bureau's 2006-2010 American
Community Survey (ACS) 5-Year Estimates." This data is summarized in Table 9A-1.
Table 9A-1. Census Derived Demographic Data
Race, Ethnicity, and Age Data (Census 2010 block-level SF1 data)*
Parameter Definition
Population Total population
White Number of whites (may include Hispanics)
African Americans Number of African Americans (may include Hispanics)
Native Americans Number of Native Americas (may include Hispanics)
Other and multiracial Number of other race and multiracial (may include Hispanics)
Other and multiracial Number of other race and multiracial (i
Minority/Non-white Hispanic Total Population less White Population
Age 0 to 4 Number of people age 0 to 4
Age 0 to 17 Number of people age 0 to 17
Age 65 and up Number of people age 65 and up
Economic and Education Date (2006-2010 ACS)*
Parameter Definition
Poverty Number of people living in households with income below the poverty line
„ r, . Number of people living in households with income below twice the
2 xPoverty , r
poverty line
Linguistic isolation Number of people linguistically isolated
Education level Adults without a high school diploma
* Census 2010 does not currently report this data for the Virgin Islands, Guam, American Samoa, and the Northern
Marianas; Census 2000 data are used for these areas.
As noted above, the EPA uses population data collected by the 2010 Census. All data is
stored at the block level. For those indicators available from the Census at the block group, but
not block level, the EPA assigns a block the same percentage as the block group of which it is a
part. For example, a block is assigned the same percentage of people living below the national
9A-3
-------
poverty line as the block group in which it is contained. Nationally, a census block contains
about 50 people on average; and a block group contains about 26 blocks on average, or about
1,350 people. (For comparison, a census tract is larger than a block group, with each tract
containing an average of 3 block groups, or about 4,300 people). For this analysis the data was
aggregated to the county level.
Data on race, ethnicity and age for all census blocks in the country except for the Virgin
Islands, Guam, American Samoa, and the Northern Marianas were obtained from the 2010
Census SF1 dataset. This dataset gives a breakdown of the population for each census block
among different racial and ethnic classifications, including: White, African American or Black,
Hispanic or Latino, American Indian or Native Alaskan, Asian, Native Hawaiian or other South
Pacific Islander, other race, and two or more races. Data on age distributions in the U.S. and
Puerto Rico were obtained at the census block level from the 2010 Census of Population and
Housing Summary File 1 (SF1) short form, Table P12. SF1 contains the information compiled
from the questions asked of all people about every housing unit. Data on poverty status,
education level, and linguistic isolation in the U.S. and Puerto Rico were obtained at the block
group level from the Census Bureau's 2006-2010 ACS.111 Data for the Virgin Islands and other
island territories (Guam, American Samoa, and the Northern Marianas) were retrieved from
similar tables, which are available through the Census' American Fact Finder internet portal,
www.factfmder.census.gov. "Minority" means a person, as defined by the U.S. Bureau of
Census, who is a: (1) Black American (a person having origins in any of the black racial groups
of Africa); (2) Hispanic person (a person of Mexican, Puerto Rican, Cuban, Central or South
American, or other Spanish culture or origin, regardless of race); (3) Asian American or Pacific
Islander (a person having origins in any of the original peoples of the Far East, Southeast Asia,
the Indian subcontinent, or the Pacific Islands); or (4) American Indian or Alaskan Native (a
person having origins in any of the original people of North America and maintain cultural
identification through tribal affiliation or community recognition). The Minority/Non-White
Hispanic field is computed by subtracting the white population from the total population.
Compared to white populations in the U.S., numerous studies have found that minority
populations live in closer proximity to pollution sources, experience worse health, and have less
ability to participate in environmental decision making. 1V
9A-4
-------
The percentage of people living below the national poverty line is defined by the Census
as the percentage of residents whose household income is at or below the poverty guidelines
updated periodically in the Federal Register by the U.S. Department of Health and Human
Services under the authority of 42 U.S.C. 9902(2).v Low income has been linked in many
studies to poor health, lack of access to health care, closer proximity to pollution sources, and
greater susceptibility to illnesses caused by exposure to air pollution. V1 Low income in this
appendix is defined as 2 times the national poverty line.
The percentage of residents whose age is less than 5 years includes infants and children -
all of whom are considered a sensitive subpopulation for many forms of environmental
contaminants, including air pollution/11 The reasons for children's increased sensitivity to air
pollution are manifold and include: still-developing respiratory and other bodily systems; smaller
body size in proportion to inhaled contaminants; and varying behavior patterns including longer
durations spent outdoors at high breathing rates compared to adults/111 These factors lead to
increased morbidity and mortality risks for children from exposure to air pollutants.1X
Furthermore, minority and low-income children are at even greater risk, as the increased
susceptibility for each of these demographic groups compounds such a child's overall
vulnerability/
The percentage of residents of age 65 years and over is considered a sensitive
subpopulation for many forms of environmental contamination, including air pollution/1 The
increased susceptibility of this subpopulation stems not only from their age, but also from their
poorer health and lower fitness levels/11 Those age 65 years and up have a higher mortality
risk—both long-term and short-term—than other populations from air pollution, as well as
increased morbidity risks including cardiovascular illness and respiratory disease/111
The percentage of residents lacking a high school diploma is considered a relevant
because lack of education may indirectly increase the effects of environmental contamination,
including air pollution. Low education may decrease access to health care and information about
environmental risks and how to respond appropriately to such risks.X1V Studies have revealed that
low education appears to be a factor in health disparities, multi-morbidity, and mortality in
general; and lower education may increase the relative risk of air pollution and premature
9A-5
-------
mortality.^ In particular, this association was established after examining the relationship
between an increase in particulate matter and mortality among persons with lower education.XV1
The percentage of residents experiencing linguistic isolation is considered a relevant
because linguistic isolation may render households less able to identify and mitigate
environmental harms by limiting both access to health care and access to information about
environmental risks and how to respond appropriately to those risks.™1 Studies have revealed
that counties with higher concentrations of immigrants and linguistically isolated households
have more hazardous waste generators and more proposed Superfund sites, compared to other
counties.™11 For example, a significant cancer risk was found between linguistically isolated
households and exposure from Toxics Release Inventory (TRI) releases in the San Francisco Bay
area.xlx Finally, immigrants may encounter discrimination and prejudice which impact
vulnerability.
9A.2 Considerations in Evaluating and Interpreting Results
This analysis characterizes the socio-demographic attributes of populations located in
areas defined by a county or a CBS A containing a county with a monitor design value greater
than 65 ppb and 70 ppb. Therefore, the results of this analysis can only be used to inform
whether there are differences in the composition of populations residing within these areas
relative to the nation as a whole. As noted earlier, the purpose of the analysis is to determine
whether populations of interest from an EJ perspective have a higher representation in areas that
exceed the proposed range of ozone standard levels, and thus may be more affected by strategies
to attain alternative standards. This analysis does not include a quantitative assessment of
exposure and/or risk for specific populations of potential interest from an EJ-perspective, and
therefore it cannot be used to draw any conclusions regarding potential disparities in exposure or
risk across populations of interest from an EJ perspective. Nor can it be used to draw
conclusions about any disparities in the health and environmental benefits that result from
strategies to attain alternative ozone standards.
In order to clearly identify disparities in risk between populations of interest, we would
need to conduct rigorous site-specific population-level exposure and risk assessments that take
9A-6
-------
into account short-term mobility (daily patterns of travel linked for example to school or work)
or long-term mobility (families moving into or out of specific block groups).
9A.3 Presentation of Results
This section presents a summary of the results for the assessment of demographic
characteristics of populations in areas with ozone monitors with measured values greater than the
levels of the proposed range of ozone standards. The results are also provided in tabular form.
As a whole, the demographic distributions within the areas of interest estimated for the
proposed range of standard levels (i.e., 65 ppb to 70 ppb) correspond well to the national
averages. Table 9A-2 presents these results and the raw data used in this assessment. The
population totals and subtotals by demographic group as well as the percentages of the
demographic groups for the nation and for the lower and upper end of the proposed range of
standard levels (65 ppb and 70 ppb) are shown. Most of the sub-populations are within a few
percentage points of the national average. The largest difference is between the national and
study area percentages for the Minority/Non-White Hispanic demographic group and that is a
difference of only 4%. Overall, these qualitative results support the determination that the
proposed rule will tend to benefit geographic areas that have a higher proportion of minority and
low income residents than the national average. Perhaps more importantly, these results provide
EPA much needed insight into how and with whom to best target efforts to inform communities
about the proposed rule and otherwise ensure their meaningful involvement in the rulemaking
process.
9A-7
-------
Table 9A-2 Summary of Population Totals and Demographic Categories for Areas of
Interest and National Perspective
Demographic
Summary
Population
White
African
American
Native Other or
American Multiracial
Minority/
Non-White
Hispanic"
Area of Interest Total
65 ppb
70ppb
% of Area of Interest
65 ppb
70 ppb
National Total
%of
National Total
221,431,286
193,316,836
Total
312,861,256
a The race Minority '/Non-White Hispanic
population.
Demographic Linguistically
Summary Isolated
153,706,027
132,112,738
69%
68%
226,405,205
72%
30,429,108
27,193,155
14%
14%
39,475,216
13%
field is computed by subtracting the
Age 0 - 4
Age 0 - 17
1,726,110
1,488,364
1%
1%
2,952,087
1%
white population
35,570,041
32,522,579
16%
17%
44,028,748
14%
from the total
Without a
Age 65+ HS Diploma
67,725,259
61,204,098
31%
32%
86,456,051
28%
Low Income
Area of Interest Total
65 ppb
70 ppb
% of Area of Interest
65 ppb
70 ppb
National Total
%of
National Total
13,072,109
12,179,896
Total
6%
6%
19,196,507
6%
14,695,948
12,849,637
7%
7%
20,465,065
7%
54,008,810
47,296,147
24%
24%
75,217,176
24%
27,163,990
23,518,071
12%
12%
40,830,262
13%
20,914,891
18,495,474
14%
15%
30,952,789
15%
67,027,700
58,296,224
30%
30%
101,429,436
32%
9A-8
-------
'2010 Census Summary File 1 Delivered via FTP, http://www2.census.gov/census_2010/04-Summary_File_l/
11 U.S. Census Bureau 2006-2010 American Community Survey 5-Year Estimates,
http://www.census.gov/acs/www/data_documentation/2010_release/
111 U.S. Census Bureau 2006-2010 American Community Survey 5-Year Estimates,
http://www.census.gov/acs/www/data_documentation/2010_release/
lv Brulle, Robert J. and David N. Fellow. 2006. Environmental Justice: Human Health and Environmental
Inequalities. Annu. Rev. Public Health 27:103-24. http://publhealth.annualreviews.org; Cutter, Susan. 2006. The
Geography of Social Vulnerability: Race, Class, and Catastrophe. Understanding Katrina. Social Sciences Research
Council. June 11. http://understandingkatrina.ssrc.org/Cutter/; Gold, Diane R. and Rosalind Wright. 2005.
Population Disparities in Asthma. Annu. Rev. Public Health 26:89-113; Nelson DA, Johnson CC, Divine GW,
Strauchman C, Joseph CL, Ownby DR. 1997. Ethnic Differences in the Prevalence of Asthma in Middle Class
Children. Ann. Allergy Asthma Immunol. 78:21- 26; O'Neill et al., 2003. Health, Wealth, and Air Pollution:
Advancing Theory and Methods, 111 Environmental Health Perspectives 16: 1861, 1865,
http://www.ehponline.org/docs/2003/6334/abstract.html; Friebele, E. 1996. The Attack of Asthma. Env. Health
Persp. 104:1, January, pp. 22-25.; Gwynn, R. Charon, and George D. Thurston. 2001. The Burden of Air Pollution:
Impacts among Racial Minorities. Environ Health Perspect. 109 (Suppl 4): 501-506
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1240572/pdf/ehpl09s-000501.pdf; American Lung Association.
2007. State of Lung Disease in Diverse Communities. August.; USEPA. 2004. Office of Environmental Justice.
Toolkit for Assessing Potential Allegations of Environmental Injustice. November 3; Bullard, Robert D. and et al.
2007. Toxic Wastes and Race at Twenty 1987-2007: Grassroots Struggles to Dismantle Environmental Racism in
the United States. Cleveland: The United Church of Christ. March.
v U.S. Department of Health & Human Services. Frequently Asked Questions Related to the Poverty Guidelines and
Poverty, http://aspe.hhs.gov/poverty/faq.shtml
vl USEPA. Office of Environmental Justice. 2004. Toolkit for Assessing Potential Allegations of Environmental
Injustice. November 3; Cutter, Susan. 2006. The Geography of Social Vulnerability: Race, Class, and Catastrophe.
Understanding Katrina. Social Sciences Research Council. June 11. http://understandingkatrina.ssrc.org/Cutter/;
Jiang Y, Hesser JE. 2009. Using Item Response Theory to Analyze the Relationship Between Health-Related
Quality of Life and Health Risk Factors. Prev Chronic Dis. 6(1); Brulle, Robert J. and David N. Fellow. 2006.
Environmental Justice: Human Health and Environmental Inequalities. 27:103-24.
http://publhealth.annualreviews.org; American Heart Association. Air Pollution, Heart Disease and Stroke.
http://www.americanheart.org/presenter.jhtml?identifier=4419; Friebele E. 1996. The Attack of Asthma. Env.
Health Persp. 104:1, January, pp. 22-25; Gwynn, R. Charon, and George D. Thurston. 2001. The Burden of Air
Pollution: Impacts among Racial Minorities. Environ Health Perspect. August; 109 (Suppl 4):501-506
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1240572/pdf/ehpl09s-000501.pdf; USEPA. 2008. Assessment and
Standards Division, Office of Transportation and Air Quality. Regulatory Impact Analysis: Control of Emissions of
Air Pollution from Locomotive Engines and Marine Compression Ignition Engines Less than 30 Liters Per Cylinder.
EPA420-R-08-001a, p. 2-57.
™ American Lung Association. 2009. Protect the Air we Breathe. February.
http://www.lungusa.org/assets/documents/PROTECT_THE_AIR_WE_BREATHE.pdf.
V111 USEPA. 2004. Air Quality Criteria for Paniculate Matter. RTF, NC. EPA/600/P-99/002aF-bF.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=87903; USEPA. Factors for Identifying and Assessing
Disproportionate Environmental Health Impacts; USEPA. Framework for Cumulative Risk Assessment; National
Resources Defense Council. Our Children at Risk: The 5 Worst Environmental Threats to Their Health.
http://www.nrdc.org/health/kids/ocar/chap4.asp
K USEPA. 2004. Air Quality Criteria for Paniculate Matter. Research Triangle Park, NC. EPA/600/P-99/002aF-bF.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=87903; Host S; et al. 2007. Short-term Associations Between
Fine and Coarse Particles and Cardiorespiratory Hospitalizations in Six French Cities. Occup Environ Med; Mar TF;
et al. 2004. An Analysis of the Association Between Respiratory Symptoms in Subjects with Asthma and Daily Air
Pollution in Spokane, Washington. Inhal Toxicol 16:809-815; Mauad T; et al. 2008. Chronic Exposure to Ambient
Levels of Urban Particles Affects Mouse Lung Development. Am J Respir Crit Care Med 178: 721-728; Peel JL; et
al. 2005. Ambient Air Pollution and Respiratory Emergency Department Visits. Epidemiology, 16:164-174;
Pinkerton KE; et al. 2004. Reduced Lung Cell Proliferation Following Short-Term Exposure to Ultrafme Soot and
Iron Particles in Neonatal Rats: Key to Impaired Lung Growth? Inhal Toxicol 1:73-81; National Resources Defense
Council. Our Children at Risk: The 5 Worst Environmental Threats to Their Health.
http://www.nrdc.org/health/kids/ocar/chap4.asp; Gauderman, et al. 2004. The Effect of Air Pollution on Lung
9A-9
-------
Development from 10 to 18 Years of Age. NEngl J Med 351:1057-1067. September 9.
http://www.nejm.org/doi/full/10.1056/NEJMoa040610?ck=nck&#articleTop; Neidell, Matthew J. 2004. Air
Pollution, Health, and Socio-economic Status: the Effect of Outdoor Air Quality on Childhood Asthma. Journal of
Health Economics 23: pp. 1209-1236.
x O'Neill et al. 2003. Health, Wealth, and Air Pollution: Advancing Theory and Methods. Environmental Health
Perspectives 16: pp.1861, 1865. http://www.ehponline.org/docs/2003/6334/abstract.html; Gold, etal. 1993. Race
and Gender Differences in Respiratory Illness Prevalence and Their Relationship to Environmental Exposures in
Children 7 to 14 Years of Age. Am Rev Respir Dis. 148(1): 10-8; Neidell, Matthew J. 2004. Air Pollution, Health,
and Socio-economic Status: the Effect of Outdoor Air Quality on Childhood Asthma. Journal of Health Economics
23: pp. 1209-1236.
xl American Lung Association. 2009. Protect the Air we Breathe. February.
xl1 USEPA. Framework for Cumulative Risk Assessment.
xm Gerber et al. 2010. Cumulative Exposure to Air Pollution and long term outcomes after first acute myocardial
infarction: A population-based cohort study. BMC Public Health 10:369. http://www.biomedcentral.com/1471-
2458/10/369; American Heart Association. Air Pollution, Heart Disease and Stroke.
http://www.americanheart.org/presenter.jhtml?identifier=4419; National Institute of Environmental Health and
Sciences. Elderly Have Higher Risk for Cardiovascular, Respiratory Disease From Fine Particle Pollution.
http://www.nih.gov/news/pr/mar2006/niehs-08.htm; National Resources Defense Council. Our Children at Risk:
The 5 Worst Environmental Threats to Their Health, http://www.nrdc.org/health/kids/ocar/chap4.asp; Franklin M;
Zeka A; Schwartz J. 2007. Association Between PM2.5 and All-Cause and Specific-Cause Mortality in 27 US
Communities. Journal of Exposure Science and Environmental Epidemiology, 17:279-87.;
http://www.scientificintegrityinstitute.org/JESEE2005.pdf; Ostro B, et al. 2006. Fine Particulate Air Pollution and
Mortality in Nine California Counties: Results from CALFINE. Environ Health Perspect, 114:29-33.;
http://cfpub.epa.gov/ncea/hero/index.cfm?action=search.view&reference_id=87991; SamoliE; etal. 2008. Acute
Effects of Ambient Particulate Matter on Mortality in Europe and North America: Results From the APHENA
Study. Environ Health Perspect 116:1480-1486.
http://cfpub.epa.gov/ncea/hero/index.cfm?action=search.view&reference_id=188455; Zeka A; Zanobetti A;
Schwartz J. 2006. Individual-Level Modifiers of the Effects of Particulate Matter on Daily Mortality. Am J
Epidemiol, 163:849-859. http://cfpub.epa.gov/ncea/hero/index.cfm?action=search.view&reference_id=88749.
X1V USEPA. 2004. National Environmental Justice Advisory Council. Ensuring Risk Reduction in Communities with
Multiple Stressors: Environmental Justice and Cumulative Risks/Impacts.
xv DeWalt, Darren A., Marylee H. Dilling, Marjorie S. Rosenthal, and Michael P. Pignone. 2007. Low Parental
Literacy Associated with Worse Asthma Care Measures in Children. Ambul Pediatr 7(1): 25-31.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1797805/pdf/nihmsl7216.pdf; Banks J, Marmot M, Oldfield Z,
Smith JP. 2006. Disease and disadvantage in the United States and in England. JAMA 295(17):2037-45; Nagel,
Gabriele, Richard Peter, Stefanie Braig, Silke Hermann, Sabine Rohrmann and Jakob Linseisen. 2008. The Impact
of Education on Risk Factors and the Occurrence of Multimorbidity in the EPIC-Heidelberg Cohort. BMC Public
Health 8:384. http://www.biomedcentral.eom/1471-2458/8/384; AdlerNE, and K Newman. 2002. Socioeconomic
Disparities in Health: Pathways and Policies. Heath Aff (Millwood) 21(2):60-76; Kan, et al. 2008. Season, Sex, Age,
and Education as Modifiers of the Effects of Outdoor Air Pollution on Daily Mortality in Shanghai, China: The
Public Health and Air Pollution in Asia (PAPA) Study. Environ Health Perspect, 116:1183 -1188; Levy, Jonathon I.,
Susan L. Greco, John D. Spengler. 2002. The Importance of Population Susceptibility for Air Pollution Risk
Assessment: A Case Study of Power Plants Near Washington, DC. Environmental Health Perspectives, Dec.
http://www.ncbi.nlm.nih.gov/pubmed/12460806 .
XV1 Krewski D, Burnett RT, Goldberg MS, Hoover K, Siemiatycki J, Jerrett M, et al. 2000. Reanalysis of the Harvard
Six Cities Study and the American Cancer Society Study of Particulate Air Pollution and Mortality. Health Effects
Institute. Cambridge, MA; Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III;
Thurston G; Calle EE; Thun MJ. 2009. Extended follow-up and spatial analysis of the American Cancer Society
study linking particulate air pollution and mortality. Health Effects Institute. Cambridge, MA. 140; Ostro B;
Broadwin R; Green S; Feng W-Y; Lipsett M. 2006. Fine particulate air pollution and mortality in nine California
counties: results from CALFINE. Environ Health Perspect, 114:29-33.; Ostro BD; Feng WY; Broadwin R; Malig
BJ; Green RS; Lipsett MJ. 2008. The impact of components of fine particulate matter on cardiovascular mortality in
susceptible subpopulations. Occup Environ Med, 65: 750-756; Zeka A; Zanobetti A; Schwartz J. 2006. Individual-
level modifiers of the effects of particulate matter on daily mortality. Am J Epidemiol, 163: 849-859.
xvu Asian Pacific Environmental Network. Environmental Justice and API Issues.
http://www.apen4ej.org/issues_api.htm; West Coast Poverty Center. 2010. Ethnic Residential Clustering and Health
9 A-10
-------
in West Coast States. Dialogues on Research and Policy. No. 2, March.
http://depts.washington.edu/wcpc/sites/default/files/DIALOGUE%20No.%202.Ethnic%20Enclaves.webFN_.pdf;
LeClere, E B., L. Jensen and A. E. Biddlecom. 1994. Health Care Utilization, Family Context, and Adaptation
among Immigrants to the United States. Journal of Health and Social Behavior, 35 (4):370-384. December; Arias,
Beatriz. 2007. School Desegregation, Linguistic Segregation and Access to English for Latino Students. Journal of
Education Controversy 2(1) Winter. http://www.wce.wwu.edu/Resources/CEP/eJournal/v002n001/a008.shtml.
xvm San Francisco Department of Public Health. 2004. Unaffordable Housing: the Costs to Public Health. June.
http://www.sfphes.org/publications/reports/UnaffordableHousingJune04.pdf; Pinderhughes, Racquel Rivera. 1996.
Social and Economic Inequality in San Francisco: A Case Study of Environmental Risks in the City's Mission
District. Based on a presentation made at the Conference on Urban Environmental Issues in the Bay Area. Hastings
West-Northwest Journal of Environmental Law and Policy. Second Annual Environmental Law Symposium. March
23. http:^ss.sfsu.edu/raquelrp/pub/1996_march_pub.html; Hunter, Lori M. 2000. The Spatial Association between
U.S. Immigrant Residential Concentration and Environmental Hazards. International Migration Review. 34(2): pp.
460-488. http://www.jstor.org/stable/2675910.
XK Pastor, Manuel, James Sadd, Rachel Morello-Frosch. 2007. Still Toxic After All These Years: Air Quality and
Environmental Justice in the San Francisco Bay Area. Center for Justice, Tolerance & Community. February.
http://cjtc.ucsc.edu/docs/bay_fmal.pdf.
9 A-11
-------
CHAPTER 10: QUALITATIVE DISCUSSION OF EMPLOYMENT IMPACTS OF AIR
QUALITY
Overview
Executive Order 13563 directs federal agencies to consider regulatory impacts on job
creation and employment: "our regulatory system must protect public health, welfare, safety, and
our environment while promoting economic growth, innovation, competitiveness, and job
creation. It must be based on the best available science". Although benefit-cost analyses do not
typically include a separate analysis of regulation-induced employment impacts,144 during
periods of sustained high unemployment, such impacts are of particular concern and questions
may arise about their existence and magnitude. This chapter discusses some, but not all, possible
types of labor impacts that may result from measures to decrease NOx emissions.
Section 10.1 describes the theoretical framework used to analyze regulation-induced
employment impacts, discussing how economic theory alone cannot predict whether such
impacts are positive or negative. Section 10.2 presents an overview of the peer-reviewed
literature relevant to evaluating the effect of environmental regulation on employment. Section
10.3 discusses employment related to installation of NOx controls on coal and gas-fired electric
generating units, industrial boilers, and cement kilns.
10.1 Economic Theory and Employment
Regulatory employment impacts are difficult to disentangle from other economic changes
affecting employment decisions over time and across regions and industries. Labor market
responses to regulation are complex. They depend on labor demand and supply elasticities and
possible labor market imperfections (e.g., wage stickiness, long-term unemployment, etc). The
unit of measurement (e.g., number of jobs, types of job hours worked, and earnings) may affect
observability of that response. Net employment impacts are composed of a mix of potential
declines and gains in different areas of the economy (the directly regulated sector, upstream and
144 Labor expenses do, however, contribute toward total costs in the EPA's standard benefit-cost analyses.
10-1
-------
downstream sectors, etc.) over time. In light of these difficulties, economic theory provides a
constructive framework for analysis.
Microeconomic theory describes how firms adjust input use in response to changes in
economic conditions.145 Labor is one of many inputs to production, along with capital, energy,
and materials. In competitive markets, firms choose inputs and outputs to maximize profit as a
function of market prices and technological constraints.146'147
Berman and Bui (2001) and Morgenstern, Pizer, and Shih (2002), adapt this model to
analyze how environmental regulations affect labor demand.148 They model environmental
regulation as effectively requiring certain factors of production, such as pollution abatement
capital, at levels that firms would not otherwise choose.
Berman and Bui (2001, pp. 274-75) model two components that drive changes in firm-
level labor demand: output effects and substitution effects.149 Regulation affects the profit-
maximizing quantity of output by changing the marginal cost of production. If regulation causes
marginal cost to increase, it will place upward pressure on output prices, leading to a decrease in
demand, and resulting in a decrease in production. The output effect describes how, holding
labor intensity constant, a decrease in production causes a decrease in labor demand. As noted by
Berman and Bui, although many assume that regulation increases marginal cost, it need not be
the case. A regulation could induce a firm to upgrade to less polluting and more efficient
equipment that lowers marginal production costs. In such a case, output could increase for
facilities that do not exit the industry. For example, improving the heat rate of a utility boiler
increases fuel efficiency, lowering marginal production costs, and thereby potentially increasing
145 See Layard and Walters (1978), a standard microeconomic theory textbook, for a discussion, in Chapter 9.
146 See Hamermesh (1993), Ch. 2, for a derivation of the firm's labor demand function from cost-minimization.
147 In this framework, labor demand is a function of quantity of output and prices (of both outputs and inputs).
148 Berman and Bui (2001) and Morgenstern, Pizer, and Shih (2002) use a cost-minimization framework, which is a
special case of profit-maximization with fixed output quantities.
149 The authors also discuss a third component, the impact of regulation on factor prices, but conclude that this effect
is unlikely to be important for large competitive factor markets, such as labor and capital. Morgenstern, Pizer and
Shih (2002) use a very similar model, but they break the employment effect into three parts: 1) a demand effect; 2) a
cost effect; and 3) a factor-shift effect.
10-2
-------
the boiler's generation. An unregulated profit-maximizing firm may not have chosen to install
such an efficiency-improving technology if the investment cost were too high.
The substitution effect describes how, holding output constant, regulation affects labor-
intensity of production. Although stricter environmental regulation may increase use of pollution
control equipment and energy to operate that equipment, the impact on labor demand is
ambiguous. Equipment inspection requirements, specialized waste handling, or pollution
technologies that alter the production process may affect the number of workers necessary to
produce a unit of output. Berman and Bui (2001) model the substitution effect as the effect of
regulation on pollution control equipment and expenditures required by the regulation and the
corresponding change in labor-intensity of production.
In summary, as output and substitution effects may be positive or negative, theory cannot
predict the direction of the net effect of regulation on labor demand at the level of the regulated
firm. Operating within the bounds of standard economic theory, however, empirical estimation
of net employment effects on regulated firms is possible when data and methods of sufficient
detail and quality are available. The literature, however, illustrates difficulties with empirical
estimation. For example, studies sometimes rely on confidential plant-level employment data
from the U.S. Census Bureau, possibly combined with pollution abatement expenditure data that
are too dated to be reliably informative. In addition, the most commonly used empirical methods
do not permit estimation of net national effects.
The conceptual framework described thus far focused on regulatory effects on plant-level
decisions within a regulated industry. Employment impacts at an individual plant do not
necessarily represent impacts for the sector as a whole. The approach must be modified when
applied at the industry level.
At the industry-level, labor demand is more responsive if: (1) the price elasticity of
demand for the product is high, (2) other factors of production can be easily substituted for labor,
(3) the supply of other factors is highly elastic, or (4) labor costs are a large share of total
10-3
-------
production costs.150 For example, if all firms in an industry are faced with the same regulatory
compliance costs and product demand is inelastic, then industry output may not change much,
and output of individual firms may change slightly.151 In this case the output effect may be small,
while the substitution effect depends on input substitutability. Suppose, for example, that new
equipment for heat rate improvements requires labor to install and operate. In this case the
substitution effect may be positive, and with a small output effect, the total effect may be
positive. As with potential effects for an individual firm, theory cannot determine the sign or
magnitude of industry-level regulatory effects on labor demand. Determining these signs and
magnitudes requires additional sector-specific empirical study. For environmental rules, much of
the data needed for these empirical studies are not publicly available, would require significant
time and resources in order to access confidential U.S. Census data for research, and also would
not be necessary for other components of a typical RIA.
In addition to changes to labor demand in the regulated industry, net employment impacts
encompass changes in other related sectors. For example, the proposed guidelines may increase
demand for pollution control equipment and services. This increased demand may increase
revenue and employment in the firms supporting this technology. At the same time, the regulated
industry is purchasing the equipment and these costs may impact labor demand at regulated
firms. Therefore, it is important to consider the net effect of compliance actions on employment
across multiple sectors or industries.
If the U.S. economy is at full employment, even a large-scale environmental regulation is
unlikely to have a noticeable impact on aggregate net national employment.152 Instead, labor
would primarily be reallocated from one productive use to another (e.g., from producing
electricity or steel to producing high efficiency equipment), and net national employment effects
150 See Ehrenberg & Smith, p. 108.
151 This discussion draws from Herman and Bui (2001), pp. 293.
152 Full employment is a conceptual target for the economy where everyone who wants to work and is available to
do so at prevailing wages is actively employed. The unemployment rate at full employment is not zero.
10-4
-------
from environmental regulation would be small and transitory (e.g., as workers move from one
job to another).153
Affected sectors may experience transitory effects as workers change jobs. Some workers
may retrain or relocate in anticipation of new requirements or require time to search for new
jobs, while shortages in some sectors or regions could bid up wages to attract workers. These
adjustment costs can lead to local labor disruptions. Although the net change in the national
workforce is expected to be small, localized reductions in employment may adversely impact
individuals and communities just as localized increases may have positive impacts.
If the economy is operating at less than full employment, economic theory does not clearly
indicate the direction or magnitude of the net impact of environmental regulation on
employment; it could cause either a short-run net increase or short-run net decrease
(Schmalansee and Stavins, 2011). An important research question is how to accommodate
unemployment as a structural feature in economic models. This feature may be important in
assessing large-scale regulatory impacts on employment (Smith 2012).
Environmental regulation may also affect labor supply. In particular, pollution and other
environmental risks may impact labor productivity or employees' ability to work.154 While the
theoretical framework for analyzing labor supply effects is analogous to that for labor demand, it
is more difficult to study empirically. There is a small emerging literature, described in the next
section that uses detailed labor and environmental data to assess these impacts.
To summarize, economic theory provides a framework for analyzing the impacts of
environmental regulation on employment. The net employment effect incorporates expected
employment changes (both positive and negative) in the regulated sector and elsewhere. Labor
demand impacts for regulated firms, and also for the regulated industry, can be decomposed into
output and substitution effects which may be either negative or positive. Estimation of net
employment effects for regulated sectors is possible when data of sufficient detail and quality are
153 Arrow et. al. 1996; see discussion on bottom of p. 8. In practice, distributional impacts on individual workers can
be important, as discussed in later paragraphs of this section.
154 E.g. Graff Zivin and Neidell (2012).
10-5
-------
available. Finally, economic theory suggests that labor supply effects are also possible. In the
next section, we discuss the empirical literature.
10.2 Current State of Knowledge Based on the Peer-Reviewed Literature
The labor economics literature contains an extensive body of peer-reviewed empirical
work analyzing various aspects of labor demand, relying on the theoretical framework discussed
in the preceding section.155 This work focuses primarily on effects of employment policies such
as labor taxes and minimum wages.156 In contrast, the peer-reviewed empirical literature
specifically estimating employment effects of environmental regulations is more limited.
Empirical studies, such as Berman and Bui (2001), suggest that net employment impacts
were not statistically different from zero in the regulated sector. Other research suggests that
more highly regulated counties may generate fewer jobs than less regulated ones (Greenstone
2002). Environmental regulations may affect sectors that support pollution reduction earlier than
the regulated industry. Rules are usually announced well in advance of their effective dates and
then typically provide a period of time for firms to invest in technologies and process changes to
meet the new requirements. When a regulation is promulgated, the initial response of firms is
often to order pollution control equipment and services to enable compliance when the regulation
becomes effective. Estimates of short-term increases in demand for specialized labor within the
environmental protection sector have been prepared for several EPA regulations in the past,
including the Mercury and Air Toxics Standards (MATS).157 Overall, the peer-reviewed
literature does not contain evidence that environmental regulation has a large impact on net
employment (either negative or positive) in the long run across the whole economy.
10.2.1 Regulated Sectors
Berman and Bui (2001) examine how an increase in local air quality regulation affects
manufacturing employment in the South Coast Air Quality Management District (SCAQMD),
which includes Los Angeles and its suburbs. From 1979 to 1992 the SCAQMD enacted some of
155 Again, see Hamermesh (1993) for a detailed treatment.
156 See Ehrenberg & Smith (2000), Chapter 4: "Employment Effects: Empirical Estimates" for a concise overview.
"U.S. EPA (201 Ib).
10-6
-------
the country's most stringent air quality regulations. Using SCAQMD's local air quality
regulations, Berman and Bui identify the effect of environmental regulations on net employment
in regulated manufacturing industries relative to other plants in the same 4-digit SIC industries
but in regions not subject to local regulations.158 The authors find that "while regulations do
impose large costs, they have a limited effect on employment" (Berman and Bui, 2001, p. 269).
Their conclusion is that local air quality regulation "probably increased labor demand slightly"
but that "the employment effects of both compliance and increased stringency are fairly precisely
estimated zeros, even when exit and dissuaded entry effects are included" (Berman and Bui,
2001, p. 269).159
A small literature examines impacts of environmental regulations on manufacturing
employment. Kahn and Mansur (2013) study environmental regulatory impacts on geographic
distribution of manufacturing employment, controlling for electricity prices and labor regulation
(right to work laws). Their methodology identifies employment impacts by focusing on
neighboring counties with different ozone regulations. They find limited evidence that
environmental regulations may cause employment to be lower within "county-border-pairs."
This result suggests that regulation may cause an effective relocation of labor across a county
border, but since one county's loss may be another's gain, such shifts cannot be transformed into
an estimate of a national net effect on employment. Moreover this result is sensitive to model
specification choices.
10.2.2 Labor Supply Impacts
The empirical literature on environmental regulatory employment impacts focuses
primarily on labor demand. However, there is a nascent literature focusing on regulation-induced
effects on labor supply. 16° Although this literature is limited by empirical challenges, researchers
have found that air quality improvements lead to reductions in lost work days (e.g., Ostro 1987).
Limited evidence suggests worker productivity may also improve when pollution is reduced.
158 Berman and Bui include over 40 4-digit SIC industries in their sample. They do not estimate the number of jobs
created in the environmental protection sector.
159 Including the employment effect of existing plants and plants dissuaded from opening will increase the estimated
impact of regulation on employment.
leo por a recent review see Graff-Zivin and Neidell (2013).
10-7
-------
Graff Zivin and Neidell (2012) used detailed worker-level productivity data from 2009 and 2010,
paired with local ozone air quality monitoring data for one large California farm growing
multiple crops, with a piece-rate payment structure. Their quasi-experimental structure identifies
an effect of daily variation in monitored ozone levels on productivity. They find "ozone levels
well below federal air quality standards have a significant impact on productivity: a 10 parts per
billion (ppb) decreases in ozone concentrations increases worker productivity by 5.5 percent."
(Graff Zivin and Neidell, 2012, p. 3654).161
This section has outlined the challenges associated with estimating regulatory effects on
both labor demand and supply for specific sectors. These challenges make it difficult to estimate
net national employment estimates that would appropriately capture the way in which costs,
compliance spending, and environmental benefits propagate through the economy. Quantitative
estimates are further complicated by the fact that macroeconomic models often have little
sectoral detail and usually assume that the economy is at full employment. The EPA is currently
seeking input from an independent expert panel on modeling economy-wide regulatory impacts,
including employment effects.162
10.3 Employment Related to Installation and Maintenance of NOx Control Equipment
This section discusses employment related to installation of NOx controls on coal and
gas-fired electric generating units, industrial boilers, and cement kilns, which are among the
highest NOx-emitting source categories in EPA's emissions inventory (see chapter 3 for more
detail on emissions). Sections 10.3.1 and 10.3.2 below contain estimates of the number of direct
short-term and long-term jobs that would be created by addition of NOx controls at these three
categories of emissions sources, for various size units. Because the apportionment of emissions
control across emissions sources in this RIA analysis is illustrative and not necessarily
representative of the controls that will be required in individual state SIPs, EPA did not estimate
161 The EPA is not quantifying productivity impacts of reduced pollution in this rulemaking using this study. In light
of this recent research, however, the EPA is considering how best to incorporate possible productivity effects in the
future.
162 For further information see: .
10-8
-------
short-term or long-term employment that would result from addition of NOx controls at these
three source categories at the national level.
10.3.1 Employment Resulting from Addition of NOx Controls at EGUs
Coal-fired EGUs are likely to apply additional NOx controls in response to State
Implementation Plans (SIPs) approved pursuant to a revised ozone standard. While many EGUs
have already installed and operate various NOx control devices, there are additional existing
coal-fired EGUs that could decrease NOx emissions by installing or upgrading their NOx
reducing systems. While all existing coal-fired EGUs already have low NOx burners, there are
EGUs that could have a selective catalytic reduction (SCR) system installed, or could improve
their NOx emissions by replacing an existing selective non-catalytic reduction (SNCR) system
with an SCR system. The EPA identified 145 existing coal-fired EGUs, with a total of 51.0 GW
of capacity, that (1) are in areas anticipated to need additional NOx reductions under an
alternative ozone standard of 65 ppb , and (b) do not already have an SCR emission control
system. (For an alternative ozone standard of 70 ppb, there are 15 EGUs so identified, with a
total of 7.4 GW of capacity.) While there are currently SNCR systems in use that could be
upgraded to an SCR system, the EPA's 2025 baseline analysis163 estimates that the remaining
SNCR systems will already be upgraded by 2025 in response to existing emission control
programs.
The EPA used a bottom up engineering analysis using data on labor productivity,
engineering estimates of the types of labor needed to manufacture, construct and operate SCRs
on EGUs. The EPA's labor estimates include not only labor directly involved with installing
SCRs on EGUs and on-site labor used to operate the SCRs once they become operational, but
also include the labor requirements in selected major upstream sectors directly involved
manufacturing the materials used in SCR systems (steel), as well as the chemicals used to
operate an SCR system (ammonia and the catalyst used to in the construction and operation of
SCR systems, including such as steel, concrete, or chemicals used to manufacture NOx controls.
163 The 2025 baseline used in this illustrative analysis incorporates the "state only" implementation option used in
the proposed carbon pollution guidelines for existing power plants and emission standards for modified and
reconstructed power plants (a.k.a. the proposed Clean Power Plan, June, 2013).
10-9
-------
This section presents an illustrative analysis of the direct labor needs to install and
operate SCRs at 3 common sizes of coal-fired EGUs: 300 MW, 500 MW and 1000 MW. As
discussed below, the illustrative analysis is for a "model plant" of each size, using consistent
assumptions about the plant's operation that impact the material and labor needs of an
representative plant such as the capacity factor, heat rate, and type of coal. The analysis does not
include an estimate of the aggregate total of the labor needed for installing and running SCRs in
any particular level of the revised ozone standard, nor does it reflect plant-specific variations in
labor needs due to regional differences in prices and labor availability, existing control
technology at the plant, etc.
The analysis draws on information from four primary sources:
• Documentation for EPA Base Case v.5.13 Using the Integrated Planning Model.
November, 2013
• "ENGINEERING AND ECONOMIC FACTORS AFFECTING THE INSTALLATION
OF CONTROL TECHNOLOGIES: An Update". By James E. Staudt, Andover
Technology Partners. December, 2011.
• "Regulatory Impact Analysis (RIA) for the final Transport Rule". June 2011
• "Regulatory Impact Analysis for the Proposed Carbon Pollution Guidelines for Existing
Power Plants and Emission Standards for Modified and Reconstructed Power Plants".
June 2013
10.3.1.1 Existing EGUs Without SCR Systems
The EPA identified 145 existing coal-fired EGU units that are estimated to continue to be
in operation in 2025 in the baseline that are located in areas considered likely to be affected by
State Implementation Plans developed for a 65 ppb alternative ozone standard. The size
distribution of the 145 units is shown in Figure 10-1. The 145 units have a total generating
capacity of 51.0 GW and are anticipated to generate 282,000 GWh of electricity in 2025. With
the current level of NOx controls installed (or anticipated to be installed by 2025 to meet existing
environmental regulations), these 145 units are estimated to emit 290.7 tons of NOx in 2025.
The following key assumptions are used to estimate the amount of labor needed to install
and operate individual SCR systems of various sizes.
10-10
-------
Size Distribution of 145 Existing Coal-Fired ECU
Units Without SCR NOx Controls
40
35
30
| 25
^ 20
M—
0 15
* ±D
10
.Illll..
Figure 10-1. Size Distribution of 145 Existing Coal-Fired EGU Units without SCR NOx
Controls
10.3.1.2 Labor Estimates for Installing and Operating Individual SCR Systems
All labor estimates in this illustrative analysis are in terms of person-years (i.e., full time
equivalents, or FTEs).
The labor involved with manufacturing and installing the SCRs is a one-time labor need,
and occurs over a 2 to 3 year construction period; the estimated FTEs during the construction
phase are presented as the cumulative amount of labor over the multi-year period. The
construction phase labor includes both labor directly involved with installing the SCR on site
(including boiler makers, general labor and engineering).
There are three types of annual labor estimated to operate an SCR, and will be needed
each year the EGU is in operation. The largest category is on-site labor at the EGU. The
estimated amounts of direct labor involved with installing SCR systems is shown in Table 10-1.
Table 10-1. Summary of Direct Labor Impacts for SCR Installation at EGUs
Plant Size
300 MW 500 MW 1000 MW
Construction Phase
(One time, Total Labor over 2-3 Year Period)
Direct Construction-related
Employment 158.7 264.4
528.8
Operation Phase (Annual Operations)
c c
10-11
-------
Plant Size
Operation and Maintenance
1.9
4.6
The key assumptions used in the labor analysis are presented in Table 10-2.
Table 10-2. Key Assumptions in Labor Analysis for EGUs
Assumptions
Capital Investment
to Install SCR
Result: FTEs to
Install an SCR
Key Factor
Utility-owned Capital
Recovery Rate for
Environmental
Retrofits (12.1%)
1,100 hours/MW
Source 300 MW
IPM 5/13 Base $86'1
Case milllon
Documentation
158.65
Staudt, 2011
500 MW 1000 MW
$ 1 3 3 million $244 million
264.42 528.85
Labor Cost (fixed
O&M) per Year
Result: FTEs per
Year
Result: Total
FTES to Operate
an SCR Annually
8.9 FTEs per $1
million of Fixed O&M
IPM analysis $218000
ofCPP
baseline
1.95
CSAPR RIA
1.95
$310,500 $513,000
2.76 4.57
2.76 4.57
10.3.2 Assessment of Employment Impacts for Individual Industrial, Commercial, and
Institutional (ICI) Boilers and Cement Kilns
Facilities other than electric power generators are likely to apply NOx controls in
response to State Implementation Plans (SIPs) approved pursuant to a revised ozone standard. In
addition to EGUs, the EPA estimated the amount and types of direct labor that might be used to
apply and operate NOX controls for ICI boilers and for cement kilns. As with EGUs, the EPA
used a bottom up engineering analysis using data on labor productivity, engineering estimates of
the types of labor needed to manufacture, construct and operate NOx controls on ICI boilers and
cement kilns. No estimates were made for labor requirements in upstream sectors such as steel,
concrete, or chemicals used to manufacture or search as inputs to NOx controls. In addition, the
numbers presented in this section are only indicative of the relative number and types of labor
10-12
-------
that might be used at these two categories of plants, without calculating an estimate of the labor
that would be required by them in the aggregate (SC&A, 2014).
10.3.2.1 ICI Boilers
There are a number of control technologies available to reduce NOx emissions from ICI
boilers. The EPA anticipates that the most commonly applied control technology for ICI boilers
that could require NOx reductions as part of an ozone SIP will be selective catalytic reduction
(SCR). The analysis calculates s labor requirements to fabricate, install, and operate different
sizes of SCR for coal, oil and natural gas ICI boilers. Estimated total labor costs are a function
of total capital costs and boiler size in EPA's Coal Utility Environmental Cost (CUECost)
model. Total SCR capital costs of ICI boilers was estimated using the EPA's Control Strategy
tool (CoST) model. Labor is estimated to be about 50% of the total capital costs of an SCR.
(SC&A, 2014).
Just over 24% of total capital costs are for labor used in SCR fabrication. This
percentage was multiplied by the total capital cost, and the resulting dollar amount was
converted into full time equivalents (FTE) based on the average annual salary of workers (as
outlined in IEC, 2011). The annual compensation came from the Bureau of Labor Statistics
(BLS). This salary number was adjusted to account for benefits also based on BLS data. The
total fabrication expenditures were divided by the average fabrication labor compensation to
estimate the number of full time equivalent workers in SCR fabrication.
The calculation of construction or installation labor is based on previous research on
labor required for SCR installation at utility boilers. (Staudt 2011). Based on that, we estimate
that 27% of SCR capital costs are spent on installation labor. We applied that percentage to the
estimates of the capital costs of SCR for ICI boilers to give us the total labor expenditures, which
we then converted to FTE based on average annual compensation provided by BLS.
Operation and Maintenance labor was estimated using the CUECost model. Maintenance
and administrative labor for SCR is estimated to be small in relation to fabrication and
construction, with the caveat that available information on which to base an estimate is sparse.
According to the approach used in the CUECost model, most utility boilers require a full time
worker to operate and maintain the equipment. ICI boilers are much smaller and so are likely to
require less than one FTE. Table 10-3 below provides summary labor estimates for SCR at
varying sized ICI boilers.
10-13
-------
Table 10-3. Summary
Plant Type
Coal-fired
Oil-fired
Natural Gas-fired
of Direct Labor
Boiler Size
(MMBtu/hr)
750
500
400
250
250
150
100
50
250
150
100
50
Impacts for Individual ICI Boilers
One-Time Employment
Impacts1 (Annual
FTEs)
19.5
15.2
13.6
10.7
9.8
7.3
5.5
3.2
10.5
11.0
8.4
6.5
Recurring Annual
Employment Impacts2
(FTEs per year)
1.2
1.1
1.0
0.9
0.9
0.9
0.8
0.8
0.9
0.9
0.9
0.8
1. Includes Fabrication and Installation Labor
2. Includes Operations, Maintenance, and Administrative Support
10.3.2.2 Cement Kilns
There are a number of technologies that can be used to control NOx emissions at cement
kilns. The analysis focused on synthetic non-catalytic reduction (SNCR) as the most likely
choice for future NOx controls at cement kilns affected by requirements in ozone SIPs.
Although SNCR is not considered an appropriate technology for wet and long dry kilns, most
new or recently constructed kilns will likely be preheater and precalciner kilns, and these kilns
will likely operate using SNCR as a control technology.
Fabrication capital cost was estimated for an SNCR system for a mid-sized preheater and
precalciner kiln (125 to 208 tons of clinker per hour). The percent of capital cost of these
systems attributable to labor is 44% based on vendor supplied estimates. (Wojichowski, 2014).
This labor cost was converted to FTE using BLS data. A similar methodology was used to
estimate installation labor. Labor costs for SNCR installation was estimated by the vendor to be
17% of the capital cost. That was converted to FTE using BLS data. This information is
summarized in Table 10-4.
10-14
-------
Table 10-4. Estimated Direct Labor Impacts for Individual SNCR Applied to a Mid-
Sized Cement Kiln (125-208 tons clinker/hr)
Kiln Type Preheater / Precalciner
Manufacturing FTE L5
Installation FTE 0.9
O&M Annual Recurring FTE 13
10.4 References
Bureau of Labor Statistics, 2011. "National Compensation Survey: Occupational Earnings in the United States,
2010." Located online at: http://www.bls.gov/ncs/ncswage2010.htm#Wage_Tables
Bureau of Labor Statistics, 2014. "Employer Costs for Employee Compensation". Released September 10, 2014.
Located online at: http://www.bls.gov/news.release/ecec.toc.htm
U.S. Environmental Protection Agency (U.S. EPA), 2002: USEPA Office of Air Quality Planning and Standards,
EPA Air Pollution Control Cost Manual, Sixth Edition, EPA/452/B-02-001, Research Triangle Park, NC,
January 2002.
U.S. Environmental Protection Agency (U.S. EPA), 2007: Alternative Control Techniques Document Update -
NOx Emissions from New Cement Kilns, EPA-453/R-07-006, prepared by EPA and Research Triangle Institute
for EPA Office of Air Quality Planning and Standards, Sector Policies and Programs Division, November 2007.
U.S. Environmental Protection Agency (U.S. EPA). 2011. Regulatory Impact Assessment for the Cross State Air
Pollution Rule. Office of Air Quality Planning and Standards, Research Triangle Park, NC. Available at
http://www.epa.gov/airtransport/pdfs/FinalRIA.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2014. Regulatory Impact Analysis for the Clean Power Plan.
Office of Air and Radiation, Washington, D.C. Available at:
http://www.epa.gOv/ttn/ecas/regdata/RIAs/l 1 ldproposalRIAfmal0602.pdf
SC&A,2014:, Assessment of Employment Impacts of NOx Controls. Memorandum prepared to Jason Price, lEc and
Ellen Kurlansky, US EPA., prepared by J. Wilson; M Mullen; and j Schreiber, SC&A, Inc. September 2014
Staudt, 2011: Staudt, James E., Engineering and Economic Factors Affecting the Installation of Control
Technologies—An Update, Andover Technology Partners, North Andover, MA, December 15, 2011.
Wojichowski, 2014: David Wojichowski, De-Nox Technologies, LLC, SNCR cost estimates provided to Maureen
Mullen, SC&A via email, September 25, 2014.
10-15
-------
United States Office of Air Quality Planning and Standards Publication No.
Environmental Protection Health and Environmental Impacts Division EPA-452/P-14-006
Agency Research Triangle Park, NC November 2014
------- |