Regulatory Impact Analysis for the Final
Revisions to the National Ambient Air Quality
Standards for Particulate Matter
On February 28, 2013, the PM RIA is being replaced to
add Appendix 3.A, which was inadvertently left out and
to correct the document number to EPA-452/R-12-005.
-------
EPA-452/R-12-005
December 2012
Regulatory Impact Analysis for the Final Revisions to the National Ambient Air Quality Standards
for Particulate Matter
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, NC 27711
-------
CONTENTS
Section Page
Executive Summary ES-1
ES.l Overview ES-1
ES.2 Existing and Revised PM Air Quality Standards ES-1
ES.2.1 Overview of the Analytical Steps in this RIA ES-2
ES.2.2 Health and Welfare Co-Benefits ES-10
ES.2.3 Cost Analysis Approach ES-13
ES.2.4 Comparison of Benefits and Costs ES-14
ES.3 Discussion and Conclusions ES-17
ES.4 Caveats and Limitations ES-21
ES.4.1 Benefits Caveats ES-21
ES.4.2 Control Strategy and Cost Analysis Caveats and Limitations ES-22
ES.5 Important Updates and Analytical Differences Between the PM NAAQS
Proposal RIAand the Final RIA ES-23
ES.6 References ES-25
Chapter 1 Introduction and Background 1-1
1.1 Synopsis 1-1
1.2 Background 1-1
1.2.1 NAAQS 1-1
1.2.2 2006 PM NAAQS 1-2
1.3 Role of this RIA in the Process of Setting the NAAQS 1-2
1.3.1 Legal Requirement 1-2
1.3.2 Role of Statutory and Executive Orders 1-3
1.3.3 The Need for National Ambient Air Quality Standards 1-4
1.3.4 Illustrative Nature of the Analysis 1-4
1.4 Overview and Design of the Final RIA 1-5
-------
1.4.1 Important Updates and Differences Between the PM NAAQS
Proposal RIA and the Final RIA 1-5
1.4.2 Existing and Revised PM Air Quality Standards 1-7
1.4.3 Health and Welfare Co-Benefits Analysis Approach 1-11
1.4.4 Cost Analysis Approach 1-14
1.5 Organization of this Regulatory Impact Analysis 1-16
1.6 References for Chapter 1 1-17
Chapter 2 Defining the PM Air Quality Problem 2-1
2.1 Synopsis 2-1
2.2 Particulate Matter (PM) Properties 2-1
2.2.1 PM2.5 2-4
2.3 References 2-11
Chapter 3 Air Quality Modeling and Analysis 3-1
3.1 Synopsis 3-1
3.2 Modeling PM2.5 Levels in the Future 3-1
3.2.1 Air Quality Modeling Platform 3-2
3.3 PM2.5 Modeling Results and Analyses 3-11
3.3.1 Calculating Future-year Design Values for 2020 Base and Control
Cases 3-15
3.3.2 Calculating Future-year Design Values for Meeting the Existing
Standards, the Revised Annual Standard, and Alternative Annual
Standard Levels 3-22
3.3.3 Estimating Changes in Annual Average PM2.5 for Benefits Inputs 3-29
3.3.4 Limitations of Using Adjusted Air Quality Data 3-31
3.3.5 Weight-of-Evidence Approach for Lincoln County, MT and Santa
Cruz, NM 3-32
3.3.6 Estimating Changes in Visibility for Analyzing Welfare Benefits 3-33
3.4 References 3-35
Appendix 3.A Additional Air Quality Modeling Information 3.A-1
3.A.1 Air Quality Modeling and Analysis 3.A-1
IV
-------
3.A.1.1 Development of Air Quality Response Ratios 3.A-1
3.A.1.2 Estimating Area NOX Emission Contributions to NH4N03 PM2.5 in
theSan Joaquin Valley 3.A-12
Chapter 4 Control Strategies 4-1
4.1 Synopsis 4-1
4.2 PM2.5 Control Strategy Analysis 4-2
4.2.1 Identify Geographic Areas 4-2
4.2.2 Developing the Control Scenario 4-6
4.2.3 Identify Known Controls Needed to Meet the Analytical Baseline 4-9
4.2.4 Identify Known Controls Needed to Meet the Revised and
Alternative Standards 4-12
4.2.5 Identify Emission Reductions Beyond Known Controls Needed to
Meet the Revised and Alternative Standards 4-13
4.3 Limitations and Uncertainties 4-14
4.4 References 4-16
Appendix4.A Additional Control Strategy Information 4.A-1
4.A.1 Control Measures for Stationary Sources 4.A-1
4.A.1.1 PM Emissions Control Technologies 4.A-1
4.A.1.2 S02 Control Measures 4.A-4
4.A.1.3 NOX Emissions Control Measures 4.A-5
Chapter 5 Human Health Benefits Analysis Approach and Results 5-1
5.1 Synopsis 5-1
5.2 Overview 5-3
5.3 Updated Methodology Presented in this RIA 5-7
5.4 Human Health Benefits Analysis Methods 5-8
5.4.1 Health Impact Assessment 5-9
5.4.2 Economic Valuation of Health Impacts 5-11
5.5 Uncertainty Characterization 5-13
5.5.1 Monte Carlo Assessment 5-14
5.5.2 Concentration Benchmark Analysis for PM2.s 5-15
-------
5.5.3 Alternative Concentration-Response Functions for PM2.5-
Related Mortality 5-15
5.5.4 Sensitivity Analyses 5-16
5.5.5 Distributional Assessment 5-17
5.5.6 Influence Analysis—Quantitative Assessment of Uncertainty 5-18
5.5.7 Qualitative Assessment of Uncertainty and Other Analysis
Limitations 5-18
5.6 Benefits Analysis Data Inputs 5-21
5.6.1 Demographic Data 5-21
5.6.2 Baseline Incidence and Prevalence Estimates 5-22
5.6.3 Effect Coefficients 5-26
5.6.4 Unquantified Human Health Benefits 5-46
5.6.5 Economic Valuation Estimates 5-48
5.6.6 Hospital Admissions and Emergency Department Valuation 5-60
5.6.7 Minor Restricted Activity Days Valuation 5-62
5.6.8 Growth in WTP Reflecting National Income Growth Over Time 5-62
5.7 Benefits Results 5-65
5.7.1 Benefits of the Revised and Alternative Annual Primary PM2.5
Standards 5-65
5.7.2 Uncertainty in Benefits Results 5-73
5.7.3 Estimated Life Years Gained and Reduction in the Percentage of
Deaths Attributable to PM2.5 5-74
5.7.4 Evaluation of Mortality Impacts Relative to Various
Concentration Benchmarks 5-80
5.7.5 Additional Sensitivity Analyses 5-87
5.8 Discussion 5-87
5.9 References 5-90
Appendix 5.A Additional Sensitivity Analyses Related to the Health Benefits Analysis 5.A-1
5.A.1 Cessation Lag Structure for PM2.5-Related Premature Mortality 5.A-1
5.A.2 Income Elasticity of Willingness to Pay 5.A-9
5.A.3 Analysis of Cardiovascular Emergency Department Visits,
Cerebrovascular Events and Chronic Bronchitis 5.A-10
5.A.4 Updating Basis for Population Projections to 2010 Census 5.A-11
VI
-------
5.A.5 Long-term PM2.5 Mortality Estimates Using Cohort Studies in California ...5.A-12
5.A.6 Analysis of Health Benefits Estimated for 2025 5.A-14
5.A.7 References 5.A-14
Appendix 5.B Comprehensive Characterization of Uncertainty in Benefits Analysis 5.B-1
5.B.I Description of Classifications Applied in the Uncertainty
Characterization 5.B-1
5.B. 1.1 Direction of Bias 5.B-2
5.B.I.2 Magnitude of Impact 5.B-2
5.B.I.3 Confidence in Analytic Approach 5.B-3
5.B.1.4 Uncertainty Quantification 5.B-4
5.B.2 Organization of the Qualitative Uncertainty Table 5.B-5
5.B.3 References 5.B-15
Chapter 6 Welfare co-Benefits of the Primary Standard 6-1
6.1 Synopsis 6-1
6.2 Introduction to Welfare Benefits 6-1
6.3 Visibility Co-Benefits Approach 6-6
6.3.1 Visibility and Light Extinction Background 6-6
6.3.2 Quantifying Light Extinction for Assessing Visibility Co-benefits 6-11
6.3.3 Visibility Valuation Overview 6-12
6.3.4 Recreational Visibility 6-14
6.3.5 Residential Visibility 6-23
6.3.6 Discussion of Visibility Co-benefits 6-32
6.4 Materials Damage Co-benefits 6-33
6.5 Climate Co-benefits 6-36
6.5.1 Climate Effects of Short Lived Climate Forcers 6-37
6.5.2 Climate Effects of Long-Lived Greenhouse Gases 6-40
6.6 Ecosystem Co-benefits and Services 6-40
6.6.1 Ecosystem Co-benefits for Metallic and Organic Constituents of
PM 6-42
6.6.2 Ecosystem Co-benefits from Reductions in Nitrogen and Sulfur
Emissions 6-46
VII
-------
6.6.3 Ecosystem Co-benefits from Reductions in Mercury Emissions 6-64
6.6.4 Vegetation Co-benefits from Reductions in Ambient Ozone 6-67
6.7 References 6-74
Appendix 6.A Additional Details Regarding the Visibility Benefits Methodology 6.A-1
6.A.1 Introduction 6.A-1
6.A.2 Converting Modeled Light Extinction Estimates 6.A-1
6.A.3 Basic Utility Model 6.A-2
6.A.4 Measure of Visibility: Environmental "Goods" Versus "Bads" 6.A-5
6.A.5 Estimating the Parameters for Visibility at Class I Areas: the y's and 6's 6.A-7
6.A.5.1 Estimating Region-Specific Recreational Visibility Parameters for
the Region Covered in the Chestnut and Rowe Study (Regions 1,
2, and 3) 6.A-10
6.A.5.2 Inferring Region-Specific Recreational Visibility Parameters for
Regions Not Covered in the Chestnut and Rowe Study (Regions
4, 5, and 6) 6.A-11
6.A.5.3 Estimating Park- and Wilderness Area-Specific Parameters 6.A-13
6.A.5.4 Derivation of Region-Specific WTP for National Parks and
Wilderness Areas 6.A-14
6.A.5.5 Derivation of Park- and Wilderness Area-Specific WTPs, Given
Region-Specific WTPs for National Parks and Wilderness Areas 6.A-14
6.A.5.6 Derivation of Park- and Wilderness Area-Specific Parameters,
Given Park- and Wilderness-Specific WTP 6.A-15
6.A.6 Estimating the Parameter for Visibility in Residential Areas: 0 6.A-17
6.A.7 Putting It All Together: The Household Utility and WTP Functions 6.A-18
6.A.8 Income Elasticity and Income Growth Adjustment for Visibility Benefits...6.A-19
6.A.9 Summary of Parameters 6.A-20
6.A.10 References 6.A-23
Appendix 6.B Illustrative Scenario of Recreational and Residential Visibility Benefits 6.B-1
6.B.I Synopsis 6.B-1
6.B.2 Recreational Benefits Results 6.B-1
6.B.3 Residential Benefits Results 6.B-5
VIM
-------
6.B.4 References 6.B-8
Chapter 7 Engineering Cost Analysis 7-1
7.1 Synopsis 7-1
7.2 Estimation of Engineering Control Costs 7-2
7.2.1 Data and Methods—Identified Control Costs (non-EGU Point
and Area Sources) 7-2
7.2.2 Identified Control Costs 7-3
7.2.3 Extrapolated Costs 7-6
7.2.4 Total Cost Estimates 7-12
7.3 Changes in Regulatory Cost Estimates overTime 7-15
7.3.1 Examples of Technological Advances in Pollution Control 7-16
7.3.2 Influence on Regulatory Cost Estimates 7-19
7.3.3 Influence of Regulation on Technological Change 7-21
7.4 Uncertainties and Limitations 7-21
7.5 References 7-22
Appendix 7.A Data to Calculate Costs to Meet 12 u,g/m3 and Sensitivity Analyses of
Extrapolated Cost Estimates 7.A-1
7.A.1 PM2.5 Emission Reductions and Costs to Meet 12 u,g/m3 7.A-1
7.A.2 Sensitivity Analyses of Extrapolated Cost Estimates 7.A-2
7.A.2.1 Sensitivity Analysis of Fixed-Cost Methodology 7.A-2
7.A.2.2 Sensitivity Analysis of Alternative Hybrid Methodology 7.A-3
Chapter 8 Comparison of Benefits and Costs 8-1
8.1 Synopsis 8-1
8.2 Comparison of Benefits and Costs 8-1
8.3 Discussion and Conclusions 8-6
8.4 References 8-9
Chapter 9 Statutory and Executive Order Reviews 9-1
9.1 Synopsis 9-1
IX
-------
9.2 Executive Order 12866: Regulatory Planning and Review 9-1
9.3 Paperwork Reduction Act 9-1
9.4 Regulatory Flexibility Act 9-1
9.5 Unfunded Mandates Reform Act 9-2
9.6 Executive Order 13132: Federalism 9-2
9.7 Executive Order 13175: Consultation and Coordination with Indian
Tribal Governments 9-3
9.8 Executive Order 13045: Protection of Children from Environmental
Health and Safety Risks 9-4
9.9 Executive Order 13211: Actions that Significantly Affect Energy Supply,
Distribution or Use 9-4
9.10 National Technology Transfer and Advancement Act 9-4
9.11 Executive Order 12898: Federal Actions to Address Environmental
Justice in Minority Populations and Low-Income Populations 9-5
Chapter 10 Qualitative Discussion of Employment Impacts of Air Quality Regulations 10-1
10.1 Introduction 10-1
10.2 Influence of NAAQS Controls on Employment 10-1
10.3 The Current State of Knowledge Based on the Peer-Reviewed Literature .... 10-2
10.3.1 Immediate and Short-Run Employment Impacts 10-3
10.3.2 Long-Term Employment Impacts on the Regulated Industry 10-4
10.4 Conclusion 10-6
10.5 References 10-7
-------
LIST OF FIGURES
Number Page
ES-1. PM RIA Analytical Flowchart ES-3
ES-2. Annual NAAQS Exceedances in "Adjusted 2020 Base Case" Scenario ES-5
ES-3. Annual NAAQS Exceedances in "Analytical Baseline" Scenario ES-6
ES-4. Monetized Benefit to Cost Comparison for 12 u.g/m3 in 2020 (7% Discount Rate).... ES-16
1-1. PM RIA Analytical Flow Diagram 1-8
2-1. Detailed Source Categorization of Anthropogenic Emissions of Primary PM2.5,
PMio and Gaseous Precursor Species S02, NOX, NH3 and VOCs for 2008 2-3
2-2. Regional and Local Contributions to Annual Average PM2.5 by Particulate S042~,
Nitrate and Total Carbon (i.e., organic plus EC) for Select Urban Areas Based on
Paired 2000-2004 IMPROVE and CSN Monitoring Sites 2-5
2-3. Regional and Seasonal Trends in Annual PM2.5 Composition from 2002 to 2007
Derived Using the SANDWICH Method. Data from the 42 monitoring locations
shown on the map were stratified by region and season including cool months
(October-April) and warm months (May-September) 2-6
2-4. RCFM (left) versus SANDWICH (right) Pie Charts Comparing the Ambient and
PM2.5 FRM Reconstructed Mass Protocols on an Annual Average Basis 2-8
2-5. Maximum County-level PM2.5 Annual Design Values Calculated Using 2006-2008
FRM 24-hr Average PM2.5 Measurements 2-10
2-6. Maximum County-level PM2.5 24-hour Design Values Calculated Using 2006-2008
FRM 24-hr Average PM2.5 Measurements 2-11
3-1. Map of the CMAQ Modeling Domain Used for PM NAAQS RIA 3-3
3-2. Flow Diagram of Process Used to Determine Future-Year Design Values and
Associated Emission Reductions for Meeting the Current, Revised and
Alternative Standard Levels 3-14
3-3. Counties that Exceed the Revised and/or Alternative Annual Standard Levels of
13, 12 and 11 u,g/m3 in the Analytical Baseline 3-25
3-5. Three-year Annual and 24-hr Design Values for the Monitor in Libby, MT 3-33
3.A-1. Location of Kings County Monitor Relative to Tulare County Border 3.A-3
3.A-2. California Modeling Domain for 12-km Simulations 3.A-13
4-1. Counties Projected to Exceed the Current PM2.5 Standard (15/35) in 2020 4-3
4-2. Counties Projected to Exceed the 13 ug/m3 Alternative Standard in the Analytical
Baseline 4-4
XI
-------
4-3. Counties Projected to Exceed the 12 ug/m3 Revised Standard in the Analytical
Baseline 4-5
4-4. Counties Projected to Exceed the 11 ug/m3 Alternative Standard in the Analytical
Baseline 4-6
5-1. Illustration of BenMAP Approach 5-11
5-2. Data Inputsand Outputs for the BenMAP Model 5-12
5-3. Estimated PM2.5-Related Premature Mortalities Avoided According to
Epidemiology or Expert-Derived PM2.5 Mortality Risk Estimate for 12 u.g/m3 in
2020 5-72
5-4. Total Monetized Benefits Using 2 Epidemiology-Derived and 12-Expert Derived
Relationships Between PM2.5 and Premature Mortality for 12 u.g/m3 in 2020 5-73
5-5. Estimated Life Years Gained as a Result of the Revised Annual Primary PM2.5
Standard of 12 u.g/m3 in 2020, Partitioned by the Age When Exposure Reduction
Occurred, Not Necessarily Age at Death 5-78
5-6. Estimate of Avoided Premature Deaths Attributed to the Revised Annual Primary
PM2.5 Standard of 12 u.g/m3 in 2020, Partitioned by the Age When Exposure
Reduction Occurred, Not Necessarily Age at Death 5-80
5-7. Relationship between the Size of the PM Mortality Estimates and the PM2.5
Concentration Observed in the Epidemiology Study 5-82
5-8. Number of Premature PM2.5-related Deaths Avoided for the Revised Annual
Primary PM2.5 Standard of 12 u.g/m3 in 2020 According to the Baseline Level of
PM2.5 and the Lowest Measured Air Quality Levels of Each Mortality Study 5-84
5-9. Number of Premature PM2.5-related Deaths Avoided for the Revised Annual
Primary PM2.5 Standard of 12 u.g/m3 in 2020 According to the Baseline Level of
PM2.5 and the Lowest Measured Air Quality Levels of Each Mortality Study 5-85
5-10. Illustration of Relative Size of County with Exceeding Monitor and Modeled Grid
Cells 5-86
5.A-1. Alternate Lag Structures for PM2.5 Premature Mortality (Cumulative) 5.A-8
5.A-2. Exponential Lag Structures for PM2.5 Premature Mortality (Cumulative) 5.A-8
6-1. Important Factors Involved in Seeing a Scenic Vista (Malm, 1999) 6-7
6-2. Visibility in Selected National Parks and Wilderness Areas in the U.S., 1992-2008 6-9
6-3. Estimated Improvement in Annual Average Visibility Levels Associated with the
CAAA Provisions in 2020 6-10
6-4. Mandatory Class 1 Areas in the U.S 6-15
6-5. Visitation Rates and Park Regions for Class 1 Areas 6-20
6-6. Residential Visibility Study City Assignment 6-29
6-7. Linkages between Categories of Ecosystem Services and Components of Human
Well-Being from Millennium Ecosystem Assessment (MEA, 2005) 6-41
6-8. Schematic of the Benefits Assessment Process (U.S. EPA, 2006c) 6-42
6-9. Schematics of Ecological Effects of Nitrogen and Sulfur Deposition 6-47
6-10. Nitrogen and Sulfur Cycling, and Interactions in the Environment 6-48
XII
-------
6-11. Areas Potentially Sensitive to Aquatic Acidification 6-51
6-12. Areas Potentially Sensitive to Terrestrial Acidification 6-53
6-13. Distribution of Red Spruce (pink) and Sugar Maple (green) in the Eastern U.S 6-54
6-14. Spatial and Biogeochemical Factors Influencing MeHg Production 6-62
6-15. Preliminary USGS Map of Mercury Methylation-Sensitive Watersheds 6-63
6-16. Visible Foliar Injury to Forest Plants from Ozone in U.S. by EPA Regions 6-71
6-17. Presence and Absence of Visible Foliar Injury, as Measured by U.S. Forest
Service, 2002 6-72
6.B-1. Recreational Visibility Benefits in Class I Areas for the Illustrative Scenario in
2020 6.B-5
6.B-2. Residential Visibility Benefits for the Illustrative Scenario in 2020 (2010$) 6.B-7
7-1. Technological Innovation Reflected by Marginal Cost Shift 7-16
8-1. Monetized Benefit to Cost Comparison for the Revised Annual Standard of 12
u.g/m3 in 2020 (7% Discount Rate) 8-3
8-2. Net Benefits for Revised Annual Standard of 12 u.g/m3 in 2020 at a 7% Discount
Rate 8-4
XIII
-------
LIST OF TABLES
Number Page
ES-1. Emission Reductions in Illustrative Emission Reduction Strategies for the Revised
and Alternative Annual Primary PM2.5 Standards, by Pollutant and Region in 2020
(tons) ES-8
ES-2. Total Monetized Benefits, Total Costs, and Net Benefits in 2020 (millions of
2010$)—Full Attainment ES-15
ES-3. Benefit-to-Cost Ratios for Alternative Standards at 3% and 7% Based on
Projected Benefits and Costs in 2020 ES-15
ES-4. Estimated Number of Avoided PM2.5 Health Impacts for Standard Alternatives-
Full Attainment in 2020 ES-17
2-1. Annual Average FRM and CSN PM2.5 N03~ and NH4N03 Concentrations at Six Sites
during 2003 2-8
3-l(a). Control Strategies and Growth Assumptions for Creating 2020 Base Case
Emissions Inventories from the 2007 Base Case for Non-EGU Point Sources 3-9
3-l(b). Control Strategies and Growth Assumptions for Creating 2020 Base Case
Emissions Inventories from the 2007 Base Case for Nonpoint and Onroad Mobile
Sources 3-10
3-l(c). Control Strategies and Growth Assumptions for Creating 2020 Base Case
Emissions Inventories from the 2007 Base Case for Nonroad Mobile Sources 3-11
3-2. Air Quality Model Simulations Used in this Regulatory Impact Analysis 3-15
3-3. Nonattainment Areas Where Episodic Residential Wood Burning Curtailment was
Applied 3-17
3-4. Estimated Number of Burn Ban Days by County Based on 2005-2009 FRM Data 3-19
3-5. Tons of Direct PM2.5 Emission Reductions beyond the Adjusted 2020 Control
Case to Meet the Current Standard Level for Counties that Exceed the Revised or
Alternative Annual Standard Levels in the Adjusted 2020 Base Case 3-23
3-6. Annual Design Values and Exceedance Category for the Highest County Monitor
in the Analytical Baseline for Counties with at Least one Monitor Above the
Revised and/or Alternative Standard Levels 3-26
3-7. Tons of Direct PM2.5 Emission Reductions Beyond the Analytical Baseline to Meet
the 13 u,g/m3 Level 3-27
3-8. Tons of Direct PM2.5 Emission Reductions Beyond the Analytical Baseline to Meet
the 12 u,g/m3 Level 3-28
3-9. Tons of Direct PM2.5 Emission Reductions Beyond the Analytical Baseline to Meet
the Alternative Standard 11 u,g/m3 Level 3-29
XIV
-------
3.A-1. Annual NOX Air Quality Ratios for Monitors in California Counties that Received a
2025 Mobile NOX Emission Adjustment 3.A-5
3.A-2. Annual and Daily S02 Air Quality Ratios for Monitors in Counties where Ratios
were Used in Adjusting Daily Design Values to Remove the Impact of S02
Controls 3.A-8
3.A-3. Annual and Daily Direct PM2.5 Air Quality Ratios for Monitors in Counties with at
Least One Monitor having an Annual Design Value Above 11 u,g/m3 in the 2020
Base Case 3.A-9
3.A-4. Contribution Weighting Factors for NOX Emissions in Counties or County Groups
in California's Central Valley as Calculated According to Equation 3.A.2 3.A-14
4-1. Controls Applied in the Revised and Alternative Standard Analysis 4-7
4-2. Number of Counties with Exceedances and Number of Additional Counties
Where Reductions Were Applied 4-13
4-3. Emission Reductions from Known Controls for the Revised and Alternative
Standards 4-13
4-4. Emission Reductions Needed Beyond Known Controls for the Revised and
Alternative Standards 4-14
4-5. Summary of Qualitative Uncertainty for Elements of Control Strategies 4-15
4.A-1. Example PM Control Measures for Non-EGU Point Source Categories 4.A-3
4.A-2. Example PM Control Measures for Nonpoint Sources 4.A-4
4.A-3. Example S02 Control Measures for Non-EGU Point 4.A-5
4.A-4. Example NOX Control Measures for Non-EGU Source Categories 4.A-7
5-1. Estimated Monetized Benefits of the of Revised and Alternative Annual PM2.5
Standards in 2020 Incremental to the Analytical Baseline (billions of 2010$) 5-2
5-2. Human Health Effects of Pollutants Potentially Affected by Strategies to Attain
the Primary PM2.5 Standards 5-4
5-3. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population 5-24
5-4. Asthma Prevalence Rates 5-25
5-5. Criteria Used When Selecting C-R Functions 5-28
5-6. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts
in the Core Analysis 5-29
5-7. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts
in the Sensitivity Analysis 5-31
5-8. Summary of Effect estimates from Associated With Change in Long-Term
Exposure to PM2.5 in Recent Cohort Studies in North America 5-35
5-9. Unit Values for Economic Valuation of Health Endpoints (2010$) 5-50
5-10. Influence of Applied VSL Attributes on the Size of the Economic Benefits of
Reductions in the Risk of Premature Death (U.S. EPA, 2006a) 5-55
5-11. Alternative Direct Medical Cost of Illness Estimates for Nonfatal Heart Attacks 5-60
xv
-------
5-12. Estimated Costs Over a 5-Year Period of a Nonfatal Myocardial Infarction (in
2010$) 5-60
5-13. Unit Values for Hospital Admissions 5-61
5-14. Elasticity Values Used to Account for Projected Real Income Growth 5-64
5-15. Adjustment Factors Used to Account for Projected Real Income Growth 5-65
5-16. Population-Weighted Air Quality Change for the Revised and Alternative Annual
Primary PM2.5 Standards Relative to Analytical Baseline 5-66
5-17. Emission Reductions in Illustrative Emission Reduction Strategies for the Revised
and Alternative Annual Primary PM2.5 Standards, by Pollutant and Region in 2020
(tons) 5-67
5-18. Estimated number of Avoided PM2.5 Health Impacts for the Revised and
Alternative Annual Primary PM2.5 Standards (Incremental to the Analytical
Baseline) 5-68
5-19. Monetized PM2.5 Health Benefits for the Revised and Alternative Annual Primary
PM2.5 Standards (Incremental to Analytical Baseline) (Millions of 2006$, 3%
discount rate) 5-69
5-20. Monetized PM2.5 Health Benefits for the Revised and Alternative Annual Primary
PM2.5 Standards (Incremental to Analytical Baseline) (Millions of 2006$, 7%
discount rate) 5-70
5-21. Total Estimated Monetized Benefits of the for Revised and Alternative Annual
Primary PM2.5 Standards (Incremental to the Analytical Baseline) (billions of
2006$) 5-71
5-22. Regional Breakdown of Monetized Benefits Results 5-71
5-23. Sum of Life Years Gained by Age Range Attributed to the Revised Annual Primary
PM2.5 Standard of 12 ug/m3 in 2020 5-77
5-24. Estimated Reduction in the Percentage of All-Cause Mortality Attributed to the
Revised Annual Primary PM2.5 Standard of 12 u.g/m3 in 2020 5-79
5-25. Estimated Reduction in Incidence of Adult Premature Mortality Occurring Above
and Below Various Concentration Benchmarks in the Underlying Epidemiology
Studies 5-83
5.A-1. Values of the Time Constant (k) for the Exponential Decay Lag Function 5.A-6
5.A-2. Sensitivity of Monetized PM2.5-Related Premature Mortality Benefits to
Alternative Cessation Lag Structures, Using Effect Estimate from Krewski et al.
(2009) 5.A-7
5.A-3. Ranges of Elasticity Values Used to Account for Projected Real Income Growth 5.A-9
5.A-4. Ranges of Adjustment Factors Used to Account for Projected Real Income
Growth 5.A-9
5.A-5. Sensitivity of Monetized Benefits to Alternative Income Elasticities 5.A-10
5.A-6. Avoided Cases of Cardiovascular Emergency Department Visits, Stroke and
Chronic Bronchitis in 2020 (95th percentile confidence intervals) 5.A-11
5.A-7. Change in Total Monetized Benefits for 2000 and 2010 Census projections for 12
ug/m3 in 2020 (2010$) 5.A-12
XVI
-------
5.A-8 Summary of Effect Estimates From Associated With Change in Long-Term
Exposure to PM2.5 in Recent Cohort Studies in California 5.A-13
5.A-9. Comparison of Health Benefits Estimated for 2000 and 2025 for 12 u.g/m3 5.A-14
5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits.... 5.B-6
6-1. Welfare Effects by Pollutants Potentially Affected by Attainment of the PM
NAAQS 6-3
6-2. Quantified and Unquantified Welfare Co-Benefits 6-4
6-3. Key Assumptions in the Light Extinction Estimates Affecting the Visibility Co-
Benefits Approach 6-12
6-4. WTP for Visibility Improvements in Class 1 Areas in Non-Studied Park Regions 6-19
6-5. Summary of Key Assumptions in Estimating Recreational Visibility Co-benefits 6-22
6-6. Summary of Residential Visibility Valuation Estimates 6-26
6-7. Summary of Key Assumptions in the Residential Visibility Co-benefits 6-30
6-8. Materials Damaged by Pollutants Affected by this Rule (U.S. EPA, 2011b) 6-36
6-9. Aquatic Status Categories 6-50
6.A-1. Available Information on WTP for Visibility Improvements in National Parks 6.A-8
6.A-2. Summary of Region-Specific Recreational Visibility Parameters to be Estimated in
Household Utility Functions 6.A-9
6.A-3. Mean Annual Household WTP for Changes in Visual Range for Recreational
Visibility (1990$) 6.A-21
6.A-4. Region-Specific Parameters for Recreational Visibility Benefits 6.A-21
6.A-5. Mean Annual Household WTP for Changes in Visual Range for Residential
Visibility 6.A-22
6.A-6. Parameters for Income Growth Adjustment for Visibility Benefits 6.A-23
6.B-1. Annual Average Visibility Improvements in the Top 10 Most Visited Class I Areas
forthe Illustrative Scenario in 2020 6.B-2
6.B-2. Recreational Visibility Benefits in Studied Regions for the Illustrative Scenario in
2020 (in millions of 2010$) 6.B-3
6.B-3. Sensitivity Analysis for Recreational Visibility Benefits outside Studied Park
Regions for the Illustrative Scenario in 2020 (in millions of 2010$) 6.B-3
6.B-4. Sensitivity Analysis for Incorporating Coarse Particles into Recreational Visibility
Benefits forthe Illustrative Scenario (in millions of 2010$) 6.B-4
6.B-5. Monetized Residential Visibility Benefits in Studied Areas in 2020 for the
Illustrative Scenario (millions of 2010$, 2020 income) 6.B-6
6.B-6. Sensitivity Analysis for Monetized Residential Visibility Benefits in Other Areas
forthe Illustrative Scenario in 2020 (in millions of 2010$) 6.B-6
6.B-7. Sensitivity Analysis for Incorporating Coarse Particles into Residential Visibility
Benefits (in millions of 2010$) 6.B-7
XVII
-------
7-1. Summary of Sectors, Emissions Reductions, and Known Annualized Control Costs
(millions of 2010$) 7-5
7-2. Partial Attainment Known Annualized Control Costs in 2020 for Revised and
Alternative Standards Analyzed (millions of 2010$) 7-5
7-3. Extrapolated Costs by Revised and Alternative Standard Analyzed(millions of
2010$) 7-10
7-4. Total Costs by Revised and Alternative Standard Analyzed (millions of 2010$),
Fixed-Cost Methodology 7-14
7-5 Total Costs by Revised and Alternative Standard Analyzed (millions of 2010$),
Hybrid Methodology 7-15
7-6. Phase 2 Cost Estimates 7-20
7-7. Summary of Qualitative Uncertainty for Modeling Elements of PM Engineering
Costs 7-22
7.A.1 PM2.s Emission Reductions and Costs to Meet 12 u,g/m3 7.A-1
7.A.2 Sensitivity Analysis of Fixed-Cost Methodology for Unknown Controls by Revised
and Alternative Standard Analyzed (millions of 2010$) 7.A-2
7.A.3 Sensitivity Analysis of Hybrid Methodology for Unknown Controls by Revised and
Alternative Standard Analyzed (millions of 2010$) 7.A-3
8-1. Total Monetized Benefits, Total Costs, and Net Benefits in 2020 (millions of
2010$)—Full Attainment 8-2
8-2. Human Health Effects of Ambient PM2.5 8-5
8-3. Welfare Co-Benefits of PM2.5 8-6
XVIII
-------
EXECUTIVE SUMMARY
ES.l Overview
Based on its review of the air quality criteria and the national ambient air quality
standards (NAAQS) for particulate matter (PM), the U.S. Environmental Protection Agency (EPA)
is making revisions to the primary standards for PM to provide requisite protection of public
health and welfare. The EPA is revising the primary annual (health-based) standard, retaining
the primary 24-hour standard, and retaining the secondary (welfare-based) NAAQS for fine
particles (generally referring to particles less than or equal to 2.5 micrometers [u.m] in
diameter—PM2.5). The EPA is retaining the current primary and secondary 24-hour PMi0
standards.
As has traditionally been done in NAAQS rulemakings, the EPA has conducted a
Regulatory Impact Analysis (RIA) to provide the public with illustrative estimates of the
potential costs and health and welfare benefits of attaining the revised annual standard along
with two alternative standards. In NAAQS rulemakings, the RIA is prepared for informational
purposes only, and the decisions related to the setting of the PM NAAQS standards are not in
any way based on consideration of the information or analyses in the RIA. The RIA fulfills the
requirements of Executive Orders 12866 and 13563 and guidelines of the Office of
Management and Budget's (OMB) Circular A-4.1
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 this RIA cannot be added to benefits and costs
from other regulations because each regulation is based on a different set of analytical
assumptions and policy decisions. The costs and benefits identified in this RIA will not be
realized until specific controls are mandated by future State Implementation Plans (SIPs) or
other Federal regulations.
ES.2 Existing and Revised PM Air Quality Standards
Two primary PM2.s standards provide public health protection from effects associated
with fine particle exposures: the annual standard and the 24-hour standard. The current annual
standard is set at a level of 15.0 u.g/m3, based on the 3-year average of annual arithmetic mean
PM2.s concentrations. The current 24-hour standard is set at a level of 35 u.g/m3, based on the
1 U.S. Office of Management and Budget. Circular A-4, September 17, 2003. Available at
http://www.whitehouse.gov/omb/circulars/a004/a-4.pdf.
ES-1
-------
3-year average of the 98th percentile of 24-hour PM2.5 concentration. In the RIA, the current
primary PM2.5 standards, including both annual and 24-hour standards, are denoted as 15/35
u.g/m3. Attainment of the 24-hour standard is analyzed only in developing the scenario of
attainment with the existing standards of 15/35 u.g/m3. All other scenarios evaluate additional
emissions reductions needed to attain alternative annual standards only.
In this PM NAAQS review, the EPA has revised and lowered the level of the primary
annual PM2.5 standard to 12 u.g/m3 in conjunction with retaining the level of the 24-hour
standard at 35 u.g/m3 and this standard is denoted as 12 u.g/m3. In addition to the revised
annual standard of 12 u.g/m3, the RIA also analyzes the benefits and costs of incremental
control strategies for two alternative annual standards of 13 u.g/m3 and 11 u.g/m3.
Currently, the existing secondary (welfare-based) PM2.5 standards are identical in all
respects to the primary standards. In this PM NAAQS review, the EPA is retaining the current
suite of secondary standards for 24-hour and annual PM2.5. Thus, while the new primary annual
standard will be revised to 12 u.g/m3, the secondary annual standard will remain at 15 u.g/m3.
Non-visibility welfare effects are addressed by this suite of secondary standards, and PM-
related visibility impairment is addressed by the secondary 24-hour PM2.5 standard, which EPA
is leaving unchanged at 35 u.g/m3. The secondary standards will thus remain at 15/35 u.g/m3.
With regard to the primary and secondary standards for particles less than or equal to
10 u.m in diameter (PMi0), the EPA is retaining the current primary and secondary 24-hour PMi0
standards, which are both set at a level of 150 u.g/m3, not to be exceeded more than once per
year on average over 3 years (U.S. EPA, 1997).2 Because the benefit-cost analysis of the
alternative PMi0 standards was conducted when the standard was promulgated in 1997, this
RIA does not repeat that analysis here.
ES.2.1 Overview of the Analytical Steps in this RIA
The goal of this RIA is to provide the best estimates of the costs and benefits of an
illustrative attainment strategy to the meet the revised annual standard. The flowchart below
(Figure ES-1) outlines the analytical steps taken to illustrate attainment with the revised annual
standard of 12 u.g/m3, and the following discussion, by primary flowchart section, describes
each of the steps taken. For important updates and analytical differences between proposal
and final, see section ES.5 of this Executive Summary.
2 U.S. Environmental Protection Agency. 1997. Regulatory Impact Analyses for the Particulate Matter and Ozone
National Ambient Air Quality Standards and Proposed Regional Haze Rule. Available at:
http://www.epa.gov/ttn/oarpg/naaqsfin/ria.html.
ES-2
-------
Figure ES-1. PM RIA Analytical Flowchart
PM RIA Analytical Flowchart
See Individual Chapters for Complete Description
Adjusted 2020 Base Case with "On the Books"
Federal and State Rules and Programs
Establishing the Analytical
Baseline
Analytical Baseline with Attainment of 15/35
^^H
M
Select Controls by Emissions Sector and Pollutant
Control Strategies
Analysis of Costs
Adjust 2020 Control Scenario to Meet Alternate Standards
Estimate Known Controls Engineering Costs
Estimate Extrapolated Costs
Calculate Full Attainment
Estimate Full Attainment
Benefits
Full Attainment
T
Net Benefits
Establishing the Analytical Baseline (Flowchart Section 1)
This section of the flowchart reflects the analytical steps taken to account for Federal
and State rules and programs currently underway, as well as to reflect attainment of the
current annual and daily standards of 15/35 u.g/m3 for the purpose of estimating the
incremental costs and benefits of attaining the revised annual standard. Detailed discussions of
the elements of this section of the flowchart are in Chapters 1 and 3 of the final RIA.
• Adjusted 2020 Base Case with "On the Books" Federal and State Rules and
Programs—The adjusted 2020 base case includes reductions expected to occur
between 2007 and 2020 from existing (i.e., "on-the-books") Federal and State
control programs. This projection reflects air quality modeling of 2020 that accounts
for major Federal and State programs along with adjustments to these national-level
modeling results to account for PM2.5 reductions expected, largely in the western
ES-3
-------
US, from the implementation of episodic burn ban programs in certain counties and
to remove the effects of atypical events such as wildfires and fireworks displays.
(See Figure ES-2 below for illustration.) Below is a list of some of the major national
rules reflected in the base case. Refer to Chapter 3, Section 3.2.1.4 for a more
detailed discussion of the rules reflected in the 2020 base case emissions inventory.
- Light-Duty Vehicle Tier 2 Rule (U.S. EPA, 1999)
- Heavy Duty Diesel Rule (U.S. EPA, 2000)
- Clean Air Nonroad Diesel Rule (U.S. EPA, 2004)
- 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,
2005a)
- Emissions Standards for Locomotives and Marine Compression-Ignition
Engines (U.S. EPA, 2008a)
- Control of Emissions for Nonroad Spark Ignition Engines and Equipment (U.S.
EPA, 2008b)
- C3 Oceangoing Vessels (U.S. EPA, 2010a)
- Boiler MACT (U.S. EPA, 2011d)
- Hospital/Medical/lnfectious Waste Incinerators: New Source Performance
Standards and Emission Guidelines: Final Rule Amendments (U.S. EPA, 2009)
- Reciprocating Internal Combustion Engines (RICE) NESHAPs (U.S. EPA, 2010b)
- Mercury and Air Toxics Standards (U.S. EPA, 2011b)
- Cross-State Air Pollution Rule (U.S. EPA, 2011a)3
3 See Chapter 3, Section 3.2.1.4 for a discussion of the role CSAPR plays in the PM2.5 RIA and the reasons we believe
CSAPR remains an appropriate proxy for this analysis.
ES-4
-------
Figure ES-2. Annual NAAQS Exceedances in "Adjusted 2020 Base Case" Scenario
24-hr PM2.5 NAAQS Exceedances as Follows:
• San Joaquin Valley (6 counties); South Coast (2 counties); Imperial, CA; Allegheny, PA; Salt Lake, UT; Lake, OR; and
Sacramento, CA
Analytical Baseline with Attainment of 15/35—The analytical baseline includes
reductions from additional controls that the EPA estimates are needed to attain the
current standards (15/35 u.g/m3) for the purpose of estimating the incremental costs
and benefits of attaining the revised annual standard of 12 u.g/m3. Determining the
level of emissions reductions needed to meet both the current annual and daily
standards is done through adjusting a county's projected design value (DV)
concentration using the geographic area and pollutant specific air quality ratios (as
described in Chapter 3 of the final RIA). In addition, for each area, it is necessary to
determine which of the standards will be "controlling," i.e., which standard will
require the most reductions to reach attainment of current standard. This is
important for establishing the analytical baseline because when the daily standard is
controlling (as is the case in much of California), the emissions reductions required
to meet the daily standard will result in reductions in the annual design value to
below the annual standard of 15 u.g/m3. As a result, the annual PM2.5 increment
needed to attain alternative standards for each county will not be relative to an
annual design value of 15 u.g/m3, rather, the increment will vary by county based on
the annual design value in that county that resulted from applying emissions
controls to meet the 35 u.g/m3 daily standard. As shown in the map below, for
counties labeled as exceeding the 12 u.g/m3 annual standard, the baseline design
values can be any value between 12 and 15 u.g/m3. Similarly, exceedances of the 11
u.g/m3and 13 u.g/m3 standards can be anywhere between 11 or 13 and 15 u.g/m3.
ES-5
-------
The analytical baseline reflects attainment of 15/35 u.g/m3 by2020. We also
modified the analytical baseline for counties in the South Coast Air Quality
Management District and the San Joaquin Valley Air Pollution Control District to
reflect reductions in mobile NOX emissions that these areas are expected to achieve
between 2020 and 2025 due to fleet turnover. These reductions in NOX emissions
are not attributable to attainment of the current or revised PM standards, but
reflect the impacts of other "on-the-books" mobile programs so that these
reductions are not included as either an incremental cost nor benefit to these area's
attaining the revised annual standard.
Figure ES-3. Annual NAAQS Exceedances in "Analytical Baseline" Scenario
Exceedance
• 13
I™
11
• 13 Level: 2 Counties (Tulare, CA and Riverside, CA)
• 12 Level: 7 Counties (All in CA)
11 Level: 23 Counties
Control Strategies (Flowchart Section 2)
This section of the flowchart reflects analytical steps taken to analyze controls and
emissions reductions needed beyond the current standard (15/35 u.g/m3) and other existing
major rules to achieve the revised standard (12 u.g/m3). We apply control options that might be
available to States for application by 2020. Detailed discussion of the elements of this section of
the flowchart is in Chapter 4 of the final RIA.
• Select Controls by Emissions Sector and Pollutant— Non-EGU point and nonpoint
control measures were applied for the revised and alternative standards' control
strategies. These controls were identified using the U.S. EPA's Control Strategy Tool.
ES-6
-------
Additional control measures were not applied to EGUs because of the extensive
nature of controls resulting from the inclusion of MATS.
Adjust 2020 Control Scenario to Meet Alternate Standards—After identifying the
known controls in the control scenario that were needed to meet the analytical
baseline, additional known controls needed to meet the revised and alternative
standards were identified. The EPA used air quality modeling results to determine
whether the control scenario was sufficient to meet the revised and alternative
standards for each geographic area. Where the control scenario modeling resulted
in design value reductions below the level needed for the revised or alternative
standards for specific geographic areas, county-specific ratios of air quality response
to emission reductions were used to determine the subset of controls that were
needed to attain the standard. Where it was determined that the control scenario
was not sufficient in attaining the standard, these same response factors were used
to calculate the amount of additional emission reductions beyond known controls
needed to meet the standard. For the revised and alternative control strategy
analysis, known controls for two sectors were used: non-EGU point and area
sources. Onroad mobile source controls were not used in the revised and alternative
standards analysis because they were applied previously in the analytical baseline
analysis where they were deemed to be most cost effective. Emission reductions
were calculated for the known control strategy analysis and the cost analysis for
emission reductions needed beyond known controls ("extrapolated" costs) for each
alternative standard being analyzed. The EPA estimates the national-scale emission
reductions for revised annual standard of 12 u.g/m3 and two alternative annual
standards (13 u.g/m3 and 11 u.g/m3) as shown in Table ES-1.
Because the rules listed above and other emissions reductions should have
substantially reduced ambient PM2.s concentrations by 2020 in the East, no
additional controls are anticipated to be needed outside of California (assuming the
absence of new sources). Specifically, our analysis estimates that in 2020 only 7
counties, all in California, will be out of attainment with the revised annual standard
of 12ug/m3. Emissions reductions are needed in more locations for the alternative
standard of llug/m3.
ES-7
-------
Table ES-1. Emission Reductions in Illustrative Emission Reduction Strategies for the Revised
and Alternative Annual Primary PM2.s Standards, by Pollutant and Region in 2020
(tons)3
13 |ig/m3 12 |ig/m3 11 |ig/m3
Directly emitted PM2.5
East
West
CA
S02
East
West
CA
NOx
East
West
CA
0
0
730
0
0
0
0
0
0
0
0
4,000
0
0
0
0
0
0
8,200
160
10,600
21,000
43
0
9
0
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. Estimates are rounded to two significant figure. Estimates are rounded
to two significant figures.
Analysis of Costs (Flowchart Section 3)
This section of the flowchart reflects analytical steps taken to estimate the costs of both
known and unknown controls. Detailed discussion of the elements of this section of the
flowchart is in Chapter 7 of the final RIA.
• Estimate Known Controls Engineering Costs, Estimate Extrapolated Costs—We
provide engineering cost estimates for the control strategies identified in Chapter 4
that include control technologies on non-EGU point sources and area sources.
Engineering costs generally refer to the capital equipment expense, the site
preparation costs for the application, and annual operating and maintenance costs.
For this analysis, we included known controls for all of the geographic areas likely to
exceed the revised and/or alternative standards. We also provide estimates for the
engineering costs of the additional emissions reductions that are needed beyond the
application of known controls to reach full attainment of the alternative standards
analyzed; the cost estimates derived from this approach are referred to as
"extrapolated" costs. By definition, no cost data currently exists for the additional
emissions reductions needed beyond known controls. We employ two
ES-8
-------
methodologies for estimating the costs of unidentified future controls, and both
approaches assume that innovative strategies and new control options make
possible the emissions reductions needed for attainment by 2020.
Full Attainment (Flowchart Section 4)
This section of the flowchart reflects analytical steps taken to estimate the costs of
attainment, the benefits of attainment and the net benefits for the revised standard of 12
u.g/m3. Detailed discussions of the elements of this section of the flowchart are in Chapter 5,
Chapter 7, and Chapter 8 of the final RIA.
• Calculate Full Attainment Costs—In Chapter 7 we present a summary of the total
national costs of attaining the revised annual standard of 12 u.g/m3 and the
alternative annual standards of 13 u.g/m3 and 11 u.g/m3 in 2020. This summary
includes the known and extrapolated costs. The total cost estimates are $53 million
(2010$) and $350 million (2010$) for the revised annual standard of 12 u.g/m3; $11
million and $100 million for the alternative annual standard of 13 u.g/m3; and $320
million and $1,700 million for the alternative annual standard of 11 u.g/m3.
• Calculate Full Attainment Benefits—Chapter 5 presents the estimated human health
benefits for the revised NAAQS. We quantify the health-related benefits of the fine
particulate matter (PM2.5)-related air quality improvements resulting from the
illustrative emissions control scenarios that reduce emissions of directly emitted
particles and precursor pollutants including S02 and NOX to reach alternative PM2.5
NAAQS levels in 2020. These benefits are relative to an analytical baseline reflecting
nationwide attainment of the current primary PM2.5 standards (i.e., annual standard
of 15 u.g/m3 and 24-hour standard of 35 u.g/m3) that includes promulgated national
regulations and illustrative emissions controls to simulate attainment with 15/35 as
well as a NOX emission adjustment to reflect expected reductions in mobile NOX
emissions between 2020 and 2025.
The estimated benefits for the revised and alternative standards are in addition to
the substantial benefits estimated for several recent implementation rules. Rules
such as the Mercury and Air Toxics Standard (MATS) and other emission reductions
will have substantially reduced ambient PM2.5 concentrations by 2020 in the East,
such that no additional controls would be needed in the East for the revised annual
standard of 12 u.g/m3. Thus, all of the estimated benefits occur in California because
this is the only State that needs additional air quality improvement beyond the
analytical baseline after accounting for air quality improvements from recent rules.
• Net Benefits—Chapter 8 compares estimates of the benefits with costs and
summarizes the net benefits of revised annual standard of 12 u.g/m3 and the
alternative annual standards of 13 u.g/m3 and 11 u.g/m3 relative to the analytical
baseline that includes recently promulgated national regulations and additional
ES-9
-------
emissions reductions needed to attain the existing 15/35 u.g/m3 standards, as well as
adjustments to NOX emissions in the San Joaquin and South Coast areas.
ES.2.2 Health and Welfare Co-Benefits
The EPA estimated impacts on human health (e.g., mortality and morbidity effects)
under full attainment of the three alternative annual PM2.5 standards. We considered an array
of health impacts attributable to changes in PM2.5 exposure and estimated these benefits using
the BenMAP model (Abt Associates, 2012), which has been used in many recent RIAs (e.g., U.S.
EPA, 2006, 2011a, 2011b), and The Benefits and Costs of the Clean Air Act 1990 to 2020 (U.S.
EPA, 2011c). The monetized benefits estimated in the core analysis include avoided premature
deaths (derived from effect coefficients in tow cohort studies [Krewski et al. (2009) and Lepeule
et al. (2012)] for adults and one for infants [Woodruff et al. (1997)] ) as well as avoided
morbidity effects for 10 non-fatal endpoints ranging in severity from lower respiratory
symptoms to heart attacks. As noted above, because California is the only state that needs
additional air quality improvement beyond the analytical baseline after accounting for expected
air quality improvements expected from recent rules, all of the benefits associated with the
revised standard of 12ug/m3 occur in California.
Since the proposed rule, 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 follow-up to the Harvard Six Cities cohort study (Lepeule et al., 2012), more recent
demographic data projections, additional hospitalization and emergency department visit
studies, inflation adjustment to 2010 dollars, and an expanded uncertainty assessment. Each of
these updates is fully described in the health benefits chapter (Chapter 5) and summarized
below in section ES.5. Compared with the proposal benefits, the estimated benefits for the
revised standard are about double due to a combination of updates to the analytic baseline
Even though the primary standards are designed to protect against adverse effects to
human health, the emission reductions will 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, such as reductions in visibility impairment, materials damage,
and ecosystem damage. Despite our attempts to quantify and monetize as many of the benefits
as possible, the welfare co-benefits associated with meeting the alternative standards are not
quantified or monetized in this analysis. Unquantified health benefits are discussed in Chapter
5, and unquantified welfare co-benefits are discussed in Chapter 6.
ES-10
-------
It is important to note that estimates of the health benefits from reduced PM2.5
exposure reported here contain uncertainties, which are described in detail in Chapter 5 and
Appendix 5b. Below are two key assumptions in the benefits analysis:
1. We assume that all fine particles, regardless of their chemical composition, are
equally potent in causing premature mortality. This is an important assumption,
because PM2.5 varies considerably in composition across sources, but the scientific
evidence is not yet sufficient to allow differentiation of effect estimates by particle
type. The Integrated Science Assessment for Particulate Matter (PM ISA), which was
twice reviewed by CASAC, concluded that "many constituents of PM2.5 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, 2009). These uncertainties are likely to be magnified in
the current analysis to the extent that the emissions controls are less diverse when
focusing on one small region of the country rather than a broader geography with
more diverse emissions sources and the application of a more diverse set of
controls.
2. We assume that health impact functions based on national studies are
representative for exposures and populations in California. In addition to the
national risk coefficients we use as our core estimates, the EPA considered the
cohort studies conducted in California specifically. Although we have not calculated
the benefits results using the cohort studies conducted in California, we provided
these risk coefficients to show how much the monetized benefits could have
changed. Most of the California cohort studies report central effect estimates similar
to the (nation-wide) all-cause mortality risk estimate we applied from Krewski et al.
(2009) and Lepeule et al. (2012) albeit with wider confidence intervals. Three cohort
studies conducted in California indicate statistically significant higher risks than the
risk estimates we applied from Lepeule et al. (2012), and four studies showed
insignificant results.
3. 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 PM2.5, including both areas that
do not meet the fine particle standard and those areas that are in attainment, down
to the lowest modeled concentrations.
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 PM2.5 concentrations that fall below the bulk
of the observed data in these studies. As noted in the preamble to the rule, the range from the
ES-11
-------
25th to 10th percentiles of the air quality data in the epidemiology studies is a reasonable
range below which we start to have appreciably less confidence in the magnitude of the
associations observed in the epidemiological studies. Concentration benchmark analyses (e.g.,
25th percentile, 10th percentile, one standard deviation below the mean,4 and lowest
measured level [LML]) provide some insight into the level of uncertainty in the estimated PM2.5
mortality benefits. Most of the estimated avoided premature deaths for this rulemaking occur
at or above the lowest measured PM2.5 concentration in the two studies that are used to
estimate mortality benefits. 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 benefits
of air quality improvements. Rather, the core benefits estimates reported in this RIA (i.e., those
based on Krewski et al. [2009] and Lepeule et al. [2012]) are the best measures because they
reflect the full range of modeled air quality concentrations associated with the emission
reduction strategies and because the current body of scientific literature indicates that a no-
threshold model provides the best estimate of PM-related long-term mortality. It is important
to emphasize that "less confidence" does not mean "no confidence."
The estimated benefits reflect illustrative control measures and emission reductions to
lower PM2.5 concentrations at monitors projected to exceed the revised and alternative annual
standards. The result is that air quality is expected to improve in counties that exceed these
standards as well as surrounding areas that do not exceed the alternative standards. In order to
make a direct comparison between the benefits and costs of the emission reduction strategies,
it is appropriate to include all the benefits occurring as a result of the emission reduction
strategies applied regardless of where they occur. Therefore, it is not appropriate to estimate
the fraction of benefits that occur only in the counties that exceed the standards because it
would omit benefits attributable to emission reduction in exceeding counties. In addition, 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 PM2.5 concentrations
can vary spatially within a county. Some grid cells in a county can be below the level of the
alternative standard even though the highest monitor exceeds the alternative standard. Thus,
emission reductions can lead to benefits in grid cells that are below the alternative standards
within an exceeding county.
A range of one standard deviation around the mean represents approximately 68% of normally distributed data
and below the mean falls between the 25th and 10th percentiles.
ES-12
-------
ES.2.3 Cost Analysis Approach
The EPA estimated total costs under partial and full attainment of the alternative PM2.5
standards. The engineering costs 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 costs associated with applying known
controls. Costs for full attainment include estimates for the engineering costs of the additional
tons of emissions reductions that are needed beyond identified controls, referred to as
extrapolated costs. By definition, no cost data currently exist for the additional emissions
reductions needed beyond known controls. We employ two methodologies for estimating the
costs of unidentified future controls: a fixed-cost methodology and a hybrid methodology; both
approaches assume either that existing technologies can be applied in particular combinations
or to specific sources that we currently can't predict or that innovative strategies and new
control options make possible the emissions reductions needed for attainment by 2020. The
two approaches, however, implicitly reflect different assumptions about technological progress
and innovation in emissions reductions strategies. The fixed-cost methodology uses a
$15,000/ton estimate for each ton of PM2.s reduced, and the hybrid methodology generates a
total annual cost curve for PM2.s for unknown future controls that might be applied in order to
move toward 2020 attainment. The hybrid methodology has the advantage of incorporating
information about how significant the needed reductions from unspecified control technologies
are relative to the known control measures and matching that information with expected
increasing per-ton cost for applying unknown controls. Employing the fixed-cost methodology,
approximately 90% of total costs for attaining the revised annual standard of 12 u.g/m3 are from
unspecified control technologies. Employing the hybrid methodology, approximately 97% of
total costs for attaining the revised annual standard of 12 u.g/m3 are from unspecified control
technologies. The EPA recognizes that the extrapolated portion of the engineering cost
estimates reflects substantial uncertainty about which sectors and which technologies might
become available for cost-effective application in the future.
The engineering cost estimates are limited in their scope. Our analysis focuses on the
emissions reductions needed for attainment of the revised and alternative standards. Also, the
amendments to the ambient air monitoring regulations will revise the network design
ES-13
-------
requirements for PM2.5 monitoring sites, resulting in moving 21 monitors to established near-
road monitoring stations by January 1, 2015. The incremental cost associated with moving
these 21 monitors is a one-time cost of $28,570. Lastly, the EPA understands that some States
will incur costs designing SIPs and implementing new control strategies to meet the revised
standard. However, the EPA does not know what specific actions States will take to design their
SIPs to meet the revised standards; therefore, we do not include 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.
ES.2.4 Comparison of Benefits and Costs
The EPA's illustrative analysis has estimated the health and welfare benefits and costs
associated with the revised annual PM NAAQS. The results for 2020 suggest there will be
significant health and welfare benefits and these benefits will outweigh the costs associated
with the illustrative control strategies in 2020. In the analysis, we estimate the net benefits of
the revised annual PM2.s standard of 12 u.g/m3 and alternative annual standards of 13 u.g/m3
and 11 u.g/m3, incremental to the 2020 analytical baseline. For the revised annual standard of
12 u.g/m3, net benefits are estimated to be $3.7 billion to $9 billion at a 3% discount rate and
$3.3 billion to $8.1 billion at a 7% discount rate in 2020 (2010 dollars). For an alternative annual
standard of 13 u.g/m3, net benefits are estimated to be $1.2 billion to $2.9 billion at the 3%
discount rate and $1.1 billion to $2.6 billion at the 7% discount rate. Net benefits of an
alternative annual PM2.s standard of 11 u.g/m3 are estimated to be $11 billion to $29 billion at a
3% discount rate and $10 billion to $26 billion at a 7% discount rate in 2020. See Table ES-2.
For the revised annual standard of 12 u.g/m3, the EPA estimates that the benefits of full
attainment exceed the costs of full attainment by 12 to 171 times at a 3% discount rate. For the
alternative annual standard of 13 u.g/m3, the EPA estimates that the benefits of full attainment
exceed the costs of full attainment by 13 to 272 times at a 3% discount rate. For the alternative
annual standards of 11 u.g/m3, the EPA estimates that the benefits of full attainment exceed the
costs of full attainment by 8 to 90 times at a 3% discount rate.
ES-14
-------
Table ES-2. Total Monetized Benefits, Total Costs, and Net Benefits in 2020 (millions of
2010$)—Full Attainment3
Alternative
Annual
Standard
(ug/m3)
13
12
11
Total
3% Discount
Ratec
$11 to $100
$53 to $350
$320 to
$1,700
Costs"
7% Discount
Rate
$11 to $100
$53 to 350
$320 to
$1,700
Monetized
3% Discount
Rate
$1,300 to
$2,900
$4,000 to
$9,100
$13,000 to
$29,000
Benefits'1
7% Discount
Rate
$1,200 to
$2,600
$3,600 to
$8,200
$12,000 to
$26,000
Net
3% Discount
Rateb
$1,200 to
$2,900
$3,700 to
$9,000
$11,000 to
$29,000
Benefits
7% Discount
Rate
$1,100 to
$2,600
$3,300 to
$8,100
$10,000 to
$26,000
a These estimates reflect incremental emissions reductions from an analytical baseline that gives an "adjustment"
to the San Joaquin and South Coast areas in California for NOX emissions reductions expected to occur between
2020 and 2025, when those areas are expected to demonstrate attainment with the revised standards. Full
benefits of the revised standards in those two areas will not be realized until 2025.
b The two cost estimates do not represent lower- and upper-bound estimates but represent estimates generated
by two different methodologies. The lower estimate is generated using the fixed-cost methodology, which
assumes that technological change and innovation will result in the availability of additional controls by 2020 that
are similar in cost to the higher end of the cost range for current, known controls. The higher estimate is generated
using the hybrid methodology, which assumes that while additional controls may become available by 2020, they
become available at an increasing cost and the increasing cost varies by geographic area and by degree of difficulty
associated with obtaining the needed emissions reductions.
c Due to data limitations, we were unable to discount compliance costs for all sectors at 3%. See Chapter 7, Section
7.2.2 for additional details on the data limitations. As a result, the net benefit calculations at 3% were computed by
subtracting the costs at 7% from the monetized benefits at 3%.
d The reduction in premature deaths each year accounts for over 90% of total monetized benefits. Mortality risk
valuation assumes discounting over the SAB-recommended 20-year segmented lag structure. Not all possible
benefits or disbenefits are quantified and monetized in this analysis. B is the sum of all unquantified 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. The range of benefits reflects the range of the central
estimates from two mortality cohort studies (i.e., Krewski et al., 2009 and Lepeule et al., 2012).
Table ES-3. Benefit-to-Cost Ratios for Alternative Standards at 3% and 7% Based on
Projected Benefits and Costs in 2020
13 ug/m3
12 u,g/m3
u,g/m3
Benefit-Cost Ratio 3%a
Benefit-Cost Ratio 7%
13 to 272
11 to 246
12 to 171
11 to 154
8 to 90
7 to 81
a Due to data limitations, we were unable to discount compliance costs for all sectors at 3%. See Chapter 7, Section
7.2.2 for additional details on the data limitations. As a result, the net benefit calculations at 3% were computed by
subtracting the costs at 7% from the monetized benefits at 3%.
ES-15
-------
Figure ES-4 reflects the range of costs based on the calculation of costs using the fixed-
cost approach and the hybrid approach. Additionally, we see the difference in the calculation of
benefits based on using various studies.
Benefits using Lepeule etal. (2012!
58.2 billion*
Benefits using Krewskietal. (2009)
13.6 billion*
Costs using Hybrid Approach
S 3 F>0 million
Costs using Fixed Costs
^53 million
Note: Relative size of benefits and costs are to scale.
Figure ES-4. Monetized Benefit to Cost Comparison for 12 ng/m3 in 2020 (7% Discount Rate)
ES-16
-------
Table ES-4. Estimated Number of Avoided PM2.5 Health Impacts for Standard Alternatives-
Full Attainment in 2020a
Alternative Annual Standards
Health Effect
Adult Mortality
Krewski et al. (2009) (adult)
Lepeuleetal. (2012) (adult)
Woodruff et al. (1997) (infant)
Non-Fatal Heart Attacks (age >18)
Peters et al. (2001)
Pooled estimate of 4 studies
Hospital admissions— respiratory (all ages)
Hospital admissions— cardiovascular (age > 18)
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 (age 6-18)
Lost work days (age 18-65)
Minor restricted-activity days (age 18-65)
13 ng/m3
140
330
0
160
17
31
43
67
280
3,500
5,100
13,000
22,000
130,000
12 ng/m3
460
1,000
1
480
52
110
140
230
870
11,000
16,000
40,000
71,000
420,000
11 ng/m3
1,500
3,300
4
1,600
170
380
480
810
2,700
34,000
49,000
120,000
230,000
1,300,000
a Incidence estimates are rounded to whole numbers with no more than two significant figures.
ES.3 Discussion and Conclusions
An extensive body of scientific evidence documented in PM ISA indicates that PM2.5 can
penetrate deep into the lungs and cause serious health effects, including premature death and
other non-fatal illnesses (U.S. EPA, 2009). As described in the preamble to the rule, the
revisions to the standards are based on an integrative assessment of an extensive body of new
scientific evidence (U.S. EPA, 2009). Health studies published since the PM ISA (e.g., Pope et al.
[2009]) confirm that recent levels of PM2.5 have had a significant impact on public health. Based
on the air quality analysis in this RIA, the EPA projects that nearly all counties with PM2.5
monitors in the United States would meet an annual standard of 12 u.g/m3without additional
Federal, State, or local PM control programs. This demonstrates the substantial progress that
the United States has made in reducing air pollution emissions over the last several decades.
ES-17
-------
Regulations such as the EPA's recent Mercury and Air Toxics Standards (MATS) and other
Federal programs such as diesel standards will provide substantial improvements in regional
concentrations of PM2.s. Our analysis shows a few areas would still need additional emissions
reductions to address local sources of air pollution, including ports and uncontrolled industrial
emissions. For this reason, we have designed the RIA analysis to focus on local controls in these
few areas. We estimate that these additional local controls would yield benefits well in excess
of costs.
The setting of a NAAQS does not compel specific pollution reductions and as such does
not directly result in costs or benefits. For this reason, NAAQS RIAs are merely illustrative. The
NAAQS RIAs illustrate the potential costs and benefits of additional steps States could take to
attain a revised air quality standard nationwide beyond rules already on the books. We base
our illustrative estimates on an array of emission control strategies for different sources. The
costs and benefits identified in this RIA will not be realized until specific controls are mandated
by SIPs or other Federal regulations. In short, NAAQS RIAs hypothesize, but do not prescribe,
the control strategies that States may choose to enact when implementing a revised NAAQS.
It is important to emphasize that the EPA does not "double count" the costs or the
benefits of our rules. Emission reductions achieved under rules that require specific actions
from sources—such as MATS—are in the baseline of this NAAQS analysis, as are emission
reductions needed to meet the current NAAQS. For this reason, the cost and benefits estimates
provided in this RIA and all other NAAQS RIAs should not be added to the estimates for
implementation rules.
In calculating the costs, the EPA assumed the application of a significant number of
unidentified future controls that would make possible the additional emissions reductions
needed for attainment in 2020. EPA used two methodologies—the fixed-cost and hybrid
methodologies—for estimating the costs of unidentified future controls, and both approaches
assume either that existing technologies can be applied in particular combinations or to specific
sources that we currently can't predict or that innovative strategies and new control options
make possible the emissions reductions needed for attainment by 2020. Estimates generated
by the two approaches do not represent lower- and upper-bound estimates but simply
represent estimates generated by two different methodologies. The fixed-cost methodology
implicitly assumes that technological change and innovation will result in the availability of
additional controls by 2020 that are similar in cost to the higher end of the cost range for
current controls. The hybrid methodology implicitly assumes that while additional controls
become available by 2020, they become available at an increasing cost and the increasing cost
ES-18
-------
varies by geographic area and by degree of difficulty associated with obtaining the needed
emissions reductions.
For the revised annual standard of 12 u.g/m3, the total cost estimates comprise between
90 and 97% extrapolated cost estimates, and the estimated total cost using the hybrid
methodology is roughly 6.5 times more than the estimated total cost using the fixed-cost
methodology. Because the hybrid methodology reflects increasing marginal costs in areas
needing a higher ratio of emissions reductions from unknown to known controls, it could be
more representative of total costs. In an effort to consider the potential fitness of the
extrapolated cost estimates, we reviewed the South Coast Air Quality Management District's
(SCAQMD) 2012 Air Quality Management Plan (AQMP), and we located data on recent emission
reduction credit (ERC) transactions in both the SCAQMD and San Joaquin Valley Air Pollution
Control District (SJV APCD). While this information provides context for the extrapolated cost
estimates, the current relationship between available controls and costs to reduce emissions
may or may not be applicable in 2020 because of changes in innovation and advances in
technology.
The SCAQMD's 2012 AQMP includes information on control measures to meet the
current 24-hour standard of 35 u.g/m3, including further PM2.s controls for under-fired
charbroilers at a cost per ton reduced of $15,000. This control cost matches the parameter used
in the fixed-cost methodology, as well as the initial value used for the hybrid methodology and
is supportive of our selection of that value. In addition, the California Air Resources Board's
2009 and 2010 Emission Reduction Offset Transaction Costs, Summary Report included PM-m
ERC prices in both the SCAQMD and the SJV APCD. To some degree, ERC transaction prices
reflect a choice between installing a more stringent control and purchasing ERCs. Between 2009
and 2010 PMin ERC prices in SJV APCD ranged from $40,000 per ton per year (tpy) to
$70,000/tpy, and PMin ERC prices in the SCAQMD ranged from $575,000/tpy to more than $1.9
million/tpy. These prices reflect both marginal costs that are higher than the fixed-cost
estimates and marginal costs that are not inconsistent with the higher cost estimates generated
using the hybrid methodology. For further discussion of the total cost estimates, refer to
Section 7.2.4 in Chapter 7 of this RIA.
Furthermore, the monetized benefits estimates presented in this RIA are not intended
to capture the full burden of PM to public health but rather represent the incremental benefits
expected upon attaining the revised annual primary standard of 12 u.g/m3. In comparison,
modeling by Fann et al. (2012) estimated that 2005 levels of air pollution were responsible for
between 130,000 and 320,000 PM2.5-related deaths, or between 6.1% and 15% of total deaths
ES-19
-------
from all causes in the continental United States. The monetized benefits associated with
attaining the proposed range of standards appear modest when viewed within the context of
the potential overall public health burden of PM2.5and ozone air pollution estimated by Fann et
al. (2012), but this is primarily because regulations already on the books will make great strides
toward reducing future levels of PM. One important distinction between the total public health
burden estimated for 2005 air pollution levels and the estimated benefits in this RIA is that
ambient levels of PM2.5 will have improved substantially by 2020, due to major emissions
reductions resulting from implementation of Federal regulations. For example, we estimate
that S02 emissions (an important PM2.5 precursor) in the United States would fall from 14
million tons in 2005 to less than 5 million tons by 2020 (a reduction of 66%). For this reason,
States will only need to achieve small air quality improvements to reach the proposed PM
standards. As shown in the recent RIA for MATS (U.S. EPA, 2011b), implementing other Federal
and State air quality actions will address a substantial fraction of the total public health burden
of PM2.5 and ozone air pollution.
The NAAQS are not set at levels that eliminate the risk of air pollution. 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 PM 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, 2010c). While benefits occurring below the standard
are assumed to be more uncertain than those occurring above the standard, the EPA considers
these to be legitimate components of the total benefits estimate. Although there are greater
uncertainties at lower PM2.5 concentrations, there is no evidence of a population-level
threshold in PM2.5-related health effects in the epidemiology literature.
Lastly, the EPA recognizes that there are uncertainties in both the cost and benefit
estimates provided in this RIA. The EPA was unable to monetize fully all of the benefits
associated with reaching these standards in this RIA, including other health effects of PM,
visibility effects, ecosystem effects, and climate effects. If the EPA were able to monetize all of
the benefits, the benefits would exceed the estimated costs by an even greater margin.
ES-20
-------
ES.4 Caveats and Limitations
EPA acknowledges several important limitations of this analysis. These include the
following:
ES.4.1 Benefits Caveats
* PM2.5 mortality benefits represent a substantial proportion of total monetized
benefits (over 98%). To characterize the uncertainty in the relationship between
PM2.5 and premature mortality, we include a set of 12 estimates of the
concentration-response function based on results of the PM2.5 mortality expert
elicitation study in addition to our core estimates. Even these multiple
characterizations omit the uncertainty in air quality estimates, baseline incidence
rates, populations exposed, chemical composition, transferability of the effect
estimate to diverse locations, and additional uncertainty around the mean estimates
expressed by the experts. As a result, the reported confidence intervals and range of
estimates give an incomplete picture about the overall uncertainty in the PM2.5
estimates. This information should be interpreted within the context of the larger
uncertainty surrounding the entire analysis.
• Most of the estimated avoided premature deaths occur at or above the lowest
measured PM2.5 concentration in the two studies used to estimate mortality
benefits. In general, we have greater confidence in risk estimates based on PM2.5
concentrations where the bulk of the health and air quality data reside and
somewhat less confidence where data density is lower.
• We analyzed full attainment in 2020, and projecting key variables introduces
uncertainty. Inherent in any analysis of future regulatory programs are uncertainties
in projecting atmospheric conditions and source-level emissions, as well as
population, health baselines, incomes, technology, and other factors.
• There are uncertainties related to the health impact functions used in the analysis.
These include within-study variability; pooling across studies; the application of C-R
functions nationwide and for all particle species; extrapolation of impact functions
across populations; and various uncertainties in the C-R function, including causality
and shape of the function at low concentrations. Therefore, benefits may be under-
or over-estimates.
• This analysis omits certain unquantified effects due to lack of data, time, and
resources. These unquantified endpoints include other health and ecosystem
effects. The EPA will continue to evaluate new methods and models and select those
most appropriate for estimating the benefits of reductions in air pollution.
• Full benefits of the revised standards in San Joaquin and South Coast will not be
realized until 2025 when those areas are expected to demonstrate attainment with
the revised standards. If we were to estimate the monetized benefits for 2025, those
ES-21
-------
benefits would be higher due to population growth, aging of the population, and
income growth over time.
ES.4.2 Control Strategy and Cost Analysis Caveats and Limitations
* The control technologies applied as part of the illustrative control strategies
represent technologies that are available currently and may not reflect emerging
devices that may be available in future years to aid in attainment of the revised
standard. In addition, the emission reductions calculated from the known controls
(control efficiencies) 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.
• The illustrative control strategy analysis estimates only one potential pathway to
attainment. The control strategies are not recommendations for how the revised
PM2.5 standard should be implemented, and States will make all final decisions
regarding implementation strategies for the revised NAAQS.
• The application of known controls is based on source information obtained from the
emissions inventory. To the extent the inventory is lacking data on baseline controls
from SIPs, we may analyze control options that are currently in place.
• The future-year emissions used as a basis for developing the control strategies in
this RIA have implicit assumptions regarding emissions growth and control, which
differ by sector. For some emission sectors, these future-year emissions may not
reflect new sources locating in these areas.
• For two areas in California (South Coast Air Quality Management District and San
Joaquin Valley Air Pollution Control District) the degree of projected nonattainment
with the revised annual standard of 12 u.g/m3 is such that these areas are not
expected to be able to demonstrate attainment with the new standard by 2020.
These areas may qualify for up to a 5-year extension of their attainment date and
are likely to have until 2025 to demonstrate attainment with the revised annual
standard.
• The control technologies applied do not reflect potential effects of technological
change that may be available in future years and the effects of "learning by doing"
are not accounted for in the emissions reduction estimates. In our analysis, we do
not have the necessary data for cumulative output, fuel sales, or emission
reductions for all sectors included in order to properly generate control costs that
reflect learning-curve impacts. We believe the effect of including these impacts
would be to lower our estimates of costs for our control strategies in 2020.
• In addition to the application of known controls, the EPA assumes the application of
unidentified future controls that make possible the additional emission reductions
needed for attainment in 2020. By definition, no cost data currently exist for
ES-22
-------
unidentified future technologies or innovative strategies and the cost estimates for
unidentified future controls reflect some uncertainty.
• 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% 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% 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 this analysis, for the revised annual standard of 12 u.g/m3 and the
alternative standard of 13 u.g/m3 we did not have any disaggregated known control
cost data; therefore, we were not able to recalculate known control costs using a 3%
discount rate.
• The EPA understands that some States will incur costs designing SIPs and
implementing new control strategies to meet the revised annual standard. However,
the EPA does not know what specific actions States will take to design their SIPs to
meet the 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.
ES.5 Important Updates and Analytical Differences Between the PM NAAQS Proposal RIA
and the Final RIA
There have been several major improvements in the analytical components the EPA
used to estimate benefits and costs between the proposal RIA (June 2012) and this RIA
accompanying the final PM NAAQS. Important updates to our emissions, air quality modeling
and ambient data, air quality ratios, population projections, as well as currency year valuation
resulted in an improved analytical base for our analysis for the final rule. Based on the
complexity and magnitude of the updates and improvements made between the PM RIA
proposal and final RIA, it would not be appropriate to perform a simple direct comparison of
results. Therefore, each analysis stands alone and must be evaluated independently as such.
Below is a list summarizing some of the analytical changes between the proposal RIA
and the final RIA. Between the proposal RIA and the final RIA, we developed the control
strategies based on an improved modeling platform and updated the approach in designing the
control strategies. The improved modeling platform updated the current and projected air
ES-23
-------
quality levels for each area across the nation and the updated approach allowed for more
effective emissions reductions for each area in both attaining the current annual and daily
standards for our analytical baseline and in attaining the revised annual standard. For example,
in the proposal RIA we controlled precursor emissions of NOX and S02 over a broader region
and these emissions reductions were not as effective in reducing design value for each area as
the direct PM2.5 emissions reductions targeted in the final RIA. These analytical improvements
resulted in different estimates of costs and benefits between the proposal RIA and the final RIA
in which we have more confidence in reflecting the approaches that States will pursue to attain
the current and revised standards.
• New modeling platform—In the modeling platform for the final rule we included
key updates to the current ambient data that generally show improved air quality
when compared to modeling for the proposed rule, although daily design values
(DVs) for some areas in California increased due largely to wildfires in 2008. To
address the increases in these areas we adjusted the ambient data for these atypical
events.
• Future air quality with "on-the-books" controls—We also project PM2.5 air quality
levels between 2007 and the 2020 base case which were generally lower in the final
RIA compared to the proposal RIA largely because the starting values for the
ambient data were lower (i.e., final RIA air quality projections showed more
improvements than proposal RIA).
• Analytical baseline with attainment of 15/35—For the final RIA, EPA's approach to
attaining the existing standards of 15/35 u.g/m3 was improved with "air quality
adjustment ratios" that were based on more focused sensitivity model runs for
(i) specific areas like California counties; (ii) influential sources like residential wood
combustion; and (iii) specific PM emission species like directly emitted PM2.5. In the
proposal RIA, we conducted the analysis using more general air quality ratios that
reflected multiple sources such as point and area and precursor PM emissions like
NOX and S02. As a result, in the final RIA, the daily standard of 35 u.g/m3 was attained
more effectively and had less impact on annual DVs because of episodic, direct PM2.5
reductions, while in the proposal RIA, the daily standard was attained less effectively
because we pursued year-around N0xand S02 reductions that necessitated more
emissions reductions and had more impact on the annual DVs.
• Incremental air quality changes— In the final RIA, the use of improved air quality
adjustment ratios resulted in more incremental air quality improvement needed to
attain the annual standard of 12 u.g/m3 in California. Thus, these larger incremental
air quality improvements needed to attain the revised standard resulted in higher
estimated health benefits in the final rule compared to the proposal RIA. As
described above, this is because the daily standard of 35 u.g/m3 was attained more
ES-24
-------
effectively and had less impact on annual DVs across the counties in California. In
addition, the focus on direct PM2.5 emissions reductions allowed for more effective
controls to attain the revised standard with fewer PM2.5 emissions reductions
thereby resulting in similar fixed costs estimates.5
• Benefits analysis—EPA incorporated several updates that affected the core benefits
estimates in this RIA. Specifically, the EPA incorporated the most recent follow-up to
the Harvard Six Cities cohort study (Lepeule et al., 2012), which decreases the high
end of the monetized benefits range by 13%. The EPA also updated the demographic
data projections to reflect the 2010 Census, which increased the monetized benefits
by 4% percent for the revised standard. Additional epidemiology studies for
hospitalizations and emergency department visits and updated survival rates for
non-fatal heart attacks did not affect the rounded benefits estimates.
• Cost estimates— In the final RIA, the EPA presents a range of costs using both the
fixed and hybrid methodologies to estimate the costs associated with unknown
controls.
• Inflation—The EPA updated the currency year in this RIA to use 2010 dollars, which
increased both the costs and the monetized benefits by approximately 8% since the
proposal.
ES.6 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.
Fann, N., A. Lamson, K. Wesson, D. Risley, S.C. Anenberg, and B.J. Hubbell. 2012. "Estimating
the National Public Health Burden Associated with Exposure to Ambient PM2.s and
Ozone. Risk Analysis." Risk Analysis 32(1): 81-95.
Krewski, D., R.T. Burnett, M.S. Goldbert, K. Hoover, J. Siemiatycki, M. Jerrett, M. Abrahamowicz,
and W.H. White. 2009. "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.
Lepeule, J., F. Laden, D. Dockery, and J. Schwartz. 2012. "Chronic Exposure to Fine Particles and
Mortality: An Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009."
Environ Health Perspect In press. Available at: http://dx.doi.org/10.1289/ehp.1104660.
Peters, A., D. W. Dockery, J. E. Muller and M. A. Mittleman. 2001. "Increased particulate air
pollution and the triggering of myocardial infarction." Circulation 103(23): 2810-5.
5 The hybrid methodology cost estimates increased between the proposal RIA and the final RIA largely because a
large amount of emissions reductions were needed from one county with a low amount of known controls.
ES-25
-------
Pope, C.A. Ill, E. Majid, and D. Dockery. 2009. "Fine Particle Air Pollution and Life Expectancy in
the United States." New England Journal of Medicine 360: 376-386.
U.S. Environmental Protection Agency (U.S. EPA). 1997. Regulatory Impact Analyses for the
Particulate Matter and Ozone National Ambient Air Quality Standards and Proposed
Regional Haze Rule. Available at: http://www.epa.gov/ttn/oarpg/naaqsfin/ria.html.
U.S. Environmental Protection Agency (U.S. EPA). 1999. Regulatory Impact Analysis—Control of
Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and
Gasoline Sulfur Control Requirements. EPA 420-R-99-023. Available at:
http://www.epa.gov/otaq/regs/ld-hwy/tier-2/documents/r99023.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2000. Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Highway Heavy-Duty Engines. EPA 420-R-00-010.
Available at: http://www.epa.gov/otaq/regs/hd-hwy/2000frm/r00010.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2004. Final Regulatory Analysis: Control of
Emissions from Nonroad Diesel Engines. EPA420-R-04-007. Available at:
http://www.epa.gov/otaq/documents/nonroad-diesel/420r04007.pdf.
U.S. Environmental Protection Agency (U.S. EPA). November 17, 2005a. Federal Register. Vol.
70, No. 221. "Control of Air Pollution From Aircraft and Aircraft Engines; Emission
Standards and Test Procedures." 40 CFR Part 87. Available at:
http://www.epa.gov/fedrgstr/EPA-AIR/2005/November/Day-17/a22704.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2005b. Regulatory Impact Analysis for the
Final Clean Air Visibility Rule or the Guidelines for Best Available Retrofit Technology
(BART) Determinations Under the Regional Haze Regulations. EPA-452/R-05-004.
Available at: http://www.epa.gov/oar/visibility/pdfs/bart_ria_2005_6_15.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2006. Regulatory Impact Analysis, 2006
National Ambient Air Quality Standards for Particulate Matter, Chapter 5. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. October. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%205-Benefits.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Locomotive Engines and Marine Compression Ignition
Engines Less than 30 Liters Per Cylinder. Available at:
http://www.regulations.gov/#!documentDetail;D=EPA-HQ-OAR-2003-0190-0938.
U.S. Environmental Protection Agency (U.S. EPA). 2008b. Control of Emissions from Marine SI
and Small SI Engines, Vessels, and Equipment. Final Regulatory Impact Analysis.
Available at: http://www.regulations.gov/#!documentDetail;D=EPA-HQ-OAR-2004-
0008-0929.
ES-26
-------
U.S. Environmental Protection Agency (U.S. EPA). 2009a. Economic Impacts of Revised MACT
Standards for Hospital/Medical/Infectious Waste Incinerators. Available at:
http://www.epa.gOV/ttn/atw/129/hmiwi/h miwi_eia_090803.pdf.
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—RTP Division. December. Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546.
U.S. Environmental Protection Agency (U.S. EPA). 2010a. Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Category 3 Marine Diesel Engines. Available at:
http://www.epa.gov/oms/regs/nonroad/marine/ci/420r09019.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2010b. Regulatory Impact Analysis:
Regulatory Impact Analysis (RIA) for Existing Stationary Spark Ignition (SI) RICE NESHAP.
Available at: http://www.epa.gOV/ttn/atw/rice/fnl_si_rice_ria.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2010c. 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 on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011a. 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:
http://www.epa.gov/airtransport/pdfs/FinalRIA.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011b. Regulatory Impact Analysis for the
Final Mercury and Air Toxics Standards. EPA-452/R-11-011. December. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/matsriafinal.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011c. 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:
http://www.epa.gov/oar/sect812/febll/fullreport.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011d. Regulatory Impact Analysis: National
Emission Standards for Hazardous Air Pollutants for Industrial, Commercial, and
Institutional Boilers and Process Heaters. February. Available at:
http://www.epa.gov/ttnecasl/regdata/RIAs/boilersriaf inalll0221_psg.pdf.
U.S. Office of Management and Budget. Circular A-4, September 17, 2003. Available at:
http://www.whitehouse.gov/omb/circulars/a004/a-4.pdf.
ES-27
-------
Woodruff, T.J., J. Grille, and K.C. Schoendorf. 1997. "The Relationship Between Selected of
postneonatal infant mortality and particulate air pollution in the United States."
Environmental Health Perspectives 105(6): 608-612.
ES-28
-------
CHAPTER 1
INTRODUCTION AND BACKGROUND
1.1 Synopsis
This chapter summarizes the purpose and results of this Regulatory Impact Analysis
(RIA). This RIA estimates the human health and welfare benefits and costs of attaining the
revised and two alternative annual particulate matter (PM) National Ambient Air Quality
Standards (NAAQS) nationwide. 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 annual PM NAAQS and to meet the
requirements of Executive Orders 12866 and 13563.
1.2 Background
1.2.1 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." PM 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 "the attainment and maintenance of which in the
judgment of the Administrator, based on [the] criteria and allowing an adequate margin of
safety, [are] 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, [are] requisite to protect the 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."
1-1
-------
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.
1.2.2 2006PM NAAQS
In 2006, the EPA's final PM rule established a 24-hour standard of 35 u.g/m3 and
retained the annual standard of 15 u.g/m3. The EPA revised the secondary standards for fine
particles by making them identical in all respects to the primary standards. Following
promulgation of the final rule in 2006, several parties filed petitions for its review. On
February 24, 2009, the U.S. Court of Appeals for the District of Columbia Circuit remanded the
primary annual PM2.s NAAQS to the EPA citing that the EPA failed to adequately explain why the
standards provided the requisite protection from both short- and long-term exposures to fine
particles, including protection for at-risk populations. The court remanded the secondary
standards to the EPA citing that the Agency failed to adequately explain why setting the
secondary PM standards identical to the primary standards provided the required protection
for public welfare, including protection from visibility impairment. In 2006, the EPA also
retained the primary and secondary 24-hour PMi0 standards, both set at a level of 150 u.g/m3,
not to be exceeded more than once per year on average over 3 years (U.S. EPA, 1997).
1.3 Role of this RIA in the Process of Setting the NAAQS
1.3.1 Legal Requirement
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 new standards. The Act requires the EPA, for each criteria pollutant, to set standards that
protect 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.
1-2
-------
1.3.2 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-4.1 In accordance with these
guidelines, the RIA analyzes the benefits and costs associated with emissions controls to attain
the revised annual PM standard, incremental to attainment of the existing standards. In
addition, this RIA analyses two alternative primary annual PM2.5 standards: one that is more
stringent than the existing standards but less stringent than the revised annual standard and
another that is more stringent than the revised annual standard.
In the current PM NAAQS review, the EPA is revising and lowering the level of the
primary annual PM2.5 standard from 15 u.g/m3 to 12 u.g/m3 in conjunction with retaining the
level of the 24-hour standard at 35 u.g/m3. Thus, the incremental benefits and costs analyzed in
this RIA result from emissions controls needed to attain a more protective annual standard,
rather than the 24-hour standard of 35 u.g/m3. In addition to the revised annual standard of 12
u.g/m3, the RIA also analyzes the benefits and costs of incremental control strategies for two
alternative annual standards (13 u.g/m3 and 11 u.g/m3).
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 cannot be added 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.
The analytical baseline for this analysis does not assume emissions controls that might
be implemented to meet the other NAAQS for 03, NOX, or S02. To the extent that some of the
estimated emissions reductions needed to meet the revised annual PM standard would be
needed to meet the current standards for 03, NOX, or S02, the costs and benefits of meeting the
revised PM annual PM standard will be overstated. We did not conduct this analysis
incremental to controls applied as part of previous NAAQS analyses (e.g., 03, NOX, or S02)
1 U.S. Office of Management and Budget. Circular A-4, September 17, 2003, available at
.
1-3
-------
because the data and modeling on which these previous analyses were based are now
considered outdated and are not compatible with the current PM2.5 NAAQS analysis.
1.3.3 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." An externality occurs when parties to a
transaction do not bear its full consequences. An environmental problem, such as pollution
generated from production of a good, which imposes health costs on those who neither
produce nor consume it, is a classic case of an externality. In the presence of externalities, a
free market does not ensure an efficient allocation of resources. Setting and implementing
primary and secondary air quality standards is one way the government can address an
externality and increase air overall public health and welfare.
1.3.4 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.
Because the illustrative goals of this RIA are somewhat different from other EPA
analyses of national rules or the implementation plans States develop, 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 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.
1-4
-------
1.4 Overview and Design of the Final RIA
The RIA evaluates the costs and benefits of hypothetical national strategies to attain the
revised annual standard of 12 u.g/m3 and two alternative annual PM standards, incremental to
attainment of the existing 15/35 u.g/m3 standards.
The EPA is retaining the current primary and secondary 24-hour PMi0 standards, which
are both set at a level of 150 u.g/m3, not to be exceeded more than once per year on average
over 3 years (U.S. EPA, 1997). Because the benefit-cost analysis of the alternative PMi0
standards was conducted when the standard was promulgated in 1997, this RIA does not repeat
that analysis here.
1.4.1 Important Updates and Differences Between the PM NAAQS Proposal RIA and the
Final RIA
There have been several major improvements in the analytical components the EPA
used to estimate benefits and costs between the proposal RIA (June 2012) and this RIA
accompanying the final PM NAAQS. Important updates to our emissions, air quality modeling
and ambient data, air quality ratios, population projections, and currency-year valuation
resulted in an improved analytical base for our analysis for the final rule. Based on the
complexity and magnitude of the updates and improvements made between the PM RIA
proposal and final RIA, it would not be appropriate to perform a simple direct comparison of
results. Therefore, each analysis stands alone and must be evaluated independently as such.
Below is a list summarizing some of the analytical changes between the proposal RIA
and the final RIA. Between the proposal RIA and the final RIA, we developed the control
strategies based on an improved modeling platform and updated the approach in designing the
control strategies. The improved modeling platform updated the current and projected air
quality levels for each area across the nation and the updated approach allowed for more
effective emissions reductions for each area in both attaining the current annual and daily
standards for our analytical baseline and in attaining the revised annual standard. For example,
in the proposal RIA we controlled precursor emissions of NOX and S02 over a broader region
and these emissions reductions were not as effective in reducing design value for each area as
the direct PM2.s emissions reductions targeted in the final RIA. These analytical improvements
resulted in different estimates of costs and benefits between the proposal RIA and the final RIA
which we have more confidence in reflecting the approaches that states will pursue to attain
the current and revised standards.
1-5
-------
New modeling platform—In the modeling platform for the final rule we included
key updates to the current ambient data that generally show improved air quality
when compared to modeling for the proposed rule, although daily design values
(DVs) for some areas in California increased due largely to wildfires in 2008. To
address the increases in these areas we adjusted the ambient data for these atypical
events.
Future air quality with "on-the-books" controls—We also project PM2.5 air quality
levels between 2007 and the 2020 base case which were generally lower in the final
RIA compared to the proposal RIA largely because the starting values for the
ambient data were lower (i.e., final RIA air quality projections showed more
improvements than proposal RIA).
Analytical baseline with attainment of 15/35—For the final RIA, the EPA's approach
to attaining the existing standards of 15/35 u.g/m3 was improved with "air quality
adjustment ratios" that were based on more focused sensitivity model runs for (i)
specific areas like California counties; (ii) influential sources like residential wood
combustion; and (iii) specific PM emission species like directly emitted PM2.5. In the
proposal RIA, we conducted the analysis using more general air quality ratios that
reflected multiple sources such as point and area and precursor PM emissions like
NOX and S02. As a result, in the final RIA, the daily standard of 35 u.g/m3 was attained
more effectively and had less impact on annual DVs because of episodic, direct PM2.5
reductions, while in the proposal RIA, the daily standard was attained less effectively
because we pursued year-around N0xand S02 reductions that necessitated more
emissions reductions and had more impact on the annual DVs.
Incremental air quality changes—\r\ the final RIA, the use of improved air quality
adjustment ratios resulted in more incremental air quality improvement needed to
attain the annual standard of 12 u.g/m3 in California. Thus, these larger incremental
air quality improvements needed to attain the revised standard resulted in higher
estimated health benefits in the final rule compared to the proposal RIA. As
described above, this is because the daily standard of 35 u.g/m3 was attained more
effectively and had less impact on annual DVs across the counties in California. In
addition, the focus on direct PM2.5 emissions reductions allowed for more effective
controls to attain the revised standard with fewer PM2.5 emissions reductions
thereby resulting in similar fixed costs estimates.2
Benefits analysis—The EPA incorporated several updates that affected the core
benefits estimates in this RIA. Specifically, the EPA incorporated the most recent
follow-up to the Harvard Six Cities cohort study (Lepeule et al., 2012), which
decreases the high end of the monetized benefits range by 13%. The EPA also
updated the demographic data projections to reflect the 2010 Census, which
2 The hybrid methodology cost estimates increased between the proposal RIA and the final RIA largely because a
large amount of emissions reductions were needed from one county with a low amount of known controls.
1-6
-------
increased the monetized benefits by 4% percent for the revised standard. Additional
epidemiology studies for hospitalizations and emergency department visits and
updated survival rates for non-fatal heart attacks did not affect the rounded benefits
estimates.
• Cost estimates— In the final RIA, the EPA presents a range of costs using both the
fixed and hybrid methodologies to estimate the costs associated with unknown
controls.
• Inflation—The EPA updated the currency year in this RIA to use 2010 dollars, which
increased both the costs and the monetized benefits by approximately 8% since the
proposal.
1.4.2 Existing and Revised PM Air Quality Standards
Two primary PM2.5 standards provide public health protection from effects associated
with fine particle exposures: the annual standard and the 24-hour standard. The existing annual
standard is set at a level of 15.0 u.g/m3, based on the 3-year average of the annual arithmetic
mean of PM2.5 concentrations. The existing 24-hour standard is set at a level of 35 u.g/m3, based
on the 3-year average of the 98th percentile of 24-hour PM2.5 concentrations. In this RIA, the
existing primary PM2.5 standards, including both the annual standard and 24-hour standard, are
denoted as 15/35 u.g/m3. In this current PM NAAQS review, the EPA has revised the level of the
primary annual PM2.5 standard to 12 u.g/m3 in conjunction with retaining the level of the 24-
hour standard at 35 u.g/m3.
Currently, the existing secondary (welfare-based) PM2.5 standards are identical in all
respects to the primary standards. In this PM NAAQS review, the EPA is retaining the current
suite of secondary standards for 24-hour and annual PM2.5. Thus, while the new primary annual
standard will be revised to 12 u.g/m3, the secondary annual standard will remain at 15 u.g/m3.
Non-visibility welfare effects are addressed by this suite of secondary standards, and PM-
related visibility impairment is addressed by the secondary 24-hour PM2.5 standard, which EPA
is leaving unchanged at 35 u.g/m3. The secondary standards will thus remain at 15/35 u.g/m3.
With regard to the primary and secondary standards for particles less than or equal to
10 u.m in diameter (PMio), the EPA is retaining the current primary and secondary 24-hour PMio
standards. Both standards are the same. The current primary and secondary 24-hour standards
are set at a level of 150 u.g/m3, not to be exceeded more than once per year on average over 3
years (U.S. EPA, 1997).
1-7
-------
1.4.2.1 Graphical Overview of the RIA Analysis
PM RIA Analytical Flowchart
See Individual Chapters for Complete Description
Adjusted 2020 Base Case with "On the Books"
Federal and State Rules and Programs
Establishing the Analytical
Baseline
Analytical Baseline with Attainment of 15/35
^^H
M
Select Controls by Emissions Sector and Pollutant
Control Strategies
Analysis of Costs
Adjust 2020 Control Scenario to Meet Alternate Standards
Estimate Known Controls Engineering Costs
Estimate Extrapolated Costs
Calculate Full Attainment
Estimate Full Attainment
Benefits
Full Attainment
T
Net Benefits
Figure 1-1. PM RIA Analytical Flow Diagram
1.4.2.2 Establishing the Analytical Baseline
The RIA is intended to evaluate the costs and benefits of reaching attainment with
alternative PM2.5 standards. In order to develop and evaluate control strategies for attaining a
more stringent primary standard, it is important to account for Federal and state rules and
programs currently underway, as well as to reflect attainment of the current annual and daily
standards of 15/35 u.g/m3. Estimating the 2020 levels after attainment of the current standards
of 15/35 u.g/m3 then allows us to estimate the incremental costs and benefits of attaining any
alternative primary standard. EPA anticipates attainment with 15/35 u.g/m3 by 2020 in all but
two areas in California, not expected to attain the current standards until 2025.
1-8
-------
The analytical baseline includes reductions already achieved as a result of national
regulations and reductions expected prior to 2020 from recently promulgated national
regulations, referred to as the base case. Reductions achieved as a result of State and local
agency regulations and voluntary programs are reflected to the extent that they are
represented in emission inventory information submitted to the EPA by State and local
agencies. Below is a list of some of the major national rules reflected in the base case. Refer to
Chapter 3, Section 3.2.1.4 for a more detailed discussion of the rules reflected in the 2020 base
case emissions inventory.
• Light-Duty Vehicle Tier 2 Rule (U.S. EPA, 1999)
• Heavy Duty Diesel Rule (U.S. EPA, 2000)
• Clean Air Nonroad Diesel Rule (U.S. EPA, 2004)
• 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, 2005a)
• Emissions Standards for Locomotives and Marine Compression-Ignition Engines (U.S.
EPA, 2008a)
• Control of Emissions for Nonroad Spark Ignition Engines and Equipment (U.S. EPA,
2008b)
• C3 Oceangoing Vessels (U.S. EPA, 2010a)
• Boiler MACT (U.S. EPA, 2011d)
• Hospital/Medical/lnfectious Waste Incinerators: New Source Performance Standards
and Emission Guidelines: Final Rule Amendments (U.S. EPA, 2009a)
• Reciprocating Internal Combustion Engines (RICE) NESHAPs (U.S. EPA, 2010b)
• Mercury and Air Toxics Standards (U.S. EPA, 2011b)
• Cross-State Air Pollution Rule (CSAPR) (U.S. EPA, 2011a)3
The analytical baseline for this analysis does not assume emissions controls that might
be implemented to meet the other NAAQS for 03, NOX, or S02. To the extent that some of the
3 See Chapter 3, Section 3.2.1.4 for a discussion of the role CSAPR plays in the PM2.5 RIA and the reasons we believe
CSAPR remains an appropriate proxy for this analysis.
1-9
-------
estimated emissions reductions needed to meet the revised annual PM standard are also
needed to meet the current standards for 03, NOX, or S02, the costs and benefits of meeting the
revised PM annual standard will be overstated. We did not conduct this analysis incremental to
controls applied as part of previous NAAQS analyses (e.g., 03, NOX, or S02) because the data
and modeling on which these previous analyses were based are now considered outdated and
are not compatible with the current PM2.s NAAQS analysis. In addition, all control strategies
analyzed in all NAAQS RIAs are hypothetical. The analysis presented here is just one scenario
that States may employ but does not prescribe how attainment must be achieved.
Most areas of the United States will be required to demonstrate attainment with the
new standard by 2020. As a result, for these areas, the correct baseline for estimating the
incremental emissions reductions that would be needed to attain the more protective standard
is a baseline with emissions projected to 2020 and adjusted to reflect the additional emissions
reductions that would be needed to attain the current 15/35 standards. For two areas in
Southern California (South Coast and San Joaquin), the degree of projected non-attainment
with the revised annual standard 12 u.g/m3 is high enough that those counties are not expected
to be able to demonstrate attainment of the new standard by 2020. Instead, those two areas
are likely to qualify for an extension of their attainment date of up to 5 years. If the areas are
granted an attainment date extension, they will have until 2025 to demonstrate attainment of
the revised annual standard of 12 u.g/m3. As a result, for these two areas, the correct baseline
for estimating the incremental emissions reductions that would be needed to attain the more
protective standard is a baseline with emissions projected to 2025 adjusted to reflect the
additional emissions reductions that would be needed to attain the current 15/35
u.g/m3standards.
This difference in attainment year is important because between 2020 and 2025
emissions from mobile sources in California are expected to be reduced because of continued
fleet turnover from older, higher emitting vehicles to newer, lower emitting vehicles. These
reductions in emissions will occur as a result of previous State rules for which costs and benefits
have already been counted and thus will not be costs and benefits attributable to meeting the
revised annual standard of 12 u.g/m3. For California, the provided future-year 2020 and 2025
emissions included most on-the-books regulations such as those for low sulfur fuel, idling of
heavy-duty vehicles, chip reflash, public fleets, trash trucks, drayage trucks, and heavy duty
trucks and buses. See Chapter 3, Section 3.2.1.4 for further details on California emission
inventories.
1-10
-------
For the purposes of this analysis, we have constructed an analytical baseline that
reflects attainment of 15/35 u.g/m3 in 2020. This analytical baseline is modified in the South
Coast Air Quality Management District and the San Joaquin Valley Air Pollution Control District
to reflect an "adjustment" for the reductions in NOX emissions that those areas are expected to
see between 2020 and 2025. These reductions in NOX are not attributable to attainment of the
current or revised PM standards but rather reflect the impacts of other programs. These NOX
emissions changes will affect baseline PM concentrations but will not affect costs or benefits of
attaining the revised annual or the alternative annual standards.
To provide the most reasonable and reliable estimates of costs and benefits of full
attainment for the nation, we construct an analytical baseline for estimating the costs and
benefits of attaining the revised annual standard of 12 u.g/m3,13 u.g/m3, and 11 u.g/m3 with the
following characteristics: (1) reflects on-the-books regulations as implemented through 2020
plus additional emissions reductions needed to meet the 15/35 u.g/m3 standard levels, and (2)
additional mobile source emissions reductions expected to occur between 2020 and 2025 for
California's South Coast and San Joaquin areas, which are likely to not demonstrate attainment
until 2025. Essentially, we are modifying the baseline in those two areas to reflect an
"adjustment" for the reductions in NOX emissions that those areas are expected to see between
2020 and 2025. This allows us to generate costs and benefits of full attainment without
overstating the costs and benefits in those two areas, which would occur if we forced costly
emissions reductions in 2020 in areas that would not have to occur until 2025 and that will be
offset because of the expected reductions in mobile source emissions due to other programs.
See Chapter 3, Section 3.2.1.4 for details on emission inventories from California.
Benefits for all areas are estimated using 2020 population data for consistency,
recognizing that full attainment costs and benefits will not actually be realized until 2025 for a
portion of the costs and benefits. The 2020 estimates of full attainment costs and benefits will
be an underestimate of benefits in 2025 because of population growth and changes in the age
distribution of the population between 2020 and 2025.
1.4.3 Health and Welfare Co-Benefits Analysis Approach
The EPA estimated impacts on human health (e.g., mortality and morbidity effects)
under full attainment of the three alternative annual PM2.s standards. We considered an array
of health impacts attributable to changes in PM2.s exposure and estimated these benefits using
the BenMAP model (Abt Associates, 2012), which has been used in many recent RIAs (e.g., U.S.
EPA, 2006, 2011a, 2011b), and The Benefits and Costs of the Clean Air Act 1990 to 2020 (U.S.
1-11
-------
EPA, 2011c). The monetized benefits estimated in the core analysis include avoided premature
deaths (derived from effect coefficients in tow cohort studies [Krewski et al. (2009) and Lepeule
et al. (2012)] for adults and one for infants [Woodruff et al. (1997)]) as well as avoided
morbidity effects for 10 non-fatal endpoints ranging in severity from lower respiratory
symptoms to heart attacks. As noted above, because California is the only state that needs
additional air quality improvement beyond the analytical baseline after accounting for expected
air quality improvements expected from recent rules, all of the benefits associated with the
revised standard of 12ug/m3 occur in California.
Since the proposed rule, 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 follow-up to the Harvard Six Cities cohort study (Lepeule et al., 2012), more recent
demographic data projections, additional hospitalization and emergency department visit
studies, inflation adjustment to 2010 dollars, and an expanded uncertainty assessment. Each of
these updates is fully described in the health benefits chapter (Chapter 5) and summarized
below in section ES.5. Compared with the proposal benefits, the estimated benefits for the
revised standard are about double due to a combination of updates to the analytic baseline
Even though the primary standards are designed to protect against adverse effects to
human health, the emission reductions will 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, such as reductions in visibility impairment, materials damage,
and ecosystem damage. Despite our attempts to quantify and monetize as many of the benefits
as possible, the welfare co-benefits associated with meeting the alternative standards are not
quantified or monetized in this analysis. Unquantified health benefits are discussed in Chapter
5, and unquantified welfare co-benefits are discussed in Chapter 6.
It is important to note that estimates of the health benefits from reduced PM2.s
exposure reported here contain uncertainties, which are described in detail in Chapter 5 and
Appendix 5b. Below are two key assumptions in the benefits analysis:
1. We assume that all fine particles, regardless of their chemical composition, are
equally potent in causing premature mortality. This is an important assumption,
because PM2.s varies considerably in composition across sources, but the scientific
evidence is not yet sufficient to allow differentiation of effect estimates by particle
type. The Integrated Science Assessment for Particulate Matter (PM ISA), which was
twice reviewed by CASAC, concluded that "many constituents of PM2.s can be linked
with multiple health effects, and the evidence is not yet sufficient to allow
1-12
-------
differentiation of those constituents or sources that are more closely related to
specific outcomes" (U.S. EPA, 2009). These uncertainties are likely to be magnified in
the current analysis to the extent that the emissions controls are less diverse when
focusing on one small region of the country rather than a broader geography with
more diverse emissions sources and the application of a more diverse set of
controls.
2. We assume that health impact functions based on national studies are
representative for exposures and populations in California. In addition to the
national risk coefficients we use as our core estimates, the EPA considered the
cohort studies conducted in California specifically. Although we have not calculated
the benefits results using the cohort studies conducted in California, we provided
these risk coefficients to show how much the monetized benefits could have
changed. Most of the California cohort studies report central effect estimates similar
to the (nation-wide) all-cause mortality risk estimate we applied from Krewski et al.
(2009) and Lepeule et al. (2012) albeit with wider confidence intervals. Three cohort
studies conducted in California indicate statistically significant higher risks than the
risk estimates we applied from Lepeule et al. (2012), and four studies showed
insignificant results.
3. 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 PM2.5, including both areas that
do not meet the fine particle standard and those areas that are in attainment, down
to the lowest modeled concentrations.
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 PM2.5 concentrations that fall below the bulk
of the observed data in these studies. As noted in the preamble to the rule, the range from the
25th to 10th percentiles of the air quality data in the epidemiology studies is a reasonable
range below which we start to have appreciably less confidence in the magnitude of the
associations observed in the epidemiological studies. Concentration benchmark analyses (e.g.,
25th percentile, 10th percentile, one standard deviation below the mean,4 and lowest
measured level [LML]) provide some insight into the level of uncertainty in the estimated PM2.5
mortality benefits. Most of the estimated avoided premature deaths for this rulemaking occur
at or above the lowest measured PM2.5 concentration in the two studies that are used to
estimate mortality benefits. There are uncertainties inherent in identifying any particular point
A range of one standard deviation around the mean represents approximately 68% of normally distributed data
and below the mean falls between the 25th and 10th percentiles.
1-13
-------
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 benefits
of air quality improvements. Rather, the core benefits estimates reported in this RIA (i.e., those
based on Krewski et al. [2009] and Lepeule et al. [2012]) are the best measures because they
reflect the full range of modeled air quality concentrations associated with the emission
reduction strategies and because the current body of scientific literature indicates that a no-
threshold model provides the best estimate of PM-related long-term mortality. It is important
to emphasize that "less confidence" does not mean "no confidence."
The estimated benefits reflect illustrative control measures and emission reductions to
lower PM2.5 concentrations at monitors projected to exceed the revised and alternative annual
standards. The result is that air quality is expected to improve in counties that exceed these
standards as well as surrounding areas that do not exceed the alternative standards. In order to
make a direct comparison between the benefits and costs of the emission reduction strategies,
it is appropriate to include all the benefits occurring as a result of the emission reduction
strategies applied regardless of where they occur. Therefore, it is not appropriate to estimate
the fraction of benefits that occur only in the counties that exceed the standards because it
would omit benefits attributable to emission reduction in exceeding counties. In addition, 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 PM2.s concentrations
can vary spatially within a county. Some grid cells in a county can be below the level of the
alternative standard even though the highest monitor exceeds the alternative standard. Thus,
emission reductions can lead to benefits in grid cells that are below the alternative standards
within an exceeding county.
1.4.4 Cost Analysis Approach
The EPA estimated total costs under partial and full attainment of the alternative PM2.s
standards. The engineering costs 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.
1-14
-------
The partial attainment cost analysis reflects the costs associated with applying known
controls. Costs for full attainment include estimates for the engineering costs of the additional
tons of emissions reductions that are needed beyond identified controls, referred to as
extrapolated costs. By definition, no cost data currently exist for the additional emissions
reductions needed beyond known controls. We employ two methodologies for estimating the
costs of unidentified future controls: a fixed-cost methodology and a hybrid methodology; both
approaches assume either that existing technologies can be applied in particular combinations
or to specific sources that we currently can't predict or that innovative strategies and new
control options make possible the emissions reductions needed for attainment by 2020. The
two approaches, however, implicitly reflect different assumptions about technological progress
and innovation in emissions reductions strategies. The fixed-cost methodology uses a
$15,000/ton estimate for each ton of PM2.5 reduced, and the hybrid methodology generates a
total annual cost curve for PM2.5 for unknown future controls that might be applied in order to
move toward 2020 attainment. The hybrid methodology has the advantage of incorporating
information about how significant the needed reductions from unspecified control technologies
are relative to the known control measures and matching that information with expected
increasing per-ton cost for applying unknown controls. Employing the fixed-cost methodology,
approximately 90% of total costs for attaining the revised annual standard of 12 u.g/m3 are from
unspecified control technologies. Employing the hybrid methodology, approximately 97% of
total costs for attaining the revised annual standard of 12 u.g/m3 are from unspecified control
technologies. The EPA recognizes that the extrapolated portion of the engineering cost
estimates reflects substantial uncertainty about which sectors and which technologies might
become available for cost-effective application in the future.
The engineering cost estimates are limited in their scope. Our analysis focuses on the
emissions reductions needed for attainment of the revised and alternative standards. Also, the
amendments to the ambient air monitoring regulations will revise the network design
requirements for PM2.s monitoring sites, resulting in moving 21 monitors to established near-
road monitoring stations by January 1, 2015. The incremental cost associated with moving
these 21 monitors is a one-time cost of $28,570. Lastly, the EPA understands that some States
will incur costs designing SIPs and implementing new control strategies to meet the revised
standard. However, the EPA does not know what specific actions States will take to design their
SIPs to meet the revised standards; therefore, we do not include 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
1-15
-------
implementation of specific technologies, especially for technologies that are not necessarily
market driven.
1.5 Organization of this Regulatory Impact Analysis
This RIA includes the following 10 chapters:
• Chapter 1: Introduction and Background. This chapter introduces the purpose of the
RIA.
• Chapter 2: Defining the PM Air Quality Problem. This chapter characterizes the
nature, scope, and magnitude of the current-year PM2.5 problem.
• 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 PM2.5 concentrations in 2020 after applying the control
strategies.
• Chapter 5: Health Benefits Analysis Approach and Results. This chapter quantifies
and monetizes the health benefits of the PM2.5-related air quality improvements
associated with the hypothetical control strategies.
• Chapter 6: Welfare Co-Benefits of the Primary Standard. This chapter describes the
welfare effects, including changes in visibility, materials damage, ecological effects
from PM deposition, ecological effects from nitrogen and sulfur emissions,
vegetation effects from ozone exposure, 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.
1-16
-------
• Chapter 10: Qualitative Discussion of Employment Impacts of Air Quality
Regulations. This chapter provides a qualitative discussion of employment impacts
of air quality regulations.
1.6 References for Chapter 1
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.
Krewski, D., R.T. Burnett, M.S. Goldbert, K. Hoover, J. Siemiatycki, M. Jerrett, M. Abrahamowicz,
and W.H. White. 2009 "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.
Lepeule, J., F. Laden, D. Dockery, and J. Schwartz. 2012. "Chronic Exposure to Fine Particles and
Mortality: An Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009."
Environ Health Perspect. In press. Available at: http://dx.doi.org/10.1289/ehp.1104660.
U.S. Environmental Protection Agency. 1997. Regulatory Impact Analyses for the Particulate
Matter and Ozone National Ambient Air Quality Standards and Proposed Regional Haze
Rule. Available at: http://www.epa.gov/ttn/oarpg/naaqsfin/ria.html.
U.S. Environmental Protection Agency (U.S. EPA). 1999. Regulatory Impact Analysis—Control of
Air Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and
Gasoline Sulfur Control Requirements. EPA 420-R-99-023. Available at:
http://www.epa.gov/otaq/regs/ld-hwy/tier-2/documents/r99023.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2000. Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Highway Heavy-Duty Engines. EPA 420-R-00-010.
Available at: http://www.epa.gov/otaq/regs/hd-hwy/2000frm/r00010.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2004. Final Regulatory Analysis: Control of
Emissions from Nonroad Diesel Engines. EPA420-R-04-007. Available at:
http://www.epa.gov/otaq/documents/nonroad-diesel/420r04007.pdf.
U.S. Environmental Protection Agency (U.S. EPA). November 17, 2005a. Federal Register. Vol.
70, No. 221. "Control of Air Pollution From Aircraft and Aircraft Engines; Emission
Standards and Test Procedures." 40 CFR Part 87. Available at:
http://www.epa.gov/fedrgstr/EPA-AIR/2005/November/Day-17/a22704.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2005b. Regulatory Impact Analysis for the
Final Clean Air Visibility Rule or the Guidelines for Best Available Retrofit Technology
(BART) Determinations Under the Regional Haze Regulations. EPA-452/R-05-004.
Available at: http://www.epa.gov/oar/visibility/pdfs/bart_ria_2005_6_15.pdf.
1-17
-------
U.S. Environmental Protection Agency (U.S. EPA). 2006. Regulatory Impact Analysis, 2006
National Ambient Air Quality Standards for Particulate Matter, Chapter 5. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. October. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%205-Benefits.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Locomotive Engines and Marine Compression Ignition
Engines Less than 30 Liters Per Cylinder. Available at:
http://www.regulations.gov/#!documentDetail;D=EPA-HQ-OAR-2003-0190-0938.
U.S. Environmental Protection Agency (U.S. EPA). 2008b. Control of Emissions from Marine SI
and Small SI Engines, Vessels, and Equipment. Final Regulatory Impact Analysis.
Available at: http://www.regulations.gov/#!documentDetail;D=EPA-HQ-OAR-2004-
0008-0929.
U.S. Environmental Protection Agency (U.S. EPA). 2009a. Economic Impacts of Revised MACT
Standards for Hospital/Medical/lnfectious Waste Incinerators. Available at:
http://www.epa.gov/ttn/atw/129/hmiwi/hmiwi_eia_090803.pdf.
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—RTP Division. December. Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546.
U.S. Environmental Protection Agency (U.S. EPA). 2010a. Regulatory Impact Analysis: Control of
Emissions of Air Pollution from Category 3 Marine Diesel Engines. Available at:
http://www.epa.gov/oms/regs/nonroad/marine/ci/420r09019.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2010b. Regulatory Impact Analysis:
Regulatory Impact Analysis (RIA) for Existing Stationary Spark Ignition (SI) RICE NESHAP.
Available at: http://www.epa.gOV/ttn/atw/rice/fnl_si_rice_ria.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011a. 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:
http://www.epa.gov/airtransport/pdfs/FinalRIA.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011b. Regulatory Impact Analysis for the
Final Mercury and Air Toxics Standards. EPA-452/R-11-011. December. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/matsriafinal.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011c. 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:
http://www.epa.gov/oar/sect812/febll/fullreport.pdf.
1-18
-------
U.S. Environmental Protection Agency (U.S. EPA). 2011d. Regulatory Impact Analysis: National
Emission Standards for Hazardous Air Pollutants for Industrial, Commercial, and
Institutional Boilers and Process Heaters. February. Available at:
http://www.epa.gov/ttnecasl/regdata/RIAs/boilersriaf inalll0221_psg.pdf.
U.S. Office of Management and Budget. Circular A-4, September 17, 2003. Available at
http://www.whitehouse.gov/omb/circulars/a004/a-4.pdf.
Woodruff, T.J., J. Grille, and K.C. Schoendorf. 1997. "The Relationship Between Selected of
postneonatal infant mortality and particulate air pollution in the United States."
Environmental Health Perspectives 105 (6): 608-612.
1-19
-------
CHAPTER 2
DEFINING THE PM AIR QUALITY PROBLEM
2.1 Synopsis
This chapter characterizes the nature, scope and magnitude of the current year
particulate matter (PM) problem. It includes a summary of the spatial and temporal distribution
of PM2.5 and the likely origin from direct emissions or atmospheric transformations of gaseous
precursors and recent design values for PM2.5.
2.2 Particulate Matter (PM) Properties
PM is a highly complex mixture of solid particles and liquid droplets distributed among
numerous atmospheric gases which interact with solid and liquid phases. Particles range in size
from those smaller than 1 nanometer (10~9 meter) to over 100 micrometer (u.m, or 10"6 meter)
in diameter (for reference, a typical strand of human hair is 70 u.m in diameter and a grain of
salt is about 100 u.m). Atmospheric particles can be grouped into several classes according to
their aerodynamic and physical sizes, including ultrafine particles (<0.1 u.m), accumulation
mode or "fine" particles (0.1 to ~3 u.m), and coarse particles (>1 u.m). For regulatory purposes,
fine particles are measured as PM2.s and inhalable or thoracic coarse particles are measured as
PMio-2.5, corresponding to their size (diameter) range in micrometers and referring to total
particle mass under 2.5 and between 2.5 and 10 micrometers, respectively. The EPA currently
has standards that measure PM2.s and PMi0.
Particles span many sizes and shapes and consist of hundreds of different chemicals.
Particles are emitted directly from sources and are also formed through atmospheric chemical
reactions; the former are often referred to as "primary" particles, and the latter as "secondary"
particles. Particle pollution also varies by time of year and location and is affected by several
weather-related factors, such as temperature, clouds, humidity, and wind. A further layer of
complexity comes from particles' ability to shift between solid/liquid and gaseous phases,
which is influenced by concentration and meteorology, especially temperature.
Particles are made up of different chemical components. The major chemical
components include carbonaceous materials (carbon soot and organic compounds), and
inorganic compounds including, sulfate and nitrate compounds that usually include ammonium,
and a mix of substances often apportioned to crustal materials such as soil and ash. As
mentioned above, particulate matter includes both "primary" PM, which is directly emitted into
the air, and "secondary" PM, which forms indirectly from emissions from fuel combustion and
2-1
-------
other sources. Primary PM consists of carbonaceous materials (soot and accompanying
organics) and includes:
• Elemental carbon, organic carbon, and crustal material directly emitted from cars,
trucks, heavy equipment, forest fires, some industrial processes and burning waste.
• Both combustion and process related fine metals and larger crustal material from
unpaved roads, stone crushing, construction sites, and metallurgical operations.
Secondary PM forms in the atmosphere from gases. Some of these reactions require
sunlight and/or water vapor. Secondary PM includes:
• Sulfates formed from sulfur dioxide (S02) emissions from power plants and industrial
facilities;
• Nitrates formed from nitrogen oxide (NOX) emissions from cars, trucks, industrial
facilities, and power plants; and
• Ammonium formed from ammonia (NH3) emissions from gas-powered vehicles and
fertilizer and animal feed operations. These contribute to the formation of sulfates
and nitrates that exist in the atmosphere as ammonium sulfate and ammonium
nitrate.1
• Organic carbon (OC) formed from reactive organic gas emissions, including volatile
organic compounds (VOCs), from cars, trucks, industrial facilities, forest fires, and
biogenic sources such as trees.1
As described above, organic carbon has both primary and secondary components. The
percentage contribution to total OC from directly emitted OC versus secondarily formed OC
varies based on location. In an urban area, near direct sources of OC such as cars, trucks, and
industrial sources, the percentage of primary OC may dominate, whereas, in a rural area with
more biogenic sources, OC may be mostly secondarily formed. In addition, emissions from
sources such as power plants and industrial facilities may have small amounts of directly
emitted PM2.5 speciated into sulfate. Figure 2-1 shows, in detail, the sources contributing to
directly emitted PM2.5 and PMio, as well as PM precursors: S02, NOX, NH3, and VOC according to
the 2008 NEI, version 2 (EPA, 2012). In Figure 2-1, EGUs stands for Electric Generating Utilities.
1 Direct NH3 and VOC emissions are not controlled as part of the control strategy analysis. Emissions of PM2.5, NOX
and SO2 are controlled in the control strategies, for a complete discussion please refer to Chapter 4.
2-2
-------
Commercial
Marine, 2%
Aircraft, 0.1% .
Nonroad
equipment,
4%
Onroad, 6%
Misc, 6%
Residential
Wood, 7%
Residential
Fossil, 0.2%
Industrial
Processes,
Agand
i Railroad, Prescribed
0.5%
Agriculture,
18%
EGUs, 6%
PM2.5(5.1 million short tons)
Onroad, 2%
Residential
Wood, 2%
Nonroad
Aircraft, 0.05%
Commercial
0.5%
Comm/lnstit
Boilers, 0.1%
PM10 (20.5 million short tons)
Onroad, 1%
Misc, 0.2%
Residential
Wood, 0.1%
Nonroad
equipment,
0.3%
Aircraft, 0.1%
Commercial
Marine, 6%
Railroad,
0.1%.,
Industrial
Boilers,
Ag and
Prescribed
_Fires,l%
.Comm/lnstit
Boilers, 1%
EGUs, 72%
Railroad, 5%
Agand
Prescribed Comm/lnstit
Fires, 1% Boilers, 1%
EGUs, 17%
Onroad, 40%
Industrial
.Boilers, 7%
ndustrial
Processes, 6%
Residential
Fossil, 2%
.Residential
Wood, 0.2%
Misc, 1%
SO2 (10.7 million short tons)
(18.0 million short tons)
Residential
Wood, 0.5%
Residential
Fossil, 1%
Industria
Processes, 2%
Industrial
Boilers, 0.3%
EGUs, 1%
Comm/lnstit
Boilers, 0.1%
Nonroad Commercial
I equipment, Marine,
.0.02%
Agriculture,
Comm/lnstit
Boilers, 0.1%
Railroad,
0.3%
Nonroad
equipment,
17%
Onroad, 20%
EGUs, 0.3%
Industrial
Boilers, 1%
Industrial
Processes,
17%
Residential
Fossil, 0.1%
esidential
Wood, 2%
Misc, 8%
NH3 (4.2 million short tons)
VOC (15.0 million short tons)
Figure 2-1. Detailed Source Categorization of Anthropogenic Emissions of Primary PM2.s, PMi0
and Gaseous Precursor Species SO2, NOX, NH3 and VOCs for 2008
2-3
-------
2.2.1 PM2.5
"Fine particles" or PM2.5 are particles with diameters that are less than 2.5 micrometers.
As discussed above, these particles are composed of both primary (derived directly from
emissions) and secondary (derived from atmospheric reactions involving gaseous precursors)
components.
2.2.1.1 Geographical Scale and Transport
Both local and regional sources contribute to particle pollution. Fine particles can be
transported long distances by wind and weather and can be found in the air thousands of miles
from where they formed. Nitrates and sulfates formed from NOX and S02 are generally
transported over wide areas leading to substantial background contributions in urban areas.
Organic carbon, which has both a primary and secondary component, can also be transported
but to a far lesser degree. In general, higher concentrations of elemental carbon and crustal
matter are found closest to the sources of these emissions.
Figure 2-2 shows how much of the PM2.5 mass can be attributed to local versus regional
sources for 13 selected urban areas (EPA, 2004).2 In each of these urban areas, monitoring sites
were paired with nearby rural sites. When the average rural concentration is subtracted from
the measured urban concentration, the estimated local and regional contributions become
apparent. We observe a large urban excess across the U.S. for most PM2.5 species but especially
for total carbon mass with Fresno, CA having the highest observed measure. Larger urban
excess of nitrates is seen in the western U.S. with Fresno, CA and Salt Lake City, UT significantly
higher than all other areas. These results indicate that local sources of these pollutants are
indeed contributing to the PM2.5 air quality problem in these areas. As expected for a
predominately regional pollutant, only a modest urban excess is observed for sulfates.
In the East, regional pollution contributes to more than half of total PM2.5
concentrations. Rural background PM2.5 concentrations are high in the East and are somewhat
uniform over large geographic areas. These regional concentrations come from emission
sources such as power plants, natural sources, and urban pollution and can be transported
hundreds of miles and reflect to some extent the denser clustering of urban areas in the East as
compared to the West. In the West, much of the measured PM2.5 concentrations tend to be
local in nature. These concentrations come from emission sources such as wood combustion
and mobile sources. In general, these data indicate that reducing regional S02 and local sources
2 The measured PM2.5 concentration is not necessarily the maximum for each urban area.
2-4
-------
of carbon in the East, and local sources of nitrate and carbon in the West will be most effective
in reducing PM2.5 concentrations.
Nitrates
Sulfates
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
|_ _|
1 | WEST
LJ EAST
-| — 1 O Regional
1 1 Contribution
1 1 • Local
Contribution
Fresno
Missoula
Salt Lake City
Tulsa
St. LOUiS
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
ID
H
~~[] WEST
EAST
H
II
U
||
II Contribution
D Local
Contribution
2 4 6 8 10
Annual Average Concentration
of Nitrates, ug/m3
12
2 4 6 8 10
Annual Average Concentration
of Sulfates, ug/m3
12
Carbon
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
WEST
EAST
D
i n Regional
Contribution
I D Local
Contribution
2 4 6 8 10 12
Annual Average Concentration
of Carbon, ug/m3
2-
Figure 2-2. Regional and Local Contributions to Annual Average PM2.s by Particulate SO4
Nitrate and Total Carbon (i.e., organic plus EC) for Select Urban Areas Based on Paired 2000-
2004 IMPROVE3 and CSNb Monitoring Sites
3 Interagency Monitoring of Protected Visual Environments (IMPROVE) http://vista.cira.colostate.edu/improve
Chemical Speciation Network (CSN)
2.2.1.2 Regional and Seasonal Patterns
The chemical makeup of particles varies across the United States, as illustrated in
Figure 2-3. For example, the higher regional emissions of S02 in the East result in higher
2-5
-------
absolute and relative amounts of sulfates as compared to the western U.S. Fine particles in
southern California generally contain more nitrates than other areas of the country. Carbon is a
substantial component of fine particles everywhere.
Cool
Warm
20-
4 -
0
02 03 04 05 06 07 02 03 04 05 06 07
Midwest
20-
16-j
—H12-
02 03 04 05 06 07
] Sulfate I I Nitrate
I J Organic Carbon
02 03 04 05 06 07
I Elemental Carbon
I Crustal
Cool
Warm
20
~16
02 03 04 05 06 07 02 03 04 05 06 07
^^^^^ Southeast
20-
16-
12-
s:
4
ol
02 03 04 05 06 07 02 03 04 05 06 07
Southern California
20-
f 16 :
112-
& 8-
(0
5 4-
n -
-
•
—
20-
m
-
m
-
=
m
-
p.
-
•
-
16-
12-
8-
4-
n -
=
•
_
=
•
~
-
=
™
pq
•
-
02 03 04 05 06 07
Northwest
North Central
Southern*
California
02 03 04 05 06 07
Midwest
f9 * . Northeast
.. •
t
Southwest
Southeast
Figure 2-3. Regional and Seasonal Trends in Annual PM2.5 Composition from 2002 to 2007
Derived Using the SANDWICH Method. Data from the 42 monitoring locations shown on the
map were stratified by region and season including cool months (October-April) and warm
months (May-September)
Fine particles can also have a seasonal pattern. As shown in Figure 2-3, PM2.5 values in
the eastern half of the United States are typically higher in warmer weather when
meteorological conditions are more favorable for the formation and build up of sulfates from
2-6
-------
higher sulfur dioxide (S02) emissions from power plants in that region. Fine particle
concentrations tend to be higher in the cooler calendar months in urban areas in the West, in
part because fine particle nitrates and carbonaceous particles are more readily formed in cooler
weather, and wood stove and fireplace use increases direct emissions of carbon.
2.2.1.3 Composition of PM2.5 as Measured by the Federal Reference Method
The speciation measurements in the preceding analyses represented data from EPA's
Speciation Trends Network, along with adjustments to reflect the fine particle mass associated
with these ambient measurements. In order to more accurately predict the change in PM2.5
design values for particular emission control scenarios, EPA characterizes the composition of
PM2.5 as measured by the Federal Reference Method (FRM). The current PM2.5 FRM does not
capture all ambient particles measured by speciation samplers as presented in the previous
sections. The FRM-measured fine particle mass reflects losses of ammonium nitrate (NH4N03)
and other semi-volatile organic compounds (SVOCs; negative artifacts). It also includes particle-
bound water (PBW) associated with hygroscopic species (positive artifacts) (Frank, 2006).
Comparison of FRM and co-located speciation sampler N03" values in Table 2-1 show that
annual average N03 retention in FRM samples for six cities varies from 15% in Birmingham to
76% in Chicago, with an annual average loss of 1 u.g/m3. The volatilization is a function of
temperature and relative humidity (RH), with more loss at higher temperatures and lower RH.
Accordingly, nitrate is mostly retained during the cold winter days, while little may be retained
during the hot summer days.
PM2.5 FRM measurements also include water associated with hygroscopic aerosol. This
is because the method derives fine particle concentrations from sampled mass equilibrated at
20-23 °C and 30-40% RH. At these conditions, the hygroscopic aerosol collected at more humid
environments will retain their particle-bound water. The water content is higher for more acidic
and sulfate-dominated aerosols. Combining the effects of reduced nitrate and hydrated aerosol
causes the estimated nitrate and sulfate FRM mass to differ from the measured ions simply
expressed as dry ammonium nitrate and ammonium sulfate. The composition of FRM mass is
denoted as SANDWICH based on the Sulfate, Adjusted Nitrate Derived Water and Inferred
Carbon approach from which they are derived. The PM2.5 mass estimated from speciated
measurements of fine particles is termed Reconstructed Fine Mass (RCFM). The application of
SANDWICH adjustments to speciation measurements at six sites is illustrated in Table 2-1 and
Figure 2-4. EPA's modeling incorporates these SANDWICH adjustments in the Model
Attainment Test Software (MATS) (Abt, 2010).
2-7
-------
Table 2-1. Annual Average FRM and CSN PM2.5 NO3 and NH4NO3 Concentrations at Six Sites
during 2003
Sampling Site
Location
Mayville, Wl
Chicago, IL
Indianapolis, IN
Cleveland, OH
Bronx, NY
Birmingham, AL
No. of
Observations
100
76
92
90
108
113
FRM
Mass
9.8
14.4
14.8
16.8
15.0
17.0
CSNa
2.5
2.8
2.5
2.9
2.4
1.1
N03-(u
FRMb
1.5
2.1
1.3
1.7
1.1
0.2
g/m3)
Difference
(CSN - FRM)
1.0
0.7
1.3
1.2
1.3
0.9
NH4NO3
CSN
3.2
3.7
3.2
3.7
3.1
1.4
(ug/m3)
FRM
1.9
2.8
1.6
2.2
1.4
0.2
Percent of
NH4NO3 in PM2.5
FRM Mass
CSN FRM
33% 19%
25% 19%
22% 11%
22% 13%
21% 9%
8% 1%
On denuded nylon-membrane filters for al sites except for Chicago, where denuded Teflon-membrane followed
by nylon filters were used.
On undenuded Teflon-membrane filters.
SANDWICH
(FRM)
Sufata Mow ^H Nhrata MOM
TGM
CfUDlal
Figure 2-4. RCFM (left) versus SANDWICH (right) Pie Charts Comparing the Ambient and PM2.5
FRM Reconstructed Mass Protocols on an Annual Average Basis3
Estimated NH4* and PBW for SANDWICH are included with their respective sulfate and nitrate mass slices.
Circles are scaled in proportion to PM2.5 FRM mass.
2-8
-------
2.2.1.4 2006-2008 Design Values
Annual and 24-hour PM2.5 design values for 2006-2008 are shown in Figures 2-5 and 2-6,
respectively. These design values were calculated using 2006-2008 FRM 24-hour average PM2.5
concentration measurements in a manner consistent with CFR Part 50.3 For the most part,
counties in the center of the U.S. have PM2.s design values that are above both 11 u.g/m3 for the
annual standard and 35 u.g/m3 for the 24-hour standard. In the East, the counties above the
current NAAQS (i.e., 15 u.g/m3 annual and 35 u.g/m3 24-hour standards) are similar. In the West,
there are fewer counties above the annual level of 15 u.g/m3 than exceed the 24-hour standard
of 35 u.g/m3.
3 These years of ambient measurements are presented here since they frame the air quality model year of 2007. As
discussed in Chapter 3, ambient measurement for the period 2005 through 2008 were used to construct 5-year
weighted average concentrations for use as the starting point for future year projections in conjunction with air
quality model predictions.
2-9
-------
Legend
| 22 counties exceed 15 ug'm3
H 36 additional counties exceed 14 ug'mS
| 60 additional counties exceed 13 ug.'mS
92 additional counties exceed 12 ug'm3
96 additional counties exceed 11 ugi'mS
208 counties with monitors are below 11 ug/m3
514 counties with monitors have PM2.5 annual design values
Figure 2-5. Maximum County-level PM2.s Annual Design Values Calculated Using 2006-2008
FRM 24-hr Average PM2.s Measurements.
2-10
-------
I 48 counties exceed 35 ug/m3
_J 463 counties with monitors are below 35 ug/m3
511 counties with monitors have PM2.5 daily design values.
Figure 2-6. Maximum County-level PM2.s 24-hour Design Values Calculated Using 2006-2008
FRM 24-hr Average PM2.s Measurements.
2.3 References
Abt Associates, 2010. User's Guide: Modeled Attainment Test Software.
http://www.epa.gov/scram001/modelingapps_mats.htm.
Frank, N.H., Retained Nitrate, Hydrated Sulfates, and Carbonaceous Mass in Federal Reference
Method Fine Particulate Matter for Six Eastern U.S. Cities, J. Air & Waste Manage. Assoc.
2006, 56, 500-511.
U.S. Environmental Protection Agency (EPA). 2004. Air quality criteria for particulate matter.
U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA/600/P-
99/002aF-bF.
U.S. Environmental Protection Agency (EPA). 2008. National Air Quality Status and Trends
Through 2007. U.S. Environmental Protection Agency. Research Triangle Park. EPA-
454/R-08-006.
2-11
-------
U.S. Environmental Protection Agency (EPA). 2009. Integrated Science Assessment for
Particulate Matter. U.S. Environmental Protection Agency. Research Triangle Park.
EPA/600/R-08/139F.
U.S. Environmental Protection Agency (EPA). 2011. Policy Assessment for the Review of the
Particulate Matter National Ambient Air Quality Standards. Office of Air Quality Planning
and Standards, Research Triangle Park, NC.
U.S. Environmental Protection Agency (EPA). 2012. 2008 National Emissions Inventory,
Version 2 Data and Documentation available from
http://www.epa.gov/ttn/chief/net/2008inventory.html. Office of Air Quality Planning
and Standards, Research Triangle Park, NC.
2-12
-------
CHAPTER 3
AIR QUALITY MODELING AND ANALYSIS
3.1 Synopsis
In order to evaluate the health and environmental impacts of trying to reach the
alterative primary standards in this final RIA, it was necessary to use models to predict
concentrations in the future. The data, tools and methodology used for projecting future-year
air quality 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.
3.2 Modeling PM2.s Levels in the Future
A national scale air quality modeling analysis was performed to estimate PM2.5
concentrations for the annual and 24-hour primary standards for the future year of 2020.1 Air
quality ratios were then developed using model responsiveness to emissions changes based on
"sensitivity" air quality modeling that was designed to determine the response of PM2.5
concentrations to reductions in emissions of S02, NOX, and directly emitted PM2.5. The air
quality ratios were used to determine the amount of emissions reductions needed to attain the
revised annual standard of 12 u.g/m3 and two alternative annual standards. The emissions
reductions were then used to estimate how air quality would change under each set of
emissions scenarios. These data were used as inputs to the calculation of expected costs and
benefits associated with the emissions and air quality changes resulting from just attaining the
revised and alternative annual standards.
As described in section 3.3, air quality modeling was used in a relative sense to project
future concentrations of PM2.5. As part of this approach air quality model predictions from a
base year simulation are coupled with predictions from the future case to calculate the relative
change (between base year and future case) in each species component of PM2.5. These
species-specific relative response factors (RRFs) are applied to the corresponding measured
concentrations to estimate future species concentrations. The future case PM2.5 annual and
daily design values are then calculated using the projected species concentrations. We used
2007 as the base year and 2020 as the future year for air quality-related analyses in this RIA. For
2020 we modeled two emissions scenarios, a 2020 base case and a 2020 control case. The 2007
and 2020 scenarios were modeled as annual model simulations. In addition to these emissions
scenarios, we also performed several emissions sensitivity model runs to quantify the response
1 In addition, we used air quality modeling to estimate light extinction in 2020 to support the analysis of the
welfare benefits of this rule.
3-1
-------
of PM2.5 to various precursor emissions. The modeling for the 2020 base case, the 2020 control
case, and sensitivity scenarios were used to inform the development of design values for the
baseline which provides for attainment of the 15/35 NAAQS and the incremental emissions
reductions needed to attain the revised 12 u.g/m3 annual standard and two alternative annual
standards, 13 u.g/m3 and 11 u.g/m3. Details on the 2007-based air quality modeling platform,
the 2007 base year and 2020 base case scenarios, and the methods and results for attaining
these NAAQS levels are provided below. Information on the 2020 control case can be found in
Chapter 4 of this RIA.
3.2.1 Air Quality Modeling Platform
The 2007-based Community Multi-scale Air Quality (CMAQ) modeling platform was used
as the tool to project future-year air quality for 2020 and to estimate the costs and benefits for
attaining the current and revised alternative NAAQS presented in this assessment. This
platform provides the most recent, complete set of base year emissions information currently
available for national scale modeling. In addition to the CMAQ model and the emissions data,
the modeling platform includes the meteorology, and initial and boundary condition data for
2007 which are inputs to this model. The CMAQ model 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.) (Appel et al., 2008; Appel et al., 2007; Byun
and Schere, 2006). Consideration of the different processes (e.g., transport and deposition) that
affect primary (directly emitted) and secondary (formed by atmospheric processes) PM at the
regional scale in different locations is fundamental to understanding and assessing the effects
of pollution control measures that affect PM, ozone and deposition of pollutants to the surface.
Because it accounts for spatial and temporal variations as well as differences in the reactivity of
emissions, CMAQ is useful for evaluating the impacts of the control strategies on PM2.s
concentrations. Version 4.7.1 of CMAQ was employed for this RIA modeling.2 CMAQ is applied
with the AER05 aerosol module, which includes the ISORROPIA inorganic chemistry (Nenes et
al., 1998) and a secondary organic aerosol module (Carlton et al., 2010). The CMAQ model is
applied with sulfur and organic oxidation aqueous phase chemistry (Carlton et al., 2008) and
the carbon-bond 2005 (CB05) gas-phase chemistry module (Yarwood et al., 2005).
2 More information is available online at: www.cmaq-model.org
3-2
-------
3.2.1.1 Air Quality Modeling Domain
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 24 vertical layers with a top at about 17,600 meters, or 50 millibars (mb). A horizontal
resolution of 12 x 12 km was used for modeling the 2007 base year and the 2020 base and
control strategy scenarios. The model simulations produce gridded air quality concentrations
on an hourly basis for the entire modeling domain.
Domain Boundary
12US2 domain
x,y origin: -2412000rp
col: 396 row:246 ^
Figure 3-1. Map of the CMAQ Modeling Domain Used for PM NAAQS RIA
3.2.1.2 Air Quality Model Inputs
CMAQ 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 2007 base year and the future year of 2020 base case and control strategy
3-3
-------
scenarios. All other inputs were specified for the 2007 base year model application and
remained unchanged for each future-year modeling scenario.
CMAQ 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 or precursors to secondary pollutants. The annual
emission inventories, described in Section 3.2.2, were preprocessed into CMAQ-ready inputs
using the SMOKE emissions preprocessing system3. Meteorological inputs reflecting 2007
conditions across the contiguous U.S. were derived from Version 3.1 of the Weather Research
Forecasting Model (WRF). 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 2007 meteorological model simulation
and evaluation are provided in a separate technical support document (EPA, 2011a).
The lateral boundary and initial species concentrations are provided by a three-
dimensional global atmospheric chemistry model, the GEOS-CHEM model version 8-02-03
(Yantosca, 2004)4. 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). This model was run for 2007 with a grid resolution of 2.0 degrees x
2.5 degrees (latitude-longitude) and 47 vertical layers. The predictions were used to provide
one-way dynamic boundary conditions at three-hour intervals and an initial concentration field
for the CMAQ simulations. A GEOS-Chem evaluation was conducted for the purpose of
validating the 2007 GEOS-Chem simulation for predicting selected measurements relevant to
their use as boundary conditions for CMAQ. This evaluation included reproducing GEOS-Chem
evaluation plots reported in the literature for previous versions of the model (Lam, 2010).
3.2.1.3 Air Quality Model Evaluation
An operational model performance evaluation for PM2.s and its related speciated
components (e.g., sulfate, nitrate, elemental carbon, organic carbon) was performed to
estimate the ability of the CMAQ modeling system to replicate 2007 measured concentrations5.
This evaluation principally comprises statistical assessments of model predictions versus
observations paired in time and space depending on the sampling period of measured data.
Details on the evaluation methodology and the calculation of performance statistics are
3 More information is available online at: www.smoke-model.org.
4 More information is available online at: http://www-as.harvard.edu/chemistry/trop/geos.
5 This operational evaluation for CMAQ included statistical and graphical comparisons of model predictions for
select PM2.5 component species to the corresponding measured data from monitoring sites in the Continuous
Speciation Network (CSN), the Interagency Monitoring of PROtected Visual Environments (IMPROVE) network, and
the Clean Air Status and Trends Network (CASTNet).
3-4
-------
provided in the Technical Support Document: Air Quality Modeling for the Final PM NAAQS
(AQMTSD, EPA, 2012a). Overall, the model performance statistics for sulfate, nitrate, organic
carbon, and elemental carbon from the CMAQ 2007 simulation are within or close to the ranges
found in other recent applications. These model performance results give us confidence that
our applications of CMAQ using this 2007 modeling platform provide a scientifically credible
approach for assessing PM2.s concentrations for the purposes of the RIA.
3.2.1.4 Emissions Inventory
The 2007 emissions inventory and the 2020 base case emissions inventory were
developed using the 2007 Version 5.0 emissions modeling platform (documentation and data
files available from http://www.epa.gov/ttn/chief/emch/index.html). The starting point for the
2007v5 platform was Version 2 of the 2008 National Emissions Inventory
(http://www.epa.gov/ttn/chief/net/2008inventory.html). The 2008 NEI v2 is the most recently
available NEI. The next NEI will be developed for 2011. Data collection for the 2011 NEI is
ongoing through the end of 2012, with the inventory due to be published in 2013. Some data in
the 2008 NEI v2 were adjusted to better represent 2007 for this analysis. For example, MOVES
2010b was used to compute onroad emissions and duplicate emissions values were removed
where they were identified. For additional details, see the Technical Support Document:
Preparation of Emissions Inventories for the Version 5.0, 2007 Emissions Modeling Platform
(EITSD, EPA, 2012b). The 2020 base case inventory is the starting point for the baseline and
control strategy modeling performed for this assessment. The above-referenced EITSD (EPA,
2012b) describes the development of the 2007 base year inventory in detail for all emissions
sectors, along with the projection methodology applied to develop the 2020 base case
inventory.
The 2020 ECU projected inventory represents demand growth, fuel resource availability,
generating technology cost and performance, and other economic factors affecting power
sector behavior. It also reflects the expected 2020 emissions effects due to environmental rules
and regulations, consent decrees and settlements, plant closures, units built or with control
devices updated since 2007, and forecast unit construction through the calendar year 2020. In
this analysis, the projected ECU emissions include the Final Mercury and Air Toxics (MATS) rule
announced on December 21, 2011 and the Final Cross-State Air Pollution Rule (CSAPR) issued
on July 6, 2011.
On August 21, 2012, the D.C. Circuit Court of Appeals issued an opinion vacating CSAPR.
In its decision, the Court also instructed EPA to "continue administering CAIR [the 2005 Clean
3-5
-------
Air Interstate Rule] pending the promulgation of a valid replacement." In the interim, the EPA
and the states are continuing to implement CAIR to address regional transport of air pollution,
as directed by the Court. The EPA has filed a petition for rehearing of the Court's decision on
CSAPR. In light of the Court's instructions, the EPA believes that it is appropriate to rely on CAIR
emission reductions as permanent and enforceable reductions until such a time as the EPA
issues a replacement transport rule.
Because of the similarity in emissions reductions associated with CSAPR and CAIR6, and
the inclusion of MATS in the RIA baseline, EPA has determined that it remains appropriate that
CSAPR continue to be used in the RIA baseline as a proxy for representing the emission
reductions required by CAIR for the purposes of the rulemaking's modeling projections for
2020.
Regarding the impact of MATS on this determination, the MATS emission rate standard
for hydrogen chloride (HCI) is expected to result in a substantial amount of new pollution
controls (scrubbers and dry sorbent injection) and upgrading of existing scrubbers that will also
significantly reduce S02 emissions from power plants. MATS implementation is projected to
reduce nationwide S02 emissions from power plants to a level more than 40 percent lower
than the S02 emissions projected under CSAPR without MATS in place (EPA-HQ-OAR-2009-
0234-20131).
In addition to these conclusions, the ECU baseline used in modeling the PM NAAQS was
based on ElA's AEO 2010 and represents a conservative approach to emission projections, given
that more recent trends in power sector economics suggest a likelihood of lower future ECU
emissions. This is supported by the results of a sensitivity analysis conducted using the
electricity demand forecast from ElA's AEO 2012 that shows slightly lower ECU emissions. It is
reasonable to expect that recent reductions in gas prices and increases in coal prices would
yield yet lower estimations of future ECU emissions in the context of this rule's analysis. The
details of this analysis can be found in the memo titled AEO 2012 Demand Sensitivity, which is
available in the docket.
The ECU emissions were developed using the Integrated Planning Model (IPM) Version
4.10 Final MATS and are documented in detail at http://www.epa.gov/airmarkt/progsregs/epa-
ipm/toxics.html. IPM is a multiregional, dynamic, deterministic linear programming model of
the U.S. electric power sector. Note that for this analysis, no further ECU control measures
6 U.S. EPA, Cross-State Air Pollution Rule Presentation, December 15, 2011, available at
http://www.epa.gov/airtransport/pdfs/CSAPRPresentation.pdf, p. 15.
3-6
-------
were selected for illustrating attainment of the current and proposed alternative standard
levels discussed in Chapter 4. Thus, the ECU emissions are unchanged between the future-year
base-case and the control strategies.
Table 3-1 provides a comprehensive list of all the control programs, growth
assumptions, and facility and unit closures information in the future year base case. The future-
year base 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. Many state and
local control programs are also applied where those programs were finalized and enough
details were available to apply reductions to the 2007 emissions data.
The 2007 and 2020 onroad mobile source emissions were developed using emissions
factors derived from the MOtor Vehicle Emission Simulator (MOVES)7 Version 2010b. The
emissions were computed by using the Sparse Matrix Operator Kernel Emissions system
(SMOKE) to combine the county-, vehicle type-, and temperature-specific emission factors and
vehicle miles traveled and vehicle population activity data while taking into account hourly
gridded temperature data. For California we received onroad emissions directly from the
California Air Resources Board (CARB) in July 2012 for 2007, 2020, and 2025. These emissions
were based on the latest available data and models from their SIP development process. We
allocated the California onroad emissions down to the hourly, grid-cell, and CMAQ model-
species level using ratios derived from the MOVES-based emissions data output from SMOKE.
The MOVES-based 2020 onroad emissions account for changes in activity data and the
impact of on-the-books national rules including: the Light-Duty Vehicle Tier 2 Rule, the Heavy
Duty Diesel Rule, the Mobile Source Air Toxics Rule, the Renewable Fuel Standard, the Light
Duty Green House Gas/Corporate Average Fuel Efficiency (CAFE) standards for 2012-2016, and
the Heavy-Duty Vehicle Greenhouse Gas Rule. The emissions do not account for the 2017 and
Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions and Corporate Average Fuel
Economy Standards; Final Rule (LD GHG), issued October 15, 2012. The LD GHG rule was not
included in this analysis because the rule was not signed at the time the modeling was
performed, and it is expected to have little impact on particulate matter emissions. The RIA for
the LD GHG (EPA, 2012c) shows that in 2030 counties are showing decreases in PM 2.5 design
values of up to 0.16 u.g/m3. The modeling indicates that the majority of the modeled counties
7More information is available online at: http://www.epa.gov/otaq/models/moves/index.htm.
3-7
-------
will experience small changes of between 0.05 u.g/m3 and -0.05 u.g/m3 in their annual PM2.5
design values due to the vehicle standards. The impacts of the rule in 2020 should be even less
than the 2030 impacts. The MOVES-based 2020 emissions 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 provided future year 2020 and 2025 emissions
included most on-the-books regulations such as those for low sulfur fuel, idling of heavy-duty
vehicles, chip reflash, public fleets, trash trucks, drayage trucks, and heavy duty trucks and
buses. The zero emission vehicle program prior to adoption of Advanced Clean Cars is included
but has a very small impact. The California emissions do not reflect the impacts of the
GHG/Smartway regulation, Advanced Clean Cars, nor the low carbon fuel standard because it is
assumed that there is no impact on criteria pollutants.
Table 3-1 provides details on the national rules included to develop all categories of
mobile source emissions. The nonroad mobile 2020 base 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. The nonroad, locomotive, and class 1
and 2 commercial marine emissions used for California were obtained from CARB, and include
nonroad rules reflected in the December 2010 Rulemaking Inventory
(http://www.arb.ca.gov/regact/2010/offroadlsilO/offroadisor.pdf), those in the March 2011
Rule Inventory, the Off-Road Construction Rule Inventory for "In-Use Diesel", cargo handling
equipment rules in place as of 2011 (see http://www.arb.ca.gov/ports/cargo/cargo.htm), rules
through 2011 related to Transportation Refrigeration Units, the Spark-Ignition Marine Engine
and Boat Regulations adopted July 24, 2008 for pleasure craft, and the 2007 and 2010
regulations to reduce emissions from commercial harbor craft. For ocean-going vessels, the
data represents 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 ECA
zone, the 2012-2015 Tier 2 NOx controls, the 2016 0.1% sulfur fuel regulation in ECA zone, and
the 2016 IMO Tier 3 NOx controls. Control and growth-related assumptions in the 2020 base
case are described in more detail in the EITSD.
All modeled 2007 and 2020 scenarios use the same year 2006 Canada emissions data.
Note that 2006 is the latest year for which Canada provided data, and no accompanying future-
year projected inventories were provided in a form suitable for this study. For Mexico, different
emissions were used for 2008 and 2018 as described in the Development of Mexico National
3-8
-------
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). All base year and projected emissions
inventories are available on the EPA's Emissions Modeling Clearinghouse website at
http://www.epa.gov/ttn/chief/emch/index.html.
Table 3-l(a). Control Strategies and Growth Assumptions for Creating 2020 Base Case
Emissions Inventories from the 2007 Base Case for Non-EGU Point Sources
Control Strategies and/or Growth Assumptions
(Grouped by Affected Pollutants or Standard and Approach)
Pollutants
Affected
Non-EGU Point (ptnonipm) Controls and Growth Assumptions
Boat Manufacturing MACT rule, national, VOC: national applied by SCC
Consent decrees on companies (based on information from the Office of Enforcement and
Compliance Assurance—OECA) apportioned to plants owned/operated by the companies
Refinery Consent Decrees: plant/SCC controls
Commercial/lnstitutional/Hospital/Medical/lnfectious Waste Incinerator Regulations
NESHAP: Portland Cement (09/09/10)—plant level based on Industrial Sector Integrated
Solutions (ISIS) policy emissions in 2013. The ISIS results are from the ISIS-Cement model runs
for the NESHAP and NSPS analysis of July 28, 2010 and include closures.
New York ozone SIP controls
Additional plant and unit closures provided by state, regional, and the EPA agencies and
additional consent decrees. Includes updates from CSAPR comments.
Reciprocating Internal Combustion Engines (RICE) NESHAP with reconsideration
Ethanol plants that account for increased ethanol production due to RFS2 mandate
State fuel sulfur content rules for fuel oil—as of July, 2012, effective only in Maine,
Massachusetts, New Jersey, New York and Vermont.
Emission reductions resulting from controls put on specific boiler units (not due to MACT) after
2005, identified through analysis of the control data gathered from the Information Collection
Request (ICR) from the Industrial/Commercial/lnstitutional Boiler NESHAP.
Emissions reductions resulting from Boiler MACT controls to specific boiler units
Plant and unit closures resulting from state submissions and industry and web postings
effective prior to January 2012
Aircraft growth via Itinerant (ITN) operations at airports to 2020
Livestock Emissions Growth from year 2008 to year 2020 (some farms in the point inventory)
Upstream adjustments to year 2020 for refineries and gasoline distribution via the Energy
Information and Security Act/Renewable Fuel Standards 2 (EISA/RFS2) impacts
VOC
VOC, CO, NOX,
PM, SO2
NOX, PM, SO2
NOX, PM, SO2
Hg, NOX, S02,
PM, HCI
VOC, NOX,
HAP VOC
All
NOX, CO, PM,
S02
All
S02
NOX, S02, HCI
NOX, CO, PM,
SO2, VOC, HCI
All
All
NH3, PM
All
3-9
-------
Table 3-l(b). Control Strategies and Growth Assumptions for Creating 2020 Base Case
Emissions Inventories from the 2007 Base Case for Nonpoint and Onroad Mobile
Sources
Control Strategies and/or Growth Assumptions Pollutants
(Grouped by Affected Pollutants or Standard and Approach Used to Apply to the Inventory) Affected
Nonpoint (nonpt sector) Controls and Growth Assumptions
Residential Wood Combustion Growth and Change-outs from year 2008 to 2020 All
State fuel sulfur content rules for fuel oil—as of July, 2012, effective only in Maine, SO2
Massachusetts, New Jersey, New York and Vermont.
Reciprocating Internal Combustion Engines (RICE) NESHAP with reconsideration NOX, CO, PM,
S02
New York, Connecticut, and Virginia ozone SIP controls VOC
Livestock Emissions Growth from year 2008 to year 2020 (some farms in the point inventory) NH3, PM
Upstream adjustments to year 2020 for refineries and gasoline distribution via the Energy All
Information and Security Act/Renewable Fuel Standards 2 (EISA/RFS2) impacts
Portable Fuel Container Mobile Source Air Toxics Rule 2 (MSAT2) inventory growth and control VOC
from year 2007 to 2020
Texas oil and gas projections to year 2020 VOC, SO2,
NOX, CO, PM
Onroad Mobile Controls
(list includes all key mobile control strategies but is not exhaustive)
National Onroad Rules: All
Tier 2 Rule: Signature date February 2000
2007 Onroad Heavy-Duty Rule: February 2009
Final Mobile Source Air Toxics Rule (MSAT2): February 2007
Renewable Fuel Standard: March 2010
Light-Duty Greenhouse Gas Emissions Standards and Corporate Average Fuel Efficiency
Standards: May 2010
Heavy (and Medium)-Duty Greenhouse Gas Emissions Standards and Fuel Efficiency Standards:
August 2011
Corporate Average Fuel Economy standards for 2008-2011
Local Onroad Programs: VOC
National Low Emission Vehicle Program (NLEV): March 1998
Ozone Transport Commission (OTC) LEV Program: January 1995
3-10
-------
Table 3-l(c). Control Strategies and Growth Assumptions for Creating 2020 Base Case
Emissions Inventories from the 2007 Base Case for Nonroad Mobile Sources
Control Strategies and/or Growth Assumptions Pollutants
(Grouped by Affected Pollutants or Standard and Approach Used to Apply to the Inventory) Affected
Nonroad Mobile Controls
(list includes all key mobile control strategies but is not exhaustive) (continued)
National Nonroad Controls: All
Clean Air Nonroad Diesel Final Rule—Tier 4: June 2004
Control of Emissions from Nonroad Large-Spark Ignition Engines and Recreational Engines
(Marine and Land Based): "Pentathalon Rule": November 2002
Clean Bus USA Program: October 2007
Control of Emissions of Air Pollution from Locomotives and Marine Compression-Ignition
Engines Less than 30 Liters per Cylinder: October 2008
Locomotive and marine rule (May 6, 2008)
Marine SI rule (October 4,1996)
Nonroad large SI and recreational engine rule (November 8, 2002)
Nonroad SI rule (October 8, 2008)
Phase 1 nonroad SI rule (July 3,1995)
Tier 1 nonroad diesel rule (June 17, 2004)
Locomotives: All
Energy Information Administration (EIA) fuel consumption projections for freight rail
Clean Air Nonroad Diesel Final Rule—Tier 4: June 2004
Locomotive Emissions Final Rulemaking, December 17,1997
Locomotive rule: April 16, 2008
Control of Emissions of Air Pollution from Locomotives and Marine: May 2008
Commercial Marine: All
Category 3 marine diesel engines Clean Air Act and International Maritime Organization
standards (April 30, 2010)—also includes CSAPR comments.
EIA fuel consumption projections for diesel-fueled vessels
Clean Air Nonroad Diesel Final Rule—Tier 4
Emissions Standards for Commercial Marine Diesel Engines, December 29,1999
Locomotive and marine rule (May 6, 2008)
Tier 1 Marine Diesel Engines, February 28, 2003
3.3 PM2.5 Modeling Results and Analyses
The air quality modeling results were used in the RIA to estimate future PM2.5
concentrations for the 2020 base case and 2020 control case as well as to calculate the air
quality ratios that were used in determining the emissions reductions to attain the existing
3-11
-------
standards of 15/35, the revised annual standard of 12 u.g/m3 and the two alternative annual
standards. These data are then used to estimate the costs and benefits of attaining these
existing and revised NAAQS levels. Consistent with EPA guidance (EPA, 2007 and EPA, 2011b),
the air quality modeling results are applied in a relative sense to estimate 2020 future design
values for PM2.5 for the 2020 base case and 2020 control case. Air quality response ratios are
calculated and used to estimate the tons of emissions reductions needed beyond the 2020
control case needed to show attainment of the existing, revised, and alternative NAAQS levels.
Based on the tons of emissions needed in each county, design values are calculated for
attaining the revised and alternative annual standard levels for input into the benefits
assessment.
The flow diagram shown in Figure 3-2 summarizes our approach for calculating future-
year design values for meeting the existing standards, the revised annual standard, and
alternative annual standard levels. Table 3-2 describes the specific air quality modeling
simulations that informed this approach. The 2020 base case simulation (Box 1) was performed
to estimate which monitors would exceed the current and alternative standard levels in 2020
based on emissions reductions expected from existing (i.e., "on-the-books") state and federal
control programs. The 2020 control case simulation (Box 3) was performed to estimate the
impact of emission reductions from additional controls beyond those of the 2020 base case in
areas with design values above the revised and alternative standard levels. As discussed below,
the 2020 base case and 2020 control case design values were adjusted to reflect PM2.5
reductions expected from the implementation of existing burn ban programs in certain counties
and to remove the effects of atypical events such as wildfires and fireworks displays(Boxes 2
and 4). To calculate future-year design values at the different standard levels, and the
associated emissions reductions, these 2020 base and control case design values were adjusted
downward using air quality response ratios, which give the change PM2.5 design value (u,g/m3)
per change in emissions by species (Boxes 5 through 9).
The air quality response ratios (hereafter referred to as air quality ratios) used to adjust
the 2020 cases to meet the standard levels were calculated based on results of several
sensitivity simulations. The sensitivity simulations, as described in Table 3-2, were defined to
isolate the changes in the (NH4)2S04, NH4N03 and direct PM2.5 associated with changes in
emissions of S02, NOX and direct PM2.5, respectively. These PM2.5 component species were
selected for reduction to meet the standard levels because they dominate the mass of PM2.5 in
the areas of concern in the 2020 cases. The sensitivity simulation referred to as "2020
NOx_PM2.5" was used in calculating the air quality ratios associated with changes in NOX and
3-12
-------
direct PM2.5 emissions. This simulation was based on anthropogenic NOX and direct PM2.5
emission reductions from non-EGU sources of 25% and 50%, respectively, relative to the 2020
base case. The sensitivity simulation referred to as "2020 S02_RWC" was used in calculating the
air quality ratios associated with changes in S02 emissions. This simulation was based on
anthropogenic S02 and residential wood combustion emissions reductions from non-EGU
sources of 25% and 100%, respectively, relative to the 2020 base case8. In the sensitivity runs,
emissions reductions for direct PM2.5 were generally applied in counties with monitors with
annual design values above 11 u.g/m3 level in the 2020 base case, while emission reductions for
NOX and S02 were generally applied in those counties as well as their adjacent counties. This
approach reflects the local impacts of direct PM2.5 emissions on air quality and the broader
geographic impacts on PM2.5 of S02 and NOX emissions reductions.
The development of the air quality response ratios used in the process of adjusting the
air quality modeling results to meet the current and alternative standard levels is described in
Appendix 3.A.I.I. The remainder of this section describes the procedures and the results from
the 2020 base case modeling and the development of the adjusted 2020 base case (Box 1 and
Box 2, respectively in Figure 3-2) and the identification of the emissions reductions estimated to
be needed to attain the 15/35 standard and annual standards of 13, 12, and 11 u.g/m3 (Boxes 4
through 9 in Figure 3-2).
The results of this sensitivity run were also used in the method to quantify the impacts on design values of
existing burn ban programs, as described in Section 3.3.1.1.
3-13
-------
1.2020 Base Case (modeled) 3.2020 Control Case (modeled^
--Accounts for controls due to existing state -- Includes "local" control measures bevond
and federal programs existing state and federal programs
2. Adjusted 2020 Base Case 4. Adjusted 2020 Control Case
--Accounts for atypical events and episodic --Accounts for atypical events, episodic1 wood
wood burning curtailments burning curtailments and removal of
--Adjustments involved use of sensitivity runs inappropriate controls
--Adjustments involved use of sensitivity rmis
5. Attainment of 15 / 3 5 Level
--Includes direct PM-. r emission reductions
beyond known controls
6. Analytical Baseline
--Accounts for 2025 mobile N Ox emission
adjustment in South Coast Air Basin and San
Joacniiii Valley
7.13 Standard Level 8.12 Standard Level 9.11 Standard Level
Figure 3-2. Flow Diagram of Process Used to Determine Future-Year Design Values and
Associated Emission Reductions for Meeting the Current, Revised and Alternative Standard
Levels
3-14
-------
Table 3-2. Air Quality Model Simulations Used in this Regulatory Impact Analysis
Simulation
Description
Purpose
2020 base case
2020 control case
2020 NOX_PM2.5
sensitivity
2020 SO2_RWC
sensitivity
2020 SJV
sensitivity
Simulation of 2020 that accounts for expected controls
due to existing state and federal programs.
Simulation of 2020 that includes emissions controls
beyond the controls of the 2020 base case in areas
with design values above the alternative standard
levels in the 2020 base case.
Simulation of 2020 where anthropogenic NOX and
PM2.5 emissions are decreased by 25% and 50%,
respectively, relative to the 2020 base case in selected
counties.
Simulation of 2020 where anthropogenic SO2 and
residential wood combustion emissions are decreased
by 25% and 100%, respectively, relative to the 2020
base case in selected counties.
Nine simulations of January 2020. Each simulation has
emission reductions relative to the 2020 base case in a
one- or two-county group in California's Central Valley.
The emission reductions in each county group are the
same as those in the 2020 NOX_PM2.5 sensitivity case.
Provides estimate of future-
year design values based on
existing controls
Provides impact of additional
known controls on design
values in target areas; provides
basis for meeting the existing,
revised and alternative
standard levels with emission
controls beyond known controls
Used in estimating the response
of air quality to changes in
emissions of NOX and direct
PM2.5
Used in estimating the response
of air quality to changes in
emissions of SO2 and residential
wood combustion
This series of simulations is
used to estimate the
contributions of emissions from
counties in the California's
Central Valley on air quality in
other counties in the Central
Valley
3.3.1 Calculating Future-year Design Values for 2020 Base and Control Cases
To predict the impact of the control strategies on future-year attainment, the air quality
model results are used in a relative sense by estimating future-year PM2.5 relative response
factors (RRFs). RRFs are ratios that are calculated from the modeled changes in PM2.5 species
concentrations between the base year (2007) and future-year (2020 base case and 2020 control
case) air quality modeling results. RRFs are calculated for each PM2.5 component (i.e. sulfates,
nitrates, organic carbon, etc.). Future-year estimates of the PM2.5 annual and 24-hour standard
design values at monitor locations are then calculated by applying the species-specific RRFs to
ambient PM2.5 concentrations from the Federal Reference Method (FRM) Network, which are
disaggregated into species concentrations through processing and interpolation of PM2.5
species data from the CSN and IMPROVE monitoring networks.
3-15
-------
To more easily apply this methodology, EPA has created software, called Modeled
Attainment Test Software (MATS) (Abt, 2012) to calculate future-year PM2.5 annual and 24-hour
standard design values. For this RIA, design values are projected from ambient Federal
Reference Method measurements during the period 2005-20099 coupled with PM2.5 species
data from IMPROVE and CSN sites for the 2006-2008 time period. In addition to calculating
projected future-year annual and 24-hour standard design values, MATS provides the amounts
of sulfate, nitrate, ammonium, elemental carbon, organic carbon and crustal matter that
comprise the annual and 24-hour standard design values for each site. These data are useful for
understanding the PM species contributing to high PM2.5 concentrations which is informative
for designing control strategies to reduce the future-year design values to the proposed
standard levels.
In order to derive 2020 design values for the purposes of the RIA, we made two
additional adjustments to the design value calculations at those monitoring sites that 1) had
observed ambient data in the base year period that reflects atypical events or highly variable
events that are difficult to predict in the future year, and 2) would be affected by existing local
episodic residential wood burning curtailment programs (e.g. "burn ban" programs) that we
were not able to simulate in the 2020 base case and control case air quality modeling. These
adjustments are described below.
3.3.1.1 Future-year Design Values Adjustments for Episodic Residential Wood Curtailment
Programs
A number of Western nonattainment areas have existing rules in place that require the
curtailment of residential wood burning (from fireplaces and woodstoves) on an episodic basis.
The burning curtailment programs ("burn bans") are implemented at the local level based on
local air quality forecasts of high PM2.5 days. The burn ban programs vary by area, but are
similar in many ways. They generally have "stage 1" (lower concentration PM2.5 days) and
"stage 2" (higher concentration PM2.5 days) level "burn ban" days with mandatory compliance
on stage 2 days. The forecast trigger level also varies by area. When the daily PM2.5 NAAQS was
lowered to 35 u.g/m3 in 2006 most areas implemented a trigger level at or below 35 u.g/m3 for a
mandatory burn ban10. There are also a number of exemptions in each area for residents who
use firewood as their sole source of heat. These programs have been strengthened in the last
few years to become mandatory and also to address the 35 u.g/m3 NAAQS. Since all of the
9 The 2005 -2007 period includes design values 2005-2007, 2006-2008, and 2007-2009.
10 Some areas previously (before 2007) had voluntary burn ban programs with relatively high trigger levels based
on the 1997 daily PM2.5 NAAQs (65 u.g/m3).
3-16
-------
identified areas have implemented or significantly strengthened their burn ban programs since
2007, we are assuming little or no reductions from a burn ban program in our 2007 base case
and large reductions (on an episodic basis) in the 2020 future year cases.
Due to the complexity of accounting for "burn bans" on specific days in the future year
modeling, we were not able to simulate the effects of "burn bans" in the 2020 base case
modeling. In this regard, the 2020 model-based design value were adjusted to reflect the
expected effects on design values of the episodic residential wood burning curtailment
programs. Using the best available information, we estimated the impacts of episodic
residential wood burning programs as a post-modeling adjustment to the 2020 base case. For
this analysis, episodic residential wood burning adjustments were made for the areas identified
in Table 3-311:
Table 3-3. Nonattainment Areas Where Episodic Residential Wood Burning Curtailment was
Applied
Nonattainment Areas Where Episodic Residential
Wood Burning Adjustments Were Applied
Chico
Los Angeles- South Coast Air Basin
Sacramento
San Francisco Bay Area
San Joaquin Valley
Yuba City-Marysville
Klamath Falls
Oakridge
Provo
Salt Lake City
Seattle-Tacoma
State
CA
CA
CA
CA
CA
CA
OR
OR
UT
UT
WA
We applied two slightly different methodologies employed to adjust the annual average
and daily average design values for burn bans in the selected areas. In both cases, the
adjustments were based on a modeling sensitivity run that zeroed-out all emissions from the
residential wood combustion category on all days of the year. Since the vast majority of
residential wood combustion emissions impacts are from primary PM2.5 emissions, we
calculated the total change in primary organic carbon, elemental carbon, and crustal PM2.5
These areas were all predicted to violate the daily NAAQS in the 2020 base case and are known to have
mandatory episodic curtailment programs. Adjustments were not applied to areas that solely violated the annual
NAAQS or did not have an existing curtailment program. The specific counties in which episodic residential wood
burning curtailment programs were applied are listed in Table 3-4.
3-17
-------
species between the base and zero-out cases to estimate the impact on PM2.5 from residential
wood combustion controls.12
Since the zero-out model run reduced all residential wood combustion emissions, we
had to scale the results of the sensitivity run to provide a realistic estimate of emissions
reduction from a burn ban program. To quantify the compliance rate of wood burning
curtailment programs we relied upon information from the Sacramento and South Coast Air
Quality Management Districts. The South Coast Air Quality Management District (SCAQMD,
2012) estimated a 75% rule effectiveness for their curtailment rule and the Sacramento
Metropolitan Air Quality Management District (SCMAQMD, 2009) estimated a 70% reduction in
residential wood combustion emissions on burn ban days in their area. Based on this
information we assumed a 70% reduction in residential wood combustion emissions on
episodic burn ban days in all areas with a mandatory burn ban program. This implies a relatively
high level of compliance, but recognizes that the program will provide less than a 100%
reduction due to non-compliance and exemptions from the rule.
For the annual NAAQS, we assumed that the burn ban programs provide reductions in
PM2.s concentrations that are commensurate with the reduction in primary PM2.s emissions13.
We also assumed that burn bans are applicable on certain days in the 1st and 4th quarters of the
year (i.e., during the residential wood combustion season). The number of days for which we
applied the burn ban was based on the observed fraction14 of measured days above 35 u.g/m3 in
the 1st and 4th quarters in the 2005-2009 base period in each affected county. For multi-county
areas, it was assumed that the burn ban control program would be applied by county (i.e. there
may be a forecasted burn ban in only a portion of a large nonattainment area). The number of
burn ban days applied per year by county is provided in Table 3-415.
12 The sensitivity run also included SO2 emissions reductions. The SO2 reductions have no impact on the organic
carbon, elemental carbon, and crustal primary PM2.5 species concentrations.
13 Since all of the adjustments are for primary PM2.5, it is assumed that emissions reductions and the change in
concentration are linear (i.e. a 50% reduction in residential wood combustion PM2.5 emissions leads to a 50%
reduction in the primary PM2.5 concentrations from residential wood combustion.)
14 FRM monitoring sites operate on different schedules (1 in 3 day, 1 in 6 day, or every day). The calculation was
based on the fraction of exceedence days during the 1st and 4th quarters. This proportionality approach normalizes
the number of high days between monitoring sites and allows a percentage of days to be applied to the modeled
days (which include all days of the year).
15 The number of burn ban days was based on the monitoring site in the county with the maximum percentage of
exceedence days (days > 35 u.g/m3).
3-18
-------
Table 3-4. Estimated Number of Burn Ban Days by County Based on 2005-2009 FRM Data
State
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
Oregon
Oregon
Utah
Utah
Washington
Nonattainment Area
Chico
Los Angeles- South Coast
Air Basin
Los Angeles- South Coast
Air Basin
Los Angeles- South Coast
Air Basin
Sacramento
San Francisco Bay Area
San Francisco Bay Area
San Francisco Bay Area
San Joaquin Valley
San Joaquin Valley
San Joaquin Valley
San Joaquin Valley
San Joaquin Valley
San Joaquin Valley
San Joaquin Valley
Yuba City-Marysville
Klamath Falls
Oakridge
Salt Lake City
Provo
Seattle-Tacoma
County
Butte
Los Angeles
Riverside
San Bernardino
Sacramento
Alameda
Santa Clara
Solano
Fresno
Kern
Kings
Merced
San Joaquin
Stanislaus
Tulare
Sutter
Klamath
Lane
Salt Lake
Utah
Pierce
Total Number of Burn Ban
Days in 1st plus 4th Quarters
18
16
20
16
20
4
8
8
42
48
40
30
20
30
38
4
16
20
16
10
16
The 2020 base case model output files were modified to replace the base case modeled
concentrations with the burn ban day concentrations on the identified number of days per year
(from Table 3-4) at each monitoring site in the 21 counties. The burn ban adjustment was
applied to an equal number of high days per quarter in the 1st and 4th quarters (i.e., half of the
burn ban days were applied to the high modeled days in the 1st quarter and half to the high
modeled days in the 4th quarter). This approach provided burn ban RRFs for the 1st and 4th
quarters. The modified 2020 base case predictions were re-run through the MATS tool to
calculate adjusted annual average design values which account for the episodic residential
wood burning curtailment programs.
A similarly representative burn ban RRF was calculated to adjust the daily design values
to account for episodic residential wood burning curtailment programs. Due to the nature of
the future year daily design value calculations, the methodology differed slightly from the
3-19
-------
annual average design value calculations. The daily design value modeled RRFs were calculated
from the change in modeled PM2.5 species on the 10% highest modeled PM2.5 days in each
quarter (i.e., the 9 highest modeled days per quarter). In this approach we assume that a burn
ban will apply to all high PM2.5 days (days > 35 u.g/m3) in the 1st and 4th quarters at each site.
Therefore, we performed the calculation by applying the 70% burn ban adjustment on the 10%
highest modeled days in the 1st and 4th quarters. The revised model data were re-run through
the MATS tool to calculate an adjusted 2020 base case daily design values which account for
the episodic residential wood burning curtailment programs. The impact on the 2020 base case
annual design values (where the burn ban adjustments were applied) ranged from 0.03 to 0.68
u.g/m3. The impact on the 2020 base case daily design values ranged from 0.1 to 13.1 u.g/m3.The
procedures for calculating 2020 control case design values that reflect the effects of the burn
ban programs are described in Section 3.3.3. Additional details on the procedures for treating
burn ban programs are provided in the AQMTSD.
3.3.1.2 Future-year Design Values Adjustments for Atypical or Unpredictable Events
Concentrations of PM2.5 at a number of monitoring sites may be influenced by atypical
or unpredictable events such as wildfires or fireworks. In the base year 2005-2009 FRM data, all
design value calculations at all sites reflect adjustments to data that EPA officially determined
have met the criteria for exclusion under the Exceptional Events Rule (EPA, 2007) during that
base year period. However, under a future year scenario it is possible that some atypical events
would qualify as exceptional events even though they did not qualify in the base 2005-2009
period. This is due to the nature of the "but for" test in the Exceptional Events Rule. The rule
states that exceptional events cannot be removed from the design value calculations unless the
monitor would not violate the NAAQS, "but for" the exceptional events. There are a number of
sites that are above the current daily PM2.5 NAAQS in the 2005-2009 period and would also
continue to violate the current NAAQS even if certain atypical event days were removed.
Therefore, those days cannot be removed from the official design value calculations for 2005-
2009 period because they do not meet the "but for" test. However, in the future year 2020
projections, we assume for analysis purposes that the impact of certain atypical or highly
variable events would meet the "but for" test. This is a reasonable assumption because at a
certain point in our modeling analysis the design value at each violating site is reduced to a
value that is slightly above the NAAQS level such that the site would attain the NAAQS "but for"
the atypical events days.
The identification of atypical event data that could affect future design value
calculations could involve an extensive data analysis exercise. It would be difficult to identify
3-20
-------
every potentially important past event for each monitoring site in the country and completely
characterize the exact nature of those days as part of this RIA. Therefore, we limited the
analysis to a small group of monitoring sites where a few atypical event days may have an
important impact on the future year design value calculations. In our analysis of potentially
important atypical events we included only monitoring sites that have 24-hour design values
predicted to violate the 35 ug/m3 daily NAAQS in the 2020 base case. At these sites we
examined the concentrations on days with daily average measurements > 35 u.g/m3. There
were several categories of potentially important atypical event days that we identified:
1. Wildfires—Summer days with high concentrations at sites in the West which
normally do not exceed the NAAQS in the summer16 [62 site-days];
2. Fireworks—High PM2.5 concentrations predominantly on July 4th or 5th [37 site-days];
3. Other unusual high data—Other site-days with very high measured PM2.5
concentrations that were much higher than concentrations on the same days at
surrounding sites [2 site-days].
Based on this assessment, we identified 101 site-days in the above categories at 25 monitoring
sites (23 of them in California) in the period 2005-200917'18. In all of the subsequent future year
design value calculations (for both the annual and daily NAAQS), the design values have been
adjusted to reflect the removal of these days. The impact on the 2005-2009 annual design
values at these 25 sites ranged from 0.08 to 0.97 u.g/m3. The impact on the daily design values
ranged from 0 to 12.9 u.g/m3.
We recalculated the 2020 base case and 2020 control case annual and daily PM2.5 design
values to reflect the removal of potential future atypical event days from the starting point
2005-2009 ambient measured data. Additional details on the methodology for adjusting the
future year design values for the purposes of this analysis are provided in the AQMTSD.
16 The vast majority of the wildfires days occurred during a well documented summer 2008 wildfire period in
Central California.
17 These were site days that were not already identified and removed from the ambient data as EPA-concurred
exceptional events.
18 The adjustments are made to the base year design values for the sole purpose of projecting ambient data to the
future year (2020). It is not appropriate to adjust the base year 2005-2009 data for the purpose of examining
current or past attainment of the NAAQS.
3-21
-------
3.3.2 Calculating Future-year Design Values for Meeting the Existing Standards, the Revised
Annual Standard, and Alternative Annual Standard Levels
The air quality ratios were used in the process of adjusting the air quality modeling
results to meet the current and alternative standard levels. The 2020 base case modeling and
the development of the adjusted 2020 base case (Box 1 and Box 2, Figure 3-2) were described
above. The procedures for determining the emissions reductions estimated to be needed to
attain the 15/35 standard and annual standards of 13, 12, and 11 u.g/m3are identified in boxes
4 through 9 in Figure 3-2. These procedures and the results are described below.
Adjusted 2020 Control Case (Box 4). Adjust design values of 2020 control case to
account for episodic wood burning curtailments and to account for atypical events and
inappropriate emissions controls. The impact of atypical events on design values was removed
from 2020 control case design values by removing these days from the ambient data used in
the future-year design value calculations in the MATS tool, as described in section 3.3.1.2,
above. To account for the impacts of wood burning curtailments in the 2020 control case, we
started with the fractional change (i.e., RRF) in speciated design values between the 2020 base
case and the 2020 control case (both cases without the effects of wood burning curtailment
programs). We then applied these species-specific RRFs to adjust the corresponding speciated
design values in the 2020 base case that reflects the application of wood burning
curtailments19.
Attainment of the 15/35 Level (Box 5). Estimate future-year design values and emission
reductions beyond the adjusted 2020 control case to meet the existing standard level (15/35).
For monitors with design values greater than 15/35 in the adjusted 2020 control case (Box 4,
Figure 3-2), additional direct PM2.5 emission reductions were applied to meet this level. The
additional direct PM2.5 emission reduction amounts were estimated using air quality ratios. The
direct PM2.5 emissions reductions needed to attain the 15/35 standard were also applied to
reduce PM2.5 design values at all attaining monitoring sites in the same county as the
nonattainment monitor. For example, the highest 24-hr design value in San Bernardino County
in the adjusted 2020 control case was 36.4 u.g/m3 at monitor 60719004. Additional emissions
reductions of 585 tons of direct PM2.5 were estimated to be required for this monitor to meet
19 We also had to adjust the 2020 modeled control case design values in certain counties to remove the impacts
from a subset of control measures. These control measures were deemed to be inappropriate for the purposes
of the 2020 control case after the 2020 base case air quality modeling was completed. To remove the impact of
these inappropriate emissions controls, the design values were adjusted either based on the air quality ratios or
on the change in the design value between the 2020 base case and 2020 control case scaled by the fraction of
the total emission reduction associated with the inappropriate controls.
3-22
-------
the 24-hr standard level20 as follows: (36.4 - 35.4) / 1.710 x 1000 = 585 tons, where 1.710 is the
24-hr direct PM2.5 air quality ratio for the monitor 60719004 (Table 3.A-3). The 585 tons of
direct PM2.5 emissions reductions in this county were estimated to reduce the highest annual
design value in San Bernardino at monitor 60710025 from 13.41 to 12.99 u.g/m3 as follows:
13.41 - (585 x 0.710 / 1000) = 12.99 u.g/m3, where 0.710 is the annual direct PM2.5 air quality
ratio for the 60710025 monitor (Table 3.A-3). The direct PM2.5 emission reduction amounts
beyond the adjusted 2020 control case that are necessary to meet the current standard level
for individual counties are listed in Table 3-5.
Table 3-5. Tons of Direct PM2.5 Emission Reductions beyond the Adjusted 2020 Control Case
to Meet the Current Standard Level for Counties that Exceed the Revised or
Alternative Annual Standard Levels in the Adjusted 2020 Base Case
FIPS Code
6019
6025
6029
6031/6107
6071
6099
42003
State Name
California
California
California
California
California
California
Pennsylvania
County Name
Fresno
Imperial
Kern
Kings/Tulare
San Bernardino
Stanislaus
Allegheny
Direct PM2.5 Emissions
(tons)
497
288
1,496
610
585
346
764
Emissions were controlled in certain counties in the 2020 control case that exceeded the
alternative annual standard of 11 u.g/m3 but that did not exceed the existing standard level.
These emissions controls are relevant for meeting the 11 u.g/m3 level (Box 9) but are not
relevant for meeting the existing standard level. Therefore annual design values in the 15/35
case are set to those of the adjusted 2020 base case for monitors in the following counties:
Jefferson, AL; Shoshone, ID; Cook, IL; Madison, IL; Klamath, OR; Lake, IN; Scott, IA; Wayne, Ml;
Milwaukee, Wl; and Harris, TX.21
20 A 24-hour design value of 35.4 ng/m3 is the highest value that meets the 24-hour standard.
21 Arrows point from Box 2 and Box 4 to Box 5 in Figure 3-2 because information from both the adjusted 2020 base
case and the adjusted 2020 control case was used in developing the set of design values that correspond to
attainment of the existing standard.
3-23
-------
Analytical Baseline (Box 6). Create analytical baseline for meeting alternative standards
that accounts for 2025 mobile NOx emission adjustment in San Joaquin Valley and South Coast
Air Basin. The goal of this RIA is to provide the best estimates of the costs and benefits of an
illustrative attainment strategy to just meet the revised annual 12 u.g/m3 standards, as well as
two alternative annual standards of 13 u.g/m3 and 11 u.g/m3, incremental to just meeting the
current standards of 15/35, and reflecting emissions projected to reflect the impact of
economic growth and implementation of state and federal emissions controls. Most areas of
the U.S. will be required to demonstrate attainment with the new standards by 2020. As a
result, for these areas, the correct baseline for estimating the incremental emissions reductions
that would be needed to attain the more protective standards is a baseline with emissions
projected to 2020 and adjusted to reflect the additional emissions reductions that would be
needed to attain the current 15/35 standards. For two areas in Southern California (South Coast
and San Joaquin), the degree of projected non-attainment with the revised annual standard of
12 u.g/m3 is high enough that those counties are not expected to be able to demonstrate
attainment with the new standard by 2020. Instead, those two areas are likely to qualify for an
(up to) five year extension of their attainment date. If the areas are granted an attainment date
extension, they will have until 2025 to demonstrate attainment with the revised annual
standard. As a result, for these two areas, the correct baseline for estimating the incremental
emissions reductions that would be needed to attain the more protective standards is a
baseline with emissions projected to 2025 adjusted to reflect the additional emissions
reductions that would be needed to attain the current 15/35 standards. This difference in
attainment year is important because between 2020 and 2025, emissions from mobile sources
in California are expected to be reduced due to continued fleet turn over from older, higher
emitting vehicles to newer, lower emitting vehicles. These reductions in emissions will occur as
a result of previous state rules for which costs and benefits have already been counted, and
thus will not be costs and benefits attributable to meeting the revised annual standard.
Modeling of two separate years is time prohibitive, and would result in two separate
years of benefits and costs which would not provide a complete picture of the nationwide costs
and benefits of just meeting the new standards in either 2020 or 2025 because of differences in
the baselines between the two years. To provide the most reasonable and reliable estimates of
costs and benefits of full attainment for the nation, we are constructing an analytical baseline
for estimating the costs and benefits of attaining the revised standard of 12 u.g/m3 and
alternative annual standards of 13 u.g/m3 and 11 u.g/m3 with the following characteristics. The
analytical baseline was developed by applying a mobile NOx emission adjustment to design
values at levels attaining 15/35. This approach allows us to generate costs and benefits of full
3-24
-------
attainment without overstating the costs and benefits in those two areas, which would occur if
we forced costly emissions reductions in 2020 in areas that would not have to be incurred until
2025, and which will be offset because of the expected reductions in mobile source emissions
due to other programs.22
The emissions adjustment is equal to 27,467 tons of NOx emissions reductions in the
South Coast Air Basin and 14,410 tons in the San Joaquin Valley. Annual design values for the
15/35 baseline were adjusted to account for these emissions reductions using the air quality
ratios listed in Table 3.A-1. Incremental costs and benefits of the revised and alternative
standards are assessed relative to this set of analytic baseline design values. Annual design
values and exceedance categories are provided for the analytic baseline in Table 3-6 and Figure
3-3 for counties with at least one monitor that exceeds a level.23
11
Figure 3-3. Counties that Exceed the Revised and/or Alternative Annual Standard Levels of
13,12 and 11 u,g/m3 in the Analytical Baseline
Benefits for all areas are estimated using 2020 population data for consistency, recognizing that full attainment
costs and benefits will not actually be realized until 2025 for a portion of the costs and benefits. The 2020
estimates of full attainment costs and benefits will be an underestimate of benefits in 2025 because of population
growth and changes in the age distribution of the population between 2020 and 2025.
23 There were two counties (Lincoln County, MT and Santa Cruz County, AZ) that exceeded alternative standard
levels in the 2020 base case for which we used a weight-of-evidence approach to determine how they would attain
these levels, as described in Section 3.3.5.
3-25
-------
Table 3-6. Annual Design Values and Exceedance Category for the Highest County Monitor
in the Analytical Baseline for Counties with at Least one Monitor Above the
Revised and/or Alternative Standard Levels
FIPS Code
6065
6107
6029
6071
6025
6037
6047
55079
6031
17119
6019
26163
1073
17031
16079
19163
48201
48141
41035
55133
18089
6063
42003
Monitor ID
60658005
61072002
60290016
60710025
60250005
60371002
60472510
550790059
60310004
171191007
60190008
261630033
10730023
170316005
160790017
191630019
482011035
481410044
410350004
551330027
180891003
60631009
420030064
State Name
California
California
California
California
California
California
California
Wisconsin
California
Illinois
California
Michigan
Alabama
Illinois
Idaho
Iowa
Texas
Texas
Oregon
Wisconsin
Indiana
California
Pennsylvania
County Name
Riverside
Tulare
Kern
San Bernardino
Imperial
Los Angeles
Merced
Milwaukee
Kings
Madison
Fresno
Wayne
Jefferson
Cook
Shoshone
Scott
Harris
El Paso
Klamath
Waukesha
Lake
Plumas
Allegheny
Annual DV
14.58
13.23
12.7
12.64
12.57
12.34
12.12
12.02
11.79
11.7
11.61
11.58
11.56
11.52
11.52
11.51
11.43
11.39
11.3
11.22
11.17
11.15
11.12
13/35
X
X
12/35
X
X
X
X
X
X
X
11/35
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
13 Standard Level (Box 7). Estimate future-year design values and emission reductions
beyond the analytical baseline to meet the alternative annual standard level of 13 |-ig/m3.
Annual PM2.5 design values at monitors in Tulare and Riverside Counties in California exceeded
the alternative standard level of 13 u,g/m3 in the analytical baseline (Table 3-6 and Figure 3-3).
The additional direct PM2.5 emission reductions required for these counties to meet this
standard level were estimated using air quality ratios. For example, the highest annual design
value in Riverside County in the analytical baseline case was 14.58 u.g/m3. Emission reductions
of 626 tons of direct PM2.5 were estimated to be required for this monitor to meet the annual
standard level of 13.04 u.g/m3 as follows: (14.58 - 13.04) / 2.459 x 1000 = 626 tons, where
2.459 is the annual direct PM2.5 air quality ratio for monitor 60658005 (Table 3.A-3). The
3-26
-------
emissions reductions by county to attain a 13 u.g/m3 standard are provided in Table 3.7. These
reductions were applied to lower the annual PM2.5 design values at all sites in the given
county.24
Table 3-7. Tons of Direct PM2.5 Emission Reductions Beyond the Analytical Baseline to Meet
the 13 u,g/m3 Level
FIPS Code
6065
6107
State Name
California
California
County Name
Riverside
Tulare
Direct PM2.5 Emissions Reductions (tons)
626
101
12 Standard Level (Box 8). Estimate future-year design values and emission reductions
beyond the analytical baseline to meet the revised annual standard level of 12 u.g/m3. Annual
PM2.5 design values at monitors in the following 7 counties in California exceeded the revised
standard level of 12 u,g/m3 in the analytical baseline (Table 3-6 and Figure 3-3): Los Angeles,
Riverside, San Bernardino, Kern, Tulare, Merced, and Imperial. The additional direct PM2.5
emission reductions required for these counties to meet the standard level of 12 u,g/m3 were
estimated using air quality ratios. For example, the highest annual design value in Riverside
County in the analytical baseline case was 14.58 u.g/m3. Emission reductions of 1,033 tons of
direct PM2.5 were estimated to be required for this monitor to meet the annual standard level
of 12.04 ug/m3 as follows: (14.58 - 12.04) / 2.459 x 1000 = 1033 tons, where 2.459 is the
annual direct PM2.5 air quality ratio for monitor 60658005 (Table 3.A-3). The emissions
reductions by county to attain a 12 u.g/m3 standard are provided in Table 3.8. These reductions
were applied to lower the annual PM2.5 design values at all sites in the given county25.
24 Emissions reductions needed in Tulare County were also applied to reduce the annual PM2.5 design value at the
monitor in Kings county, which is combined with Tulare in our analysis, as discussed in Appendix 3.A.I.I.
25 For Kings and Tulare Counties, the maximum of the emission reductions required for the individual counties was
applied to monitors in both counties using the air quality ratios since these counties are combined in our analysis,
as discussed in Appendix 3.A.I.I.
3-27
-------
Table 3-8. Tons of Direct PM2.5 Emission Reductions Beyond the Analytical Baseline to Meet
the 12 u,g/m3 Level3
FIPS Code
6037
6065
6025
6029
6107
6047
6071
State Name
California
California
California
California
California
California
California
County Name
Los Angeles
Riverside
Imperial
Kern
Tulare
Merced
San Bernardino
Direct PM2.5 Emissions Reductions (tons)
743
1,033
294
418
635
19
844
a See Appendix Chapter 7.A for additional details on known and unknown emissions reductions and costs, by
county, for 12 |J.g/m3.
11 Standard Level (Box 9). Estimate future-year design values and emission reductions
beyond the analytical baseline to meet the alternative annual standard level of 11 |-ig/m3.
Annual PM2.5 design values at monitors in 23 counties exceeded the alternative standard level
of 11 u,g/m3 in the analytical baseline (Table 3-6 and Figure 3-3). As discussed above, annual
design values in the analytical baseline do not reflect the emission controls of the 2020 control
case for counties with monitors that did not exceed the current standard level in the 2020 base
case. To estimate the emission reductions beyond the known controls needed to meet the
alternative standard level of 11 u,g/m3 in these counties, we started with annual design values
for the adjusted 2020 control case (Box 4 of Figure 3-2). The additional direct PM2.5 emission
reductions required for these counties to meet the alternative standard level were then
estimated using air quality ratios. For example, the annual design value at the high monitor in
Jefferson, AL was 11.56 u,g/m3 in the adjusted 2020 base case and 11.11 u,g/m3 in the adjusted
2020 control case. The additional direct PM2.5 emission reductions needed beyond the emission
reductions of the 2020 control case for this monitor to meet the 11 u,g/m3 level were estimated
using air quality ratios as follows: (11.11 - 11.04) / 0.561 x 1000 = 125 tons, where 0.561 is the
direct PM2.5 air quality ratio for monitor 10730023. Annual PM2.5 design values associated with
emission reductions estimated in this way in (Table 3-9) were calculated for the counties with
exceedance monitors.
3-28
-------
Table 3-9. Tons of Direct PM2.5 Emission Reductions Beyond the Analytical Baseline to Meet
the Alternative Standard 11 u,g/m3 Level3
FIPS Code
6037
6065
1073
6019
6025
6029
6031/6107
6071
6047
6063
17031
17119
18089
16079
41035
42003
19163
26163
55079
55133
48141
48201
State Name
California
California
Alabama
California
California
California
California
California
California
California
Illinois
Illinois
Indiana
Idaho
Oregon
Pennsylvania
Iowa
Michigan
Wisconsin
Wisconsin
Texas
Texas
County Name
Los Angeles
Riverside
Jefferson
Fresno
Imperial
Kern
Kings/Tula re
San Bernardino
Merced
Plumas
Cook
Madison
Lake
Shoshone
Klamath
Allegheny
Scott
Wayne
Milwaukee
Waukesha
El Paso
Harris
Tons of Direct PM2.5
3,222
1,440
125
325
850
1,051
1,168
2,252
255
44
427
1,687
0
61
25
154
188
870
455
55
158
123
aFor the following counties, the emission reductions listed are relative to the adjusted 2020 control case design
values rather than the analytical baseline: Jefferson, AL; Shoshone, ID; Cook, IL; Madison, IL; Klamath, OR; Lake, IN;
Scott, IA; Wayne, Ml; Milwaukee, Wl; and Harris, TX.
3.3.3 Estimating Changes in Annual A verage PM2.$ for Benefits Inputs
The calculation of health benefits for the revised annual standard of 12 u.g/m3 and the
two alternative annual standards uses spatial surfaces of gridded annual average PM2.5
concentrations for the analytical baseline and spatial surface reflecting attainment of each
different standard. The spatial surface for each case covers the U.S. portion of the air quality
3-29
-------
modeling domain. To create the spatial field for the analytical baseline we started with a spatial
surface for the 2020 control case reflecting the removal of atypical events. The 2020 control
case spatial surface was adjusted using the projected annual design values for the analytical
baseline to create the spatial surface for the baseline. The spatial surface for the 2020 control
case was also adjusted to reflect attainment of the different standards using the annual design
values for each standard. Details of this process are described below.
The spatial surface for the 2020 control case (with removal of potential future atypical
events) was developed using MATS by calculating species-specific RRFs at every grid cell within
the modeling domain for the 2020 control case and applying these RRFs to ambient data that
have been interpolated to cover all grid cells in the modeling domain. The basic spatial
interpolation technique, called Voronoi Neighbor Averaging (VNA), was applied for annual
design values for the 2020 control case and each standard to create spatial fields of annual
PM2.5 for each of these cases. As part of this technique, VNA uses the inverse distance squared
weighted average of the annual design values at monitoring sites that are nearest to the center
of each model grid cell. We then calculate the ratio of annual PM2.5 for each standard level to
annual PM2.5 for the 2020 control case for each grid cell in the VNA fields. These gridded ratios
are then multiplied by the gridded annual concentrations from the MATS outputs for the 2020
control case. That is, a spatial surface was calculated by adjusting the 2020 control case using a
multiplicative factor calculated as the ratio of the gridded design values for attainment of each
standard to the gridded design values of the 2020 control case where the design value gridded
spatial fields are based on the nearest neighbor monitor locations (weighted by distance). This
approach is shown mathematically in the equation below.
Adjusted AQi •= VNA InterpQlQted AQu from 4/temgt/Ve - XMATSAQ,;
11 VNA Interpolated AQu from 2020
where ij refers to column / and row 7 of the modeling domain. This approach aims to estimate
the change in population exposure associated with attaining an alternate NAAQS, relying on
data from the existing monitoring network and the inverse distance squared variant of the VNA
interpolation method to adjust the MATS gridded concentrations such that each area attains
the standard alternatives. Using the VNA spatial averaging technique, the annual average PM2.5
spatial surfaces are smoothed to minimize sharp gradients in PM2.5 concentrations in the spatial
fields due to changes in the monitor concentrations26. Because the VNA approach interpolates
26 For the purposes of estimating benefits, this smoothed surface was then clipped to grid cells within 50 km of
monitors whose design values were changed as a result of the standard level.
3-30
-------
monitor values, it is most reliable in areas with a denser monitoring network. In areas with a
sparser monitoring network, there is less observed monitoring data to support the VNA
interpolation and we have less confidence in the air quality values further away from the
location of monitoring sites. To the extent that any bias in the interpolated values is present,
the ratio of the interpolated values should be relatively insensitive to this bias and the adjusted
air quality values should be unaffected.
3.3.4 Limitations of Using Adjusted Air Quality Data
Due to time constraints, design values and PM2.5 surfaces at the analytical baseline level
and the alternative standard levels were based on adjusted fields derived from the modeled
2020 base case and 2020 control case, rather than directly on air quality simulation results.
While a credible technical basis exists for the adjustment procedures used in this analysis, there
are important limitations to the approaches used to estimate the response of air quality to
emissions changes. For instance, air quality ratios are calculated with results from a limited
number of CMAQ sensitivity runs and are based on the assumption that the monitor design
values would decrease with additional emissions reductions of S02, NOX and direct PM2.5 similar
to how the CMAQ sensitivity runs predicted changes in air quality concentrations. The
uncertainty of this assumption will increase with increasing emissions reductions needed to
estimate attainment. In addition, the model response to emissions changes are analyzed at a
county-level or within a small group of counties, and we assume that air quality concentrations
at a monitor will decrease linearly with emissions reductions in a county (e.g., direct PM2.5
emission reductions) or a group of counties (e.g., S02 and NOX emissions reductions). Because
of the more local influence of changes in directly emitted PM2.5 emissions on air quality, it is
also particularly difficult for the air quality ratio approach to estimate well how the design value
at a monitor in a county would respond to changes in direct PM2.5 emissions in a county
without knowing the location of the source (e.g., extrapolated emissions reductions) relative to
the location of the monitor.
The exact impact of using this methodology to estimate the emissions reductions
needed for attainment and the associated effect on the cost and benefits is uncertain and may
vary from monitor-to-monitor. We do not believe that this methodology tends towards any
general trend and does not always result in either an underestimation or overestimation of the
costs and benefits of attaining the alternative standards.
3-31
-------
3.3.5 Weight-of-Evidence Approach for Lincoln County, MT and Santa Cruz, NM
There were two counties that exceeded alternative standard levels in the 2020 base
case for which we used a weight-of-evidence approach to determine how they would attain
these levels. These counties are Lincoln County, MT and Santa Cruz County, AZ.
Lincoln County's PM2.5 air quality problem is dominated by residential wood combustion
emissions of PM2.5, and the County has few additional emissions sources to control. The Lincoln
County monitor is situated in the City of Libby in a valley that is subject to wintertime
temperature inversions (Figure 3-4). These temperature inversions, which suppress air mixing
and dilution of PM2.5, combined with resident's reliance on wood burning for home heating can
produce poor PM2.5 air quality. However, since 2005, Libby has successfully implemented a
woodstove change-out program that has resulted in consistent improvements in PM2.5 air
quality in recent years (Figure 3-5). The success of this program and the downward trend in
annual design values at the Libby monitor suggests that Libby will meet the revised and
alternative standard levels in 2020. Since residential wood combustion emissions in Libby and
the emission reductions due to the wood-stove change-out program are not fully captured in
our emission inventory, our modeled estimates of future-year design values are not reliable at
this site. However, our weight-of-evidence considerations suggest that Lincoln County would
likely attain the alternative standard levels in 2020 based on on-the-books control programs.
Figure 3-4. City of Libby in Lincoln County, Montana (Image taken from Google Earth)
3-32
-------
3
20 0
•
S- 15.0
Lincoln Co. MT(Libby)
^*V
****"•— •— •
' ' ' *~~-*— t » i
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
16.0
Lincoln Co. MT (Libby)
*— • i
**--» « i -^
^^ * ~* «»
2002 2003 2004 2005 2006 2007 2003 2009 2010 2011
Figure 3-5. Three-year Annual and 24-hr Design Values for the Monitor in Libby, MT
Santa Cruz, AZ had a 24-hr design value of 29.7 u,g/m3 and an annual design value of
12.65 u,g/m3 in the 2020 base case. However, Santa Cruz has few local emissions sources and
therefore relatively low emissions available for control. Total emissions of S02, NOx and direct
PM2.5 were 65, 688, and 542 tons, respectively, in Santa Cruz County in the 2020 base case.
Total emissions of S02, NOx and direct PM2.5 for the Mexican State of Sonora, which borders
Santa Cruz, were much greater: 100,089, 53,518 and 27,641 tons, respectively. The lack of
substantial local controllable emissions in Santa Cruz and the large impact of emissions from
Sonora, Mexico on air quality in Santa Cruz suggest that emissions from Mexico make meeting
the alternative standards for this county impractical in our analysis. Cross-border impacts of
Mexican emissions on Santa Cruz County have been recognized previously. On September 25,
2012, in a Federal Register Notice, EPA Region IX approved a State Implementation Plan (SIP)
revision submitted by the Arizona Department of Environmental Quality. As indicated in the
Notice, EPA Region IX reviewed three years of air quality data from Arizona and determined
that the Nogales nonattainment area in Santa Cruz County is attaining the National Ambient Air
Quality Standard for PMio, but for international emissions sources in Nogales, Sonora, Mexico.
Our weight-of-evidence considerations suggest that Santa Cruz would likely not require
emissions reductions in addition to those of on on-the-books control programs to attain the
alternative standard levels.
3.3.6 Estimating Changes in Visibility for Analyzing Welfare Benefits
Changes in visibility were calculated in order to assess both recreational and residential
visibility welfare benefits. The visibility calculations for the welfare benefits assessment are
based on annual average light extinction (bext) values, converted to units of visual range (km).
Since we are interested in providing visibility estimates throughout the US, we utilize gridded,
speciated PM2.5 data that is produced by MATS (Abt, 2012) along with future-year design values
for the annual NAAQS.
3-33
-------
The gridded species data used to calculate the visibility values is somewhat different
than the gridded data used to calculate health benefits. The gridded PM2.5 data used as input to
BenMAP for health benefits is based on adjusted species data using the SANDWICH technique
(Frank, 2006). The PM2.5 species data is adjusted to match the nature of the PM2.5 FRM filter
data that is used as the basis for determining attainment of the PM2.5 NAAQS. For example, in
the spatial fields used in BenMAP, the nitrate data has been adjusted to account for
volatilization, a particle bound water component is added to the sulfate and nitrate
concentrations, and the organic carbon is calculated as the difference between the measured
FRM PM2.5 mass and the sum of the rest of the PM2.5 species. For visibility calculations, we use
the "raw" PM2.5 species data, as measured by IMPROVE and CSN monitors. Equation 3.1 shows
the "old" IMPROVE equation which is used to calculate visibility in Mm"1. Note that the coarse
PM component of the "old" IMPROVE equation was excluded here because this term is not
used in calculating visibility spatial fields.
bext = 3 x /(RH)x [Sulfate] + 3 x /(RH)x [Nitrate] + 4 x [Organic Mass]
+ 10 x [Elemental Carbon] + 1 x [Fine Soil] + 10
The mass concentrations of the components indicated in brackets are in units of u.g/m3, and
/(RH) is the unitless water growth term that depends on relative humidity. The final term in the
equation is known as the Rayleigh scattering term and accounts for light scattering by the
natural gases in unpolluted air. Since IMPROVE does not include ammonium ion monitoring, the
assumption is made that all sulfate is fully neutralized ammonium sulfate and all nitrate is
assumed to be ammonium nitrate.
The visibility values are calculated from observed concentrations for each of the PM
species for each calendar quarter. Using the "old" IMPROVE equation (without the coarse mass
component), and with quarterly averaged climatological average relative humidity [f(RH)]
values, we calculate a quarterly average light extinction (bext) value from the IMPROVE and
CSN data for the 2006-2008 base period which has been interpolated to the CMAQgrid using
gradient adjusted spatial fields (eVNA). The observed sulfate and nitrate concentrations are
assumed to be fully neutralized by ammonium and the organic carbon is multiplied by 1.4 to
derive organic carbon mass. The interpolated gridded 2006-2008 ambient data is projected to
2020 using modeled RRFs. CMAQ derived quarterly average RRFs for sulfate, nitrate, elemental
carbon, organic carbon, and crustal components are multiplied by the gridded light extinction
components to get future year quarterly average visibility. The four quarterly average total light
3-34
-------
extinction values (for each grid cell) are then averaged together to get annual average visibility.
The procedure was repeated for both the 2020 base case and 2020 control case scenarios.
The gridded field of 2020 base case and control case annual average visibility is used to
calculate residential visibility benefits in the following manner. The visibility data at Class I areas
is extracted from the gridded data to calculate recreational visibility benefits. The Class I area
visibility is based on the visibility calculated at the grid cell which contains the centroid of each
of the 149 Class I areas in the continental United States.
3.4 References
Abt Associates, 2010. User's Guide: Modeled Attainment Test Software.
http://www.epa.gov/scram001/modelingapps mats.htm
Appel, K.W., Bhave, P.V., Gilliland, A.B., Sarwar, G., Roselle, S.J., 2008. Evaluation of the
community multiscale air quality (CMAQ) model version 4.5: Sensitivities impacting
model performance; Part II—particulate matter. Atmospheric Environment 42, 6057-
6066.
Appel, K.W., Gilliland, A.B., Sarwar, G., Gilliam, R.C., 2007. Evaluation of the Community
Multiscale Air Quality (CMAQ) model version 4.5: Sensitivities impacting model
performance Part I—Ozone. Atmospheric Environment 41, 9603-9615.
Byun, D.W., and K.L. Schere. 2006. "Review of the Governing Equations, Computational
Algorithms, and Other Components of the Models-3 Community Multiscale Air Quality
(CMAQ) Modeling System." J. Applied Mechanics Reviews 59(2):51-77.
Carlton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.D., Sarwar, G., Pinder, R.W., Pouliot, G.A.,
Houyoux, M., 2010. Model Representation of Secondary Organic Aerosol in CMAQv4.7.
Environmental Science & Technology 44, 8553-8560.
Carlton, A.G., Turpin, B.J., Altieri, K.E., Seitzinger, S.P., Mathur, R., Roselle, S.J., Weber, R.J.,
2008. CMAQ Model Performance Enhanced When In-Cloud Secondary Organic Aerosol
is Included: Comparisons of Organic Carbon Predictions with Measurements.
Environmental Science & Technology 42, 8798-8802.
ERG. 2009. Development of Mexico National Emissions Inventory Projections for 2008, 2012,
AND 2030.
ERG. 2009. Technical memorandum: "Mexico 2018 Emissions Projections for Point, Area, On-
Road Motor Vehicle and Nonroad Mobile Sources."
IMPROVE, (2006), "Spatial and Seasonal Patterns and Temporal Variability of Haze and its
Constituents in the United States: Report IV." Cooperative Institute for Research in the
3-35
-------
Atmosphere, ISSN: 0737-5352-74. Also available at:
http://vista.cira.colostate.edu/improve/Publications/Reports/2006/PDF/IMPROVE Report IV.p
df
Lam, Y-F., Fu, J.S., Jacob, D.J., Jang, C, Dolwick, P., 2010. 2006-2008 GEOS-Chem for CMAQ
Initial and Boun
Chapel Hill, NC.
Initial and Boundary Conditions. 9th Annual CMAS Conference, October 11-13, 2010,
Nenes, A., Pandis, S.N., Pilinis, C., 1998. ISORROPIA: A new thermodynamic equilibrium model
for multiphase multicomponent inorganic aerosols. Aquatic Geochemistry 4, 123-152.
South Coast Air Quality Management District, 2012. Revised Draft 2012 Air Quality
Management Plan. Also available at:
http://www.aqmd.gov/aqmp/2012aqmp/RevisedDraft/RevisedDraft2012AQMP-Main-clean.pdf
Sacramento Metropolitan Air quality Management District, 2009. Staff Report: Rule 421,
Mandatory Episodic Curtailment of Wood and Other Solid Fuel Burning. Also available
at: http://www.airqualitv.org/notices/Rules2009/200907Rule421WorkshopsStaffReport.pdf
U.S. Environmental Protection Agency (EPA). 2007. Guidance on the Use of Models and Other
Analyses for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.s, and
Regional Haze. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. Environmental Protection Agency (EPA). 2011a. Meteorological Model Performance for
Annual 2007 Simulations. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. EPA-454/R-11-007.
U.S. Environmental Protection Agency (EPA). 2011b. Update to the 24 Hour PM2.5 NAAQS
Modeled Attainment Test. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. Also available at:
http://www.epa.gov/ttn/scram/guidance/guide/Update to the 24-
hour PM25 Modeled Attainment Test.pdf
U.S. Environmental Protection Agency (EPA). 2012a. Technical Support Document: Air Quality
Modeling for the Final PM NAAQS RIA. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. November 2012.
U.S. Environmental Projections Agency (EPA). 2012b. Technical Support Document: Preparation
of Emissions Inventories for the Version 5, 2007-based Platform. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. November, 2012.
Yantosca, B., 2004. GEOS-CHEMv7-01-02 User's Guide, Atmospheric Chemistry Modeling
Group, Harvard University, Cambridge, MA, October 15, 2004.
Yarwood, G., Rao, S., Yocke, M., Whitten, G.Z., 2005. Updates to the Carbon Bond chemical
mechanism: CB05. Final Report to the US EPA, RT-0400675, December 8, 2005.
3-36
-------
APPENDIX 3.A
ADDITIONAL AIR QUALITY MODELING INFORMATION
3.A.1 Air Quality Modeling and Analysis
This appendix provides supplemental information for the air quality modeling analysis in
Chapters.
3.A.1.1 Development of Air Quality Response Ratios
The air quality response ratios (hereafter referred to as air quality ratios) used to adjust
the 2020 cases to meet the standard levels were calculated based on results of several
sensitivity simulations. The sensitivity simulations, as described in Table 3-4, were defined to
isolate the changes in the (NH4)2S04, NH4N03 and direct PM2.5 associated with changes in
emissions of S02, NOX and direct PM2.5, respectively. These PM2.5 component species were
selected for reduction to meet the standard levels because they dominate the mass of PM2.5 in
the areas of concern in the 2020 cases. The sensitivity simulation referred to as "2020
NOx_PM2.5" was used in calculating the air quality ratios associated with changes in NOX and
direct PM2.5 emissions. This simulation was based on anthropogenic NOX and direct PM2.5
emission reductions from non-EGU sources of 25% and 50%, respectively, relative to the 2020
base case. The sensitivity simulation referred to as "2020 S02_RWC" was used in calculating the
air quality ratios associated with changes in S02 emissions. This simulation was based on
anthropogenic S02 and residential wood combustion emissions reductions from non-EGU
sources of 25% and 100%, respectively, relative to the 2020 base case.1 In the sensitivity runs,
emissions reductions for direct PM2.5 were generally applied in counties with monitors with
annual design values above 11 u.g/m3 level in the 2020 base case, while emission reductions for
NOX and S02 were generally applied in those counties as well as their adjacent counties. This
approach reflects the local impacts of direct PM2.5 emissions on air quality and the broader
geographic impacts on PM2.5 of S02 and NOX emissions reductions.
In calculating air quality ratios, a "county group" associated with each monitor was
defined for estimating the change in emissions associated with a given change in the design
value at the monitor. For the development of direct PM2.5 air quality ratios, the county group
included just the county containing the monitor because of the relatively local nature of the
impacts of direct PM2.5 emissions on ambient PM2.5 concentrations. For the development of
NOX and S02 air quality ratios, the county group was generally defined as the county containing
1 The results of this sensitivity run was also used in the method to quantify the impacts on design values of existing
burn ban programs, as described in Section 3.3.1.1.
3.A-1
-------
the nonattainment monitor plus the adjacent counties (i.e., counties that border the county
with the nonattainment monitor). This multi-county group approach was used for NOx and S02
in view of the more widespread impacts on (NH4)2S04, NH4N03 of local emissions reductions of
NOx and S02 compared to direct PM2.5. Note that this same general approach was used in the
design of the 2020 sensitivity simulations and in the 2020 control case (see Chapter 4).
However, there were exceptions to this approach in certain areas in California where
meteorological conditions affect the relationships between emissions and pollutant
concentrations on a broader geographic scale within the South Coast Air Basin and within the
San Joaquin Valley Air Basin. In the South Coast Air Basin, the county group for NOX and S02
emission reductions was defined to include all counties in the air basin (i.e., Orange,
Los Angeles, Riverside, and San Bernardino). For counties in the San Joaquin Valley Air Basin,
the total NOX emission change that contributed to PM2.5 changes at a monitor in a given county
was estimated using the weighted contribution of emissions changes in area counties as
derived from the 2020 SJV simulations (see Appendix 3.A.1.2 for details).
In adjusting design values of the 2020 control case to meet different standard levels,
Kings County and Tulare County in California were considered as a single area.2 These counties
share an east-west border and experience similar air quality due to their relative positions in
the San Joaquin Valley. Also, direct PM2.5 emissions are much smaller in Kings than in Tulare,
and the Kings County monitor is close to the Tulare border (Figure 3.A-1) such that Tulare
emissions have a large impact on design values in Kings County.
2 To group these counties into a single area, the emission reductions needed for the Kings and Tulare monitors to
meet the standard individually was first determined. Then the maximum of the individual emission reductions
was selected and was used to adjust the design values at monitors in both counties using the air quality ratios.
3.A-2
-------
Figure 3.A-1. Location of Kings County Monitor Relative to Tulare County Border.
Air quality ratios for emissions of direct PM2.5, S02 and NOX were calculated using
information from the sensitivity simulations on the response of air quality at monitors to
emission changes within the county groups. Below are the steps we followed in calculating the
air quality ratios:
Step 1: Calculate the fractional change in speciated annual and 24-hr design values for
the 2020 sensitivity cases relative to the 2020 base case. Speciated annual and quarterly 24-hr
RRFs were calculated for the 2020 NOX_PM2.5 and 2020 S02_RWC sensitivity simulations
relative to the 2020 base case using MATS (Abt, 2010) for configurations where the 2020 base
case was used as the reference case and the 2020 sensitivity cases were used as the control
cases. The fractional change in the direct PM2.5, (NH4)2S04and NH4N033 components of the
design value for the 2020 sensitivity cases relative to the 2020 base case was then calculated as
(RRF-1)4 for a given monitoring site.
Step 2: Calculate the fractional change in emissions in the relevant county group for the
2020 sensitivity cases relative to the 2020 base case. The fractional changes in emissions of
3The (NH4)2SO4 and NH4NO3 components are computed using the SO4, NO3, NH4 and water fraction from MATS as
described in EPA guidance (EPA, 2007). The direct PM2.5 design value component is computed by summing the
elemental carbon, organic carbon and crustal portions of the design value.
4 For daily air quality ratios, a representative RRF was calculated as a weighted average of the quarterly 24-hr RRFs,
where the weighting factors were the fractions of high 24-hr concentration days that occurred in the quarter in
the 2020 control case.
3.A-3
-------
direct PM2.55, S02 and NOX between the 2020 base case and 2020 sensitivity cases were
determined for the county group relevant to a given monitor. County emission groups for NOX
and S02 for the monitors considered are listed in Tables 3.A-1 and 3.A-2.
Step 3: Calculate the ratio of fractional change in speciated design value to fractional
change in emissions for the sensitivity cases. The ratio of the fractional change in speciated
design values (Step 1) to fractional change in county group emissions (Step 2) was calculated.
Specifically, we calculated the fractional change in the direct PM2.5, (NH4)2S04and NH4N03
components of the annual and daily standard design values per fractional change in direct
PM2.5, S02 and NOX emissions, respectively, in the county group between the 2020 sensitivity
cases and the 2020 base case.
Step 4: Calculate the ratio of the speciated design values to emissions for the 2020
control case. Using air quality and emission data from the 2020 control case, we calculated the
ratio of direct PM2.5, (NH4)2S04 and NH4N03 to the emissions of direct PM2.5, S02 and NOX,
respectively, in the relevant county group for the 2020 control case.
Step 5: Calculate air quality ratios using results of Steps 3 and 4. Air quality ratios were
calculated by multiplying the ratios from Step 3 by the ratios from Step 4 for each 2020
sensitivity case, individually. The overall calculation of air quality ratios for PM2.5 component
specie /' and emission speciey is given by equation 3-1, where DVt indicates the PM2.5
component design value.
Air Quality Ratio= RRFt-\ —DV,— ^^QQQ (3.A.1)
AEmission . /Emission . Emission .
\ J I J J SensitivityCase \ J J ControlCase
Air quality ratios give an estimate of how PM2.5 design value components (u.g/m3) would
change if 1000 tons of direct PM2.5, S02 and/or NOX emissions were reduced in the county group
in which the monitor is located. Annual air quality ratios that relate changes in the NH4N03
component of the design value to changes in NOX emissions are listed in Table 3.A-1 for
counties in the South Coast Air Basin and San Joaquin Valley of California that received a mobile
NOX emission adjustment equal to the change in mobile NOX emissions from the year 2020 to
2025. Annual and daily air quality ratios that relate changes in the (NH4)2S04 component of the
design value to changes in S02 emissions are listed in Table 3.A-2 for monitors in counties
where air quality ratios were used in adjusting daily design values for remove the impact of
5 Direct PM2.5 emissions are computed as the sum of emissions of elemental carbon, primary organic carbon, and
unspeciated PM2.5 mass.
3.A-4
-------
inappropriate S02 controls. Annual and daily air quality ratios that relate changes in the direct
PM2.5 component of the design value to changes in direct PM2.5 emissions are listed in Table
3.A-3 for monitors in counties with at least one monitor with an annual design value above 11
u.g/m3 in the 2020 base case
Table 3.A-1. Annual NOX Air Quality Ratios for Monitors in California Counties that Received a
2025 Mobile NOX Emission Adjustment
Annual NOX Air
Quality Ratio
3 ,
Monitor FIPS
ID Code State Name County Name
(ug/m Change
in NO3 per 1000
tons NOX)
County Emission Group
60190008 6019 California Fresno
60195001 6019 California Fresno
60195025 6019 California Fresno
60290010 6029 California Kern
60290014 6029 California Kern
60290016 6029 California Kern
0.047 Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
0.046 Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
0.046 Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
0.043 Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
0.042 Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
0.042 Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
(continued)
3.A-5
-------
Table 3.A-1. Annual NOX Air Quality Ratios for Monitors in California Counties that Received a
2025 Mobile NOX Emission Adjustment (continued)
Monitor
ID
60310004
60370002
60371002
60371103
60371201
60371301
60371602
60372005
60374002
60374004
60379033
60472510
60590007
60592022
60651003
60652002
FIPS
Code
6031
6037
6037
6037
6037
6037
6037
6037
6037
6037
6037
6047
6059
6059
6065
6065
State Name
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
County Name
Kings
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Merced
Orange
Orange
Riverside
Riverside
Annual NOX Air
Quality Ratio
(ug/m3 Change
in NO3 per 1000
tons NOX)
0.049
0.007
0.006
0.006
0.003
0.005
0.007
0.005
0.004
0.004
0.003
0.029
0.006
0.005
0.012
0.000
County Emission Group
Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
(continued)
3.A-6
-------
Table 3.A-1. Annual NOX Air Quality Ratios for Monitors in California Counties that Received a
2025 Mobile NOX Emission Adjustment (continued)
Monitor FIPS
ID Code State Name County Name
Annual NOX Air
Quality Ratio
(ug/m3 Change
in NO3 per 1000
tons NOX)
County Emission Group
60655001 6065 California Riverside 0.000
60658001 6065 California Riverside 0.012
60658005 6065 California Riverside 0.011
60710025 6071 California San Bernardino 0.009
60710306 6071 California San Bernardino 0.004
60712002 6071 California San Bernardino 0.011
60718001 6071 California San Bernardino 0.000
60719004 6071 California San Bernardino 0.009
60771002 6077 California San Joaquin 0.018
60990005 6099 California Stanislaus 0.041
61072002 6107 California Tulare 0.055
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Los Angeles, Orange, Riverside, San
Bernardino
Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
Weighted contributions from Kern,
Kings/Tulare, Fresno/Madera, Merced,
Stanislaus, San Joaquin, Alameda, and
Sacramento
3.A-7
-------
Table 3.A-2. Annual and Daily SO2 Air Quality Ratios for Monitors in Counties where Ratios
were Used in Adjusting Daily Design Values to Remove the Impact of SO2
Controls
Annual SO2 Air
Quality Ratio
DailySO2 Air
Quality Ratio
(ug/m Change (u.g/m Change
FIPS County
Monitor ID Code State Name Name
in SO4 per
1000 tons SO2)
in SO4 per
1000 tons SO2)
County Emission Group
60990005 6099 California Stanislaus 0.123
420030008 42003 Pennsylvania Allegheny 0.026
420030064 42003 Pennsylvania Allegheny 0.028
420030067 42003 Pennsylvania Allegheny 0.017
420030095 42003 Pennsylvania Allegheny 0.021
420031008 42003 Pennsylvania Allegheny 0.019
420031301 42003 Pennsylvania Allegheny 0.027
420033007 42003 Pennsylvania Allegheny 0.021
0.468 Stanislaus, San Joaquin,
Merced
0.084 Allegheny, Armstrong,
Beaver, Butler,
Washington,
Westmoreland
0.151 Allegheny, Armstrong,
Beaver, Butler,
Washington,
Westmoreland
0.050 Allegheny, Armstrong,
Beaver, Butler,
Washington,
Westmoreland
0.066 Allegheny, Armstrong,
Beaver, Butler,
Washington,
Westmoreland
0.033 Allegheny, Armstrong,
Beaver, Butler,
Washington,
Westmoreland
0.062 Allegheny, Armstrong,
Beaver, Butler,
Washington,
Westmoreland
0.068 Allegheny, Armstrong,
Beaver, Butler,
Washington,
Westmoreland
3.A-8
-------
Table 3.A-3. Annual and Daily Direct PM2.5 Air Quality Ratios for Monitors in Counties with at
Least One Monitor having an Annual Design Value Above 11 u,g/m3 in the 2020
Base Case
Monitor ID
10730023
10731005
10731009
10731010
10732003
10732006
10735002
10735003
60190008
60195001
60195025
60250005
60250007
60251003
60290010
60290014
60290016
60310004
60370002
60371002
60371103
60371201
60371301
60371602
60372005
60374002
60374004
60379033
FIPS
Code
1073
1073
1073
1073
1073
1073
1073
1073
6019
6019
6019
6025
6025
6025
6029
6029
6029
6031
6037
6037
6037
6037
6037
6037
6037
6037
6037
6037
State Name
Alabama
Alabama
Alabama
Alabama
Alabama
Alabama
Alabama
Alabama
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
County Name
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Fresno
Fresno
Fresno
Imperial
Imperial
Imperial
Kern
Kern
Kern
Kings
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Annual PM2.5 Air Quality
Ratio (u.g/ms Change in
Direct PM2.5 per 1000 tons
PM2.5)
0.561
0.257
0.107
0.221
0.602
0.383
0.257
0.195
1.751
1.534
1.717
1.801
1.523
1.612
1.341
1.531
1.579
1.277
0.367
0.404
0.404
0.279
0.419
0.401
0.322
0.325
0.299
0.119
Daily PM2.5 Air Quality
Ratio (u.g/ms Change in
Direct PM2.5 per 1000 tons
PM2.5)
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
5.714
4.825
4.921
6.594
5.309
5.270
4.344
4.475
4.892
4.919
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
(continued)
3.A-9
-------
Table 3.A-3. Annual and Daily Direct PM2.5 Air Quality Ratios for Monitors in Counties with at
Least One Monitor having an Annual Design Value Above 11 ug/m3 in the 2020
Base Case (continued)
Monitor ID
60472510
60631006
60631009
60651003
60652002
60655001
60658001
60658005
60710025
60710306
60712002
60718001
60719004
60771002
60990005
61072002
160790017
170310022
170310050
170310052
170310057
170310076
170312001
170313301
FIPS
Code
6047
6063
6063
6065
6065
6065
6065
6065
6071
6071
6071
6071
6071
6077
6099
6107
16079
17031
17031
17031
17031
17031
17031
17031
State Name
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
California
Idaho
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
County Name
Merced
Plumas
Plumas
Riverside
Riverside
Riverside
Riverside
Riverside
San
Bernardino
San
Bernardino
San
Bernardino
San
Bernardino
San
Bernardino
San Joaquin
Stanislaus
Tulare
Shoshone
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Annual PM2.5 Air Quality
Ratio (u.g/ms Change in
Direct PM2.5 per 1000 tons
PM2.5)
4.233
2.428
2.518
1.620
0.930
0.797
2.089
2.459
0.710
0.305
0.619
0.353
0.606
1.789
2.449
1.875
7.675
0.330
0.298
0.356
0.324
0.281
0.256
0.307
Daily PM2.5 Air Quality
Ratio (u.g/ms Change in
Direct PM2.5 per 1000 tons
PM2.5)
17.925
N/A
N/A
3.223
2.463
1.885
3.627
5.039
1.423
0.439
1.180
1.674
1.710
8.486
8.955
4.222
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
(continued)
3.A-10
-------
Table 3.A-3. Annual and Daily Direct PM2.5 Air Quality Ratios for Monitors in Counties with at
Least One Monitor having an Annual Design Value Above 11 ug/m3 in the 2020
Base Case (continued)
Monitor ID
170314007
170314201
170316005
171191007
171192009
171193007
180890006
180890027
180890031
180891003
180892004
180892010
191630015
191630018
191630019
261630001
261630015
261630016
261630019
261630025
261630033
261630036
261630038
261630039
410350004
420030008
420030064
420030067
FIPS
Code
17031
17031
17031
17119
17119
17119
18089
18089
18089
18089
18089
18089
19163
19163
19163
26163
26163
26163
26163
26163
26163
26163
26163
26163
41035
42003
42003
42003
State Name
Illinois
Illinois
Illinois
Illinois
Illinois
Illinois
Indiana
Indiana
Indiana
Indiana
Indiana
Indiana
Iowa
Iowa
Iowa
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Michigan
Oregon
Pennsylvania
Pennsylvania
Pennsylvania
County Name
Cook
Cook
Cook
Madison
Madison
Madison
Lake
Lake
Lake
Lake
Lake
Lake
Scott
Scott
Scott
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Klamath
Allegheny
Allegheny
Allegheny
Annual PM2.5 Air Quality
Ratio (u.g/ms Change in
Direct PM2.5 per 1000 tons
PM2.5)
0.200
0.205
0.374
0.332
0.443
0.417
0.419
0.320
0.367
0.401
0.404
0.397
1.106
1.051
1.492
0.404
0.502
0.423
0.335
0.241
0.483
0.336
0.381
0.406
3.994
0.358
0.519
0.222
Daily PM2.5 Air Quality
Ratio (u.g/ms Change in
Direct PM2.5 per 1000 tons
PM2.5)
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
1.463
4.060
0.657
(continued)
3.A-11
-------
Table 3.A-3. Annual and Daily Direct PM2.5 Air Quality Ratios for Monitors in Counties with at
Least One Monitor having an Annual Design Value Above 11 ^g/m3 in the 2020
Base Case (continued)
Monitor ID
420030095
420031008
420031301
420033007
481410037
481410044
482010058
482011035
550790010
550790026
550790043
550790059
550790099
551330027
FIPS
Code
42003
42003
42003
42003
48141
48141
48201
48201
55079
55079
55079
55079
55079
55133
State Name
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Texas
Texas
Texas
Texas
Wisconsin
Wisconsin
Wisconsin
Wisconsin
Wisconsin
Wisconsin
County Name
Allegheny
Allegheny
Allegheny
Allegheny
El Paso
El Paso
Harris
Harris
Milwaukee
Milwaukee
Milwaukee
Milwaukee
Milwaukee
Waukesha
Annual PM2.5 Air Quality
Ratio ug/m3 Change in
Direct PM2.5 per 1000 tons
PM2.5)
0.263
0.172
0.405
0.397
1.608
2.209
0.188
0.408
1.566
1.602
1.674
1.869
1.689
3.297
Daily PM2.5 Air Quality
Ratio (u,g/ms Change in
Direct PM2.5 per 1000 tons
PM2.5)
0.931
0.645
1.409
1.752
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
3.A.1.2 Estimating Area NOX Emission Contributions to NH4NO3 PM2.s in the San Joaquin
Valley
We conducted 9 air quality model simulations for January 2020 for a domain centered
on California (Figure 3.A-2) that is a subset of the continental U.S. domain used for the 2020
base case modeling. One of the simulations reflected the 2020 base case emission scenario, and
the other 8 simulations had NOX and direct PM2.5 emissions reductions relative to the 2020 base
case in a one- or two-county group that matched the emissions reductions in that group in the
2020 NOx_PM2.5 sensitivity run. The purpose of these simulations was to estimate the impact of
NOX emission reductions in a given area of California's Central Valley on NH4N03 PM2.5 in other
areas of the valley. The month of January was selected for this analysis because high NH4N03
PM2.5 episodes occur during winter months in the Central Valley.
3.A-12
-------
106
1 79
Figure 3.A-2. California Modeling Domain for 12-km Simulations
Equation 3.A.2 was used to estimate the fractional contribution of NOX emissions from
one area of California's Central Valley on another (i.e., Wigrpjgrp).
(3.A.2)
igrpjgrp max(ACNO3jigrp/AEmissNOXjjgrp}
where AC^o3,tgrp is the change in average nitrate PM2.5 concentration at a given monitor in the
igrp county group between the simulation with 2020 base case emissions and the simulation
with NOX emissions reductions in thejgrp county group, and AEmissNOxjgrp is the change in NOX
emissions in thejgrp county group between the simulations with 2020 base case emissions and
the simulation with NOX emissions reductions in thejgrp county group. Note that Equation
3.A.2 normalizes each Aconcentration-to-Aemission ratio for a given county group (numerator)
by the maximum Aconcentration-to-Aemission ratio associated with that county group
(denominator). The fraction of NOX emissions from a given county or county group that impacts
PM2.5 nitrate in another county or county group as estimated according to Equation 3.A.2 is
given in Table 3.A-4.
3.A-13
-------
Table 3.A-4. Contribution Weighting Factors for NOX Emissions in Counties or County Groups
in California's Central Valley as Calculated According to Equation 3.A.2
County or
Cou nty
Group
Kern
Kings/Tulare
Fresno/
Madera
Merced
Stanislaus
San Joaquin
Sacramento
Alameda
Weight of
Kern
Emissions
1
0.11
0.06
0.02
0.01
0.02
0.01
0.04
Weight of
Kings/
Tula re
Emissions
0.94
1
0.41
0.07
0.03
0.04
0.03
0.14
Weight of
Fresno/
Madera
Emissions
0.6
0.89
1
0.23
0.05
0.08
0.04
0.13
Weight of
Merced
Emissions
0.47
0.56
0.65
1
0.37
0.39
0.12
0.24
Weight of
Stanislaus
Emissions
0.4
0.4
0.48
0.92
1
0.7
0.22
0.59
Weight of
San
Joaquin
Emissions
0.35
0.31
0.38
0.51
0.56
0.66
0.25
1
Weight of
Sacramento
Emissions
0.29
0.24
0.35
0.4
0.39
1
1
0.66
Weight of
Alameda
Emissions
0.07
0.05
0.04
0.04
0.06
0.09
0.01
0.03
3.A-14
-------
3-37
-------
CHAPTER 4
CONTROL STRATEGIES
4.1 Synopsis
As discussed in previous chapters, the EPA is revising the PM2.5annual standard to a level
of 12 u.g/m3and is retaining the 24-hour standard of 35 u.g/m3. Pursuant to Executive Order
12866 and 13563 as well as the guidelines of OMB Circular A-4.1, the EPA assessed the
incremental costs of hypothetical control strategies to attain the revised standard. EPA also
estimated the incremental costs of attaining a less stringent alternative annual standard of 13
u.g/m3 and a more stringent alternative annual standard of 11 u.g/m3. This chapter documents
the emission control measures we applied to simulate attainment with the revised and
alternative PM2.5 annual standards.
The EPA analyzed the impact that additional emissions controls across numerous sectors
would have on predicted ambient PM2.5 concentrations incremental to an analytical baseline,
which includes the current PM2.5 standard of 15/35 u.g/m3 as well as other major rules such as
MATS. Thus, the analysis for the revised and alternative standards focuses specifically on
incremental improvements beyond the current standard and other existing major rules, and
uses control options that might be available to states for application by 2020. The hypothetical
control strategies presented in this RIA represent illustrative options for emissions reductions
that achieve national attainment of the revised standard as well as the alternative standards.
The hypothetical control strategies are not recommendations for how the revised PM2.5
standard should be implemented, and states will make all final decisions regarding
implementation strategies for the revised NAAQS.
The traditional analytical approach to a NAAQS RIA is to perform air quality modeling for:
• base case projections, then
• baseline attainment strategy for the current standard, then
• revised and alternative control strategies incremental to the baseline strategy
Each subsequent model run would build on what was learned from the previous run. Because
of the short timeframe for this analysis, we were limited to performing air quality modeling for
the base case in parallel with air quality modeling for a single control scenario, along with
several limited sensitivity runs (see Chapter 3 for a detailed description of the air quality
modeling runs performed for this analysis). The following steps were taken by the EPA to
4-1
-------
analyze the impacts and costs of the control scenario incremental to attainment of the current
standard of 15/35 u.g/m3 and beyond other existing major rules:
1. Identify geographic areas in the U.S. likely to exceed the revised or alternative
standards in the year 2020 using the base case projections.
2. Develop a hypothetical control scenario for these areas and generate a control case
2020 emissions inventory. These control measures will serve as the basis for the
"known" controls in this analysis.
3. Perform air quality modeling to assess the air quality impacts of the hypothetical
control scenario (as mentioned, the analyses were performed in parallel with base
case air quality modeling).
4. Adjust results to remove controls deemed as inappropriate for application to a
specific source (e.g., controls likely to already be in place or controls estimated to
have a very high cost but little impact on emissions).
5. Calculate the portion of the hypothetical control scenario control measures that are
attributed to meeting the current standard of 15/35 u.g/m3. These are the known
"analytical baseline" reductions. Estimate any additional emission reductions
beyond the known controls that are needed to meet the current standard (15/35
u.g/m3). Costs of known controls incremental to (i.e., over and above) the analytical
baseline reductions are attributed to the costs of meeting the revised and
alternative standards.
6. Estimate the additional emission reductions incremental to the analytical baseline
and beyond the known controls that are needed to meet the revised and/or
alternative standards. These are referred to as emission reductions needed beyond
known controls (i.e., extrapolated tons).
7. Calculate the total costs of reductions from emission reductions from known
controls and emission reductions needed beyond known controls (extrapolated
costs) incremental to the analytical baseline.
This chapter discusses the steps listed above that were key to conducting the control
strategy analysis for year 2020 for the revised and alternative standards.
4.2 PM2.5 Control Strategy Analysis
4.2.1 Identify Geographic Areas
The first step in the analysis was to identify the geographic areas likely to exceed the
revised and/or alternative standards in 2020 for the base case. For a detailed description of the
process used to identify these geographic areas, see Chapter 3. For the revised standard, there
4-2
-------
were seven counties projected to exceed the standard, all in California. For the hypothetical
control scenario for the revised standard we also identified and applied control measures in
one county adjacent to one of the projected exceedance counties in order to reduce transport
emissions. For a description of the areas included in the final analysis of the revised and
alternative standards, see Figures 4-1 through 4-4.
Legend
569 counties with monitors have PM 2.5 design values of which:
^^| 13 counties are projected to exceed 15/35
^J 5 adjacent counties are controlled but do not exceed 15/35
f 552 counties with monitors do not exceed 15/35
Out of the 5 adjacent counties one has no monitor
Figure 4-1. Counties Projected to Exceed the Current PM2.5 Standard (15/35) in 2020
4-3
-------
Legend
569 counties with monitors have PM 2-5 design values of which
| 2 counties are projected to exceed 13
^ 1 adjacent county is controlled but does not exceed 13
^\ 566 counties with monitors are projected to be below 13
Figure 4-2. Counties Projected to Exceed the 13 ug/m3 Alternative Standard in the
Analytical Baseline
4-4
-------
Legend
569 counties with monitors nave PM 25 design values of which
| 7 counties are projected to exceed 12
^\ 1 adjacent county is controlled outdoes not exceed 12
^ 561 counties with monitors are projected to be below 12
0 220 440
Figure 4-3. Counties Projected to Exceed the 12 ug/m3 Revised Standard in the Analytical
Baseline
4-5
-------
Legend
569 counties with monitors have PM 2.5 design values of which
| 23 counties are projected to exceed 11
"^ 1 adjacent county is controlled but does not exceed 11
| 545 counties with monitors are projected to be below 11
Figure 4-4. Counties Projected to Exceed the 11 ug/m3 Alternative Standard in the
Analytical Baseline
4.2.2 Developing the Control Scenario
The U.S. EPA used monitoring and emissions inventory information, including
monitoring speciation, to identify the pollutants that were the primary contributors to PM2.5
exceedances at the subject monitors. This allowed us to select a set of control measures
tailored to the conditions for each area.
Non-EGU point, nonpoint (area), and onroad mobile control measures were applied for
the control strategy for demonstrating attainment of the current standard (15/35 u.g/m3). Non-
EGU point and nonpoint control measures were applied for the revised and alternative
standards control strategies. These controls were identified using the U.S. EPA's Control
4-6
-------
Strategy Tool1 (CoST). These controls are summarized in Appendix 4.A. Additional control
measures were not applied to electric generating units (EGUs) due to the extensive nature of
controls resulting from the inclusion of MATS.
Nonpoint and onroad mobile source emissions data are generated at the county level,
and therefore controls for these emissions sectors were applied at the county level. Non-EGU
point source controls are applied to individual point sources. Nonpoint source controls were
applied to NOX, S02, and PM2.5. The analysis for non-EGUs applied NOX, S02, and PM2.5 controls
to the following source categories: industrial boilers, commercial and institutional boilers,
sulfuric acid plants (both stand alone and at other facilities such as copper and lead smelters),
primary metal plants (iron and steel mills, lead smelters), mineral products (primarily cement
kilns), and petroleum refineries. Among the control measures applied were: wet FGD scrubbers
and spray dryer absorbers (SDA) for S02 reductions, fabric filters for PM2.5 reductions, and SCR
and low NOX burners for NOX. Table 4-1 lists the major controls applied to each sector.
Table 4-1. Controls Applied in the Revised and Alternative Standard Analysis
Sector/Pollutant
Non-EGU Point
PM2.5
S02
Control Measure
Diesel Particulate Filter
Dry Electrostatic Precipitator (ESP)
Fabric Filters
Venturi Scrubber
Wet Electrostatic Precipitator (ESP)
Coal Washing
Flue Gas Desulfurization (FGD)
15/35 13 12 11
X XX
X
X XX
X
X
X
NOX
Spray Dry Absorber
Sulfur Recovery and/or Tail Gas Treatment
Wet FGDs X
Biosolid Injection X
Low NOX Burner (LNB) X
Low NOX Burner (LNB) + Selective Catalytic Reduction (SCR) X
Non-Selective Catalytic Reduction (NSCR) X
OXY-Firing X
1 See http://www.epa.gov/ttn/ecas/cost.htm for a description of CoST.
4-7
-------
Sector/Pollutant
Non-Point (Area)
PM2.5
S02
NOX
Onroad Mobile
NOX
Control Measure
SCR + Steam Injection
Selective Catalytic Reduction (SCR)
Selective Non-Catalytic Reduction (SNCR)
ESPs for Commercial Cooking
Low NOX Burners for Residential Natural Gas
Substitute chipping for open burning
Substitute landfilling for open burning
Chemical Additives to Waste
Fuel Switching for Stationary Source Fuel Combustion
Low Sulfur Home Heating Fuel
Low NOX Burners for Residential Natural Gas
Water heater + Low NOX Burner Space Heaters
Elimination of Long Duration Truck Idling (diesel trucks)
Continuous Inspection and Maintenance (gasoline cars)
15/35 13
X
X
X
X X
X
X
X
X
X
X
X
X
X
12 11
X X
X X
X
X
X
*
3 |
Note that control measures indicated in the table for 13,12, and 11 ng/m levels are incremental to control
3
measures indicated for the 15/35 ng/m analysis.
To more accurately depict available controls, the EPA employed a decision rule in which
controls were not applied to any non-EGU or area sources with 50 tons/year of emissions or
less. Furthermore, controls were not applied to sources unless at least 5 tons/year of emission
reductions were achieved. This decision rule is the same rule we employed for sources in the
previous PM2.5 NAAQS RIA completed in 2006. The reason for applying this decision rule is
based on a finding that most point sources with emissions of this level or less had controls
already in place. This decision rule helps fill gaps in information regarding existing controls on
non-EGU sources. An additional constraint was applied in the control strategy selection that
concerned cost of control. No control measures were applied that cost greater than
$20,000/ton of emission reduction. We did not include known controls at an annual cost of
more than $20,000 per ton because either (i) the remaining emissions sources were relatively
small sources, or we believe they are already controlled, or (ii) the equations in CoST were not
applicable to these remaining emissions sources. Note that there were potential controls
available at an annual cost of more than $20,000 per ton for ten of the geographic areas
included in the analysis. The application of these control strategies results in some, but not all,
geographic areas reaching attainment for the alternative PM2.5 standards.
4-8
-------
As stated above, because of the tight time constraints for this analysis, we performed air
quality modeling for a single control scenario and then used the results of sensitivity analyses to
identify the subset of controls and associated emission reductions in the control scenario that
were needed to meet the current baseline. We then used a similar approach to determine the
subset of additional controls and emissions reductions incremental to those applied in the
analytical baseline that were needed to meet the revised and alternative standards.
4.2.3 Identify Known Controls Needed to Meet the Analytical Baseline
The analytical baseline includes reductions already achieved as a result of national
regulations, reductions expected prior to 2020 from recently promulgated national regulations,
adjustments for expected emission reductions in two areas (South Coast and San Joaquin
Valley, CA) not expected to reach attainment until 2025, and reductions from additional
controls which the U.S. EPA estimates need to be included to attain the current standard
(15/35). Reductions achieved as a result of state and local agency regulations and voluntary
programs are reflected to the extent that they are represented in emission inventory
information submitted to the U.S. EPA by state and local agencies. Two steps were used to
develop the analytical baseline. First, the reductions expected in national PM2.5 concentrations
from national rules promulgated prior to this analysis were considered (referred to as the base
case). Below is a list of some of the major national rules reflected in the base case. Refer to
Chapter 3, Section 3.2.1.4 for a more detailed discussion of the rules reflected in the 2020 base
case emissions inventory.
• Light-Duty Vehicle Tier 2 Rule (U.S. EPA, 1999)
• Heavy Duty Diesel Rule (U.S. EPA, 2000)
• Clean Air Nonroad Diesel Rule (U.S. EPA, 2004)
• 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)
• Emissions Standards for Locomotives and Marine Compression-Ignition Engines (U.S.
EPA, 2008)
• Control of Emissions for Nonroad Spark Ignition Engines and Equipment (U.S. EPA,
2008)
• C3 Oceangoing Vessels (U.S. EPA, 2010)
4-9
-------
• Hospital/Medical/lnfectious Waste Incinerators: New Source Performance Standards
and Emission Guidelines: Final Rule Amendments (U.S. EPA, 2009)
• Reciprocating Internal Combustion Engines (RICE) NESHAPs (U.S. EPA, 2010)
• Mercury and Air Toxics Standards (U.S. EPA, 2011)
• Cross-State Air Pollution Rule (U.S. EPA, 2011)2
Most areas of the U.S. will be required to demonstrate attainment with the new
standards by 2020. As a result, for these areas, the correct baseline for estimating the
incremental emissions reductions that would be needed to attain the more protective
standards is a baseline with emissions projected to 2020 and adjusted to reflect the additional
emissions reductions that would be needed to attain the current 15/35 u.g/m3 standard. For
two areas in Southern California (South Coast and San Joaquin), the degree of projected non-
attainment with the revised annual standard of 12 u.g/m3 is high enough that those counties are
not expected to have to demonstrate attainment with the new standards by 2020. Instead,
those two areas will likely have until 2025 to demonstrate attainment with the revised annual
standard of 12 u.g/m3. As a result, for these two areas, the correct baseline for estimating the
incremental emissions reductions that would be needed to attain the more protective
standards is a baseline with emissions projected to 2025 adjusted to reflect the additional
emissions reductions that would be needed to attain the current 15/35 u.g/m3 standard. The
following paragraphs describe the steps we followed for this analysis.
This difference in attainment year is important because between 2020 and 2025,
emissions from mobile sources in California are expected to be reduced due to continued fleet
turn over from older, higher emitting vehicles to newer, lower emitting vehicles. These
reductions in emissions will occur as a result of previous EPA and California rules for which costs
(and benefits) have already been counted and thus will not be attributed to meeting the revised
or alternative standards3.
Modeling of two separate years is time prohibitive and would result in two separate
years of benefits and costs which would not provide a complete picture of the nationwide costs
and benefits of just meeting the new standards in either 2020 or 2025 because of differences in
the baselines between the two years. To provide the most reasonable and reliable estimates of
2 See Chapter 3, Section 3.2.1.4 for a discussion of the role CSAPR plays in the PM25RIA.
3 See Chapter 3, Section 3.2.1.4 for a listing of the EPA and California rules reflected in the 15/35 analysis.
4-10
-------
costs and benefits of full attainment for the nation, we constructed an analytical 2020 baseline
for estimating the costs and benefits of attaining the selected annual standard of 12 u.g/m3 and
alternative standards of 13 u.g/m3 and 11 u.g/m3 with the following characteristics: (1) reflects
"on the books" regulations as implemented through 2020 for all nonattainment counties, and
(2) mobile source emissions reductions expected to occur between 2020 and 2025 for
California's South Coast and San Joaquin areas, which are likely to not demonstrate attainment
until 2025. Essentially, we are adjusting emissions for the two areas in California to reflect
future emissions reductions that they will achieve prior to their 2025 attainment date. This
allows us to generate costs and benefits of full attainment without overstating the costs and
benefits in those two areas which would occur if forced to apply costly emissions reductions in
2020 in areas that would not have to be incurred until 2025.
The 2020 analytical baseline for this analysis presents one scenario of future year air
quality based upon specific control measures, promulgated federal rules such as MATS, years of
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 employ other strategies and/or other federal rules may be adopted that would also
help in achieving attainment. The number of counties that will be part of the designations
process may be different than the number of counties projected to exceed as part of this
analysis.
A map of the country is presented in Figure 4-1 which shows the counties projected to
exceed the current PM2.5 standard of 15/35 u.g/m3 in the year 2020. Control measures were
identified in the control scenario run that were needed for these counties in the analytical
baseline analysis to meet the current PM2.5 standard. In addition, control measures were
applied to five California counties adjacent to exceedance counties in order to address
transport coming from these adjacent counties.
The additional known controls included in the analytical baseline to simulate attainment
with the current PM2.5 NAAQS are listed in Table 4-1; details regarding the individual controls
are provided in Appendix 4.A. Controls were applied to directly emitted PM2.5 and the PM2.5
precursors of NOX and S02 given that nitrate, sulfate, and primary PM2.5 species usually
dominate measured PM2.5 based on speciation data measured at the Chemical Speciation
Network (CSN) sites. Control measures that directly reduced emissions of PM2.5 in proximity to
the exceeding monitors were determined to be most effective at bringing areas into
attainment, with NOX and S02 controls supplementing the PM2.5 controls depending upon the
monitor speciation data. PM2.5 control measures were applied in the county containing the
4-11
-------
exceeding monitor for the non-EGU point and area source emissions. If additional emissions
control was needed, S02 and NOX control measures were applied within the county exceeding
and in a small number of cases, an adjacent county or counties.
For the analytical baseline, there were several geographic areas that did not reach
attainment with known controls. For these geographic areas, we estimated the additional
emission reductions needed beyond identified known controls for PM2.5 to attain the standard.
4.2.4 Identify Known Controls Needed to Meet the Revised and Alternative Standards
After identifying the known controls in the control scenario that were needed to meet
the analytical baseline, additional known controls needed to meet the revised and alternative
standards were identified. The EPA used air quality modeling results to determine whether the
control scenario was sufficient to meet the revised and alternative standards for each
geographic area. Where the control scenario modeling resulted in design value reductions
below the level needed for the revised or alternative standards for specific geographic areas,
county-specific ratios of air quality response to emission reductions were used to determine the
subset of controls that were needed to attain the standard. Where it was determined that the
control scenario was not sufficient in attaining the standard, these same response factors were
used to calculate the amount of additional emission reductions beyond known controls needed
to meet the standard. For the revised and alternative control strategy analysis, known controls
for two sectors were used: non-EGU point and area sources. Onroad mobile source controls
were not used in the revised and alternative standards analysis because they were applied
previously in the analytical baseline analysis where they were deemed to be most cost
effective.
In the revised and alternative standards analysis, PM2.5 controls were sought first
because they were generally more cost-effective. If it was determined that additional controls
were needed, S02 and NOX control measures were selected depending on the chemistry of each
specific geographic area.
It should be noted that while PM2.5 controls were applied only within the counties with
monitors projected to exceed the alternative standard being analyzed, S02 and NOX controls
were applied in the exceeding county as well as a small number of adjacent counties because of
the transport of NOX and S02 across counties. Table 4-2 shows the number of exceeding
counties and the number of adjacent counties to which controls were applied for the revised
and alternative standards. Table 4-3 shows the emission reductions from known controls for
the revised and alternative standards analyzed.
4-12
-------
Table 4-2. Number of Counties with Exceedances and Number of Additional Counties
Where Reductions Were Applied
Revised/Alternative
Standard
13
12
11
Number of Counties with
Exceedances
2
7
23
Number of Additional Counties
Where Reductions Were Applied
1
1
1
Table 4-3. Emission Reductions from Known Controls for the Revised and Alternative
Standards3
Emission Reductions in 2020 (annual tons/year)
Revised/
Alternative
Standard
13
12
11
East
West
CA
Total
East
West
CA
Total
East
West
CA
Total
Region PM2.5 SO2 NOX
— — —
— — —
53 - -
53 — —
— — —
— — —
803 - -
803 — —
3,400 21,000 9
75 43 -
930 - -
4,400 21,000 9
a Estimates are rounded to two significant figures.
4.2.5 Identify Emission Reductions Beyond Known Controls Needed to Meet the Revised
and Alternative Standards
There were several areas where known controls did not achieve enough emission
reductions to attain the revised or alternative standards in 2020. To complete the analysis, the
EPA then estimated the additional emission reductions beyond known controls needed to reach
attainment. For information on the methodology used to develop those estimates, see Chapter
4-13
-------
3, Section 3.3.2. Table 4-4 shows the emission reductions needed from unknown controls for
the alternative standards analyzed.
Table 4-4. Emission Reductions Needed Beyond Known Controls for the Revised and
Alternative Standards3
Emission Reductions in 2020 (annual tons/year)
Revised/
Alternative
Standard
13
12
11
East
West
CA
Total
East
West
CA
Total
East
West
CA
Total
Region PM2.5 SO2 NOX
— — —
— — —
674 - -
674 — —
— — —
— — —
3,190 - -
3,190 — —
4,800 - -
86 - -
9,700 - -
15,000 — —
a Estimates are rounded to two significant figures.
4.3 Limitations and Uncertainties
The 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. We outline, and
qualitatively assess the impact of, those limitations and uncertainties that are most significant.
A number of limitations and uncertainties are associated with the analysis of emission
control measures are listed in Table 4-5.
4-14
-------
Table 4-5. Summary of Qualitative Uncertainty for Elements of Control Strategies
Potential Source of Uncertainty
Magnitude Degree of
of Impact Confidence
Direction of on in Our Ability to
Potential Monetized Analytical Assess
Bias Costs3 Approach Uncertainty0
Uncertainties Associated with PM Concentration Changes
Projections of future levels of emissions and emissions Both
reductions necessary to achieve the NAAQS
Responsiveness of air quality model to changes in Both
precursor emissions from control scenarios
Air quality model chemistry, particularly for formation Both
of ambient nitrate concentrations
Post-processing of air quality modeled concentrations Both
to estimate future-year PM2.5design value and spatial
fields of PM2.5 concentrations
Medium Medium
Medium-
high
Medium
High
Medium
High
High
Tierl
Tierl
Tierl
Tierl
Uncertainties Associated with Control Strategy Development
Control Technology Data
• Technologies applied may not reflect most current
emerging devices that may be available in future
years
• Control efficiency data is dependent upon
equipment being well maintained.
• Area source controls assume a constant estimate of
emission reductions, despite variability in extent
and scale of application.
Both
Medium-
high
High
Tier 2
Control Strategy Development
• States may develop different control strategies than
the ones illustrated
• Lack of data on analytical baseline controls from
current SIPs
• Timing of control strategies may be different than
envisioned in RIA
• Controls are applied within the county with the
exceeding monitor. In some cases, additional
known controls are also applied in adjacent
contributing counties.
• Emissions growth and control from new sources
locating in these analysis areas is not included.
Both Medium- Medium-
high high
TierO
4-15
-------
Potential Source of Uncertainty
Direction of
Potential
Bias
Magnitude
of Impact
on
Monetized
Costs3
Degree of
Confidence
in Our
Analytical
Approach13
Ability to
Assess
Uncertaintyc
Technological Change
• Emission reductions do not reflect potential effects
of technological change that may be available in
future years
• Effects of "learning by doing" are no accounted for
in the emission reduction estimates
Likely over-
estimate
Medium-
high
Low
TierO
Emission Reductions from Unidentified Controls
• emission control cut points for each pollutant
Both
High
Low
Tierl
Magnitude of Impact
High—If error could influence the total costs by more than 25%
Medium—If error could influence the total costs by 5%-25%
Low—If error could influence the total costs by less than 5%
Degree of Confidence in Our Analytic Approach
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
Low—Limited data exists to support the selected approach
Ability to Assess Uncertainty (using WHO Uncertainty Framework)
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
Future expected emissions are difficult to predict because they depend on many independent factors. Emission
inventories are aggregated from many spatially and technically diverse sources of emissions, so simplifying
assumptions are necessary to make estimating the future tractable.
4.4 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.
4-16
-------
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/sea rch/Regs/contentStreamer?obiectld=09000064800be20
3&disposition=attachment&contentType=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.
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/Julv/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). 2007. Guidance on the Use of Models and
other Analyses for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.s, 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). 2008. Control of Emissions for Nonroad Spark
Ignition Engines and Equipment. Office of Transportation and Air Quality. Available at
http://www.epa.gov/otaq/regs/nonroad/marinesi-equipld/bondfrm.pdf.
U.S. Environmental Protection Agency (U.S. EPA).2008. Emissions Standards for Locomotives
and Marine Compression-Ignition Engines. Office of Transportation and Air Quality.
Available at http://www.epa.gov/nonroad/420f08004.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/lnfectious 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). 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/sea rch/Regs/contentStreamer?obiectld=0900006480ae43a
6&disposition=attachment&contentType=pdf.
4-17
-------
U.S. Environmental Protection Agency (U.S. EPA). 2011. 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). 2011. Cross-State Air Pollution Rule. Office of
Atmospheric Programs, Washington, D.C. Available at
http://www.gpo.gov/fdsvs/pkg/FR-2011-08-08/pdf/2011-17600.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011. Mercury and Air Toxics Standards.
Office of Atmospheric Programs, Washington, D.C. Available at
www.gpo.gov/fdsys/pkg/FR-2012-02-16/pdf/2012-806.pdf.
4-18
-------
APPENDIX 4.A
ADDITIONAL CONTROL STRATEGY INFORMATION
4.A.1 Control Measures for Stationary Sources
This appendix describes measures that were employed in this analysis to illustrate a
hypothetical scenario for controlling emissions of PM and precursors from non-EGU point and
nonpoint (area) source categories to attain the baseline or to attain the revised or alternative
standards for PM2.5. Most of the control measures available are add-on technologies but some
other technologies and practices that are not add-on in nature can reduce emissions of PM and
PM precursors.
4.A.1.1 PM Emissions Control Technologies1
This section summarizes control measures focused on reduction of PM2.s from non-EGU
point and nonpoint sources. However, it should be noted that PMi0 will also be reduced by
these measures. The amount of PMi0 reduction varies by the fraction of PMi0 in the inlet
stream to the control measure and the specific design of the measure.
4.A.1.1.1 PM Control Measures for Non-EGU Point Sources
Most control measures for non-EGU point sources are add-on technologies. These
technologies include: fabric filters (baghouses), ESPs, and wet PM scrubbers. Fabric filters
collect particles with sizes ranging from below 1 micrometer to several hundred micrometers in
diameter at efficiencies in excess of 99%, and this device is used where high-efficiency particle
collection is required. A fabric filter unit consists of one or more isolated compartments
containing rows of fabric bags in the form of round, flat, or shaped tubes, or pleated cartridges.
Particle-laden gas passes up (usually) along the surface of the bags then radially through the
fabric. Particles are retained on the upstream face of the bags, and the cleaned gas stream is
vented to the atmosphere. The filter is operated cyclically, alternating between relatively long
periods of filtering and short periods of cleaning. Dust that accumulates on the bags is removed
from the fabric surface when cleaning and deposited in a hopper for subsequent disposal.
ESPs use electrical forces to move particles out of a flowing gas stream and onto
collector plates. The particles are given an electrical charge by forcing them to pass through a
corona, a region in which gaseous ions flow. The electrical field that forces the charged particles
to the plates comes from electrodes maintained at high voltage in the center of the flow lane.
1 The descriptions of add-on technologies throughout this section are taken from the EPA Air Pollution Control Cost
Manual, Sixth Edition. This is found on the Internet at http://epa.gov/ttn/catc/products.htmlffcccinfo.
4.A-1
-------
Once particles are on the collector plates, they must be removed without re-entraining them
into the gas stream. This is usually accomplished by rapping the plates mechanically which
loosens the collected particles from the collector plates, allowing the particles to slide down
into a hopper from which they are evacuated. This removal of collected particles is typical of a
"dry" ESP. A "wet" ESP operates by having a water flow applied intermittently or continuously
to wash away the collected particles for disposal. The advantage of wet ESPs is that there are
no problems with rapping re-entrainment or with "back coronas" (unintended injection of
positively charged ions which reduces the charge on particles and lowers the collection
efficiency). The disadvantage is that the collected slurry must be handled more carefully than a
dry product, adding to the expense of disposal. ESPs capture particles with sizes ranging from
below 1 micrometer to several hundred micrometers in diameter at efficiencies from 95 to up
to 99% and higher.
Wet PM scrubbers remove PM and acid gases from waste gas streams of stationary
point sources. The pollutants are removed primarily through the impaction, diffusion,
interception and/or absorption of the pollutant onto droplets of liquid. The liquid containing
the pollutant is then collected for disposal. Collection efficiencies for wet scrubbers vary by
scrubber type, and with the PM size distribution of the waste gas stream. In general, collection
efficiency decreases as the PM size decreases. Collection efficiencies range from in excess of
99% for venturi scrubbers to 40% to 60% for simple spray towers. Wet scrubbers are generally
smaller and more compact than fabric filters or ESPs, and have lower capital cost and
comparable operation and maintenance (O&M) costs. Wet scrubbers, however, operate with a
higher pressure drop than either fabric filters or ESPs, thus leading to higher energy costs. In
addition, they are limited to lower waste gas flow rates and operating temperatures than fabric
filters or ESPs, and also generate sludge that requires additional treatment or disposal. This RIA
only applies wet scrubbers to fluid catalytic cracking units (FCCUs) at petroleum refineries.
In addition, we also examined additional add-on control measures specifically for steel
mills. Virtually all steel mills have some type of PM control measure, but there is additional
equipment that in many cases could be installed to further reduce emissions. Capture hoods
that route PM emissions from a blast furnace casthouse to a fabric filter can provide 80% to
90% additional emission reductions from a steel mill. Other capture and control systems at
blast oxygen furnaces (BOFs) can also provide 80% to 90% additional reductions.
4.A-2
-------
Table 4.A-1 lists some of these technologies. For more information on these
technologies, refer to the EPA Air Pollution Control Cost Manual.2
Table 4.A-1. Example PM Control Measures for Non-EGU Point Source Categories
Control Measure
Fabric Filters3
ESPs— wet or dry3
Wet Scrubbers
Secondary Capture and Control
Systems— Capture Hoods for Blast
Oxygen Furnaces
CEM Upgrade and Increased
Monitoring Frequency
Sector(s) to which Control
Measure Can Apply
Industrial Boilers, Iron and Steel
Mills, Pulp and Paper Mills
Industrial Boilers, Iron and Steel
Mills, Pulp and Paper Mills
Industrial Boilers, Iron and Steel
Mills
Coke Ovens
Non-EGUswithan ESP
Control
Efficiency
(percent)
98 to 99.9
95 to 99.9
40 to 99
80 to 90
5 to 7
Average
Annualized
Cost/Ton
$2,000-$ 100,000
$1,000-$20,000
$750-$2,800
$5,000
$600-$5,000
3 CoST contains equations to estimate capital and annualized costs for ESP and FF installation and operation. The
average annualized cost/ton estimates presented here for these control measures are outputs from our
modeling, not inputs. They also reflect applications of control where there is no PM control measure currently
operating except if the control measure is an upgrade (e.g., ESP upgrades).
4.A.1.1.2PM Control Measures for Nonpoint Sources
Specific controls exist for a number of stationary nonpoint sources. Nonpoint source PM
controls at stationary sources include:
• catalytic oxidizers on conveyorized charbroilers at restaurants (up to 80% reduction
of PM), and
• diesel particulate filters, applied to existing diesel-fueled compression-ignition (C-l)
engines (up to a 90reduction in fine PM).
Diesel particulate filters are being applied to new C-l engines as part of a NSPS that was
implemented beginning in 2006.
Please refer to Footnote 1.
4.A-3
-------
Table 4.A-2. Example PM Control Measures for Nonpoint Sources3
Control Measures
Catalytic oxidizers for conveyorized charbroilers
Replace open burning of wood waste with
chipping for landfill disposal
Sectors to which
These Control
Measures Can Apply
Restaurants
Residential waste
sources
Control
Efficiency
(percent)
83
Near 100
Average
Annualized
Cost/ton
$1,300
$3,500
a The estimates for these control measures reflect applications of control where there is no PM nonpoint source
control measure currently operating. Also, the control efficiency is for total PM, and thus accounts for PM10 and
PM2.s. Data for these measures is available in the CoST Control Measures Documentation Report at
http://www.epa.gov/ttn/ecas/models/CoST CMDB Document 2010-06-09.pdf.
4.A.1.2 SO2 Control Measures
4. A. 1.2. ISO 2 Control Measures for Non-EGU Point Sources
The S02 emission control measures used in this analysis are similar to those used in the
PM2.5 RIA prepared about four years ago. Flue gas desulfurization (FGD) scrubbers can achieve
95-98% control of S02for Non-EGU point sources and for utility boilers. Spray dryer absorbers
(SDA) are another commonly employed technology, and SDA can achieve up to 90% or more
control of S02. For specific source categories, other types of control technologies are available
that are more specific to the sources controlled. Table 4.A-3 lists some of these technologies.
For more information on these technologies, please refer to the CoST control measures
documentation report.3
3 For a complete description of the control technologies used in CoST, please refer to the report at
http://www.epa.gov/ttn/ecas/models/CoST CMDB Document 2010-06-09.pdf.
4.A-4
-------
Table 4.A-3. Example SO2 Control Measures for Non-EGU Point3
Control Measure
Wet and Dry FGD
scrubbers and SDA
Sectors to Which These Control
Measures Can Be Applied
ICI boilers— all fuel types, kraft pulp
mills, Mineral Products (e.g., Portland
cement plants (all fuel types), primary
metal plants, petroleum refineries
Control
Efficiency
(percent)
95-FGD
scrubbers,
90-forSDA
Average Annualized
Cost/Ton (2006$)
$800-$8,000-FGD
$900-$7,000-SDA
Increase percentage Sulfur recovery plants
sulfur conversion to meet
sulfuric acid NSPS (99.7%
reduction)
Sulfur recovery and/or tail Sulfuric Acid Plants
gas treatment
Cesium promoted catalyst Sulfuric Acid Plants with Double-
Absorption process
75-95
95-98
50%
$4,000
$1,000-$4,000
$1,000
Sources: CoST control measures documentation report, May 2008, NESCAUM Report on Applicability of NOX,
SO2, and PM Control Measures to Industrial Boilers, November 2008 available at
http://www.nescaum.org/documents/ici-boilers-20081118-final.pdf, and Comprehensive Industry Document on
Sulphuric Acid Plant, Govt. of India Central Pollution Control Board, May 2007. The estimates for these control
measures reflect applications of control where there is no SO2 control measure currently operating except for
the Cesium promoted catalyst.
4.A.1.2.2S02 Control Technology for Nonpoint Sources
Fuel switching from high to low-sulfur fuels is the predominant control measure
available for S02 nonpoint sources. For home heating oil users, our analyses included switching
from a high-sulfur oil (approximately 2,500 parts per million (ppm) sulfur content) to a low-
sulfur oil (approximately 500 ppm sulfur). A similar control measure is available for oil-fired
industrial boilers. For more information on these measures, please refer to the CoST control
measures documentation report.4
4.A.1.3 NOX Emissions Control Measures
4.A. 1.3.1 A/0X Control Measures for Non-EGU Point Sources
This section describes available measures for controlling emissions of NOX from non-EGU
point sources. In general, Iow-N0x burners (LNB) are often applied as a control technology for
industrial boilers and for some other non-EGU sources because of their wide applicability and
cost-effectiveness. While all controls presented in this analysis are considered generally
Please refer to Footnote 3.
4.A-5
-------
technically feasible for each class of sources, source-specific cases may exist where a control
technology is in fact not technically feasible.
Several types of NOX control technologies exist for non-EGU sources: selective catalytic
reduction (SCR), selective noncatalytic reduction (SNCR), natural gas reburn (NCR), coal reburn,
and Iow-N0x burners. The two control measures chosen most often were LNB and SCR because
of their breadth of application. 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 are not
feasible 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 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 (Wl). NOX control measures available for cement kilns include those available to
industrial boilers, namely LNB, SCR, and SNCR. In addition, mid-kiln firing (MKF), ammonia-
based SNCR, and biosolids injection can be used on cement kilns where appropriate. 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 emissions at such plants. LNB, SCR, and SCR combined with steam injection (SI) are
available measures for combustion turbines. Finally, SNCR is an available control technology at
incinerators. Table 4.A-4 lists typical examples of the control measures available for these
categories. For more information on these measures, please refer to the CoST control measures
documentation report.5
5 Please refer to Footnote 3.
4.A-6
-------
Table 4.A-4. Example NOX Control Measures for Non-EGU Source Categories3
Control Measures
Sectors to Which These Control Measures Control Efficiency Average Annualized
Apply (percent) Cost/ton
LNB
LNB+FGR
Industrial boilers—all fuel types, Petroleum
refineries, Cement manufacturing, Pulp
and Paper mills
Petroleum refineries
SNCR (urea-based or Industrial boilers—all fuel types, Petroleum
not)
SCR
refineries, Cement manufacturing, pulp
and paper mills, incinerators
Industrial boilers—all fuel types, Petroleum
refineries, Cement manufacturing, pulp
and paper mills, Combustion turbines
25 to 50%
55
45 to 75
80 to 90
$200 to $1,000
$4,000
$1,000 to $2,000
$2,000 to 7,000
OXY-Firing
NSCR
MKF
Biosolids Injection
SCR + SI
Glass manufacturing
Stationary internal combustion engines
Cement manufacturing— dry
Cement manufacturing— dry
Industrial boilers— all fuel types
85
90
25
23
95
$2,500 to 6,000
500
-$460 to 720
$300
$2,700
Source: CoST control measures documentation report (June 2010). Note: a negative sign indicates a cost savings
from application of a control measure. The estimates for these control measures reflect applications of control
where there is no NOX control measure currently operating except for post-combustion controls such as SCR and
SNCR. For these measures, the costs presume that a NOX combustion control (such as LNB) is already operating
on the unit to which the SCR or SNCR is applied.
4.A-7
-------
CHAPTER 5
HUMAN HEALTH BENEFITS ANALYSIS APPROACH AND RESULTS
5.1 Synopsis
This chapter presents the estimated human health benefits for the revised National
Ambient Air Quality Standards (NAAQS) for particulate matter (PM). In this chapter, we quantify
the health-related benefits of the fine particulate matter (PM2.5)-related air quality
improvements resulting from the illustrative emission control scenarios that reduce emissions
of directly emitted particles and precursor pollutants including S02 and NOX to reach alternative
PM2.5 NAAQS levels in 2020.1
These benefits are relative to an analytical baseline reflecting nationwide attainment of
the current primary PM2.5 standards (i.e., annual standard of 15 u.g/m3 and 24-hour standard of
35 u.g/m3) that includes promulgated national regulations and illustrative emission controls to
simulate attainment with 15/35 as well as a NOX emission adjustment to reflect expected
reductions in mobile NOX emissions between 2020 and 2025. We project PM2.5 levels in certain
areas that would exceed the revised annual standard of 12 u.g/m3 as well as alternative annual
standards of 13 and 11 u.g/m3 after simulated attainment with 15/35 in the analytical baseline.
Table 5-1 summarizes the total monetized benefits of the revised and alternative PM2.5
standards in 2020. These estimates reflect the sum of the economic value of estimated PM2.5
mortality impacts identified and the value of all morbidity impacts.
The estimated benefits for the revised and alternative standards are in addition to the
substantial benefits estimated for several recent implementation rules (U.S. EPA, 2009a, 2011c,
2011d, 2011e). Rules such as the Mercury and Air Toxics Standard (MATS) and other emission
reductions will have substantially reduced ambient PM2.5 concentrations by 2020 in the East,
such that no additional controls would be needed in the East for the revised annual standard of
12 u.g/m3. Thus, all of the estimated benefits occur in California because this is the only state
that needs additional air quality improvement beyond the analytical baseline after accounting
for air quality improvements from recent rules. Because of the national focus of many of the
inputs to the benefits model, benefits estimated for any particular location have greater
uncertainty than benefits occurring nationally. Compared with the proposal benefits, the
estimated benefits for the revised standard are about double due to a combination of updates
1 The estimates in this chapter reflect incremental emissions reductions from an analytical baseline that gives an
adjustment to the San Joaquin and South Coast areas in California for NOx emissions reductions expected to
occur between 2020 and 2025, when those areas are expected to demonstrate attainment with the revised
standards. Full benefits of the revised standards in those two areas will not be realized until 2025.
5-1
-------
to the analytic baseline (See Chapter 3). We believe that all of these updates improve our
estimate of benefits for the revised annual primary standard.
Table 5-1. Estimated Monetized Benefits of the of Revised and Alternative Annual PM2.5
Standards in 2020 Incremental to the Analytical Baseline (billions of 2010$)a
Benefits Estimate 13 |ig/ms 12 |ig/ms 11 |ig/ms
Economic value of avoided PM2.5.related morbidities and premature deaths using PM2.5 mortality estimate from
Krewskietal. (2009)
3% discount rate $1.3 + B $4.0 +B $13 + B
7% discount rate $1.2 + B $3.6 +B $12 + B
Economic value of avoided PM2.5.related morbidities and premature deaths using PM2.5 mortality estimate from
Lepeuleetal. (2012)
3% discount rate $2.9 + B $9.1+B $29 + B
7% discount rate $2.6 + B $8.2 +B $26 + B
3 Rounded to two significant figures. Avoided premature deaths account for over 98% of monetized benefits here,
which are discounted over the SAB-recommended 20-year segmented lag. It was not all possible to quantify all
benefits due to data limitations in this analysis. "B" is the sum of all unquantified health benefits and welfare co-
benefits.
As we describe in detail below, we estimate PM-related health impacts using
concentration-response relationships drawn from the epidemiological literature. 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 in the benefits estimates. Likewise, we are less confident
in the risk we estimate from simulated PM2.5 concentrations that fall below the bulk of the
observed data in these studies. As noted in the preamble to the rule, the range from the 25th to
10th percentiles of the air quality data in the epidemiology studies is a reasonable range below
which we start to have appreciably less confidence in the magnitude of the associations
observed in the epidemiological studies. Most of the estimated avoided premature deaths
occur at or above the lowest measured PM2.5 concentration in the two studies that are used to
estimate mortality benefits.
In addition to PM2.5 benefits, implementation of emissions controls to reach some of the
alternative PM2.5 standards would reduce other ambient pollutants, such as S02, N02, and
ozone. 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 other pollutants. In addition, due to data and
methodology limitations, we were unable to estimate additional health benefits associated
with exposure to PM2.5 or the additional co-benefits from improvements in welfare effects (i.e.,
5-2
-------
non-health effects) associated with emission reductions to attain the primary standard, such as
visibility. We describe the unquantified health benefits in this chapter and the unquantified
welfare co-benefits in Chapter 6.
5.2 Overview
This chapter contains a subset of the estimated health benefits of the revised and
alternative PM2.5 standards in 2020 that the EPA was able to quantify, given the available
resources and methods. The analysis in this chapter aims to characterize the benefits of the air
quality changes resulting from the implementation of new PM standards by answering two key
questions:
1. What are the health effects of changes in ambient particulate matter
resulting from reductions in directly emitted PM2.s and precursors due to the
attainment of a new PM2.s standard?
2. What is the economic value of these effects?
In this analysis, we consider an array of health impacts attributable to changes in PM2.s.
The Integrated Science Assessment for Particulate Matter ("PM ISA")(U.S. EPA, 2009b) identifies
the human health effects associated with ambient particles, which include premature mortality
and a variety of morbidity effects associated with acute and chronic exposures. Table 5-2
summarizes human health categories contained within the core benefits estimate as well as
those categories that were unquantified due to limited data or resources. It is important to
emphasize that the list of unquantified benefit categories is not exhaustive, nor is quantification
of each effect complete. To identify human health benefits that we could quantify with
confidence we selected endpoints that were classified as causal or likely causal in the PM ISA,
and we excluded effects not identified as having at least a causal, likely causal, or suggestive
relationship with the affected pollutants in the most recent comprehensive scientific
assessment, such as an ISA. This decision does not imply that additional relationships between
these and other human health and environmental co-benefits and the affected pollutants do
not exist. Due to this decision criterion, we excluded some effects that were identified in
previous lists of unquantified benefits in other RIAs (e.g., UVb exposure).
The benefits analysis in this chapter 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. We employ several techniques to characterize this uncertainty, which are
described in detail in Section 5.4.
5-3
-------
Table 5-2. Human Health Effects of Pollutants Potentially Affected by Strategies to Attain
the Primary PM2.s Standards
Benefits Category
Specific Effect
Effect Has Effect Has
Been Been
Quantified Monetized More Information
Improved Human Health
Reduced incidence Adult premature mortality based on cohort
of premature study estimates and expert elicitation
estimates (age >25 or age >30)
mortality from
exposure to PM2s
Infant mortality (age <1)
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)
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
•/
•/
•/
•/
•/
•/
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
PM ISAb
PM ISAb
PM ISA
b'c
PM ISA
b'c
(continued)
5-4
-------
Table 5-2. Human Health Effects of Pollutants Potentially Affected by Strategies to Attain
the Primary PM2.s Standards (continued)
Benefits Category
Specific Effect
Effect Has Effect Has
Been Been
Quantified Monetized More Information
Reduced incidence Premature mortality based on short-term
of mortality from study estimates (all ages)
exposure to ozone
Premature mortality based on long-term study
estimates (age 30-99)
Hospital admissions—respiratory causes (age
>65)
Hospital admissions—respiratory causes (age
Ozone ISA
Ozone ISA
Ozone ISA
Ozone ISA
Emergency department visits for asthma (all
ages)
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., premature
aging of lungs)
Cardiovascular and nervous system effects
Reproductive and developmental effects
Ozone ISA
Ozone ISA
Ozone ISA
Ozone ISA
Ozone ISAb
Ozone ISA0
Ozone ISA0
Reduced incidence Asthma hospital admissions (all ages)
of morbidity from
exposure to NO Chronic lung disease hospital admissions (age
2 >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
hyperresponsivenessand inflammation, lung
function, other ages and populations)
NO2 ISA
NO, ISA
NO, ISA
NO, ISA
NO, ISA
NO, ISA
b,c
NO, ISA
b,c
(continued)
5-5
-------
Table 5-2. Human Health Effects of Pollutants Potentially Affected by Strategies to Attain
the Primary PM2.s Standards (continued)
Benefits Category
Reduced incidence
of morbidity from
exposure to SO2
Specific Effect
Respiratory hospital admissions (age > 65)
Asthma emergency department visits (all
ages)
Asthma exacerbation (asthmatics age 4-12)
Acute respiratory symptoms (age 7-14)
Premature mortality
Other respiratory effects (e.g., airway
hyperresponsivenessand inflammation, lung
function, other ages and populations)
Effect Has Effect Has
Been Been
Quantified Monetized More Information
- - SO2 ISAd
- - SO2 ISAd
- - SO2 ISAd
- - SO2 ISAd
- - S02 ISAb'c
- - S02 ISAb'c
Reduced incidence Neurologic effects—IQ loss
of morbidity from
exposure to
methylmercury
(through role of
sulfate in
methylation)
Other neurologic effects (e.g., developmental
delays, memory, behavior)
Cardiovascular effects
Genotoxic, immunologic, and other toxic
effects
IRIS; NRC, 2000
IRIS; NRC, 2000b
b,c
IRIS; NRC, 2000
IRIS; NRC, 2000b'c
a We quantify these benefits in a sensitivity analysis, but not in the core analysis.
We assess these benefits qualitatively because we do not have sufficient confidence in available data or
methods.
c 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.
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 implementation rules,
such as the recent MATS rule (U.S. EPA, 2011d). The NAAQS RIAs illustrate the potential costs
and benefits of attaining a revised air quality standard nationwide based on an array of
emission reduction strategies for different sources including known and unknown controls,
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 the estimated costs and benefits are not intended to be added to the
5-6
-------
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 addressing emissions from
coal and oil-fired electricity generating units (U.S. EPA, 2011d). 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. Subsequent implementation rules
will be reflected in the baseline for the next PM 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 PM2.5 standards
in 2006, and since publication of the proposal RIA for this rulemaking. Below we note the
aspects of this analysis that differ from the proposal RIA (U.S. EPA, 2012a):
1. Incorporation of the newest Harvard Six Cities mortality study. In 2012, Lepeule et al.
published an extended analysis of the Six Cities cohort. Compared to the study it
replaces (Laden et al., 2006), this new analysis follows the cohort for a longer time
and includes more years of PM2.5 monitoring data. The all-cause PM2.5 mortality risk
coefficient drawn from Lepeule et al. (2012) is roughly similar to the Laden et al.
(2006) risk coefficient applied in the EPA's recent analyses of long-term PM2.5
mortality and has narrower confidence intervals.
2. Updated demographic data. We updated the population demographic data in
BenMAP to reflect 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).
3. Incorporation of new morbidity studies. Since the publication of the PM ISA (U.S.
EPA, 2009) the epidemiological literature has produced several new studies
examining the association between short-term PM2.5 exposure and respiratory and
cardiovascular hospitalizations, respiratory and cardiovascular emergency
department visits, and stroke. Upon careful evaluation of this new literature
identified in the Provisional Assessment of Recent Studies on Health Effects of
Particulate Matter Exposure ("Provisional Assessment")(U.S. EPA, 2012b) we added
several new studies to our health impact assessment.
5-7
-------
4. 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 replace the survival rates from Rosamond et al. (1999).
5. 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 (previously the LML analysis).
Although the list above identifies the major changes implemented since the proposal
RIA, the EPA has also updated several additional components of the benefits analysis since the
2006 PM NAAQS RIA (U.S. EPA, 2006a). 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 PM2.5-related health effects and began using 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 PM2.5 mortality expert elicitation.
In the N02 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. In the proposed CSAPR (previously
the "Transport Rule") (U.S. EPA, 2010g), we incorporated the "lowest measured level"
assessment to help characterize uncertainty in estimates of benefits of reductions in PM2.5at
lower baseline concentrations of PM2.5. In the final CSAPR (U.S. EPA, 2011c), we updated the
baseline incidence rates for hospital admissions and emergency department visits and asthma
prevalence rates. We direct the reader to each of these RIAs for more information on these
changes. In the proposal RIA for this NAAQS review (U.S. EPA, 2012a), we updated the
American Cancer Society cohort study to Krewski et al. (2009), updated health endpoints in the
core and sensitivity analyses, incorporated new morbidity studies, updated the median wage
data in the cost-of-illness studies, and expanded the uncertainty assessment.
5.4 Human Health Benefits Analysis Methods
We follow a "damage-function" approach in calculating total benefits of the modeled
changes in environmental quality.2 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
2 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. See discussion of hedonic studies in Chapter 6.
5-8
-------
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 value 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, a 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 (HIA) is limited to those health
effects that are directly linked to ambient levels of air pollution and specifically to those linked
to PM2.5.
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.
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 PM2.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 (BenMAP) 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 (Abt
Associates, 2010). 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
5-9
-------
quality (e.g., U.S. EPA, 2011d). For this assessment, the HIA is limited to those health effects
that are directly linked to ambient PM2.5 concentrations. There may be other indirect health
impacts associated with implementing emissions controls, such as occupational health
exposures.
The HIA approach used in this analysis involves three basic steps: (1) utilizing projections
of PM2.5 air quality3 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 (Hubbell et al., 2009) to this change in population exposure.
A typical health impact function might look as follows:
Ay = 1 - exy0 • Pop (5.1)
where y0 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 3 is the effect coefficient drawn from
the epidemiological study. Figure 5-1 provides a simplified overview of this approach.
3 Projections of ambient PM2.5 concentrations for this analysis were generated using the Community Multiscale Air
Quality model (CMAQ). See Chapters of this RIAfor more information on the air quality modeling.
5-10
-------
Baseline Air Quality
Post-Policy Scenario Air Quality
Incremental Air Quality
Improvement
Background
Incidence
Rate
Effect
Estimate
Figure 5-1. Illustration of BenMAP 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 u.g/m3 for PM2.s).
These relative risks can be used to develop risk coefficients that relate a unit reduction in 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
calculated 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
5-11
-------
$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.
For some health effects, such as hospital admissions, WTP estimates are generally not
available. In these cases, we use the cost of treating or mitigating the effect. For example, for
the valuation of hospital admissions we use the avoided medical costs as an estimate of the
value of avoiding the health effects causing the admission. These cost-of-illness (COI) estimates
generally (although not necessarily in every case) understate the true value of reductions in risk
of a health effect. They tend to reflect the direct expenditures related to treatment but not the
value of avoided pain and suffering from the health effect.
We use the BenMAP model version 4.0.52 (Abt Associates, 2010) to estimate the health
impacts and monetized health benefits for the proposed standard. Figure 5-2 shows the data
inputs and outputs for the BenMAP model.
Census
Population Data
Modeled Baseline
and Post-Control
Ambient PM2.s
2020 Population
Projections
Woods & Poole
Population
Projections
PM2.5 Incremental Air
Quality Change
PM2.5 Health
Functions
Economic
Valuation
Functions
PM2.5-Related Health
Impacts
Background
Incidence and
Prevalence Rates
Monetized PM2.5-
related Benefits
Blue identifies a user-selected input within the BenMAP model
Green identifies a data input generated outside of the BenMAP
Figure 5-2. Data Inputs and Outputs for the BenMAP Model
5-12
-------
5.5 Uncertainty Characterization
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
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. 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 make up alternative
assumptions simply for the purpose of probabilistic uncertainty characterization when there is
no scientific literature to support alternate assumptions.
To characterize uncertainty and variability, we follow an approach that combines
elements from two recent analyses by the EPA (U.S. EPA, 2010b; 2011a), 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 5c. Data limitations prevent us
5-13
-------
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 six quantitative analyses
described in more detail below:
1. A Monte Carlo assessment that accounts for random sampling error and between
study variability in the epidemiological and economic valuation studies;
2. A concentration benchmark assessment that characterizes the distribution of
avoided PM2.5-related deaths relative to specific concentrations in the long-term
epidemiological studies used to estimate PM2.5-related mortality;
3. The quantification of PM-related mortality using alternative PM2.5 mortality effect
estimates drawn from two long-term cohort studies and an expert elicitation;
4. Sensitivity analyses of several aspects of PM-related benefits;
5. Distributional analyses of PM2.5-related benefits by location, race, income, and
education; and
6. An analysis of the influence of various parameters on total monetized benefits.
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 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.
5-14
-------
5.5.2 Concentration Benchmark Analysis for PM2.5
We include a concentration benchmark assessment, which identifies the baseline (i.e.,
pre-rule) annual mean PM2.5 levels at which populations are exposed and specific
concentrations in the two long-term cohort studies we use to quantify mortality impacts. This
analysis characterizes avoided PM2.5-related deaths relative to the 10th and 25th percentiles of
the air quality data used the Krewski et al. (2009) study as well as the lowest measured level
(LML) of the Krewski et al. (2009) and Lepeule et al. (20) studies.
5.5.3 Alternative Concentration-Response Functions for PM2.5-Related Mortality
We assign the greatest economic value to the reduction in PM2.5 related mortality risk.
Therefore, it is particularly important to attempt to characterize the uncertainties associated
with reductions in premature mortality. To better understand the concentration-response
relationship between PM2.5 exposure and premature mortality, the EPA conducted an expert
elicitation in 2006 (Roman et al., 2008; lEc, 2006).4 In general, the results of the expert
elicitation support the conclusion that the benefits of PM2.5 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. Thus, we include the expert elicitation results as well as standard
errors approaches to provide insights into the likelihood of different outcomes and about the
state of knowledge regarding the benefits estimates. In this analysis, we present the results
derived from the expert elicitation as 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 PM2.5 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 PM2.5 benefits may be, and
the EPA's independent Science Advisory Board (SAB) recommended refinements to the way
EPA presented the results of the elicitation (U.S. EPA-SAB, 2008). Therefore, we began to
Expert elicitation is a formal, highly structured and well documented process whereby expert judgments, usually
of multiple experts, are obtained (Ayyub, 2002).
5-15
-------
present the cohort-based studies (Krewski et al., 2009; Laden et al., 2006)5 as our core
estimates in the proposal RIA for the Portland Cement NESHAP (U.S. EPA, 2009a). Using
alternate relationships between PM2.5and premature mortality supplied by experts, higher and
lower benefits estimates are plausible, but most of the expert-based estimates of the mean
PM2.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 or other recent multi-state
cohort studies conducted in North America, but we have not estimated benefits using the effect
coefficients from these studies. 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 PM2.5 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.
Even these multiple characterizations with confidence intervals 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 parameters and the identification of reasonable upper
and lower bounds for input distributions characterizing uncertainty in additional model
elements. As a result, the reported confidence intervals and range of estimates give an
incomplete picture about the overall uncertainty in the estimates. This information should be
interpreted within the context of the larger uncertainty surrounding the entire analysis.
5.5.4 Sensitivity Analyses
For some aspects of uncertainty, we have sufficient data to conduct sensitivity analyses.
In this analysis, we performed five such analyses for the revised standard level. In particular,
we:
1. Assessed the sensitivity of the economic value of reductions in the risk of PM2.5-
related death according to differing assumptions regarding the lag between PM2.5
exposure and premature death. The timing of such premature deaths affects the
magnitude of the discounted PM2.5-related mortality benefits. In this sensitivity
assessment, we consider 6 alternative cessation lags.
2. Characterized the sensitivity of the economic value of the health endpoints valued
using willingness-to-pay estimates to a higher and a lower assumption regarding
5 We have since updated the 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).
5-16
-------
income elasticity. As we discuss below, economic theory argues that individual
willingness to pay increases as personal income grows. The relationship between
growth in personal income and willingness-to-pay to reduce mortality and morbidity
risk is characterized by the income growth factor.
3. Summarized the avoided cases of certain health endpoints for which we either
lacked an appropriate economic value (cardiovascular hospital admissions and
stroke) or in which we no longer had sufficient confidence to retain in our primary
benefits estimate (chronic bronchitis).
4. Assessed the sensitivity of the benefits results to the new population data from the
2010 Census.
5. Assessed the sensitivity of the benefits results to an analysis year of 2025.
5.5.5 Distributional Assessment
In an Appendix to the proposal RIA, we characterized the distribution of PM2.5-related
benefits based on the geographic distribution of race and education in areas where the
illustrative emission reduction strategies would reduce PM2.5 concentrations. In that
assessment, we aimed to answer two key questions:
1. What was the estimated distribution of PM2.5-related mortality risk based on the
race and education characteristics of the population living within areas projected to
exceed alternative combinations of the proposed primary PM2.5 standards?
2. How would air quality improvements within these counties change the distribution
of risk among populations of different races and educational attainment?6
That assessment was generally consistent with the distributional assessments
performed in support of MATS (U.S. EPA, 2011c), with one key difference. The environmental
justice analyses accompanying the MATS RIA applied CMAQ-modeled PM2.5 predictions that
represent the change in air quality after the implementation of each rule. By contrast, the
proposal RIA aimed to illustrate the potential benefits and costs of attaining alternative primary
PM2.s standards; the states will ultimately implement attainment strategies, which may differ
greatly from the least-cost strategy the EPA modeled here. Alternative emission reduction
strategies—particularly those that maximize benefits to human health and provide a more
equitable distribution of risk—are also available to the states, though not modeled here (Fann
6 In this analysis we assess the change in risk among populations of different race and educational attainment. As
we discuss further in the methodology, we consider this last variable because of the availability of education-
modified PM2.5 mortality risk estimates.
5-17
-------
et al., 2012b). Due in part to time constraints, the EPA did not perform such an analysis for the
final RIA, and instead refers readers to the Appendix noted above.
5.5.6 Influence Analysis—Quantitative Assessment of Uncertainty
In the past few years, the EPA has initiated several projects to improve the
characterization of uncertainty for benefits analysis. In particular, the EPA recently completed
the first phase of a quantitative uncertainty analysis of benefits, hereafter referred to as the
"Influence Analysis" (Mansfield et al., 2009). The Influence Analysis diagramed the uncertain
components of each step within the benefits analysis process, identified plausible ranges for a
sensitivity analysis, and assessed the sensitivity to total benefits to changes in each component.
Although this analysis does not quite fulfill the goal of a full probabilistic assessment, it
accomplished the necessary first steps and identified the challenges to accomplishing that goal.
Below are some of the preliminary observations from the first phase of the project.
• The components that contribute the most to uncertainty of the monetized benefits
and mortality incidence (in order of importance) are the value-of-a-statistical-life
(VSL), the concentration-response (C-R) function for mortality, and change in PM2.s
concentration.
• The components that contribute the least to uncertainty of the monetized benefits
and mortality incidence are population, morbidity valuation, and income elasticity.
• The choice of a C-R function for mortality affects the mortality incidence and
monetized benefits more than other sources of uncertainty within each C-R
function.
• Alternative cessation lag structures for mortality have a moderate effect on the
monetized benefits.
• Because the health impact function is essentially linear, the key components show
the same sensitivity across all mortality C-R functions even if the midpoints differ
significantly from one expert to another.
5.5.7 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 revised and alternative annual standards:
5-18
-------
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) (SAB-HES) (U.S. EPA- SAB, 2010a) and the National
Academies of Science (NAS) (NRC, 2002). The benefits estimates are subject to a number of
assumptions and uncertainties. For example, the key assumptions underlying the estimates for
premature mortality, which account for over 98% of the total monetized benefits in this
analysis, include the following:
1. We assume that all fine particles, regardless of their chemical composition, are
equally potent in causing premature mortality. This is an important assumption,
because PM2.s varies considerably in composition across sources, but the scientific
evidence is not yet sufficient to allow differentiation of effect estimates by particle
type. The PM ISA, which was twice reviewed by Clean Air Scientific Advisory
Committee (SAB-CASAC), concluded that "many constituents of PM2.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). These uncertainties are likely to be magnified
in the current analysis to the extent that the emissions controls are less diverse
when focusing on one small region of the country rather than a broader geography
with more diverse emissions sources and the application of a more diverse set of
controls.
2. We assume that health impact functions based on national studies are
representative for exposures and populations in California. In addition to the
national risk coefficients we use as our core estimates, the EPA considered the
cohort studies conducted in California specifically. Although we have not calculated
the benefits results using the cohort studies conducted in California, we provided
these risk coefficients to show how much the monetized benefits could have
changed. Most of the California cohort studies report central effect estimates similar
to the (nation-wide) all-cause mortality risk estimate we applied from Krewski et al.
(2009) and Lepeule et al. (2012) albeit with wider confidence intervals. Three cohort
studies conducted in California indicate statistically significant higher risks than the
risk estimates we applied from Lepeule et al. (2012), and four studies showed
insignificant results.
3. 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 PM2.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.
4. 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
5-19
-------
some of the incidences of premature mortality related to PM2.5 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.
5. To characterize the uncertainty in the relationship between PM2.5and premature
mortality (which account for over 98% of total monetized benefits in this analysis),
we include a set of twelve estimates based on results of the expert elicitation study
in addition to our core estimates. Even these multiple characterizations omit the
uncertainty in air quality estimates, 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 PM2.5 estimates. This information should be
interpreted within the context of the larger uncertainty surrounding the entire
analysis.
As previously described, we strive to monetize as many of the benefits anticipated from
the revised 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 benefits.
Specifically, only certain benefits attributable to the health impacts associated with exposure to
ambient fine particles have been monetized in this analysis. Data and methodological
limitations prevented the EPA from quantifying or monetizing the benefits from several
important health benefit categories from emission reduction strategies to reach the revised
annual standard in this RIA, including potential co-benefits from reducing ozone exposure, N02
exposure, S02 exposure, and methylmercury exposure (see section 5.6.5 for more information).
If we could fully monetize all of the benefit categories, the total monetized benefits would
exceed the costs by an even greater margin than we currently estimate.
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
analyses (U.S. EPA, 2010b, 2011a). 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 Appendix 5B.
5-20
-------
5.6 Benefits Analysis Data Inputs
In Figure 5-2, 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 MATS RIA (U.S. EPA, 2011d) and the proposal RIA (U.S. EPA, 2012a).
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 proposal RIA incorporated WP projections relative to a baseline using
2000 Census data. An analysis exploring the sensitivity of population and health impact
estimates to this update can be found in Appendix 5.A. 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:
• 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.
5-21
-------
• 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 10 u.g/m3 decrease in daily PM2.5 levels might be
associated with a decrease in hospital admissions of 3%. 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 5-3 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 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
5-22
-------
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 revised rates first applied in the CSAPR RIA (U.S. EPA, 2011c). In addition, we have
revised the baseline incidence rates for acute myocardial infarction. These revised rates are
more recent (AHRQ, 2007), which provides 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 5-4 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, 2011c).
5-23
-------
Table 5-3. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population
Rates
Endpoint
Parameter
Value
Source
Mortality
Hospitalizations
ER Visits
Cerebrovascular
events
Chronic Bronchitis0
Daily or annual mortality rate
projected to 2020a
Daily hospitalization rate
Daily ER visit rate for asthma and
cardiovascular events
Age-, cause-, and county-
specific rate
Age-, region-, state-,
county- and cause-specific
rate
Age-, region-, state-,
county- and cause-specific
rate
Incidence of new cerebrovascular 0.0015751
events among populations 50-79
Annual prevalence rate per
person
• Aged 18-44 • 0.0315
• Aged 45-64 • 0.0549
• Aged 65 and older • 0.0563
Annual incidence rate per person 0.00378
Nonfatal Myocardial
Infarction (heart
attacks)
Asthma
Exacerbations
Acute Bronchitis
Lower Respiratory
Symptoms
Upper Respiratory
Symptoms
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
Age-, region-, state-, and
county-specific rate
• 0.173
• 0.145
• 0.074
0.043
Daily lower respiratory symptom 0.0012
incidence among childrend
Daily upper respiratory symptom
incidence among asthmatic
children
0.3419
CDC WONDER (2004-
2006)
U.S. Census bureau, 2000
2007 HCUP data files"
2007 HCUP data files
Table 3 of Miller etal.
(2007)
American Lung Association
(2010a, Table 4).
Abbey et al. (1993,
TableS)
2007 HCUP data files;"
adjusted by 0.93 for
probability of surviving
after 28 days (Rosamond
etal., 1999)
Ostro et al. (2001)
American Lung Association
(2002c, Table 11)
Schwartz et al. (1994,
Table 2)
Pope etal. (1991, Table 2)
(continued)
5-24
-------
Table 5-3. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population (continued)
Rates
Endpoint
Parameter
Value
Source
Work Loss Days
School Loss Days
Minor Restricted-
Activity Days
Daily WLD incidence rate per
person (18-65)
• Aged 18-24 • 0.00540
• Aged 25-44 • 0.00678
• Aged 45-64 • 0.00492
Rate per person per year, 9.9
assuming 180 school days per
year
Daily MRAD incidence rate per 0.02137
person
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)
Mortality rates are only available at 5-year increments.
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).
Assessed in sensitivity analysis only. The rate numbers may be slightly different from those in Table 4 because
we received more current estimates from ALA.
Lower respiratory symptoms are defined as two or more of the following: cough, chest pain, phlegm, and
wheeze.
Table 5-4. Asthma Prevalence Rates
Population Group
All Ages
<18
5-17
18-44
45-64
65+
African American, 5-17
African American, <18
Value
0.0780
0.0941
0.1070
0.0719
0.0745
0.0716
0.1776
0.1553
Asthma Prevalence Rates
Source
American Lung Association (2010b, Table 7)
American Lung Association (2010b, Table 9)
American Lung Association3
Calculated by ALA for U.S. EPA, based on NHIS data (CDC, 2008).
5-25
-------
5.6.3 Effect Coefficients
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 SAB-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, 2011a). In addition, we have
included more recent epidemiology studies from the PM ISA (U.S. EPA, 2009b) and the
Provisional Assessment (U.S. EPA, 2012b).7 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
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 PM2.5and adverse human health effects. We evaluated epidemiological
studies using the selection criteria summarized in Table 5-5. 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
7 The peer-reviewed studies in the Provisional Assessment have not yet undergone external review by the SAB. The
new studies from the PM ISA and Provisional Assessment for health endpoints not previously quantified in
EPA's RIAs are presented in a sensitivity analysis in Appendix 5B, but these new endpoints have not been
incorporated into the core benefits analysis.
5-26
-------
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 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 effects 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 effects 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.8 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.
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 RIA.
5-27
-------
Table 5-5. Criteria Used When Selecting C-R Functions
Consideration
Comments
Peer-Reviewed
Research
Study Type
Study Period
Population Attributes
Study Size
Study Location
Pollutants Included in
Model
Measure of PM
Economically Valuable
Health Effects
Non-overlapping
Endpoints
Peer-reviewed research is preferred to research that has not undergone the peer-review
process.
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.
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.
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).
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.
When modeling the effects of ozone and PM (or other pollutant combinations) jointly, it
is important to use properly specified impact functions that include both pollutants.
Using single-pollutant models in cases where both pollutants are expected to affect a
health outcome can lead to double-counting when pollutants are correlated.
For this analysis, impact functions based on PM2.5 are preferred to PM10 because of the
focus on reducing emissions of PM2.5 precursors, and because air quality modeling was
conducted for this size fraction of PM. Where PM2.5 functions are not available, PM10
functions are used as surrogates, recognizing that there will be potential downward
(upward) biases if the fine fraction of PM10 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.
5-28
-------
The specific studies from which effect estimates for the core analysis are drawn are
included in Table 5-6. We highlight in blue those studies that have been added since the
benefits analysis conducted for the MATS RIA (U.S. EPA, 2011d), and we highlight those studies
in red that have been added since the proposal RIA (U.S. EPA, 2012a). 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 5-7 summarizes those
health endpoints and studies we have included as in sensitivity analyses.
Table 5-6. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts
in the Core Analysisa
Endpoint
Study
Study Population
Risk Estimate
(95th Percentile
Confidence Interval)3
Premature Mortality
Premature
mortality—cohort
study, all-cause
Premature mortality,
total exposures
Premature
mortality—all-cause
Krewski et al. (2009)
Lepeule et al. (2012)
PM2.5 Expert Elicitation (Roman et al.,
2008)
Woodruff etal. (1997)
Krewski et al. (2009)
Lepeule et al. (2012)
PM2.5 Expert Elicitation
Premature
mortality-
cohort study,
all-cause
Premature
mortality, total (Roman et al., 2008)
exposures
Premature
Woodruff etal. (1997)
mortality—all-
cause
Chronic Illness
Nonfatal heart
attacks
Nonfatal heart
attacks (cont'd)
Peters et al. (2001)
Pooled estimate:
Pope et al. (2006)
Sullivan etal. (2005)
Zanobetti et al. (2009)
Zanobetti and Schwartz (2006)
Nonfatal heart
attacks
Peters et al. (2001)
Pooled estimate:
Pope et al. (2006)
Sullivan etal. (2005)
Nonfatal heart Zanobetti et al. (2009)
attacks (cont'd) Zanobetti and Schwartz
(2006)
Hospital Admissions
Respiratory
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)
Babin et al. (2007)-ICD 493 (asthma)
Sheppard (2003)-ICD493 (asthma)
Respiratory Zanobetti et al. (2009) —
ICD 460-519 (All
respiratory)
Kloog etal. (2012)-ICD
460-519 (All Respiratory
Moolgavkar (2000)-ICD
490-496 (Chronic lung
disease)
Babin et al. (2007)-ICD
493 (asthma)
Sheppard (2003)-ICD 493
(asthma)
(continued)
5-29
-------
Table 5-6. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts
in the Core Analysisa (continued)
Endpoint
Cardiovascular
Asthma-related ER
visits
Study
Pooled estimate:
Zanobetti et al. (2009)-ICD 390-459 (all
cardiovascular)
Peng et al. (2009)-ICD 426-427; 428; 430-438;
410-414; 429; 440-449 (Cardio-, cerebro- and
peripheral vascular disease)
Peng et al. (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 390-429 (all
cardiovascular)
Pooled estimate:
Mar et al. (2010)
Slaughter etal. (2005)
Glad et al. (2012)
Risk Estimate
Study (95th Percentile
Population Confidence Interval)3
>64 years
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
Acute bronchitis
Asthma
exacerbations
Work loss days
Acute respiratory
symptoms
Upper respiratory
symptoms
Lower respiratory
symptoms
Dockeryetal. (1996)
Pooled estimate:
Ostro et al. (2001) (cough, wheeze and
shortness of breath) b
Mar et al. (2004) (cough, shortness of breath)
Ostro (1987)
Ostro and Rothschild (1989) (Minor restricted
activity days)
Pope etal. (1991)
Schwartz and Neas (2000)
8-12 years OR = 1.50 (0.91-2.47)
per 14.9 u.g/m3
6-18 years b QR = 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/m
18-65 years p=0.0046 (0.00036)
18-65 years P=0.00220 (0.000658)
Asthmatics, 9- 1.003 (1-1.006) per
11 years 10 u.g/m3
OR =1.11 (1.58-1.58)
7-14 years . 3
per 15 u.g/m
Studies highlighted in blue represent updates incorporated since the RIA for MATS (U.S. EPA, 2011d). Studies
highlighted in red represent updates incorporated since the proposal RIA (U.S. EPA, 2012a).
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).
5-30
-------
Table 5-7. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts
in the Sensitivity Analysis3
Endpoint
Chronic Illness
Chronic bronchitis
Stroke
Hospital Admissions
Cardiovascular ED Visits
Study
Abbey etal. (1995)
Miller etal. (2007)
Metzger et al. (2004)
Tolbert et al. (2007)
Mathes et al. (2011)
Study Population
>26 years
50-79 years
0-99
0-99
40-99
a Studies highlighted in blue represent updates incorporated since the RIA for MATS (U.S. EPA, 2011d). Studies
highlighted in red represent updated incorporated since the proposal RIA (U.S. EPA, 2012a).
5.6.3.1 PM2,5 Premature Mortality Effect Coefficients
Core Mortality Effect Coefficients for Adults. A substantial body of published scientific
literature documents the association between elevated PM2.5 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 this review
of the PM standards, which was twice reviewed by the SAB-CASAC (U.S. EPA-SAB, 2009b,
2009c), 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 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"
5-31
-------
(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 in PM2.5.
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 PM2.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 Effect 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 (e.g., U.S.
EPA, 2006a, 2010c, 2011c, 2011d). The PM ISA concludes that the ACS and Six Cities cohorts
provide the strongest evidence of the association between long-term PM2.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
5-32
-------
urban areas (c) addressing confounding by spatial autocorrelation by incorporating ecological,
or community-level, co-variates; (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 PM2.5 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, 2010b) 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, 2010b). The SAB-CASAC also provided extensive peer review of the risk
assessment and supported the use of effect estimates from this study (U.S. EPA-SAB, 2009a,
2010b, c).
Consistent with the Quantitative Health Risk Assessment for Particulate Matter (U.S.
EPA, 2010b) which was reviewed by the SAB-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 10u.g/m3 increase in PM2.5). The relative risk estimate (1.06 per 10u.g/m3 increase in
PM2.5) 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 PM2.5 exposure and increased risk of all-cause, cardiovascular and lung cancer
mortality. The authors also concluded that the concentration-response relationship was linear
5-33
-------
down to PM2.5 concentrations of 8 |-ig/m3, and that mortality rate ratios for PM2.5 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
u.g/m3 increase in PM2.5). 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 PM2.5-related premature mortality impacts are several key
assumptions, which are described in further detail later in this chapter. 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 5.6.6.1). Second, following conclusions of the PM ISA,
we assume that all fine particles are equally potent in causing premature mortality (see section
5.7.2). 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 5.7.4).
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 PM2.5 and the risk of premature death. Many of
these additional cohort studies are described in the PM ISA (U.S. EPA, 2009) and the Provisional
Assessment (U.S. EPA, 2012b) (and thus not summarized here).9'10 Table 5-8 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 PM2.5 expert elicitation and the
derivation of effect estimates from the expert elicitation results are described in detail in the
9 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 H6C cohort studies have
been recommended by the SAB as appropriate for benefits analysis of national rulemakings.
10 In this chapter, 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.
5-34
-------
2006 PM2.5 NAAQS RIA (U.S. EPA, 2006a), the elicitation summary report (lEc, 2006) and Roman
et al. (2008), and so we summarize here 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.
Table 5-8. Summary of Effect estimates from Associated With Change in Long-Term
Exposure to PM2.5 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
Eftimetal.
(2008)
Zeger et al.
(2008)c
Krewski et
al. (2009)d
Puettetal.
(2009)b
Grouse et
al. (2011)d'e
Puettetal.
(2011)'
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
(ug/m3)
7.5
10
<14.1
3.4
6
<9.8
5.8
5.8
1.9
<14.4
8
Mean
(ug/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 u.g/m3 Change in PM2.5
(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
(1.17-4.16)
N/A
N/A
1.24
(1.19-1.29)
2.02
(1.07-3.78)
N/A
N/A
N/A
Low socio-economic status (SES) men only. Used traffic proximity as a surrogate of exposure.
Women only.
c 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.
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
5-35
-------
among the adult U.S. population resulting from reductions in ambient annual average PM2.5
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 PM2.5-mortality relationship designed to build twelve
individual distributions for the coefficient (or slope) of the C-R function relating changes in
annual average PM2.5 exposures to annual, adult all-cause mortality. The elicitation also
provided useful information regarding uncertainty characterization in the PM2.5-mortality
relationship. 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 PM2.5 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"11 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)
11 These criteria are substantively similar to EPA's study selection criteria identified in Table 5-5 of this chapter.
5-36
-------
• 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 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]).
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 5-8.
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
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
5-37
-------
Table 5-8. In addition, we now present the full distributions of the expert-derived results in a
probabilistic graphic (rather than cascading the expert-derived results throughout the results
tables as done in prior RIAs). We do not combine the expert results in order to preserve the
breadth and diversity of opinion on the expert panel (Roman et. al., 2006). 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 PM2.5 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.12 The PM ISA states that less evidence is available
regarding the potential impact of PM2.5 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, 2004a). 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
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
12 For the purposes of this analysis, we only calculate benefits for infants age 0-1, not all children under 5 years
old.
5-38
-------
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, 2004a).
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 PM2.5 and infant mortality in
California. The SAB-HES also noted, "when PMi0 results are scaled to estimate PM2.5 impacts,
the results yield similar risk estimates." Consistent with the Costs and Benefits of the Clean Air
Act (U.S. EPA, 2011a), we continue to rely on the earlier 1997 study in part due to the national-
scale of the earlier study.
5.6.3.2 Nonfatal Acute Myocardial Infarctions (AMI) (Heart Attacks)
Nonfatal heart attacks have been linked with short-term exposures to PM2.5 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.,
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 PM2.5 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 PM2.5-related hospitalization for AMIs in King County, WA; Pope et al.
(2006), based in Wasatch Range, UT; Zanobetti and Schwartz (2006), based in Boston, MA; 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
5-39
-------
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.s 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 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).
5.6.3.3 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.
The two main groups of hospital admissions estimated in this analysis are respiratory
admissions and cardiovascular admissions. There is not much evidence linking PM2.s with other
5-40
-------
types of hospital admissions. Both asthma- and cardiovascular-related visits have been linked to
PM2.5 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 PM2.5 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 64.13 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 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.
13 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.
5-41
-------
To estimate avoided incidences of respiratory hospital admissions associated with PM2.5,
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 PM2.5 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 PM2.5 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
PM2.5 and asthma hospitalizations in Seattle, Washington, among children 0 to 18.
5.6.3.4 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 PM2.5. The sources
for the effect estimates used to quantify these effects are described below.
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,14 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 Dockery
et al. (1996). 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.
We estimate three types of acute respiratory symptoms: 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
using an effect estimate from Schwartz and Neas (2000). Incidences of upper respiratory
symptoms in asthmatic children aged 9 to 11 are estimated using an effect estimate developed
14 See http://www.nlm.nih.gov/medlineplus/ency/article/001087.htm, accessed April 2012.
5-42
-------
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.
Minor restricted activity days (MRAD) 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 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.s
exposure and acute respiratory symptoms was available in the PM ISA (U.S. EPA, 2009), but
proved to be unsuitable for use in this benefits analysis. In particular, the best available study
(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.
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
MRADs. Finally, note also the important distinction 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.
To characterize asthma exacerbations in children, we selected two studies (Ostro et al.,
2001; Mar et al., 2004) that followed panels of asthmatic children. Ostro et al. (2001) followed a
5-43
-------
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
PM2.5, 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 semiarid 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 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 asthma exacerbation incidence
estimate. 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.
5.6.3.5 Effect Coefficients Selected for the Sensitivity Analyses
Chronic Bronchitis. Chronic bronchitis is characterized by mucus in the lungs and a
persistent wet cough for at least 3 months a year for several years in a row. Chronic bronchitis
affects an estimated 5% of the U.S. population (ALA, 1999). A limited number of studies have
estimated the impact of air pollution on new incidences of chronic bronchitis. Schwartz (1993)
and Abbey et al. (1995) provide evidence that long-term PM2.s exposure gives rise to the
5-44
-------
development of chronic bronchitis in adults in the United States; these remain the two most
recent studies observing a relationship between long-term exposure to PM2.5 and the onset of
chronic bronchitis in the U.S. The absence of newer studies finding a relationship between long-
term PM2.5 exposure and chronic bronchitis argues for moving this endpoint from the core
benefits analysis to a sensitivity analysis. In their review of the scientific literature on chronic
obstructive pulmonary disease (COPD), which includes chronic bronchitis and emphysema, the
American Thoracic Society concluded that air pollution is "associated with COPD, but sufficient
criteria for causation were not met" (Eisner et al., 2010).
Stroke. The PM ISA (U.S. EPA, 2009) includes several new studies that have examined
the relationship between PM2.5 exposure and cerebrovascular events (U.S. EPA, 2009). Time-
series studies have generally been inconsistent with several studies showing positive
associations (Dominici et al., 2006; Metzger et al., 2004; Lippman et al., 2000; Lisabeth et al.,
2008). Several other studies have demonstrated null or negative associations (Anderson et al.,
2001; Barnett et al., 2006; Peel et al., 2007). In general, these studies examined cerebrovascular
disease as a group, though a few studies partition ischemic and hemorrhagic strokes separately
(Lisabeth et al., 2008). A key limitation of these time-series studies is that they use hospital
discharge rates as the diagnosis and relatively short lags (0-2 days)—this is problematic, as
discharge rates are an imperfect diagnosis and strokes may occur several days before admission
to the hospital.
Longer-term prospective cohort studies of PM2.5 and stroke include Miller et al. (2007).
Miller et al. (2007) estimated the change in risk among post-menopausal women enrolled in the
Women's Health Initiative (U.S. EPA, 2009b). After adjusting for age, race, smoking status,
educational level, household income, body-mass index, diabetes, hypertension, and
hypercholesterolemia, hazard ratios were estimated for the first cardiovascular event. Because
this study considers first-time cardiovascular events, a key challenge to incorporating this study
into the core health impact assessment is matching the baseline incidence rate correctly, and
we have approximated this information using the data in the study.
In addition, Wellenius et al. (2012) examined the association of PM2.5 with neurologist
confirmed ischemic stroke in Boston adults in a time-stratified case-crossover study. A key
feature of this study is that it included the time of stroke symptom onset for most patients.
Similar to the challenge with Miller et al. (2007), we do not have baseline incidence rates, and
we do not have sufficient data from the study to approximate it. Three factors argue for
treating this endpoint in the sensitivity analysis: (1) the epidemiological literature examining
PM-related cerebrovascular events is still evolving; (2) there are special uncertainties associated
5-45
-------
with quantifying this endpoint; (3) we have not yet identified an appropriate method for
estimating the economic value of this endpoint.
Cardiovascular Emergency Department Visits. A large number of recent U.S.-based
studies provide support for an association between short-term increases in PM2.5 and increased
risk of ED visits for ischemic heart diseases (U.S. EPA, 2009b). Both Metzger et al. (2004) and
Tolbert et al. (2007) published interim results from the Study of Particles and Health in Atlanta
(SOPHIA), finding a relationship between PM2.5 exposure and cardiovascular emergency
department visits. These cardiovascular emergency department visits are distinct from
cardiovascular hospital admissions and non-fatal heart attacks. To ensure no double-counting,
we excluded ICD-9-411 (ischemic heart disease) from the baseline incidence rates for
cardiovascular emergency department visits. Mathes et al. (2011) find relationships between
PM2.5 levels and cardiovascular emergency department visits in New York City. The principal
challenge to incorporating these studies is the absence of readily-available economic valuation
estimates for cardiovascular emergency department visits. Until we develop an approach for
estimating the economic value of this endpoint, we will quantify these ED visits in a sensitivity
analysis only.
5.6.4 Unqualified Human Health Benefits
The illustrative emission reduction strategies to reach the revised and alternative annual
standards described in Chapter 4 would reduce emissions of directly emitted particles, as well
as S02, and NOX for an alternative standard for 11 u.g/m3. The extent to which down wind
exposure to secondary pollutants ozone, and mercury would actually be reduced would depend
on the specific control strategy that States would use to reduce PM2.5 in a given area as well as
local geographic and meteorological conditions. Although we have quantified many of the
health benefits associated with reducing exposure to PM2.5, as shown in Table 5-2, we are
unable to quantify the health benefits associated with reducing the potential for ozone
exposure, S02 exposure, N02 exposure or contamination of local water bodies with mercury
due to the absence of air quality modeling data for these pollutants in this analysis. Although
the method we applied simulated the impact of attaining the revised and alternative annual
standards on ambient levels of PM2.5, this method does not simulate how the illustrative
emission reductions would affect ambient levels of ozone, S02, or N02. Furthermore, the air
quality modeling conducted for this analysis did not assess mercury, so we are unable to
estimate mercury deposition associated with the illustrative controls or subsequent
bioaccumulation and exposure. Below we provide a qualitative description of these health
5-46
-------
benefits. In general, previous analyses have shown that the monetized value of these additional
health benefits is much smaller than PM2.5-related benefits (U.S. EPA, 2010a, 2010c, 2010d).
Reducing NOX emissions also reduces ozone concentrations in most areas. Reducing
ambient ozone concentrations is associated with significant human health benefits, including
mortality and respiratory morbidity (U.S. EPA, 2008a, 2010d). Epidemiological researchers have
associated ozone exposure with adverse health effects in numerous toxicological, clinical and
epidemiological studies (U.S. EPA, 2006b; U.S. EPA, 2012c). When adequate data and resources
are available, the EPA generally quantifies several health effects associated with exposure to
ozone (e.g., U.S. EPA, 2008a, 2010d, 2011a, 2011c). These health effects include respiratory
morbidity such as asthma attacks, hospital and emergency department visits, school loss days,
as well as premature mortality. The scientific literature suggests that exposure to ozone is also
associated with chronic respiratory damage and premature aging of the lungs, but the EPA has
not quantified these effects in benefits analyses previously.
Following an extensive evaluation of health evidence from epidemiologic and laboratory
studies, the Integrated Science Assessment for Sulfur Dioxide—Health Criteria (S02 ISA)
concluded that there is a causal relationship between respiratory health effects and short-term
exposure to S02(U.S. EPA, 2008c). The immediate effect of S02on the respiratory system in
humans is bronchoconstriction. Asthmatics are more sensitive to the effects of S02 likely
resulting from preexisting inflammation associated with this disease. A clear concentration-
response relationship has been demonstrated in laboratory studies following exposures to S02
at concentrations between 20 and 100 ppb, both in terms of increasing severity of effect and
percentage of asthmatics adversely affected. Based on our review of this information, we
identified four short-term morbidity endpoints that the S02 ISA identified as a "causal
relationship": asthma exacerbation, respiratory-related emergency department visits, and
respiratory-related hospitalizations. The differing evidence and associated strength of the
evidence for these different effects is described in detail in the S02 ISA. The S02 ISA also
concluded that the relationship between short-term S02 exposure and premature mortality was
"suggestive of a causal relationship" because it is difficult to attribute the mortality risk effects
to S02 alone. Although the S02 ISA stated that studies are generally consistent in reporting a
relationship between S02 exposure and mortality, there was a lack of robustness of the
observed associations to adjustment for pollutants. We did not quantify these benefits due to
data constraints.
Epidemiological researchers have associated N02 exposure with adverse health effects
in numerous toxicological, clinical and epidemiological studies, as described in the Integrated
5-47
-------
Science Assessment for Oxides of Nitrogen—Health Criteria (N02 ISA) (U.S. EPA, 2008b). The
N02 ISA provides a comprehensive review of the current evidence of health and environmental
effects of N02. The N02 ISA concluded that the evidence "is sufficient to infer a likely causal
relationship between short-term N02 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%
increase in risks for ED visits and hospital admissions and higher risks for respiratory symptoms.
The N02 ISA concluded that the relationship between short-term N02 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 N02 alone. Although the N02 ISA stated that studies
consistently reported a relationship between N02 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.
5.6.5 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
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.
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. 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 5-9. All values are in constant year 2006 dollars, adjusted for growth
in real income for WTP estimates out to 2020 using projections provided by Standard and
5-48
-------
Poor's, which is discussed in further detail in Section 5.6.8. 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 PM2.5-related related endpoints.
5.6.5.1 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 $5.4 million estimated in the Kochi et al. (2006) meta-analysis. However, the
Agency neither changed its official guidance on the use of VSL in rule-makings nor subjected the
interim estimate to a scientific peer-review process through SAB or other peer-review group.
5-49
-------
Table 5-9. Unit Values for Economic Valuation of Health Endpoints (2010$)a
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 25^14
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
$8,000,000
$98,000
$110,000
$120,000
$200,000
$98,000
$97,000
$110,000
$110,000
$190,000
$97,000
2020 Income Level
$9,600,000
$98,000
$110,000
$120,000
$200,000
$98,000
$97,000
$110,000
$110,000
$190,000
$97,000
Derivation of Distributions of Estimates
The EPA currently recommends a central VSL of $4.8m (1990$, 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;
$21,000at 7% discount rate)
(continued)
-------
Table 5-9. Unit Values for Economic Valuation of Health Endpoints (2010$)a (continued)
Central Estimate of Value Per Statistical Incidence
Health Endpoint
2000 Income Level
2020 Income Level
Derivation of Distributions of Estimates
Hospital Admissions
Chronic Lung Disease (18-64)
$21,000
$21,000
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).
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
$42,000
$41,000
$42,000
$41,000
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+)
$36,000
$36,000 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
$430
$430
No distributional information available. Simple average of two unit
COI values (2000$):
(1) $310, from Smith et al. (1997) and
(2) $260, from Stanford et al. (1999).
(continued)
-------
Table 5-9. Unit Values for Economic Valuation of Health Endpoints (2010$)a (continued)
Central Estimate of Value Per Statistical Incidence
Health Endpoint
2000 Income Level
2020 Income Level
Derivation of Distributions of Estimates
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms
(URS)
$31
$33
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)
$20
$21
NJ
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
$54
$58
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 have a uniform distribution between $16
and $71 (2000$).
(continued)
-------
Table 5-9. Unit Values for Economic Valuation of Health Endpoints (2010$)a (continued)
Central Estimate of Value Per Statistical Incidence
Health Endpoint
2000 Income Level
2020 Income Level
Derivation of Distributions of Estimates
Respiratory Ailments Not Requiring Hospitalization (continued)
Acute Bronchitis
$450
$480
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
(U.S. median = $150)
Variable
(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)
u>
Minor Restricted Activity Days
(MRADs)
$64
$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 $16.00) and be less than that for a WLD. The triangular
distribution acknowledges that the actual value is likely to be closer
to the point estimate than either extreme.
All estimates are rounded to two significant digits. Unrounded estimates in 2000$ are available in the Appendix J of the BenMAP user manual (Abt
Associates, 2012).
-------
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)15 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 $6.3 million (2000$).16 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 recently
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).
Draft guidance responding to SAB recommendations will be developed shortly.
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 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.
15 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.
16 In this analysis, we adjust the VSL to account for a different currency year (2010$) and to account for income
growth to 2020. After applying these adjustments to the $6.3 million value, the VSL is $8.9M.
5-54
-------
Although there are several differences between the labor market studies the EPA uses
to derive a VSL estimate and the PM2.5 air pollution context addressed here, those differences in
the affected populations and the nature of the risks imply both upward and downward
adjustments. Table 5-10 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 5-10. 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
Life Expectancy/Health Status
Attitudes Toward Risk
Income
Voluntary vs. Involuntary
Catastrophic vs. Protracted Death
Uncertain, perhaps overestimate
Uncertain, perhaps overestimate
Underestimate
Uncertain
Uncertain, perhaps underestimate
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 premature mortality, we assume that there is a "cessation" lag between PM
exposures and the total realization of changes in health effects. We assumed for this analysis
that some of the incidences of premature mortality related to PM2.5 exposures occur in a
distributed fashion 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% of mortality reductions in the first year, 50% over years 2 to 5, and 20% over the years 6 to
20 after the reduction in PM2.5 (U.S. EPA-SAB, 2004c). Additional cessation lag structures are
described and assessed in Appendix 5.A of this RIA. To take this into account in the valuation of
5-55
-------
reductions in premature mortality, we discount the value of premature mortality occurring in
future years using rates of 3% and 7%.17 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 benefits using a 7% discount rate are only approximately 10% less than the
monetized benefits using a 3% discount rate. Further discussion of this topic appears in the
EPA's Guidelines for Preparing Economic Analyses (U.S. EPA, 2010e).
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 has identified valuation of
mortality-related benefits as the largest contributor to the range of uncertainty in monetized
benefits (Mansfield et al., 2009).18 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. 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
17 The choice of a discount rate, and its associated conceptual basis, is a topic of ongoing discussion within the
federal government. To comply with 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.
18 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.
5-56
-------
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 the PM NAAQS 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 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
5-57
-------
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).
5-58
-------
5.6.5.2 Nonfatal Myocardial Infarctions Valuation
We were not able to identify a suitable WTP value for reductions in the risk of nonfatal
heart attacks. 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 5-11):
• 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 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.
5-59
-------
Table 5-11. Alternative Direct Medical Cost of Illness Estimates for Nonfatal Heart Attacks3
Study Direct Medical Costs (2010$) Over an x-Year Period, for x =
Wittels et al. (1990) $160,000b 5
Russell et al. (1998) $33,000° 5
Average (5-year) costs $98,000 5
Eisenstein et al. (2001) $74,000° 10
All estimates rounded to two significant digits. Unrounded estimates in 2000$ are available in appendix J of the
BenMAP user manual (Abt Associates, 2012).
Wittels et al. (1990) did not appear to discount costs incurred in future years.
° 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 $85,000, and added it to the 5-year opportunity cost estimate. The
resulting estimates are given in Table 5-12.
Table 5-12. Estimated Costs Over a 5-Year Period of a Nonfatal Myocardial Infarction (in
2010$)a
Age Group
0-24
25-44
45-54
55-65
>65
Opportunity Cost
$0
$12,000°
$17,000°
$100,000°
$0
Medical Costb
$98,000
$98,000
$98,000
$98,000
$98,000
Total Cost
$98,000
$110,000
$120,000
$200,000
$98,000
All estimates rounded to two significant digits, so estimates may not sum across columns. Unrounded estimates
in 2000$ 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).
From Cropper and Krupnick (1990), using a 3% discount rate for illustration.
5.6.6 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
5-60
-------
Classification of Diseases (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 year 2010$ using the CPI-U
"all items" (Abt Associates, 2012). The resulting national average lost daily wage is $148. The
total cost-of-illness estimate for an ICD code-specific hospital stay lasting n days, then, was the
mean hospital charge plus $148 multiplied by n. In general, the mean length of stay has
decreased since the 2000 database used in previous version of BenMAP while the mean
hospital charge has increased. We provide the rounded unit values in 2010$ for the COI
functions used in this analysis in Table 5-13.
Table 5-13. Unit Values for Hospital Admissions
Age Range
Total Cost of Illness
End Point
HA, Chronic Lung Disease
HA, Asthma
HA, All Cardiovascular
HA, All Cardiovascular
HA, All Respiratory
ICD Codes
490^196
493
390-429
390-429
460-519
min.
18
0
18
65
65
max.
64
64
64
99
99
Mean Hospital
Charge (2010$)
$19,000
$14,000
$40,000
$37,000
$31,000
Mean Length
of Stay (days)
3.9
3.0
4.1
4.9
6.1
(unit value in
2010$)
$21,000
$16,000
$42,000
$41,000
$36,000
* All estimates rounded to two significant digits. Unrounded estimates in 2000$ 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 $155; in
2010$, that cost was $464 (using the CPI-U for medical care to adjust to 2010$). The second
estimate comes from Stanford et al. (1999), who reported the cost of an average asthma-
related emergency department visit at $335, based on 1996-1997 data. A simple average of the
two estimates yields a unit value of $388.
5-61
-------
5.6.7 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 $68 (2010$). Although Ostro
and Rothschild (1989) statistically linked ozone and minor restricted activity days, it is likely that
most MRADs associated with ozone and PM2.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.
5.6.8 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
empirical evidence that the income elasticity19 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
19 Income elasticity is a common economic measure equal to the percentage change in WTP for a 1% change in
income.
5-62
-------
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 have been 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 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 2013. 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 2020 is presented below.
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,
5-63
-------
WTP is more elastic with respect to income. The elasticity values used to adjust estimates of
benefits in 2020 are presented in Table 5-14.20
Table 5-14. Elasticity Values Used to Account for Projected Real Income Growth3
Benefit Category Central Elasticity Estimate
Minor Health Effect 0.14
Severe and Chronic Health Effects 0.45
Premature Mortality 0.40
3 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 2020 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 2020, 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 2010.21 We used projections of
real GDP (in chained 1996 dollars) provided by Standard and Poor's (2000) for the years 2010 to
2020.22
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 5-15. 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
20 We expect that the WTP for improved visibility in Class 1 areas would also increase with growth in real income
(see Chapter 6).
21 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.
22 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.
5-64
-------
growth over the 20-year cessation lag. Our approach could underestimate the benefits for the
later years of the lag.
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 underpredict 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 5-15. Adjustment Factors Used to Account for Projected Real Income Growth3
Benefit Category 2020
Minor Health Effect 1.07
Severe and Chronic Health Effects 1.22
Premature Mortality 1.20
Based on elasticity values reported in Table 5-3, U.S. Census population projections, and projections of real GDP
per capita.
5.7 Benefits Results
5.7.1 Benefits of the Revised and Alternative Annual Primary PM2.s Standards
Applying the impact and valuation functions described previously in this chapter to the
estimated changes in PM2.5 yields estimates of the changes in physical damages (e.g.,
premature mortalities, cases of acute bronchitis and hospital admissions) and the associated
monetary values for those changes. Not all known 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 PM-related health benefits plus B, the sum of the non-
5-65
-------
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.
Table 5-16 shows the population-weighted air quality change for the alternative
standards averaged across the continental U.S. Tables 5-17 through 5-22 present the benefits
results for the annual PM2.5 standards. These benefits are relative to a 2020 analytical baseline
reflecting attainment nationwide of the current primary PM2.5 standards (i.e., 15/35) that
includes promulgated national regulations and illustrative emission controls to simulate
attainment with 15/35 as well as an adjustment to NOX emissions to reflect expected reductions
in mobile NOX emissions between 2020 and 2025.23 Figure 5-3 graphically displays the total
monetized benefits of the revised annual primary standard of 12 u.g/m3, using alternative
concentration-response functions at discount rates of 3% and 7%. Figure 5-4 graphically
displays the cumulative distribution of total monetized benefits using the 2 epidemiology-
derived and the 12 expert-derived relationships between PM2.5and mortality for the revised
standard, which provides the full range of uncertainty within and across the expert-derived
relationships.
Table 5-16. Population-Weighted Air Quality Change for the Revised and Alternative Annual
Primary PM2.s Standards Relative to Analytical Baseline
Standard Population-Weighted Air Quality Change
13 |Jg/m3 0.014 ng/m3
12 |Jg/m3 0.043 |Jg/m3
11 |jg/m3 0.207 ng/m3
23 The estimates in this chapter reflect incremental emissions reductions from an analytical baseline that gives "an
adjustment" to the San Joaquin and South Coast areas in California for NOx emissions reductions expected to
occur between 2020 and 2025, when those areas are expected to demonstrate attainment with the revised
standards. Full benefits of the revised standards in those two areas will not be realized until 2025.
5-66
-------
Table 5-17. Emission Reductions in Illustrative Emission Reduction Strategies for the Revised
and Alternative Annual Primary PM2.s Standards, by Pollutant and Region in 2020
(tons)3
Directly emitted PM2.5
East
West
CA
S02
East
West
CA
NOx
East
West
CA
13 |Jg/m3
0
0
730
0
0
0
0
0
0
12 u.g/m3
0
0
4,000
0
0
0
0
0
0
11 u.g/m3
8,200
160
10,600
21,000
43
0
9
0
0
See Chapter 4 for more information on the illustrative emission reduction strategies. The emissions in this table
reflect both known and unknown controls.
5-67
-------
Table 5-18. Estimated number of Avoided PM2.s Health Impacts for the Revised and
Alternative Annual Primary PM2.s Standards (Incremental to the Analytical
Baseline)3
Revised and Alterative Annual Standards
(95th percentile confidence interval)
Health Effect
Avoided Mortality
Krewski et al. (2009)
(adult mortality age 30+)
Lepeule et al. (2012)
(adult mortality age 25+)
Woodruff etal. (1997)
(infant mortality)
Avoided Morbidity
Non-fatal heart attacks
Peters et al. (2001)
(age >18)
Pooled estimate of 4 studies
(age >18)
Hospital admissions— respiratory
(all ages)b
Hospital admissions— cardiovascular
(age > 18)
Emergency department visits for
asthma (all ages)b
Acute bronchitis
(ages 8-12)b
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)
13 ng/m3
140
(100-190)
330
(180-480)
0
(0-1)
160
(49-260)
17
(8-41)
31
(-9-58)
43
(20-76)
67
(-22-140)
280
(-36-580)
3,500
(1500-5500)
5,100
(1300-8900)
13,000
(270-81000)
22,000
(19000-25000)
130,000
(110,000-150,000)
12 ng/m3
460
(320-590)
1,000
(560-1,500)
1
(1-2)
480
(150-800)
52
(24-130)
110
(-30-200)
140
(66-240)
230
(-74-470)
870
(-110-1,800)
11,000
(4,900-17,000)
16,000
(4,100-28,000)
40,000
(850-250,000)
71,000
(61,000-81,000)
420,000
(350,000-490,000)
11 ng/m3
1,500
(1,000-1,900)
3,300
(1,800-4,800)
4
(2-6)
1,600
(480-2,600)
170
(78-410)
380
(-100-720)
480
(230-0,840)
810
(-260-1,600)
2,700
(-350-5,500)
34,000
(15,000-53,000)
49,000
(12,000-86,000)
120,000
(2,600-770,000)
230,000
(190,000-260,000)
1,300,000
(1,100,000-1,600,000)
All incidence estimates are rounded to whole numbers with a maximum of two significant digits. These
estimates reflect incremental emissions reductions from an analytical baseline that gives "an adjustment" to the
San Joaquin and South Coast areas in California for NOx emissions reductions expected to occur between 2020
and 2025, when those areas are expected to demonstrate attainment with the revised standards. Full benefits of
the revised standards in those two areas will not be realized until 2025. Additional health endpoints, such as
cardiovascular emergency department visits, are only quantified in a sensitivity analysis in Table 5.A-6 because
we do not yet have a valuation estimate for this endpoint.
The negative estimates at the 5th percentile confidence estimates for these morbidity endpoints reflect the
statistical power of the study used to calculate these health impacts. These results do not suggest that reducing
air pollution results in additional health impacts.
5-68
-------
Table 5-19. Monetized PM2.s Health Benefits for the Revised and Alternative Annual Primary
PM2.5 Standards (Incremental to Analytical Baseline) (Millions of 2006$, 3%
discount rate)3
Revised and Annual Standards
(95th percentile confidence interval)
Health Effect
Avoided Mortality13
Krewski et al. (2009)
(adult mortality age 30+)
Lepeule et al. (2012)
(adult mortality age 25+)
Woodruff etal. (1997)
(infant mortality)
Avoided Morbidity
Non-fatal heart attacks
Peters et al. (2001)
(age >18)
Pooled estimate of 4 studies
(age >18)
Hospital admissions— respiratory
(all ages)0
Hospital admissions— cardiovascular
(age > 18)
Emergency department visits for asthma
(all ages)
Acute bronchitis
(ages 8-12)c
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)
13 ng/m
$1,300
($120-$3,500)
$2,900
($250-$8,100)
$3.4
($0.29-$10)
$18
($3.0-$46)
$2.0
($0.43-$6.8)
$0.86
(-$0.22-$1.6)
$1.70
($0.85-$2.8)
$0.03
(-$0.0052-$0.061)
$0.13
(-$0.0060-$0.37)
$0.08
($0.025-$0.10)
$0.17
($0.038-$0.42)
$0.7
($0.027-$5.2)
$3.3
($2.90-$3.70)
$8.8
($4.70-$13)
12 ng/m3
$4,000
($370-$11,000)
$9,000
($800-$26,000)
$11
($0.91-$32)
$55
($9.1-$140)
$6.0
($1.3-$21)
$3.0
(-$0.8-$5.5)
$5.3
($2.7-$9.2)
$0.10
(-$0.018-$0.21)
$0.42
(-$0.019-$1.2)
$0.24
($0.078-$0.47)
$0.54
($0.12-$1.30)
$2.30
($0.085-$16)
$11
($9.4-$12)
$29
($15-$43)
11 ng/m
$13,000
($1,200-$35,000)
$29,000
($2,600-$82,000)
$35
($3.0-$100)
$180
($31-$460)
$20.0
($4.4-$68)
$11
(-$2.7-$20)
$18
($10-$32)
$0.34
(-$0.063-$0.73)
$1.30
(-$0.059-$3.5)
$0.71
($0.24-$1.4)
$1.6
($0.36-$4.0)
$7.0
($0.26-$49)
$35
($30-$39)
$91
($48-$ 140)
All estimates are rounded to two significant digits. 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. These estimates reflect incremental emissions reductions from an
analytical baseline that gives "an adjustment" to the San Joaquin and South Coast areas in California for NOx emissions
reductions expected to occur between 2020 and 2025, when those areas are expected to demonstrate attainment with the
revised standards. Full benefits of the revised standards in those two areas will not be realized until 2025.
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.
The negative estimates at the 5th percentile confidence estimates for this morbidity endpoint reflects the statistical power of
the study used to calculate these health impacts. These results do not suggest that reducing air pollution results in additional
health impacts.
5-69
-------
Table 5-20. Monetized PM2.s Health Benefits for the Revised and Alternative Annual Primary
PM2.5 Standards (Incremental to Analytical Baseline) (Millions of 2006$, 7%
discount rate)3
Revised and Alterative Annual Standards
(95th percentile confidence interval)
Health Effect
Avoided Mortality13
Krewski et al. (2009)
(adult mortality age 30+)
Lepeule et al. (2012)
(adult mortality age 25+)
Woodruff etal. (1997)
(infant mortality)
Avoided Morbidity
Non-fatal heart attacks
Peters et al. (2001)
(age >18)
Pooled estimate of 4 studies
(age >18)
Hospital admissions— respiratory
(all ages)0
Hospital admissions— cardiovascular
(age > 18)
Emergency department visits for asthma
(all ages)
Acute bronchitis
(ages 8-12)c
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)
13 ng/rn
$1,100
($110-$3,100)
$2,600
($230-$7,300)
$3.4
($0.29-$100)
$18
($2.8-$46)
$1.9
($0.40-$6.7)
$0.86
(-$0.22-$1.6)
$1.70
($0.85-$2.8)
$0.03
(-$0.0052-$0.061)
$0.13
(-$0.0060-$0.37)
$0.08
($0.025-$0.10)
$0.17
($0.038-$0.42)
$0.7
($0.027-$5.2)
$3.3
($2.90-$3.70)
$8.8
($4.70-$13)
12 ng/m
$3,600
($330-$9,800)
$8,100
($720-$23,000)
$11
($0.91-$32)
$54
($8.4-$140)
$5.9
($1.2-$20)
$3.0
(-$0.8-$5.5)
$5.3
($2.7-$9.2)
$0.10
(-$0.018-$0.21)
$0.42
(-$0.019-$1.2)
$0.24
($0.078-$0.47)
$0.54
($0.12-$1.30)
$2.30
($0.085-$16)
$11
($9.4-$12)
$29
($15-$43)
11 ng/m3
$11,000
($l,100-$3 1,000)
$26,000
($2,300-$74,000)
$35
($3.0-$100)
$180
($28-$450)
$20
($4.1-$67)
$11
(-$2.7-$20)
$18
($10-$32)
$0.34
(-$0.063-$0.73)
$1.3
(-$0.059-$3.5)
$0.71
($0.24-$1.4)
$1.6
($0.36-$4.0)
$7.0
($0.26-$49)
$35
($30-$39)
$91
($48-$ 140)
All estimates are rounded to two significant digits. 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. These estimates reflect incremental emissions reductions from an
analytical baseline that gives "an adjustment" to the San Joaquin and South Coast areas in California for NOx emissions
reductions expected to occur between 2020 and 2025, when those areas are expected to demonstrate attainment with the
revised standards. Full benefits of the revised standards in those two areas will not be realized until 2025.
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.
The negative estimates at the 5th percentile confidence estimates for this morbidity endpoint reflects the statistical power of
the study used to calculate these health impacts. These results do not suggest that reducing air pollution results in additional
health impacts.
5-70
-------
Table 5-21. Total Estimated Monetized Benefits of the for Revised and Alternative Annual
Primary PM2.s Standards (Incremental to the Analytical Baseline) (billions of
2006$) a'b
Benefits Estimate 13 |ig/m3 12 |ig/m3 11 |ig/m
Economic value of avoided PM2.5.related morbidities and premature deaths using PM2.5 mortality estimate from
Krewskietal. (2009)
3% discount rate $1.3+B $4.0+B $13+B
7% discount rate $1.2 + B $3.6+B $12+B
Economic value of avoided PM2.5.related morbidities and premature deaths using PM2.5 mortality estimate from
Lepeuleetal. (2012)
3% discount rate $2.9 + B $9.1+B $29+B
7% discount rate $2.6 + B $8.2+B $26+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 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.
b These estimates reflect incremental emissions reductions from an analytical baseline that gives "an adjustment"
to the San Joaquin and South Coast areas in California for NOx emissions reductions expected to occur between
2020 and 2025, when those areas are expected to demonstrate attainment with the revised standards. Full
benefits of the revised standards in those two areas will not be realized until 2025.
Table 5-22. Regional Breakdown of Monetized Benefits Results
Revised and Alterative Annual Standards
Region
East3
California15
Rest of West
13 ug/m3
0%
100%
0%
12 ug/m3
0%
100%
0%
11 ug/m3
23%
77%
<1%
Includes Texas and those states to the north and east. Several recent rules such as MATS and CSAPR will have
substantially reduced PM2.5 levels by 2020 in the East, thus few additional controls would be needed to reach 12
For 12 and 13 u.g/m3, all of the benefits occur in California because this highly populated area is where the most
air quality improvement beyond the analytical baseline is needed to reach these levels. These estimates reflect
incremental emissions reductions from an analytical baseline that gives "an adjustment" to the San Joaquin and
South Coast areas in California for NOx emissions reductions expected to occur between 2020 and 2025, when
those areas are expected to demonstrate attainment with the revised standards. Full benefits of the revised
standards in those two areas will not be realized until 2025.
5-71
-------
$14,000
$12,000
$10,000
= $6,000
i
$4,000
$2,000
$0
3% Discount Rate
• 7% Discount Rate
Krewski et al
PM2.5 mortality benefits estimates derived from 2 epidemiology functions and 12 expert functions
Figure 5-3. Estimated PM2.5-Related Premature Mortalities Avoided According to
Epidemiology or Expert-Derived PM2.s Mortality Risk Estimate for 12 u.g/m3 in 2020a
aThese estimates reflect incremental emissions reductions from an analytical baseline that gives "an adjustment"
to the San Joaquin and South Coast areas in California for NOx emissions reductions expected to occur between
2020 and 2025, when those areas are expected to demonstrate attainment with the revised standards. Full
benefits of the revised standards in those two areas will not be realized until 2025.
5-72
-------
100%
90%
80%
70%
CO
•a
o
o 60%
o
50%
40%
.2
i 30%
20%
10%
ExpertA
Experts
ExpertC
Expert D
Expert E
Expert F
ExpertG
Expert H
Expert I
ExpertJ
Expert K
Expert L
Krewskietal.
Lepeuleetal.
$5
$10 $15 $20
Monetized Benefits (Billions of 2010$)
$25
$30
Figure 5-4. Total Monetized Benefits Using 2 Epidemiology-Derived and 12-Expert Derived
Relationships Between PM2.s and Premature Mortality for 12 u.g/m3 in 2020a
These estimates reflect incremental emissions reductions from an analytical baseline that gives "an adjustment"
to the San Joaquin and South Coast areas in California for NOx emissions reductions expected to occur between
2020 and 2025, when those areas are expected to demonstrate attainment with the revised standards. Full
benefits of the revised standards in those two areas will not be realized until 2025.
5.7.2 Uncertainty in Benefits Results
Mortality benefits account for 98% of total monetized benefits, in part because we are
unable to quantify most of the non-health benefits. The next largest benefit is for reductions in
chronic illness (nonfatal heart attacks), although this value is more than an order of magnitude
lower than for premature mortality. Hospital admissions for respiratory and cardiovascular
causes, MRADs and work loss days account for the majority of the remaining benefits. The
remaining categories each account for a small percentage of total benefit; however, they
represent a large number of avoided incidences affecting many individuals. A comparison of the
incidence table to the monetary benefits table reveals that there is not always a close
correspondence between the number of incidences avoided for a given endpoint and the
monetary value associated with that endpoint. For example, we estimate almost 1,000 times
more work loss days would be avoided than premature mortalities, yet work loss days account
5-73
-------
for only a very small fraction of total monetized benefits. 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.
PM2.5 mortality benefits represent a substantial proportion of total monetized benefits
(over 98% in this analysis), and these estimates have the following key assumptions and
uncertainties.
1. We assume that all fine particles, regardless of their chemical composition, are
equally potent in causing premature mortality. This is an important assumption,
because PM2.5 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 PM2.5 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).
2. 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 PM2.5, including
both areas that do not meet the fine particle standard and those areas that are
in attainment, down to the lowest modeled concentrations.
3. 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 PM2.5 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.
Given that reductions in premature mortality dominate the size of the overall monetized
benefits, more focus on uncertainty in mortality-related benefits gives us greater confidence in
our uncertainty characterization surrounding total benefits.
5.7.3 Estimated Life Years Gained and Reduction in the Percentage of Deaths A ttributable to
PM2.5
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
5-74
-------
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 "[f]rom 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 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 years that might be gained. We also estimate the reduction in the percentage of
deaths attributed to PM2.5 resulting from the illustrative emission reduction strategies to reach
the revised annual primary standard. The EPA included similar estimates of life years gained in a
previous assessment of PM2.5 benefits (U.S. EPA, 2006a, 2011a), 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 at the outset. Life expectancy
varies by age. 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).
5.7.3.1 Estimated Life Years Gained
For estimating the potential life years gained by reducing exposure to PM2.5 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. Hubbell (2006)
estimated that reducing exposure to PM2.5from air pollution regulations may result in an
average gain of 15 years of life for those adults prematurely dying from PM2.5 exposure. In
contrast, Pope et al. (2009) estimated changes in average life expectancy at birth over a twenty
year period, suggesting that reducing exposure to air pollution may increase average life
5-75
-------
expectancy at birth by approximately 7 months, which was 15% of the overall increase in life
expectancy at birth from 1980 through 2000. These results are not necessarily inconsistent
because they are reporting different metrics. Because life expectancy is an average of the entire
population (including both those whose deaths would likely be attributed to PM exposure as
well as those whose deaths would not), average life expectancy changes associated with PM
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 PM exposure.
To estimate the potential distribution of life years gained for population subgroups
defined by the age range at which their reduction in PM2.s exposure is modeled to occur we use
standard life tables available from the CDC (2003) and the following formula:
Total Life Years = I?=1L£j x Mt (5.2)
where LEj is the average remaining life expectancy for age interval i, Mj is the estimated change
in number of deaths in age interval i, and n is the number of age intervals.
To get Mj (the estimate the number of avoided premature deaths attributed to changes
in PM2.s exposure in 2020), 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 5-23
summarizes the modeled number of life years gained by reducing PM2.s exposure to 12 u.g/m3
in 2020. We then calculated the average number of life years gained per avoided premature
mortality. Figure 5-5 shows the potential life years gained as a result of meeting a primary
standard of 12 u.g/m3 in 2020, partitioned by the age when exposure reduction occurred, not
necessarily age at death.
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, but half of the life years would occur in populations younger than 65.
This is because the younger populations have the potential to lose more life years per death
than older populations based on changes in PM2.s exposure in 2020. We estimate that the
average individual who would otherwise have died prematurely from PM exposure would gain
16 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 PM or whether that
5-76
-------
susceptibility was in and of itself caused by PM exposure (for a more complete discussion of
this issue, see Kunzli et al., 2001).
Table 5-23. Sum of Life Years Gained by Age Range Attributed to the Revised Annual Primary
PM2.5 Standard of 12 ug/m3 in 2020a'b
Age Rangeb
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
Krewski et al. (2009) Risk
Coefficient c
—
210
550
1,000
1,600
1,700
1,300
560
7,000
15.0
Lepeule et al. (2012)
Risk Coefficient
610
470
1,200
2,300
3,500
3,900
2,900
1,300
16,000
16.0
a Estimates rounded to two significant figures.
Because we assume that there is a "cessation" lag between PM exposures and the reduction in the risk of
premature death, there is uncertainty regarding the specific ages that people die relative to the change in
exposure.
c The youngest age in the population cohort of this study is 30.
5-77
-------
Age
Calculated using Krewski et al. (2009) risk
coefficientb
Calculated using Lepeule et al. (2012) risk
coefficient
Figure 5-5. Estimated Life Years Gained as a Result of the Revised Annual Primary PM2.s
Standard of 12 u.g/m3 in 2020, Partitioned by the Age When Exposure Reduction Occurred,
Not Necessarily Age at Death3
a As shown in these charts, slightly more than half of the avoided premature deaths occur in populations age 75-
99, but slightly more than half of the avoided life years occur in populations age <65 due to the fact that the
younger populations would lose more life years per death than older populations. Results would be similar for
other standard levels on a percentage basis. Because we assume that there is a "cessation" lag between PM
exposures and the reduction in the risk of premature death, there is uncertainty regarding the specific ages that
people die relative to the change in exposure.
b The youngest age in the population cohort of this study used to estimate PM2.5 mortality incidence is 30.
5.7.3.2 Percent of PM-related Mortality Reduced
To estimate the percentage of all-cause mortality attributed to reduced PM2.5 exposure
in 2020 as a result of the illustrative emission reduction strategies, we use Mj 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 5-24 shows the reduction in all-cause mortality
attributed to reducing PM2.5 exposure to the revised annual standard of 12 u.g/m3 in 2020.
Figure 5-6 shows the percentage of avoided premature deaths attributed to meeting the
revised primary annual standard of 12 u.g/m3 in 2020, partitioned by the age when exposure
reduction occurred, not necessarily age at death.
5-78
-------
Table 5-24. Estimated Reduction in the Percentage of All-Cause Mortality Attributed to the
Revised Annual Primary PM2.s Standard of 12 u.g/m3 in 2020a'b
Age Rangeb
25-29
30-34
35^14
45-54
55-64
65-74
75-84
85-99
Krewski et al. (2009) Risk Coefficient"
—
0.35%
0.32%
0.32%
0.33%
0.33%
0.32%
0.29%
Lepeule et al. (2012) Risk Coefficient
0.80%
0.78%
0.73%
0.73%
0.73%
0.73%
0.73%
0.65%
Rounded to two significant figures. Results would be similar for other standard levels on a percentage basis.
These estimates reflect incremental emissions reductions from an analytical baseline that gives "an adjustment"
to the San Joaquin and South Coast areas in California for NOx emissions reductions expected to occur between
2020 and 2025, when those areas are expected to demonstrate attainment with the revised standards. Full
benefits of the revised standards in those two areas will not be realized until 2025.
b Because we assume that there is a "cessation" lag between PM exposures and the reduction in the risk of
premature death, there is uncertainty regarding the specific ages that people die relative to the change in
exposure.
The youngest age in the population cohort of this study is 30.
The relative distributions of the potential number of life years gained (Figure 5-5) and
the estimated avoided mortalities (Figure 5-6) would be similar across alternative annual
standards. Because reduction in PM exposure is not associated with an immediate
improvement in chronic health conditions, we assume that there is a "cessation" lag between
PM exposures and the reduction in the risk of premature death. There is uncertainty regarding
the specific ages at which people "avoid" death relative to the change in exposure. While the
structure of the lag is uncertain, some studies suggest that most of the premature deaths are
avoided within the first 3 years after the change in exposure, while others are unable to identify
conclusively a critical window of exposure (U.S. EPA, 2004c; Schwartz, 2008; Krewski et al.
2009). These studies did not examine whether the cessation lag was modified by either age at
the time when exposure is reduced or the extent of cumulative lifetime exposure.
5-79
-------
Age
30'34 35-44 Age
.45-54
Calculated using Krewski et al. (2009) risk
coefficientb
Calculated using Lepeule et al. (2012) risk
coefficient.
Figure 5-6. Estimate of Avoided Premature Deaths Attributed to the Revised Annual
Primary PM2.s Standard of 12 u.g/m3 in 2020, Partitioned by the Age When Exposure
Reduction Occurred, Not Necessarily Age at Death3
a As shown in these charts, slightly more than half of the avoided premature deaths occur in populations age 75-
99, but slightly more than half of the avoided life years occur in populations age <65 due to the fact that the
younger populations would lose more life years per death than older populations. Results would be similar for
other standard levels on a percentage basis. Because we assume that there is a "cessation" lag between PM
exposures and the reduction in the risk of premature death, there is uncertainty regarding the specific ages that
people die relative to the change in exposure.
b The youngest age in the population cohort of this study is 30.
5.7.4 Evaluation of Mortality Impacts Relative to Various Concentration Benchmarks
In this analysis, we estimate the number of avoided PM2.5-related deaths occurring due
to PM2.5 reductions down to various PM2.5 concentration benchmarks, including the Lowest
Measured Level (LML) of each long-term PM2.5 mortality study. This analysis is one of several
sensitivities that the EPA has historically performed that characterize the uncertainty associated
with the PM-mortality relationship and the economic value of reducing the risk of premature
death (Roman et al., 2008; U.S. EPA, 2006a, 2011a; Mansfield, 2009).
Our review of the current body of scientific literature indicates that a log-linear no-
threshold model provides the best estimate of PM-related long-term mortality. The PM ISA
(U.S. EPA, 2009b), which was twice reviewed by the EPA's Clean Air Scientific Advisory
Committee (U.S. EPA-SAB, 2009a, 2009b), concluded that the evidence supports the use of a
no-threshold log-linear model while also recognizing potential uncertainty about the exact
5-80
-------
shape of the concentration-response function.24 Consistent with this finding, we estimate
benefits associated with the full range of PM2.5 exposure in conjunction with sensitivity analyses
to recognize the potential uncertainty at lower concentrations. Specifically, we incorporated a
LML assessment, a method the EPA has employed in several recent RIAs (U.S. EPA, 2010g,
2011c, 2011d). In addition, we have incorporated an assessment using specific concentration
benchmarks identified in the EPA's Policy Assessment for Particulate Matter (U.S. EPA, 2011b).
These two approaches summarize the distribution of avoided PM2.5-related mortality
impacts relative to baseline (i.e., pre-rule) annual mean PM2.5 levels. The LML approach
compares the percentage of avoided premature deaths estimated to occur above and below
the minimum observed air quality level of each long-term cohort study we use to quantify PM.
In the air quality benchmark approach, we summarize the impacts occurring at different points
in the distribution of the air quality data used in these same epidemiology studies.
Our confidence in the estimated number of premature deaths avoided (but not in the
existence of a causal relationship between PM and premature mortality) diminishes as we
estimate these impacts in locations where PM2.5 levels are below the LML. This interpretation is
consistent with the Policy Assessment (U.S. EPA, 2011b) and advice from SAB-CASAC during
their peer review (U.S. EPA-SAB, 2010d). As noted in the preamble to the final rule, the Policy
Assessment (U.S. EPA, 2011b) concludes that the range from the 25th to the 10th percentile is a
reasonable range of the air quality distribution below which we start to have appreciably less
confidence in the magnitude of the associations observed in the epidemiological studies. 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 at are used to estimate the benefits. Likewise, we are less confident in
the risk we estimate from simulated PM2.5 concentrations that fall below the bulk of the
observed data in these studies. 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.
For these reasons, we consider the LML as well as one standard deviation below the
mean25 air quality levels when characterizing the distribution of mortality impacts. It is
24 For a summary of the scientific review statements regarding the lack of a threshold in the PM2.5-mortality
relationship, see the Technical Support Document (TSD) entitled Summary of Expert Opinions on the Existence of
a Threshold in the Concentration-Response Function for PM2.5-related Mortality (U.S. EPA, 2010f).
25 A range of one standard deviation around the mean represents approximately 68 percent of normally
distributed data, and, below the mean falls between the 25th and 10th percentiles.
5-81
-------
important to emphasize that "less confidence" does not mean "no confidence." In addition,
while we may have less confidence in the magnitude of the risk, we still have high confidence
that PM2.5 is causally associated with risk at those lower air quality concentrations. To clarify
this concept, Figure 5-7 graphically displays the spectrum of confidence using illustrative
concentration benchmarks from the major epidemiology studies cited in this chapter.
Most Confidence
Mean of PM2.5 data in epidemiology study
1 standard deviation below the mean PM2.5 data in epidemiology study
Below LML of PM2.5 data in epidemiology study (extrapolation)
Less Confidence
Figure 5-7. Relationship between the Size of the PM Mortality Estimates and the PM2.s
Concentration Observed in the Epidemiology Study
Although these types of concentration benchmark analyses (e.g., LML, one standard
deviation below the mean, etc.) provide some insight into the level of uncertainty in the
estimated PM2.5 mortality benefits, the EPA does not view these concentration benchmarks as a
concentration threshold below which we would not quantify health benefits of air quality
improvements Rather, the core benefits estimates reported in this RIA (i.e., those based on
Krewski et al. [2009] and Lepeule et al. [2012]) are the best measures because they reflect the
full range of modeled air quality concentrations associated with the emission reduction
strategies. In reviewing the Policy Assessment, SAB-CASAC confirmed that "[although there is
increasing uncertainty at lower levels, there is no evidence of a threshold (i.e., a level below
which there is no risk for adverse health effects)" (U.S. EPA-SAB, 2010d). In addition, in
reviewing the Costs and Benefits of the Clean Air Act (U.S. EPA, 2011a), the SAB-HES noted that
"[t]his [no-threshold] decision is supported by the data, which are quite consistent in showing
effects down to the lowest measured levels. Analyses of cohorts using data from more recent
years, during which time PM concentrations have fallen, continue to report strong associations
with mortality" (U.S. EPA-SAB, 2010a). Therefore, the best estimate of benefits includes
5-82
-------
estimates below and above these concentration benchmarks but uncertainty is higher in the
magnitude of health benefits estimated at lower concentrations, with the lowest confidence
below the LML. Estimated health impacts reflecting air quality improvements below and above
these concentration benchmarks are appropriately included in the total benefits estimate. In
other words, our higher confidence in the estimated benefits above these concentration
benchmarks should not imply an absence of confidence in the benefits estimated below these
concentration benchmarks.
We estimate that most of the avoided PM-related impacts we estimate in this analysis
occur among populations exposed at or above the LML of the Lepeule et al. (2012) study, while
a majority of the impacts occur at or above the LML of the Krewski et al. (2009) study. We show
the estimated reduction in incidence of premature mortality above and below the LML or air
quality benchmarks of these studies in Tables 5-25, and we graphically display the distribution
of PM2.5-related mortality impacts for the selected standard in Figures 5-8 and 5-9.
Table 5-25. Estimated Reduction in Incidence of Adult Premature Mortality Occurring Above
and Below Various Concentration Benchmarks in the Underlying Epidemiology
Studies3
Allocation of Reduced Mortality Incidence
Revised and
Alternative
Standard Level
13 ug/m3
12 ug/m3
11 ug/m3
Epidemiology
Study
Krewski et al.
(2009)
Lepeule et al.
(2012)
Krewski et al.
(2009)
Lepeule et al.
(2012)
Krewski et al.
(2009)
Lepeule et al.
(2012)
Total Reduced
Mortality
Incidence
140
330
460
1,000
1,500
3,300
Below 1 Std.
Dev.
Below AQ Mean
79
(54%)
N/A
200
(43%)
N/A
690
(47%)
N/A
At or Above 1
Std. Dev. Below
AQMean
66
(46%)
N/A
260
(57%)
N/A
770
(53%)
N/A
Below
LML
6
(4%)
130
(38%)
14
(3%)
310
(30%)
27
(2%)
1,000
(31%)
At or
Above
LML
140
(96%)
200
(62%)
440
(97%)
720
(70%)
1,400
(98%)
2,300
(69%)
Mortality incidence estimates are rounded to whole numbers and two significant digits, so estimates may not
sum across columns. 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 u.g/m3 and one standard deviation below the mean is
11.0 u.g/m3. For Lepeule et al., the LML is 8 u.g/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.
5-83
-------
25%
20%
LMLof Krewskiet
al. (2009) study
LMLof Lepeuleet
al. (2012) study
15%
10%
QJ
f
5%
0%
<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 18-19 19-20
Baseline Annual Mean PM25 Level (ug/m3)
Of total PM2.5-Related deaths avoided for 12 |ig/mB:
97% occur among populations exposed to PM2.5 levels at or above the LML of the Krewski et al. (2009) study.
70% occur among populations exposed to PM2.5 levels at or above the LML of the Lepeule et al. (2012) study.
Figure 5-8. Number of Premature PM2.5-related Deaths Avoided for the Revised Annual
Primary PM2.s Standard of 12 u.g/m3 in 2020 According to the Baseline Level of PM2.s and the
Lowest Measured Air Quality Levels of Each Mortality Study
5-84
-------
100%
90%
80%
LMLof Lepeuele et
al. (2012) study
70%
60%
a.
"
LMLof Krewskiet
al. (2009) study
50%
til
40%
'is 30%
"5
3
u 20%
10%
0% I—
7 8 9 10 11 12 13 14
Baseline Annual Mean PM25 Level (ug/m3)
15
16
17
18
19
20
Of total PM2.5-Related deaths avoided for 12 |ig/m :
97% occur among populations exposed to PM2.5 levels at or above the LML of the Krewski et al. (2009) study.
70% occur among populations exposed to PM2.5 levels at or above the LML of the Lepeule et al. (2012) study.
Figure 5-9. Number of Premature PM2.5-related Deaths Avoided for the Revised Annual
Primary PM2.5 Standard of 12 u.g/m3 in 2020 According to the Baseline Level of PM2.5 and the
Lowest Measured Air Quality Levels of Each Mortality Study
When interpreting these LML and air quality benchmarks results, it is important to
understand that the avoided PM2.5 deaths are estimated to occur from PM2.5 reductions in the
baseline air quality simulation, which assumes that 15/35 is already met. When simulating
attainment with revised 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 PM levels 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 PM2.s concentrations vary spatially within a
county. Therefore, there may be a small number of grid cells with concentrations slightly
greater than 15 u.g/m3 in the gridded baseline even though all monitors could meet an annual
standard of 15 u.g/m3. 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, emission
5-85
-------
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 alternative standard. Emission reduction strategies
designed to reduce PM2.5 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 emission reduction strategies, it is appropriate to include all the benefits occurring as a
result of the emission 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 emission reductions in
exceeding counties. Figure 5-10 provides an illustration of this concept.
Monitors that meet alternative standard eve
* Monitors that exceed alternative standard level
Gridded PM2.5 concentrations
Below alternative standard level
Between alternative standard level and current standard
Above current standard
County boundary
As this illustration shows, because 12km grid cells are much smaller than counties and because PM2.5 concentrations vary within
a county, there can be modeled grid cells with PM2.5 concentrations below an standard level even when the county exceeds
that level. Because we model benefits using grid cells, this is a key reason why the LML graphs show benefits at levels below the
alternative standards. In addition, emission reductions in an exceeding county can have benefits in a neighboring county that
does not exceed.
Figure 5-10. Illustration of Relative Size of County with Exceeding Monitor and Modeled
Grid Cells
5-86
-------
While the LML of each study is important to consider when characterizing and
interpreting the overall level of PM2.5-related benefits, as discussed earlier in this chapter, the
EPA believes that both of the cohort-based mortality estimates are suitable for use in air
pollution health impact analyses. When estimating PM-related premature deaths avoided using
risk coefficients drawn from the Lepeule et al. (2012) analysis of the Harvard Six Cities and the
Krewski et al. (2009) analysis of the ACS cohorts there are innumerable other attributes that
may affect the size of the reported effect estimates—including differences in population
demographics, the size of the cohort, activity patterns and particle composition among others.
The LML assessment presented here provides a limited representation of one key difference
between the two studies.
5.7.5 Additional Sensitivity Analyses
The details of these sensitivity analyses are provided in Appendix 5.A, and summarized
here. The use of an alternate lag structure would change the PM2.5-related mortality benefits
discounted at 3% discounted by between 10% and -27%; when discounted at 7%, these
benefits change by between 22% and -52%. When applying higher and lower income growth
adjustments, the monetary value of PM2.5-related premature mortality changes between 33%
and -14%; the value of acute endpoints changes between 8% and -4%. Using the updated
population projection data, the rounded estimates of total monetized benefits increases by
4.4% for 12 u.g/m3. These sensitivity analysis results would be similar on a percentage basis for
the alternative annual standards.
5.8 Discussion
The analysis in this Chapter demonstrates the potential for significant health benefits of
the illustrative emission controls applied to simulate attainment with the revised annual
primary PM2.5 standard. We estimate that by 2020 the emissions reductions to reach the
revised annual primary standard would have reduced the number of PM2.5-related premature
mortalities and produce substantial non-mortality benefits. This rule promises to yield
significant welfare impacts as well (see Chapter 6), though the quantification of those
endpoints is absent in this RIA. Even considering the quantified and unquantified uncertainties
identified in this chapter, we believe that the implementing the revised annual PM2.5 standard
of 12 u.g/m3 would have substantial public health benefits that outweigh the costs (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
5-87
-------
response relationships and our use of the expert elicitation-derived PM2.5 mortality functions.
Others, including the 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 revised PM NAAQS.
There are important differences worth noting in the design and analytical objectives of
NAAQS RIAs compared to RIAs for implementation rules, such as the recent MATS rule (U.S.
EPA, 2011d). 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, 2011d). 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. Subsequent implementation rules will be
reflected in the baseline for the next PM 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 PM2.5 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 PM2.5 concentrations throughout the entire area. In fact, the
Quantitative Risk and Exposure Assessment for Particulate Matter (U.S. EPA, 2010b) shows how
different standard levels would affect the entire distribution of PM2.5 concentrations, and thus
people's exposures and risk, across 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
5-88
-------
adequate margin of safety, taking into consideration effects on susceptible populations based
on the scientific literature. The risk analysis prepared in support of this PM 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, 2010b). 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 PM2.s concentrations, there is no evidence of a threshold in
PM2.5-related health effects in the epidemiology literature. 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 revised and alternative annual standards are in addition
to the substantial benefits estimated for several recent implementation rules (U.S. EPA, 2009a,
2011c, 2011d, 2011e). Rules such as MATS and other emission reductions will have substantially
reduced ambient PM2.s concentrations by 2020 in the East, such that no additional controls
would be needed to reach 12 u.g/m3 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 PM NAAQS RIA (U.S. EPA, 2006a) for the year 2020 and even
smaller than the mortality risks estimated for the current year in the PM REA (U.S. EPA, 2010b).
In addition, because our analytical baseline excludes benefits associated with attaining the
current annual and daily standards as well as the mobile NOx emissions anticipated by 2025,
including the benefits associated with those air quality improvements would result in higher
benefits than we have estimated here.
For the revised annual standard of 12 u.g/m3, all of the estimated benefits occur in
California because this is the only state that needs additional air quality improvement beyond
the analytical baseline after accounting for the air quality improvements from recent rules.
Because all of the monetized human health benefits are projected to occur in California, we
have considered the cohort studies conducted in California specifically in addition to the
national risk coefficients we use as our core estimates. Although we have not calculated the
benefits results using these cohort studies, we provided the risk coefficients from these
California cohorts to show how much the monetized benefits could have changed if we used
these cohort studies. Most of the California cohort studies report central effect estimates
similar to the (nation-wide) all-cause mortality risk estimate we applied from Krewski et al.
5-89
-------
(2009) and Lepeule et al. (2012) albeit with wider confidence intervals. Three cohort studies
conducted in California indicate statistically significant higher risks than the risk estimates we
applied from Lepeule et al. (2012), and four studies showed insignificant results.
5.9 References
Abbey, David E et al. 1995. "Chronic Respiratory Symptoms Associated with Estimated Long-
Term Ambient Concentrations of Fine Particulates less than 2.5 Microns in aerodynamic
diameter (PM2.5) and Other Air Po\\utar\ts." Journal of Exposure Analysis and
Environmental Epidemiology 5(2): 137-158.
Abbey, D.E., N. Nishino, W.F. McDonnell, R.J. Burchette, S.F. Knutsen, W. Lawrence Beeson, and
J.X. Yang. 1999. "Long-term inhalable particles and other air pollutants related to
mortality in nonsmokers." American Journal of Respiratory and Critical Care Medicine
159(2):373-382.
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 on the Internet at
.
Abt Associates, Inc. 2010. Environmental Benefits and Mapping Program (Version 4.0).
Bethesda, MD. Prepared for U.S. Environmental Protection Agency Office of Air Quality
Planning and Standards. Research Triangle Park, NC. Available on the Internet at
.
Adams PF, Hendershot GE, Marano MA. 1999. "Current Estimates from the National Health
Interview Survey, 1996." Vital Health Stat 10(200):1-212.
Agency for Healthcare Research and Quality (AHRQ). 2000. HCUPnet, Healthcare Cost and
Utilization Project. Rockville, MD. Available on the Internet at
.
Agency for Healthcare Research and Quality (AHRQ). 2007. HCUPnet, Healthcare Cost and
Utilization Project. Rockville, MD. Available on the Internet at
.
Agency for Healthcare Research and Quality (AHRQ). 2009. HCUPnet, Healthcare Cost and
Utilization Project. Rockville, MD. Available on the Internet 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.
5-90
-------
Alberini, Anna & Milan Scasny. 2011. "Context and the VSL: Evidence from a Stated Preference
Study in Italy and the Czech Republic," Environmental & Resource Economics, European
Association of Environmental and Resource Economists 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 on the Internet at .
American Lung Association (ALA). 2010a. Trends in COPD (Chronic Bronchitis and Emphysema):
Morbidity and Mortality. American Lung Association Epidemiology and Statistics Unit
Research and Program Services Division. February. Table 4. Based on NHIS data (CDC,
2008). Available at http://www.lungusa.org/finding-cures/our-research/trend-
reports/copdtrend-report.pdf.
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/finding-cures/our-research/trend-reports/asthma-trend-
report.pdf.
Anderson HR, Bremner SA, Atkinson RW, Harrison RM, Walters S. 2001. "Particulate matter and
daily mortality and hospital admissions in the west midlands conurbation of the UK:
associations with fine and coarse particles, black smoke and sulphate." Occup Environ
Med 58:504-510.
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." Environ Health 6: 9.
Barnett, Adrian G. etal. 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.
5-91
-------
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 on the Internet
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). 2011. United States Life Tables, 2007
National Vital Statistics Reports. Volume 59, Number 9. September. Available on the
Internet at .
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).
Chen L, Jennison BL, Yang W, and Omaye ST. 2000. "Elementary school absenteeism and air
pollution." Inhal Toxicol 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.
Grouse DL, Peters PA, van Donkelaar A, Goldberg MS, Villeneuve PJ, Brion 0, 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."
Environ Health Perspect. In press, http://dx.doi.org/10.1289/ehp.1104049.
Dockery, D.W., C.A. Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, B.C. Ferris, and F.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.
Dominici, F., R.D. Peng, M.L. Bell, L. Pham, A. McDermott, S.L. Zeger, J.M. Samet. 2006. "Fine
Particulate Air Pollution and Hospital Admission for Cardiovascular and Respiratory
Diseases." Journal of the American Medical Association 295:1127-1134.
Eftim SE, Samet JM, Janes H, McDermott A, Dominici F. 2008. "Fine particulate matter and mor-
tality: a comparison of the Six Cities and American Cancer Society cohorts with a
Medicare cohort." Epidemiology 19:209-216.
5-92
-------
Eisenstein, E.L, L.K. Shaw, KJ. 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.
Eisner M, N Anthonisen, D Coultas et al. 2010. "An official American Thoracic Society public
policy statement: novel risk factors and the global burden of chronic obstructive
pulmonary disease," Am J Respir Crit Care Med 182: 693-718.
Fann N, Lamson A, Wesson K, Risley D, Anenberg SC, Hubbell BJ. 2012a. "Estimating the
National Public Health Burden Associated with Exposure to Ambient PM2.5and Ozone.
Risk Analysis," Risk Analysis. 32(1): 81-95.
Fann N, Roman HR, Fulcher CF, Gentile M, Wesson K, Hubbell BJ, Levy Jl. 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.
Freeman (III), AM. 1993. The Measurement of Environmental and Resource Values: Theory and
Methods. Washington, DC: Resources for the Future.
Can, 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."/Amer/con
Journal of Epidemiology. In press, doi: 10.1093/aje/kwr424.
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.O. 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.
Vol 62 (2): 103-108.
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 21(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.
5-93
-------
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 on the Internet at
.
Hubbell, B., A. Hallberg, D.R. McCubbin, and E. Post. 2005. "Health-Related Benefits of Attaining
the 8-Hr Ozone Standard." Environmental Health Perspectives 113:73-82. ICD, 1979
Hubbell BL. 2006. "Implementing QALYs in the analysis of air pollution regulations."
Environmental and Resource Economics 34:365-384.
Industrial Economics, Incorporated (lEc). 2006. Expanded Expert Judgment Assessment of the
Concentration-Response Relationship Between PM2,5 Exposure and Mortality. Prepared
for: Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC. September. Available on the Internet at
.
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.
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.
5-94
-------
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.
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, 0. Chanel, P. Filliger, M. Herry, F. HorakJr., 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.
Kunzli, N., S. Medina, R. Kaiser, P. Quenel, F. HorakJr, 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(ll):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.
Lave, L.B., and E.P. Seskin. 1977. Air Pollution and Human Health. Baltimore: Johns Hopkins
University Press for Resources for the Future.
Lepeule J, Laden F, Dockery D, Schwartz J 2012. "Chronic Exposure to Fine Particles and
Mortality: An Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009."
Environ Health Perspect. Jul;120(7):965-70.
Levy Jl, Baxter LK, Schwartz J. 2009. "Uncertainty and variability in health-related damages from
coal-fired power plants in the United States." Risk Anal. 29(7) 1000-1014.
Lichtenstein, S. and P. Slovic, eds. 2006. The construction of preference. New York: Cambridge
University Press.
Lipfert, F.W., H. Mitchell Perry Jr., J. Philip Miller, Jack D. Baty, Ronald E. Wyzg, and Sharon E.
Carmody. 2000. "The Washington University-EPRI Veterans' Cohort Mortality Study:
Preliminary Results." Inhalation Toxicology 12:41-74.
Lipfert, F.W., S.C. Morris, and R.E. Wyzga. 1989. "Acid Aerosols—the Next Criteria Air
Pollutant." Environmental Science & Technology 23(11):1316-1322.
Lipfert, F.W.; Perry, H.M., Jr.; Miller, J.P.; Baty, J.D.; Wyzga, R.E.; Carmody, S.E. 2003. "Air
Pollution, Blood Pressure, and Their Long-Term Associations with Mortality." Inhalation
Toxicology. 15, 493-512.
5-95
-------
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.
Lippmann M, Ito K, Nadas A, Burnett RT. 2000. "Association of particulate matter components
with daily mortality and morbidity in urban populations." Res Rep Health Efflnst. 95:5-
72, discussion 73-82.
Lisabeth, L.D. et al. 2008. "Ambient air pollution and risk for ischemic stroke and transient
ischemic attack." Annals of Neurology. July (64): 53—59.
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
-------
Mustafic H, Jabre P, Caussin C et al. 2012. "Main Air Pollutants and Myocardial Infarction: A
Systematic Review and Meta-analysis"7/4/W/4. 2012;307(7):713-721.
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 (lEc) Memorandum to Jim DeMocker,
U.S. Environmental Protection Agency, Office of Air and Radiation, Office of Policy
Analysis and Review. March 31. Available on the Internet at
.
Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory Analysis. Washington,
DC. Available on the Internet at
Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens, and M. 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.
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.
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 Particulate Matter and Acute Respiratory
Symptoms among New York City Area Adolescents." Environ Health Perspect.
Sep;118(9):1338-43.
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. 2007. "Ambient air
pollution and cardiovascular emergency department visits in potentially sensitive
groups." Am J Epidemiol 165: 625-633.
5-97
-------
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." Environ Health
Perspect 117(6): 957-63.
Peng, R. D., H. H. Chang, et al. 2008. "Coarse particulate matter air pollution and hospital
admissions for cardiovascular and respiratory diseases among Medicare patients. "JAMA
299(18): 2172-9.
Peters A. 2005. "Particulate matter and heart disease: evidence from epidemiological studies."
ToxicolAppI Pharmacol. Sep 1;207(2 Suppl):477-82.
Peters A. 2006. "When a myocardial infarction comes out of the not-so-blue air." Circulation.
Dec5;114(23):2430-l
Peters, A., D. W. Dockery, J. E. Muller and M. A. Mittleman. 2001. "Increased particulate air
pollution and the triggering of myocardial infarction." Circulation. Vol. 103 (23): 2810-5.
Poloniecki, J.D., R.W. Atkinson., A.P. de Leon., and H.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., 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.
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 Particulate
Air Pollution." Journal of the American Medical Association 287:1132-1141.
Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. "Respiratory Health and
PMio Pollution: A Daily Time Series Analysis." American Review of Respiratory Diseases
144:668-674.
Pope, CA III, E Majid, D Dockery. 2009. "Fine Particle Air Pollution and Life Expectancy in the
United States." New England Journal of Medicine 360: 376-386.
Pope, C. A., Ill, J. B. Muhlestein, et al. 2006. "Ischemic heart disease events triggered by short-
term exposure to fine particulate air pollution." Circulation 114(23): 2443-8.
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.
5-98
-------
Puett RC, Hart JE, Suh H, Mittleman M, Laden F. 2011. "Particulate Matter Exposures, Mortality
and Cardiovascular Disease in the Health Professionals Follow-up Study." Environ Health
Perspect. August; 119(8): 1130-1135.
Puett RC, Hart JE, Yanosky JD, Paciorek C, Schwartz J, Suh H, et al. 2009. "Chronic fine and
coarse particulate exposure, mortality and coronary heart disease in the Nurses' Health
Study." Environ Health Perspect 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 Particulate Matter in the U.S." Environ. Sci. Technol.,
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., Loomes, 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, E.G. Hamel, and S.C. Hartz. 1998. "Direct Medical Costs of
Coronary Artery Disease in the United States." American Journal of Cardiology
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.
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.
5-99
-------
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.
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, J., and T. Stamland. 2002. "Skill and the Value of Itfe." Journal of Political Economy
110:1168-1197.
Slaughter, James C et al. 2005. "Association between particulate matter and emergency room
visits, hospital admissions and mortality in Spokane, Washington." Journal of Exposure
Analysis and Environmental Epidemiology (2005) 15, 153-159.
doi:10.1038/sj.jea.7500382
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.
Standard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter. Available on the
Internet at http://:www.standardandpoors.com
Stanford, R., T. Mclaughlin and L. J. Okamoto. 1999. "The cost of asthma in the emergency
department and hospital." Am J Respir Crit Care Med. 160 (1): 211-5.
Sullivan, J., L. Sheppard, et al. 2005. "Relation between short-term fine-particulate 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." Environ. Sci. Technol. 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." J Expo Sci Environ Epidemiol 17
Suppl 2:529-35.
5-100
-------
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 31(l):73-95.
U.S. Bureau of Economic Analysis (U.S. BEA). 2004. BEA Economic Areas (EAs). U.S. Department
of Commerce. Available on the Internet at
.
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 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 on the
Internet 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 on the Internet 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 on the Internet
at
.
U.S. Environmental Protection Agency (U.S. EPA). 2004. Air Quality Criteria for Particulate
Matter. National Center for Environmental Assessment, Office of Research and
Development, U.S. Environmental Protection Agency, Research Triangle Park, NC
EPA/600/P-99/002bF. Available on the Internet at
.
5-101
-------
U.S. Environmental Protection Agency (U.S. EPA). 2006a. Regulatory Impact Analysis, 2006
National Ambient Air Quality Standards for Particulate 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). 2006b. Air Quality Criteria for Ozone and
Related Photochemical Oxidants (Final). EPA/600/R-05/004aF-cF. Washington, DC: U.S.
EPA. February. Available on the Internet at
http://cfpub.epa.gov/ncea/CFM/recordisplay.cfm?deid=149923.U.S. Environmental
Protection Agency (U.S. EPA). 2008. 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). 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 on the Internet at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=194645
U.S. Environmental Protection Agency (U.S. EPA). 2008c. Integrated Science Assessment for
Sulfur Oxides—Health Criteria (Final Report). National Center for Environmental
Assessment, Research Triangle Park, NC. September. Available on the Internet 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 on the Internet 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—RTP Division. December. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010a. Final Regulatory Impact Analysis (RIA)
for the N02 National Ambient Air Quality Standards (NAAQS). Office of Air Quality
Planning and Standards, Research Triangle Park, NC. January. Available on the Internet
at .
5-102
-------
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 on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2010c. Final Regulatory Impact Analysis (RIA)
for theS02National Ambient Air Quality Standards (NAAQS). Office of Air Quality
Planning and Standards, Research Triangle Park, NC. June. Available on the Internet 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 on the Internet 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 on the Internet
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 on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011a. 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). 2011b. Policy Assessment for the Review of
the Particulate Matter National Ambient Air Quality Standards. EPA-452/D-11-003. April.
Available on the Internet at
.
5-103
-------
U.S. Environmental Protection Agency (U.S. EPA). 2011c. 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 on the
Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2011d. Regulatory Impact Analysis for the
Final Mercury and Air Toxics Standards. EPA-452/R-11-011. December. Available on the
Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2011e. Integrated Science Assessment for
Ozone and Related Photochemical Oxidants (Second External Review Draft). EPA/600/R-
10/076B. U.S. Environmental Protection Agency, National Center for Environmental
Assessment—RTP Division. September. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011e. 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 on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2012a. 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. June. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2012b. Provisional Assessment of Recent
Studies on Health Effects of Particulate Matter Exposure. EPA/600/R-12/056A. National
Center for Environmental Assessment—RTP Division. December.
U.S. Environmental Protection Agency (U.S. EPA). 2012c. Integrated Science Assessment for
Particulate Ozone and Related Photochemical Oxidants (Third External Review Draft).
EPA/600/R-10/076C. National Center for Environmental Assessment—RTP Division.
June. Available on the Internet 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 on the Internet at
.
5-104
-------
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 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
.
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). 2007. SAB
Advisory on EPA's Issues in Valuing Mortality Risk Reduction. EPA-SAB-08-001. October.
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 on the Internet at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2009a.
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 on the Internet at
.
5-105
-------
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2009b. Review
ofEPA''s Integrated Science Assessment for Particulate Matter (First External Review
Draft, December 2008). EPA-COUNCIL-09-008. May. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2009c. Review
of Integrated Science Assessment for Particulate Matter (Second External Review Draft,
July 2009). EPA-CASAC-10-001. November. Available on the Internet 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 on the Internet 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 on the Internet at
.
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010c. 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). 2010d. CASAC
Review of Policy Assessment for the Review of the PM NAAQS—Second External Review
Draft (June 2010). EPA-CASAC-10-015. Available on the Internet 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-11-011 July. Available on the Internet at
.
Viscusi, V.K., and J.E. Aldy. 2003. "The Value of a Statistical Life: A Critical Review of Market
Estimates throughout the \Nor\d." Journal of Risk and Uncertainty 27(l):5-76.
5-106
-------
Viscusi, W.K., W.A. Magat, and J. Huber. 1991. "Pricing Environmental Health Risks: Survey
Assessments of Risk-Risk and Risk-Dollar Trade-Offs for Chronic Bronchitis." Journal of
Environmental Economics and Management 21:32-51.
Wellenius, G.A., M.R. Burger, B.A. Coull, J. Schwartz, H.H. Suh, P. Koutrakis, G. Schlaug, D.R.
Gold, M.A. Mittlemen. 2012."Ambient air pollution and the risk of acute ischemic
stroke." Arc/j. Intern. Med., 172 (3), pp. 229-234.
Wittels, E.H., J.W. Hay, and A.M. Gotto, Jr. 1990. "Medical Costs of Coronary Artery Disease in
the United States." American Journal of Cardiology 65(7):432-440.
Woodruff, T.J., J. Grille, and K.C. Schoendorf. 1997. "The Relationship Between Selected of
postneonatal infant mortality and particulate air pollution in the United States."
Environmental Health Perspectives. Vol. 105 (6): 608-612.
Woodruff TJ, Parker JD, Schoendorf KC. 2006. "Fine particulate matter (PM2.5) air pollution and
selected causes of postneonatal infant mortality in California." Environmental Health
Perspectives 114(5):786-90.
Woods & Poole Economics Inc. 2007. Population by Single Year of Age. CD-ROM. Woods &
Poole Economics, Inc. Washington, D.C.
World Health Organization (WHO). 1977. International Classification of Diseases, 9th Revision
(ICD-9). Geneva: WHO.
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
.
Zanobetti, A., M. Franklin, et al. 2009. "Fine particulate air pollution and its components in
association with cause-specific emergency admissions." Environmental Health 8: 58-
60.Zeger S; Dominici F; McDermott A; Samet J. 2008. "Mortality in the Medicare
population and chronic exposure to fine particulate air pollution in urban centers (2000-
2005)." Environ Health Perspect 116:1614-1619.
Zanobetti A. and Schwartz, J. 2006. "Air pollution and emergency admissions in Boston, MA." J
Epidemiol Community Health 60(10): 890-5.
5-107
-------
APPENDIX 5.A
ADDITIONAL SENSITIVITY ANALYSES RELATED TO THE HEALTH BENEFITS ANALYSIS
The 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 approved
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 supplements our main analysis of benefits with five additional sensitivity
calculations. The supplemental estimates examine sensitivity to assumptions about both
physical effects (i.e., the structure of the cessation lag; estimates of the number of avoided
cerebrovascular events, cardiovascular emergency department visits and cases of chronic
bronchitis; and alternate effect estimates for cohorts in California) and valuation issues (i.e., the
appropriate income elasticity, updated cost-of-illness estimates). We conducted these
sensitivity analyses for the selected annual standard of 12 u.g/m3 as an illustrative example.
These supplemental estimates are not meant to be comprehensive. Rather, they reflect some
of the key issues identified by EPA or commenters as likely to have a significant impact on total
benefits, or they are health endpoints for which the health data are still evolving, or for which
we lack an appropriate method to estimate the economic value. The individual income growth
and lag adjustments in the tables should not simply be added together because (1) there may
be overlap among the alternative assumptions, and (2) the joint probability among certain sets
of alternative assumptions may be low.
5.A.1 Cessation Lag Structure for PM2.5-Related Premature Mortality
Based in part on prior advice from the EPA's independent Science Advisory Board (SAB),
EPA typically assumes that there is a time lag between reductions in particulate matter (PM)
exposures in a population and the full realization of reductions in premature mortality. Within
the context of benefits analyses, this term is often referred to as "cessation lag." The existence
of such a lag is important for the valuation of reductions in premature mortality because
economic theory suggests that dollar-based representations of health effect incidence changes
occurring in the future should be discounted to reflect time preferences for consumption in the
5.A-1
-------
population (e.g., people generally prefer to consume now rather than later and will generally
give up greater consumption in the future for earlier consumption).
Over the last 15 years, there has been a continuing discussion and evolving advice
regarding the timing of changes in health effects following changes in ambient air pollution. It
has been hypothesized that some reductions in premature mortality from exposure to ambient
PM2.5 will occur over short periods of time in individuals with compromised health status, but
other effects are likely to occur among individuals who, at baseline, have reasonably good
health that will deteriorate because of continued exposure. The SAB-HES has recognized this
lack of direct evidence. However, in early advice, they also note that "although there is
substantial evidence that a portion of the mortality effect of PM is manifest within a short
period of time, i.e., less than one year, it can be argued that, if no lag assumption is made, the
entire mortality excess observed in the cohort studies will be analyzed as immediate effects,
and this will result in an overestimate of the health benefits of improved air quality. Thus some
time lag is appropriate for distributing the cumulative mortality effect of PM in the population"
(EPA-SAB-COUNCIL-ADV-00-001,1999, p. 9). In more recent advice, the SAB-HES suggests that
appropriate lag structures may be developed based on the distribution of cause-specific deaths
within the overall all-cause estimate (EPA-SAB-COUNCIL-ADV-04-002, 2004). They suggest that
diseases with longer progressions should be characterized by longer-term lag structures, while
air pollution impacts occurring in populations with existing disease may be characterized by
shorter-term lags.
A key question is the distribution of causes of death within the relatively broad
categories analyzed in the long-term cohort studies. Although it may be reasonable to assume
the cessation lag for lung cancer deaths mirrors the long latency of the disease, it is not at all
clear what the appropriate lag structure should be for cardiopulmonary deaths, which include
both respiratory and cardiovascular causes. Some respiratory diseases, such as chronic
obstructive pulmonary disease, may have a long period of progression, while others, such as
pneumonia, have a very short duration. In the case of cardiovascular disease, there is an
important question of whether air pollution is causing the disease, which would imply a
relatively long cessation lag, or whether air pollution is causing premature death in individuals
with preexisting heart disease, which would imply very short cessation lags (in theory, air
pollution may both cause cardiovascular disease and cause premature death in those with
preexisting cardiovascular disease). The SAB-HES provides several recommendations for future
research that could support the development of defensible lag structures, including using
disease-specific lag models and constructing a segmented lag distribution to combine
5.A-2
-------
differential lags across causes of death (EPA-SAB-COUNCIL-ADV-04-002, 2004). The SAB-HES
indicated support for using "a Weibull distribution or a simpler distributional form made up of
several segments to cover the response mechanisms outlined above, given our lack of
knowledge on the specific form of the distributions" (EPA-SAB-COUNCIL-ADV-04-002, 2004, p.
24). However, they noted that "an important question to be resolved is what the relative
magnitudes of these segments should be, and how many of the acute effects are assumed to be
included in the cohort effect estimate" (EPA-SAB-COUNCIL-ADV-04-002, 2004, p. 24-25). Since
the publication of that report in March 2004, EPA has sought additional clarification from this
committee. In its follow-up advice provided in December 2004, this SAB suggested that until
additional research has been completed, EPA should assume a segmented lag structure
characterized by 30% of mortality reductions occurring in the first year, 50% occurring evenly
over years 2 to 5 after the reduction in PM2.5, and 20% occurring evenly over the years 6 to 20
after the reduction in PM2.5 (EPA-COUNCIL-LTR-05-001, 2004). The distribution of deaths over
the latency period is intended to reflect the contribution of short-term exposures in the first
year, cardiopulmonary deaths in the 2- to 5-year period, and long-term lung disease and lung
cancer in the 6- to 20-year period. Furthermore, in their advisory letter, the SAB-HES
recommended that EPA include sensitivity analyses on other possible lag structures. In this
appendix, we investigate the sensitivity of premature mortality-reduction related benefits to
alternative cessation lag structures, noting that ongoing and future research may result in
changes to the lag structure used for the main analysis.
In previous advice from the SAB-HES, they recommended an analysis of 0-, 8-, and 15-
year lags, as well as variations on the proportions of mortality allocated to each segment in the
segmented lag structure (EPA-SAB-COUNCIL-ADV-00-001, 1999, (EPA-COUNCIL-LTR-05-001,
2004). The 0-year lag is representative of EPA's assumption in previous RIAs. The 8- and 15-year
lags are based on the study periods from the Pope et al. (1995) and Dockery et al. (1993)
studies, respectively.1 However, neither the Pope et al. nor Dockery et al. studies assumed any
lag structure when estimating the relative risks from PM exposure. In fact, the Pope et al. and
Dockery et al. analyses do not supporting or refute the existence of a lag. Therefore, any lag
structure applied to the avoided incidences estimated from either of these studies will be an
assumed structure. The 8- and 15-year lags implicitly assume that all premature mortalities
occur at the end of the study periods (i.e., at 8 and 15 years).
1 Although these studies were conducted for 8 and 15 years, respectively, the choice of the duration of the study
by the authors was not likely due to observations of a lag in effects but is more likely due to the expense of
conducting long-term exposure studies or the amount of satisfactory data that could be collected during this time
period.
5.A-3
-------
In addition to the simple 8- and 15-year lags, we have added several additional
sensitivity analyses examining the impact of assuming different allocations of mortality to the
segmented lag of the type suggested by the SAB-HES. The first alternate lag structure assumes
that more of the mortality impact is associated with chronic lung diseases or lung cancer and
less with acute cardiopulmonary causes. This illustrative lag structure ("alternate segmented")
is characterized by 20% of mortality reductions occurring in the first year, 50% occurring evenly
over years 2 to 5 after the reduction in PM2.5, and 30% occurring evenly over the years 6 to 20
after the reduction in PM2.5. The second alternate lag structure ("5-year distributed") assumes
the 5-year distributed lag structure used in previous analyses, which is equivalent to a three-
segment lag structure with 50% in the first 2-year segment, 50% in the second 3-year segment,
and 0% in the 6- to 20-year segment. The third alternate lag structure assumes a smooth
negative exponential relationship between the reduction in exposure and the reduction in
mortality risk, which is described in more detail below.
In 2004, SAB-HES (U.S. EPA-SAB, 2004) urged EPA to consider using smoothed lag
distributions, incorporating information from the smoking cessation literature. In June 2010,
the SAB-HES provided additional advice regarding alternate cessation lags (U.S. EPA-SAB, 2010).
For PM2.5-related benefits, the SAB-HES continued to support the previous 20-year distributed
lag as the main estimate, while recommending that EPA further examine additional exponential
decay functions. Specifically, the SAB-HES suggested varying the rate constant with the risk
coefficient from in the cohort studies. EPA intends to incorporate these new alternate cessation
lag for PM2.5-related benefits in the final PM NAAQS RIA.
In response to these suggestions, EPA identified Roosli et al. (2005) as model that
combines empirical data on the relationship between changes in exposure and changes in
mortality and the timing of the cessation of those effects for the smooth decay function.2
Because an exponential model is often observed in biological systems, Roosli et al. (2005)
developed a dynamic model that assumes that mortality risks decrease exponentially after
exposure termination. This model assumes the form risk=exp"kt, where k is the time constant
and t is the time after t0. The relative risk from air pollution (RR) at a given time (t) can be
calculated from the excess relative risk (ERR) attributable to air pollution from PM cohort
studies (ERR=RR-R0), as follows:
2 In the 2006 PM NAAQS RIA (U.S. EPA, 2006), EPA applied equations and the time constant from a conference
presentation by Roosli et al. (2004). We have updated this sensitivity analysis in this assessment to reflect the
published version in Roosli et al. (2005) and generated additional time constants.
5.A-4
-------
flfl(t) = ERR X exp~kt + R0 (5.A.1)
where R0 is the baseline relative risk in the absence of air pollution (Ro=l). After cessation of
exposure, mortality will start to decline and approach the baseline level. The change in
mortality (AM), in units of percent-years, can be derived from Equation (5. A.I) as follows:
AM = ERRx t-f*ERR Xexp~ktdt (5.A.2)
Integrating Equation (5.A.2) gives:
-kt (5.A.3)
K. K.
In order to calculate values for the time constant, k, we applied the AM values from the
two intervention studies that provide data on the time course of the change in mortality along
with the ERR values from cohort studies on PM2.5-related mortality. We applied the
intervention studies by Clancy et al. (2002), which analyzed the change in mortality following
the ban of coal sales in Dublin, and by Pope et al. (1992), which examined the change in
mortality resulting from the closure of a steel mill in the Utah Valley. We applied effect
estimates from the American Cancer Society (ACS) cohort by Krewski et al. (2002)3 and the Six
Cities cohort by Laden et al. (2006). 4 Applying combinations of these studies to equation 5.A.3
generates four estimates of k that range from 0.05 to 1.24. For additional context, the time
constant calculated using on a smoking cessation study (i.e., Leksell and Rabl (2001)) is in the
middle of this range (k=0.10). For this sensitivity analysis, we applied a time constant of k=0.45
as a reasonable parameter for the exponential decay function, but we acknowledge the range
of estimates that we could have chosen. This k constant is calculated as the average of the
average k constants corresponding to each cohort study.5 Table 5.A.1 provides the time
constants for each of these combinations and averages, and Figure 5. A. 2 illustrates the
exponential decay lag structures.
3 The relative risk coefficient from Krewski et al. (2009) (1.06 per 10 u.g/m3 change in average PM2.5 exposure for
all-cause mortality) is the same as the risk estimate from Pope et al. (2002).
4 We have not updated this analysis to reflect the newest Six Cities cohort from Lepeule et al. (2012). While the
relative risk coefficient from Lepeule et al. (2012) is slightly less than the relative risk coefficient from Laden et al.
(2006), this difference is unlikely to have a substantial difference in the value of k.
5 The general approach for calculating the time constants based on the combination of the intervention study and
cohort study is consistent with the 812 analysis (U.S. EPA, 2011), which was reviewed by SAB. However, in this
analysis we have applied a single time constant (k=0.45) rather than presenting the monetized benefits results for
every exponential lag function applying the various time constants.
5.A-5
-------
Table 5.A-1. Values of the Time Constant (k) for the Exponential Decay Lag Function
Value of k PM2.5 Cohort Study Intervention Study
0.05 H6C-Laden et al. (2006) Dublin-Clancy et al. (2002)
0.15 ACS-Krewski et al. (2009) Dublin-Clancy et al. (2002)
0.37 H6C-Laden et al. (2006) Utah Valley-Pope et al. (1992)
1.24 ACS-Krewski etal. (2009) Utah Valley-Pope et al. (1992)
0.70 Average k for ACS—Krewski et al. (2009)
0.21 Average k for H6C-Laden et al. (2006)
0.45 Average of average k for each cohort study
The estimated impacts of alternative lag structures on the monetary benefits associated
with reductions in PM-related premature mortality (estimated using the effect estimate from
Krewski et al. (2009)) are presented in Table 5.A-2. These monetized estimates are calculated
using the value of a statistical life (i.e., $6.3 million per incidence adjusted for inflation and
income growth) and are presented for both a 3 and 7% discount rate over the lag period). The
choice of mortality risk study and mortality valuation approach are described in detail in
Chapter 5 of this RIA. Figure 5.A.1 illustrates the cumulative distributions of the cessation lags
applied in this appendix. Because we applied an income adjustment factor specific to the
analysis year (see section 5.6.8 of this RIA), we do not adjust for income growth over the 20-
year cessation lag. This approach could underestimate the benefits for the later years of the
lag.
The results of this sensitivity analyses demonstrate that because of discounting of
delayed benefits, the lag structure may also have a large impact on monetized benefits,
reducing benefits by 27% if an extreme assumption that no effects occur until after 15 years is
applied at a 3% discount rate and 53% at a 7% discount rate. However, for most reasonable
distributed lag structures, differences in the specific shape of the lag function have relatively
small impacts on overall benefits. For example, the overall impact of moving from the previous
5-year distributed lag to the segmented lag recommended by the SAB-HES in 2004 in the main
estimate is relatively modest, reducing benefits by approximately 5% when a 3% discount rate
is used and 9% when a 7% discount rate is used. If no lag is assumed, benefits are increased by
approximately 10% relative to the segmented lag at a 3% discount rate and 22% at a 7%
discount rate.
5.A-6
-------
Table 5.A-2. Sensitivity of Monetized PM2.5-Related Premature Mortality Benefits to
Alternative Cessation Lag Structures, Using Effect Estimate from Krewski et al.
(2009)
Alternative
SAB Segmented
(Main estimate)
No lag
8-year
15-year
Alternative
Segmented
5-Year Distributed
Exponential Decay
(k=0.45)
Lag Structures for PM-Related Premature Mortality
30% of incidences occur in 1st year, 50% in years 2 to 5,
and 20% in years 6 to 20
3% discount rate
7% discount rate
Incidences all occur in the first year
3% discount rate
7% discount rate
Incidences all occur in the 8th year
3% discount rate
7% discount rate
Incidences all occur in the 15th year
3% discount rate
7% discount rate
20% of incidences occur in 1st year, 50% in years 2 to 5,
and 30% in years 6 to 20
3% discount rate
7% discount rate
50% of incidences occur in years 1 and 2 and 50% in years
2 to 5
3% discount rate
7% discount rate
Incidences occur at an exponentially declining rate
3% discount rate
7% discount rate
12
Value
(billion
2006$)ab
$4.0
$3.6
$4.4
$4.4
$3.6
$2.7
$2.9
$1.7
$3.8
$3.3
$4.2
$3.9
$4.2
$3.9
ug/m3
Percent
Difference from
Base Estimate
N/A
N/A
10.4%
22.5%
-10.3%
-23.7%
-27.0%
-52.5%
-3.2%
-6.6%
4.9%
9.4%
5.0%
9.9%
Dollar values rounded to two significant digits. The percent difference using the monetized benefits estimated
using the Lepeule et al. (2012) study would be identical, but the value would be approximately 2.3 times higher.
5.A-7
-------
100%
o,
ro
£
o
75%
SAB segmented lag (mam estimate)
Alternate Segmented Lag
Exponential decay model
5-yeardistributed lag
50%
25%
1st 2nd 3rd 4th 5th ah 7th 8th 9th 10th 11th 12th 13th 14th 15th 16th 17th 18th 19th 20th
year year year year year year year year year year year year year year year year year year year year
Year Following Reduction in PM2.5
Figure 5.A-1. Alternate Lag Structures for PM2.s Premature Mortality (Cumulative)
100%
75%
50%
25%
k=0.05 (H6C/Dublin)
k=0.37 (H6C/Utah)
k=0.15 (ACS/Dublin)
k=1.24 (ACS/Utah)
k=0.70 (Average ACS)
k=0.21 (Average H6C)
k=0.45 (Overall Average)
SAB Lag (Main Estimate)
Year Following Reduction in PM2.s
Figure 5.A-2. Exponential Lag Structures for PM2.s Premature Mortality (Cumulative)
5.A-8
-------
5.A.2 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. Table 5.A-3 lists the ranges
of elasticity values used to calculate the income adjustment factors, while Table 5.A-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 5.A-5.
Table 5.A-3. Ranges of Elasticity Values Used to Account for Projected Real Income Growth3
Benefit Category Lower Sensitivity Bound Upper Sensitivity Bound
Minor Health Effect 0.04 0.30
Premature Mortality 0.08 1.00
Derivation of these ranges can be found in Kleckner and Neumann (1999). COI estimates are assigned an
adjustment factor of 1.0.
Table 5.A-4. Ranges of Adjustment Factors Used to Account for Projected Real Income
Growth3
Benefit Category Lower Sensitivity Bound Upper Sensitivity Bound
Minor Health Effect 1.018 1.147
Premature Mortality 1.037 1.591
3 Based on elasticity values reported in Table C-4, U.S. Census population projections, and projections of real GDP
per capita.
5.A-9
-------
Table 5.A-5. Sensitivity of Monetized Benefits to Alternative Income Elasticities3
Benefit Category
Minor Health
Effect
Premature
Mortality b
Benefits
Incremental to Analytical Baseline (Millions of 2006$)
12 ug/m3
No adjustment
$30
$3,600
Lower Sensitivity Bound Upper Sensitivity Bound
$31 $35
$3,800 $5,800
All estimates rounded to two significant digits.
Using mortality effect estimate from Krewski et
(2012) or a 7% discount rate would show the same proportional range.
b Using mortality effect estimate from Krewski et al. (2009) and 3% discount rate. Results using Lepeule et al.
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 2020 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 much less pronounced, ranging from 96% to 108% of the main estimate for minor
effects.
5.A.3 Analysis of Cardiovascular Emergency Department Visits, Cerebrovascular Events and
Chronic Bronchitis
Below we summarize the results of a sensitivity analysis of three health endpoints:
cardiovascular emergency department visits, cerebrovascular events (stroke) and chronic
bronchitis (Table 5.A-6). While in the benefits chapter we provide a full description of the
rationale for treating these endpoints only in a sensitivity analysis, it is worth summarizing
these reasons here. In the case of cardiovascular emergency department visits, we lack the
necessary economic valuation functions to quantify the monetary value of these avoided cases.
We treat cerebrovascular events as a sensitivity analysis for three reasons: (1) the
epidemiological literature examining PM-related cerebrovascular events is still evolving; (2)
there are special uncertainties associated with quantifying this endpoint; (3) we have not yet
identified an appropriate means for estimating the economic value of this endpoint. Finally, we
now quantify chronic bronchitis in a sensitivity analysis because of the absence of newer
studies finding a relationship between long-term PM2.5 exposure and this endpoint, and the
relative weakness of the study available to quantify this endpoint.
5.A-10
-------
To quantify cardiovascular hospital admissions, we apply risk coefficient drawn from
three epidemiology studies: Metzger et al. (2004) (RR= 1.033, 95th percentile confidence
intervals 1.01-1.056 per 10 ug/m3 PM2.5, age 0-99), Tolbert et al. (2007) (RR= 1.005, 95th
percentile confidence intervals 0.993-1.017 per 10 u.g/m3 PM2.5, age 0-99), and Mathes et al.
(2011) (excess risk =0.8%, 95th percentile confidence intervals 0.0%-1.6% per 10 u.g/m3 PM2.5,
age 40-99) . To estimate cerebrovascular events, we apply a risk coefficient drawn from Miller
et al. (2007) (RR= 1.28, 95% confidence intervals 1.02-1.61 per 10 u.g/m3 PM2.5). To estimate
chronic bronchitis, we use a risk coefficient drawn from Abbey et al. (1995) (RR= 1.81, 95%
confidence intervals 0.98-3.25 per 45 u.g/m3 PM2.5). Additional information, including the
rationale for incorporating these new endpoints into the analysis, the baseline incidence rates
for these endpoints, and the prevalence rate for chronic bronchitis are described in Chapter 5
ofthisRIA.
Table 5.A-6. Avoided Cases of Cardiovascular Emergency Department Visits, Stroke and
Chronic Bronchitis in 2020 (95th percentile confidence intervals)3
Endpoint 12 |ig/ms
Cardiovascular emergency department visits
Metzger et al. (2004) 440
(ages 0-99) (160-720)
Tolbert et al. (2007) 62
(ages 0-99) (-72-200)
Mathes et al. (2011) 101
(ages 40-99) (8-193)
Stroke
Miller etal. (2007) 130
(ages 50-79) (20-230)
Chronic Bronchitis
Abbey et al. (1995) 360
(ages 27-99) (39-670)
3 All estimates rounded to two significant digits.
5.A.4 Updating Basis for Population Projections to 2010 Census
In this RIA, we updated the population demographic data in BenMAP to reflect 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) that were projected from the 2000 Census. We use
5.A-11
-------
projections based on economic forecasting models developed by Woods and Poole, Inc.
(Woods and Poole, 2012). Table 5.A-7 provides the results of a sensitivity analysis using the
older population projections compared to the newer projections.
Table 5.A-7. Change in Total Monetized Benefits for 2000 and 2010 Census projections for 12
ug/m3 in 2020 (2010$)a
Total Monetized Benefits
(3% discount rate)
Projections from 2000
Census
$3.8 to $8.7 billion
Projections from 2010
Census
$4.0 to $9.1 billion
Percent Change
+4.4%
a Percent change is based on the unrounded estimates.
5.A.5 Long-term PM2.s Mortality Estimates Using Cohort Studies in California
In Chapter 5, we described the multi-state cohort studies we used to estimate the PM2.s-
related mortality (i.e., Krewski et al., 2009; Lepeule et al., 2012), as well as summarized the
effect estimates for additional cohort studies. In this appendix, we provide additional
information about cohort studies in California.6 As shown in Table 5.8 in the health benefits
chapter, all of the monetized human health benefits associated with the illustrative control
strategy to attain the revised annual standard of 12 u.g/m3 are projected to occur in California.
For this reason, we determined that it was appropriate to consider the sensitivity of the
benefits results using effect estimates for cohorts in California specifically. Although we have
not calculated the benefits results using these cohort studies, it is possible to use the effect
estimates themselves to determine how much the monetized benefits in California would have
changed if we used effect estimates from the California cohorts. Each of the California cohort
studies are summarized in the PM ISA (U.S. EAP, 2009) or the Provisional Assessment (U.S. EPA,
2012). Table 5.A.8 provides the effect estimates from each of these cohort studies for all-cause,
cardiovascular, cardiopulmonary, and ischemic heart disease (IHD) mortality for each of the
California cohort studies.
6 In addition to cohorts studies conducted in California, we have also identified a cross-sectional studies (Hankey
et al., 2012). However, we have not summarized that study here.
5.A-12
-------
Table 5.A-8 Summary of Effect Estimates From Associated With Change in Long-Term
Exposure to PM2.s in Recent Cohort Studies in California
Hazard Ratios per 10 |ig/ms Change in PM2.5
(95th percentile confidence interval)
Authors
McDonnell et al. (2000)a
Jerrett et al. (2005)b
Chen etal. (2005)°
Enstrom et al. (2005)d
Krewski et al. (2009)e
Ostro et al. (2010)°
Ostro et al. (2011)c'f
Lipsett et al. (2011)°
Cohort
Adventist Health Study (AHS)
cohort (age > 27)
Subset of the ACS cohort living
in the Los Angeles
metropolitan area (age > 30)
Adventist Health Study (AHS)
cohort living in San Francisco,
South Coast (i.e., Los Angeles
and eastward), and San Diego
air basins (age >25)
California Prevention Study
(age >65)
Subset of the ACS cohort living
in the 5-county Los Angeles
Metropolitan Statistical Area
(age > 30)
California Teacher's study.
Current and former female
public school professionals
(age > 22)
California Teacher's study.
Current and former female
public school professionals
(age > 22)
All Causes
1.09
(.98-1.24)
1.15
(1.03-1.29)
N/A
1.04
(1.01-1.07)
1.42
(1.26-1.27)
1.84
(1.66-2.05)
1.06
(0.96-1.16)
1.01
(0.95-1.09)
Cardiopulmonary
N/A
1.10
(0.94-1.28)
N/A
N/A
1.11
(0.95-1.23)
2.05
(1.80-2.36)
1.19
(1.05-1.36)
N/A
Ischemic
Heart Disease
N/A
1.32
(1.03-1.29)
1.42
(1.06-1.90)
N/A
1.32
(1.06-1.64)
2.89
(2.27-3.67)
1.55
(1.24-1.93)
1.20
(1.02-1.41)
Table 3, adjusted for 10 ng/m change in PM2.5.
Table 1. 44 individual-level co-variates + all social (i.e., ecologic) factors specified (principal component analysis).
c Women only.
Represents deaths occurring from 1973-1982, but no significant associations were reported with deaths in later
time periods. The PM ISA (U.S. EPA, 2009) concludes that the use of average values for California counties as
exposure surrogates likely leads to significant exposure error, as many California counties are large and quite
topographically variable.
Table 23. 44 individual-level co-variates + all social (i.e., ecologic) factors specified.
Erratum Table 2.
As shown in Table 5.A.8, most of the cohort studies conducted in California report
central effect estimates similar to the (nationwide) all-cause mortality risk estimate we applied
5.A-13
-------
from Krewski et al. (2009) and Lepeule et al. (2012) albeit with wider confidence intervals.
Three cohort studies conducted in California indicate statistically significant higher risks than
the risk estimates we applied from Lepeule et al. (2012), and four studies showed insignificant
results.
5.A.6 Analysis of Health Benefits Estimated for 2025
In this RIA, we assumed an analysis year of 2020 for estimating costs and benefits with
an adjustment to the San Joaquin and South Coast areas in California for NOx emissions
reductions expected to occur between 2020 and 2025, when those areas are expected to
demonstrate attainment with the revised standards. Full benefits of the revised standards in
those two areas will not be realized until 2025. Because of population growth, population aging,
and income growth over time, the health benefits estimated for an analysis year of 2025 would
be higher. We have only conducted this sensitivity analysis for the revised standard of 12
u.g/m3, for which the health benefits occur entirely in California and are almost entirely in the
South Coast and San Joaquin air basins. Table 5.A-9 provides the comparison of the health
benefits (including avoided premature mortality and the total monetized health benefits)
estimated for 2020 and 2025.
Table 5.A-9. Comparison of Health Benefits Estimated for 2000 and 2025 for 12 |ig/m
3 a
2020 2025 Percent Difference
Avoided Premature 460 to 1,000 510 to 1,200 +12%
Mortality
Total Monetized Benefits $4.0 to $9.1 billion $4.5 to $10 billion +12%
(3% discount rate, 2010$)
Percent change is based on the unrounded estimates.
5.A.7 References
Abbey, David E et al. 1995. "Chronic Respiratory Symptoms Associated with Estimated Long-
Term Ambient Concentrations of Fine Particulates less than 2.5 Microns in aerodynamic
diameter (PM2.5) and Other Air Pollutants." Journal of Exposure Analysis and
Environmental Epidemiology 5(2): 137-158.
Agency for Healthcare Research and Quality (AHRQ). 2000. HCUPnet, Healthcare Cost and
Utilization Project. Rockville, MD. Available on the Internet at
.
5.A-14
-------
Agency for Healthcare Research and Quality (AHRQ). 2007. HCUPnet, Healthcare Cost and
Utilization Project. Rockville, MD. Available on the Internet at
.
Chen LH, Knutsen SF, Shavlik D, Beeson WL, Petersen F, Ghamsary M, et al. 2005. "The associa-
tion between fatal coronary heart disease and ambient particulate air pollution: are
females at greater risk?" Environ Health Perspect 113:1723-1729.
Clancy, L, P. Goodman, H. Sinclair, and D.W. Dockery. 2002. "Effect of Air-pollution Control on
Death Rates in Dublin, Ireland: An Intervention Study." Lancet Oct 19;360(9341):1210-4.
Desvousges, W.H., F.R. Johnson, and H.S. Banzhaf. 1998. Environmental Policy Analysis With
Limited Information: Principles and Applications of the Transfer Method (New Horizons
in Environmental Economics.) Edward ElgarPub: London.
Dockery, D.W., C.A. Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, B.C. Ferris, and F.E.
Speizer. 1993. "An Association between Air Pollution and Mortality in Six U.S. Cities."
New England Journal of Medicine 329(24):1753-1759.
Enstrom JE 2005. "Fine particulate air pollution and total mortality among elderly Californians,
1973-2002." Inhal Toxicol 17: 803-816.
Hankey, S., J. D. Marshall, et al. (2011). "Health Impacts of the Built Environment: Within-Urban
Variability in Physical Inactivity, Air Pollution, and Ischemic Heart Disease Mortality."
Environ Health Perspect 120(2).
Jerrett M, Burnett RT, Ma R, Pope CA III, Krewski D, Newbold KB, et al. 2005. "Spatial analysis of
air pollution and mortality in Los Angeles." Epidemiology 16(6):727-736.
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
.
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.
Laden F, Schwartz J, Speizer FE, Dockery DW. 2006. "Reduction in fine particulate air pollution
and mortality: extended follow-up of the Harvard Six Cities study." Am J Respir Crit Care
Med 173:667-672.
Lepeule J, Laden F, Dockery D, Schwartz J. 2012. "Chronic Exposure to Fine Particles and
Mortality: An Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009."
Environ Health Perspect. Jul;120(7):965-70.
5.A-15
-------
Lipsett MJ, Ostro BD, Reynolds P et al. 2011. "Long-term exposure to air pollution and
cardiorespiratory disease in the California teachers study cohort/Mm J Respir Crit Care
Med 184: 828-35.
Mathes, R.W., K. Ito, and T. Matte. 2011. Assessing Syndromic Surveillance of Cardiovascular
Outcomes from Emergency Department Chief Complaint Data in New York City. PloS
ONE. Vol6(2): 1-10.
McDonnell WF; Nishino-lshikawa N; Petersen FF; Chen LH; Abbey DE. 2000. "Relationships of
mortality with the fine and coarse fractions of long-term ambient PM10 concentrations
in nonsmokers." J Expo Sci Environ Epidemiol, 10: 427-436.
Metzger, K. B., P. E. Tolbert, et al. (2004). "Ambient air pollution and cardiovascular emergency
department visits." Epidemiology 15(1): 46-56.
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.
Ostro B, Lipsett M, Reynolds P, Goldberg D, Hertz A, Garcia C, et al. 2010. "Long-Term Exposure
to Constituents of Fine Particulate Air Pollution and Mortality: Results from the
California Teachers Study." Environ Health Perspect 118:363-369.
Ostro B, Reynolds P, Goldberg D, Hertz A, et al. 2011. "Assessing Long-Term Exposure in the
California Teachers Study." Environ Health Perspect June; 119(6): A242-A243.
Peters, A., D. W. Dockery, J. E. Muller and M. A. Mittleman. 2001. "Increased particulate air
pollution and the triggering of myocardial infarction." Circulation. Vol. 103 (23): 2810-5.
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.
Roosli M, Kunzli N, Braun-Fahrlander C. 2004. "Use of Air Pollution 'Intervention-Type' Studies
in Health Risk Assessment." 16th Conference of the International Society for
Environmental Epidemiology, New York, August 1-4, 2004.
Roosli M, Kunzli N, Braun-Fahrlander C, Egger M. 2005. "Years of life lost attributable to air
pollution in Switzerland: dynamic exposure-response model." International Journal of
Epidemiology 34(5):1029-35.
Tolbert, P. E., M. Klein, et al. 2007. "Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta." J Expo Sci Environ Epidemiol 17
Suppl 2:529-35.
5.A-16
-------
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—RTP Division. December. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2012. Provisional Assessment of Recent
Studies on Health Effects of Particulate Matter Exposure. EPA/600/R-12/056A. National
Center for Environmental Assessment—RTP Division. December.
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. Parti. EPA-SAB-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 (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: An Advisory by a Special Panel of the Advisory Council on Clean
Air Compliance Analysis. EPA-SAB-COUNCIL-ADV-01-004. September. Available on the
Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (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: Advisory by the Health
Effects Subcommittee of the Advisory Council on Clean Air Compliance Analysis. 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). 2004. 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
.
5.A-18
-------
APPENDIX 5.B
COMPREHENSIVE CHARACTERIZATION OF UNCERTAINTY IN BENEFITS ANALYSIS
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 analysis 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., by using monte carlo assessments, concentration benchmark analyses, alternative
concentration-response functions, sensitivity analyses, distributional assessments, and
influence 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 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 of PM2.5-related impacts (U.S. EPA, 2010b, 2011). 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, 2004, 2011a, 2011b).
The assessments presented here are largely consistent with those previous peer-reviewed
assessments.
5.B.I Description of Classifications Applied in the Uncertainty Characterization
Table 5.B-1 catalogs the most significant sources of uncertainty in the PM benefits
analysis and then characterizes four dimensions of that uncertainty. The first two dimensions
focus on the nature of the uncertainty. The third and fourth dimensions focus on the extent to
which the 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;
5.B-1
-------
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.
5.B.1.1 Direction of Bias
The "direction of bias" column in Table 5.B-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 Particulate Matter (hereafter, "PM ISA") (U.S. EPA, 2009) . 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."
5. B.1.2 Magnitude of Impact
The "magnitude of impact" column in Table 5.B-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, 2010b, 2011a), but we have slightly revised the
category names and the cut-offs here.1 The definitions used here are provided below.
• 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%.
1 In The Benefits and Costs of the Clean Air Act from 1990 to 2020 (U.S. EPA, 2011a), EPA applied a classification of
"potentially major" if a plausible alternative assumption or approach could influence the overall monetary benefit
estimate by 5% 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. 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 are likely to influence the interpretation of risk in the context of
the PM NAAQS review.
5.B-2
-------
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 PM2.5-related mortality benefits comprise over 98% 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 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 which we classified as high or medium-high
impact are: the causal relationship between PM2.s and mortality, regional differences in PM2.s
mixtures, shape of the concentration-response function, mortality valuation, and cessation lag.
The classification of these uncertainties as "high magnitude" is generally consistent with the
results of EPA's Influence Analysis (Mansfield et al., 2008), the Quantitative Health Risk
Assessment for Particulate Matter (U.S. EPA, 2010b), and the Benefits and Costs of the Clean Air
Act 1990 to 2020 (U.S. EPA, 2011).
5.B.1.3 Confidence in Analytic Approach
The "confidence in analytic approach" column of Table 5.B-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 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).2 The three categories used to characterize the
degree of confidence are:
• High—the current evidence is plentiful and strongly supports the selected approach;
2 We have applied the same classification as The Benefits and Costs of the Clean Air Act from 1990 to 2020 (U.S.
EPA, 2011a) 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.
5.B-3
-------
• 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 staff's professional
judgment based on the volume and consistency of supporting evidence, much of which has
been evaluated in the PM ISA (U.S. EPA, 2009) and EPA's independent Science Advisory Board
(SAB). The PM ISA evaluated the entire body of scientific literature on PM science and was
twice peer-reviewed by EPA's Clean Air Scientific Advisory Committee (CASAC). In general, we
regard a conclusion in the PM 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.
5.B.1.4 Uncertainty Quantification
The column of Table 5.B-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 development in order to assess quantitatively. Specifically, EPA applied this
framework in the Quantitative Risk Assessment for Particulate Matter (U.S. EPA, 2010b), and it
has been recommended in EPA guidance documents assessing air toxics-related risk and
Superfund site risks (U.S. EPA, 2004b and 2001, respectively). Ultimately, the tier decision is
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.
5.B-4
-------
• 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 a 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 PM-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).
5.B.2 Organization of the Qualitative Uncertainty Table
Table 5.B-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.
5.B-5
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 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
00
cr>
Variation in effect
estimates reflecting
differential toxicity of
particle components
and regional
differences in PM2 5
composition
(mixtures)
Either direction, depending on
the species.
PM composition and the size
distribution of those particles
vary within and between areas
due to source characteristics.
Any specific location could have
higher or lower contributions of
certain PM species and other
pollutants than the national
average, meaning potential
regional differences in health
impact of given control
strategies. Depending on the
toxicity of each PM species
reduced in the control
strategies, assuming equal
toxicity could over or
underestimate benefits.
Potentially High
Epidemiology studies examining
regional differences in PM2.5-related
health effects have found differences in
the magnitude of those effects, and
composition remains one potential
explanatory factor (PM ISA, section
2.3.2).
In addition to differences in the
contribution of any given species to the
baseline concentrations, use of different
control strategies would have a differing
magnitude of the effect in different
regions. Depending on the extent of the
differences in toxicity and the exact mix
if species controlled, different control
strategies could have a differing
magnitude of the effect in different
regions.
Medium
Consistent with SAB advice, we assume that
all fine particles, regardless of their chemical
composition, are equally potent in causing
premature mortality (U.S. EPA-SAB, 2010,
pg. 18). The PM ISA concluded many
compounds can be linked with multiple
health effects and the evidence is not yet
sufficient to allow differentiation of effects
estimates by particle type (pg. 2-17).
We also use national risk coefficients with no
local variations due to differential exposure.
The PM ISA states that available evidence
and the limited amount of city-specific
speciated PM2.5 data does not allow
differentiation of PM effects in different
locations (pg. 2-17). Using national risk
coefficients is supported by SAB (U.S. EPA-
SAB, 2010) and NAS (NRC, 2002).
Tier 2 (sensitivity analysis)
Regional differences in hazard
ratios from studies conducted
in California shown in Table
5.A-8. The hazard ratios from
the California studies range
from -83% to +1300%
compared to the national
estimate applied from Krewski
etal. (2009).
Causal relationship
between PM2.5
exposure and
premature mortality
Overestimate, if PM25does not
have a causal relationship with
premature mortality.
High
Mortality generally dominates
monetized benefits, so small
uncertainties could have large impacts
on the total monetized benefits.
High
Our approach is consistent with the PM ISA,
which concluded that premature mortality is
causally related to PM25exposure. This
conclusion is based on the consistency of the
effects observed across epidemiology
studies and biological plausibility) (pg. 2-9,
2-11).In addition, in the PM25expert
elicitation, 10 of 12 experts provided
likelihood of causality of 90% or higher
(Roman etal., 2008).
Tier 3 (probabilistic)
Each expert in the PM25 expert
elicitation had the opportunity
to specify the likelihood of a
causal relationship into their
function (Roman et al., 2008).
Using these expert-derived
functions is a probabilistic
assessment of causality (see
Figure 5-4).
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
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 (cont'd)
Causal relationship
between PM2.5
exposure and
premature mortality
Overestimate, if PM2.5does not have a
causal relationship with premature
mortality.
High
Mortality generally dominates
monetized benefits, so small
uncertainties could have large
impacts on the total monetized
benefits.
High
Our approach is consistent with the PM
ISA, which concluded that premature
mortality is causally related to PM2 5
exposure. This conclusion is based on the
consistency of the effects observed across
epidemiology studies and biological
plausibility) (pg. 2-9, 2-11).In addition, in
the PM2 5 expert elicitation, 10 of 12
experts provided likelihood of causality of
90% or higher (Roman et al., 2008).
TierS (probabilistic)
Each expert in the PM25 expert
elicitation had the opportunity
to specify the likelihood of a
causal relationship into their
function(Roman et al., 2008).
Using these expert-derived
functions is a probabilistic
assessment of causality (see
Figure 5-4).
Shape of the C-R
functions, particularly
at low concentrations
oo
Either
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.
Medium-High
Krewski et al. (2009) considered
alternative model forms and
found that the choice of
functional relationship can
make a considerable difference
in the predicted risk at lower
concentrations. Specifically,
they found a 58% increase in
risk at lower concentrations for
all-cause mortality. 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.
High
Consistent with the PM ISA, we assume a
log-linear no-threshold model for the
concentration-response function. In
previous RIAs (between 2006 and 2009),
we assumed a threshold in the mortality
relationship, which shifted the slope of
the function to account for all health
effects occurring above the threshold
concentration.
The PM ISA concluded that the studies
overall support the use of a no-threshold
log-linear model for PM2 5-related
mortality (pg. 2-25). Our approach also
follows recommendations from the SAB
(U.S. EPA-SAB, 2010, pg. 13).
Tier 3 (probabilistic)
The experts in the PM25 expert
elicitation specified the shape of
the C-R function (Roman et al.,
2008). Only expert K assumed a
threshold, but experts B, F, and L
included different slopes at
lower concentrations. Using the
expert-derived functions is a
probabilistic assessment of the
shape of the C-R function (see
Figure 5-4). Also, the
concentration benchmark
assessment shows the
premature deaths estimated at
various concentrations.
Specifically, 92% of the
monetized benefits occur at or
above the lowest annual
concentration of PM in the
Krewski et al (2009) study.
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
Potential
Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach Uncertainty Quantification
00
CO
Exposure error Underestimate (generally)
in epidemiology The PM ISA states that the results from
studies the Krewski et al. (2009) and Jerrett et
al. (2005) studies suggest that exposure
error can underestimate effect
estimates (pg. 7-90). The PM ISA states
that exposure error can potentially bias
an estimate of a health effect endpoint
towards the null or increase the size of
confidence intervals (pg. 3-152). The
PM ISA states that reducing exposure
error can result in stronger associations
between pollutants and effect
estimates than generally observed in
studies having less exposure detail (pg.
7-90).
Uncertainties Associated with Concentration-Response Functions (cont'd)
Low-Medium
Although this underestimation is well
documented, including in the PM ISA,
SAB has not suggested an approach to
adjust for this bias.
Medium
Recent analyses reported in Krewski et
al. (2009) demonstrate the potentially
significant effect (approximately 18%
increase in hazard ratio for all-cause
mortality in Los Angeles by improving
the exposure assessment) 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.
Tier 1 (qualitative)
(No quantitative method
available)
Modification of
Mortality C-R
function by
socio-economic
status (SES)
Underestimate for ACS cohort (Krewski
et al., 2009) because of the
demographics of the cohort (NRC, 2002,
pg. 101).
Medium for ACS cohort
Medium
We have not modified the function for
SES in the core analysis.
Tier 2 (sensitivity analysis)
Unknown for Six Cities cohort (Lepeule
et al., 2012), but educational
attainment in cohort is likely more
representative of the general
population (NRC, 2002, pg. 101).
Experts suggested that the educational
attainment may also modify the risk in
the Six Cities cohort, but estimates are
not available, (IEC, 2006, pg. 3-16).
Using both cohorts may balance any
potential bias.
Krewski et al. (2009) found that
educational attainment, which is a
surrogate for SES, modifies the risk
coefficient (i.e., ranging from -8% for
individuals with more than Grade 12 to
+37% for individual with less than
Grade 12 after controlling for ecologic
covariates compared to the national
estimate). The overall impact would
depend on the mixture of educational
attainment in the target population.
Unknown magnitude for Six Cities
cohort.
The PM ISA concluded that there is
evidence that SES, measured using
surrogates such as educational
attainment, modifies the association
between PM and health effects (pg. 8-
15), but gender (pg. 8-6) and race (pg.
8-7) do not seem to modify the
association between PM and health
effects. The PM ISA also concluded that
some evidence suggests that Hispanic
ethnicity may modify the relationship
(pg. 8-7).
Effect modification for
educational attainment
evaluated in the distributional
analysis in Appendix 5A of the
proposal RIA. For 12/35, the
percent of premature deaths
ranged from 4.4% for individuals
with more than Grade 12
education up to 6.5% for
individuals with less than Grade
12 education.
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty Quantification
00
ID
Confounding by
individual risk factors,
other than SES—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.
Uncertainties Associated with Concentration-Response Functions (cont'd)
Medium
Medium
Because mortality dominates
monetized benefits, even a small
amount of confounding could have
medium impacts on total monetized
benefits.
To minimize confounding effects, to the
extent possible, we use risk coefficients that
control for 44 individual and 7 ecologic
factors from Krewski et al. (2009). Although
Krewski et al. (2000, 2009) found that
ecologic covariates did not exert a
significant confounding influence on PM-
related mortality, they highlighted the "vital
need for further study of the role that
ecologic covariates have in the association
between air pollution and mortality."
We use risk coefficients that control for 3
individual factors (e.g., BMI, smoking, and
education) from Lepeule et al. (2012).
Tier 2 (sensitivity analysis)
(Quantitative methods
available but not assessed in
this analysis.)
Confounding and
effect modification by
co-pollutants
Either, depending upon the
pollutant.
Disentangling the health
responses of combustion-related
pollutants (i.e., PM, SOX, NOX,
ozone, and CO is a challenge.
Ambient PM may be an indicator
of complex mixtures that share
emission sources (e.g., traffic and
power generation). The PM ISA
states that co-pollutants may
mediate the effects of PM or PM
may influence the toxicity of co-
pollutants (pg. 1-16). Alternately,
effects attributed to one
pollutants may be due to another.
Medium
Because this uncertainty could affect
mortality and because mortality
generally dominates monetized
benefits, even a small uncertainties
could have medium impacts on total
monetized benefits.
Medium
When modeling effects of pollutants jointly,
we apply multi-pollutant effect estimates
when those estimates are available to avoid
double-counting when those estimates are
available and satisfy other selection criteria.
The PM ISA states that evidence from the
limited number of studies suggests that
gaseous co-pollutants do not confound the
PM2.5-mortality relationship (pg. 2-11).
EPA's current approach to modeling has
been supported during peer-reviews by SAB
(U.S. EPA-SAB, 2010) and NAS (NRC, 2002).
Tier 1 (qualitative)
(No quantitative method
available)
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
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 (Cont'd)
Not including short-
term exposure studies
in PM mortality
calculations
Underestimate
Long-term PM exposure studies
likely capture a large part of the
impact of short-term peak exposure
on mortality; however, the extent
of overlap between the two study
types is unclear (NRC, 2002, pg.
116).
Medium
If short-term mortality is not fully
captured within the cohort mortality
estimates, then the benefits could be
underestimated.
Medium
Consistent with the NAS, we assume
that long-term cohort studies capture
most of the mortality benefits.
However, NAS acknowledges that the
effects of short-term exposures are
unlikely to be fully captured in the
cohort studies (NRC, 2002, pgs. 108,
116).
Tier 1 (qualitative)
(No quantitative method
available)
00
I-*
o
Impact of historical
exposure on mortality
effect estimates
Either
Long-term studies of mortality
suggest that different periods of PM
exposure can produce different
effects estimates, raising the issue
of uncertainty in relation to
determining which exposure
window to use when estimating
mortality benefits.
Low-Medium
Krewski et al. (2009) evaluated exposure
windows for PM-related mortality and
suggested that differences in effect
estimates are associated with the use of
different exposure periods (with the
more recent period having larger
estimates) but those differences were
small. Lepeule et al. (2012) reported
similar findings for different exposure
periods. Mortality generally dominates
monetized benefits, so small
uncertainties could have medium
impacts on total monetized benefits.
Medium
We do not make adjustments for
temporal variation in exposure.
The PM ISA concludes that the overall
evidence for determining the
appropriate exposure window suggests
that the health benefits from reducing
air pollution do not require a long
latency period and would be expected
within a few years after intervention
(pg. 7-95).
Tier 2 (sensitivity analysis)
(Quantitative methods available
but not assessed in this
analysis.)
Application of C-R
relationships only to
the original study
population
Underestimate
Estimating health effects for only to
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 would
have a small impact on total monetized
benefits. With respect to adult
mortality—the baseline rate is
significantly lower under age 25, so
lowering the age range to 18 would
have minimal impact on total monetized
benefits.
High
For mortality, we estimate health
effects only for the ages matching the
original study population (i.e., 25+ or
30+). 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.)
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on Monetized
Benefits
Confidence in Analytical Approach Uncertainty Quantification
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, 2011c).
TierS (probabilistic)
Assessed uncertainty in
mortality valuation using a
Weibull distribution.
o
Cessation lag structure
for long term PM
mortality
Underestimate
Recent studies (Schwartz, 2008;
Puett et al., 2009; Lepeule et al.,
2012) estimate that the majority of
the risk occurs within 2 years of
reduced exposure. 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-High
Although the cessation lag does not
affect the number of premature deaths
attributable to PM2.5 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.
Medium
Consistent with SAB advice, we
estimate that 30% of mortality
reductions in the first year, 50% over
years 2 to 5, and 20% over the years 6
to 20 after the reduction in PM2.5
(U.S. EPA-SAB, 2004c).
The PM ISA concludes that health
benefits from reducing air pollution
would be expected within a few years
of intervention (pg. 7-95). Despite
recent studies providing new evidence
of the timing of mortality risk
reduction after changes in exposure,
the SAB did not suggest changes to the
default cessation lag applied in EPA's
main benefits estimates in the most
recent review (U.S. EPA-SAB, 2010).
Tier 2 (sensitivity analysis)
As shown in Appendix 5-A, the
use of an alternate lag structure
would change the monetized
benefits by +10% to -52%
depending on the discount rate
and lag structure assumed.
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on
Monetized Benefits
Confidence in Analytical Approach
Uncertainty
Quantification
Uncertainties Associated with Economic Valuation (cont'd)
Income growth
adjustments
Either
Income growth increases willingness-to-
pay (WTP) valuation estimates, including
mortality, overtime. 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. (lEc, 2012).
Medium
Income growth from 1990 to 2020
increases mortality valuation by 20%.
Alternate estimates for this
adjustment vary by 20% (IEC, 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 5-A,
the use of alternate income
growth adjustments would
change the monetized
benefits by +33% to -14%.
Morbidity valuation
o
Underestimate
Morbidity benefits such as hospital
admissions and heart attacks are
calculated using cost-of-illness (COI)
estimates, which are generally half the
willingness-to-pay to avoid the illness
(Alberini and Krupnick, 2000). In
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 AMI. 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 5.9).
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
Potential
Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on Monetized
Benefits
Confidence in Analytical Approach
Uncertainty
Quantification
Uncertainties Associated with Baseline Incidence and Population Projections
Population
estimates and
projections
Either
The monetized benefits would change
in the same direction as the over- or
underestimate in population
projections in areas where exposure
changes.
Low-Medium
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.
Medium
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.
Tier 1 (qualitative)
(No quantitative
method available)
1/1
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
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)
(continued)
-------
Table 5.B-1. Summary of Qualitative Uncertainty for Key Modeling Elements in PM2.5 Benefits (continued)
Potential Source of
Uncertainty
Direction of Potential Bias
Magnitude of Impact on Monetized
Benefits
Confidence in Analytical Approach Uncertainty Quantification
Uncertainties Associated with Omitted Benefits Categories
Unquantified PM
health benefit
categories, such as
pulmonary function,
cerebrovascular
events, low birth
weight, or cancer
Underestimate
EPA has not included monetized
estimates of these benefits
categories in the core benefits
estimate.
Medium
Although the potential magnitude is
unknown, including all of the additional
morbidity endpoints associated with
PM2 5exposure that are currently not
monetized could increase the total
benefits by a moderate amount.
Low
Current data and methods are
insufficient to develop (and value)
national quantitative estimates of
these health effects. The PM ISA
determined that respiratory morbidity
(e.g., decreases in lung function) is
causally associated with PM2 5
exposure (pg. 2-12).
Tier 2 (sensitivity analysis)
In Table 5.A-6, EPA estimates
avoided incidence of strokes,
cardiovascular emergency
department visits, and chronic
bronchitis.
Uncertainties Associated with Estimated Exposure Changes
o
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)
-------
5.B.3 References
Alberini, Anna and Alan Krupnick. 2000. "Cost-of-lllness 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.
Industrial Economics, Incorporated (lEc). 2006. Expanded Expert Judgment Assessment of the
Concentration-Response Relationship Between PM2.5 Exposure and Mortality. Prepared
for: Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC. September. Available on the Internet at
Jerrett M, Burnett RT, Ma R, et al. 2005. "Spatial analysis of air pollution and mortality in Los
Angeles." Epidemiology 16:727-736.
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.
Lepeule J, Laden F, Dockery D, Schwartz J 2012. "Chronic Exposure to Fine Particles and
Mortality: An Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009."
Environ Health Perspect. 120: 965-970.
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
-------
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." Environ. Sci. Technol.,
42(7):2268-2274.
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
3A). 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, 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). 2009. Integrated Science Assessment for
Particulate Matter (Final Report). EPA-600-R-08-139F. National Center for
Environmental Assessment—RTP Division. December. 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 Particulate 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). 2011a. 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
.
5.B-16
-------
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
.
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
.
5.B-17
-------
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2011a. 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). 2011b. 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). 2011c. 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.B-18
-------
CHAPTER 6
WELFARE CO-BENEFITS OF THE PRIMARY STANDARD
6.1 Synopsis
The emission reductions to meet the revised primary standard will have "welfare" co-
benefits in addition to human health benefits, including changes in visibility, materials damage,
ecological effects from PM deposition, ecological effects from nitrogen and sulfur emissions,
vegetation effects from ozone exposure, ecological effects from mercury deposition, and
climate effects.1 Despite our goal to quantify and monetize as many of the benefits as possible
for the revised primary standard, the welfare co-benefits of the revised primary standard
remain unquantified and nonmonetized in this RIA due to data, methodology, and resource
limitations. Specifically, we do not have air quality model runs for the regulatory baseline and
the alternative standard levels that would allow us to calculate the visibility co-benefits of
attaining the revised primary standard even though we have a complete methodology for
estimating these co-benefits. However, using the approach described in this chapter, we
provide the results of an illustrative analysis in Appendix 6.B using the 2020 base case and 2020
control case simulations that were used to develop the air quality ratios.
6.2 Introduction to Welfare Benefits
Illustrative emission reduction strategies to attain the revised and alternative annual
primary standard have numerous documented effects on environmental quality that affect
human welfare. The Clean Air Act defines welfare effects to include any non-health effects,
including direct damages to property, either through impacts on material structures or by
soiling of surfaces, direct economic damages in the form of lost productivity of crops and trees,
indirect damages through alteration of ecosystem functions, and indirect economic damages
through the loss in value of recreational experiences or the existence value of important
resources. The EPA's Integrated Science Assessments for Particulate Matter (hereafter, "PM
ISA") (U.S. EPA, 2009b), N0x/S0x—Ecological Criteria (U.S. EPA, 2008), and Ozone (U.S. EPA,
2012a) identify numerous physical and ecological effects known to be causally linked to these
pollutants. This chapter describes these individual effects and how we would quantify and
monetize them if there is enough data to do so. These welfare effects include changes in
visibility, materials damage, ecological effects from PM deposition, ecological effects from
1 While we understand that "welfare" can include health and non-health effects in the economic sense, we use the
term "welfare" in this RIA in the same context as the definitions in the Clean Air Act for NAAQS. In practice,
welfare benefits represent non-health effects.
6-1
-------
nitrogen and sulfur emissions, vegetation effects from ozone exposure, ecological effects from
mercury deposition, and climate effects.
These welfare co-benefits are associated with reductions in emissions of specific
pollutants resulting from illustrative emission reduction strategies to attain the revised and
alternative annual primary standard, not the form or intent of any specific standard. Even
though the primary standards are designed to protect against adverse effects to human health,
the emission reductions have welfare co-benefits in addition to the direct human health
benefits.
The impacts of the illustrative emission reduction strategies can be grouped into four
categories: directly emitted PM (e.g., metals, organic compounds, dust), reductions of PM2.5
precursors (e.g., NOX, SOX, VOCs), other ancillary reductions from the illustrative emission
reduction strategies (e.g., mercury and C02), and secondary co-pollutant formation from PM
precursors (e.g., ozone from NOX and VOCs). Regardless of the category, these emission
changes are anticipated to affect ambient concentrations and deposition, and consequently
affect public welfare. It is therefore appropriate and reasonable to include all the benefits and
co-benefits associated with these emission reductions to provide a comprehensive
understanding of the likely public impacts of attaining the revised or alternative annual
standards. Table 6-1 shows the welfare effects associated with the various pollutants (either
directly or as a precursor to secondary formation of PM or ozone) that would be reduced by the
illustrative emission reduction strategies to attain the revised and alternative annual standard.
Based on the EPA's previous analyses, we believe the welfare co-benefits associated
with these non-health benefit categories could be significant (U.S. EPA, 2011b). Despite our
goal to quantify and monetize as many of the benefits and co-benefits as possible, welfare co-
benefits of the revised primary standard remain unquantified and nonmonetized in this RIA due
to data, methodology, and resource limitations. Therefore, the total benefits would be larger
than we have estimated. The monetized value of these unquantified effects is represented by
adding an unknown "B," which includes both unmonetized health benefits and welfare co-
benefits, to the aggregate total for the cost-benefit comparison. These unquantified benefits
and co-benefits may be substantial, although the magnitude is highly uncertain. We include a
qualitative description of the anticipated welfare effects in this chapter to characterize the type
and potential extent of those co-benefits, as identified in Table 6-2.
6-2
-------
Table 6-1. Welfare Effects by Pollutants Potentially Affected by Attainment of the PM
NAAQS
. .... Atmospheric and _
Atmospheric Effects _ . . _„ Deposition Effects
Deposition Effects
Pollutant ,, . . Ecosystem
... ...... Vegetation Vegetation ., 4 . . _„ * . ..... t.
Visibility . . Materials _.. Effects— Acidification Nitrogen Mercury
Injury Injury „ Climate ,„ . . _ . . .....
Impairment ,' . ._ . Damage (Organics (freshwater) Enrichment Methylation
(S°2) (0z°"e) & Metals)
Direct •/ S •/ •/
PM2.5
NOX -S •/ S S •/ •/
SO2 •/ S •/ •/ •/ •/
VOCs S •/ S •/
PM10 S •/ S
Hg •/ •/
C02 S
•s = Welfare category affected by this pollutant.
The remainder of this chapter is organized as follows: Section 6.3 provides a qualitative
discussion of the visibility co-benefits and describes our approach to estimate those visibility
co-benefits if we had the data to do so. Sections 6.4 through 6.6 provide qualitative co-benefits
for the unquantified benefits categories of materials damage, climate, and ecosystem effects.
References are provided in Section 6.7. Additional information regarding technical details of the
visibility co-benefits approach is provided in Appendix 6A. The illustrative visibility co-benefits
results for the specific modeled scenario (not the revised standard scenario) are provided in
Appendix 6.B.
6-3
-------
Table 6-2. Quantified and Unquantified Welfare Co-Benefits
Benefits Category
Effect Has
Been
Specific Effect Quantified
Effect Has
Been More
Monetized Information
Improved Environment
Reduced visibility impairment
Reduced climate effects
Reduced effects on materials
Reduced effects from PM
deposition (metals and
organics)
Reduced vegetation and
ecosystem effects from
exposure to ozone
a
Visibility in Class 1 areas in SE, SW,
and CA regions
Visibility in Class 1 areas in other —
regions
Visibility in 8 cities —
Visibility in other residential areas —
Global climate impacts from CO2 —
Climate impacts from ozone and —
PM
Other climate impacts (e.g., other —
GHGs, other impacts)
Household soiling —
Materials damage (e.g., corrosion, —
increased wear)
Effects on Individual organisms —
and ecosystems
Visible foliar injury on vegetation —
Reduced vegetation growth and —
reproduction
Yield and quality of commercial —
forest products and crops
Damage to urban ornamental —
plants
Carbon sequestration in terrestrial —
ecosystems
Recreational demand associated —
with forest aesthetics
Other non-use effects
Ecosystem functions (e.g., water —
cycling, biogeochemical cycles,
net primary productivity, leaf-gas
exchange, community
composition)
a
Section 6.3,
Appendix 6.B
a
Section 6.3,
Appendix 6. B
a
Section 6.3,
Appendix 6. B
a
Section 6.3,
Appendix 6.B
— Section 6.5, SCC
TSDb
— Section 6.5,
Ozone ISA, PM
ISA0
- Section 6.5, IPCCC
— Section 6.4, PM
ISA0
— Section 6.4, PM
ISA0
— Section 6.6.1, PM
ISA0
— Section 6.6.4,
Ozone ISA0
— Section 6.6.4,
Ozone ISA
— Section 6.6.4,
Ozone ISAb'd
— Section 6.6.4,
Ozone ISA0
— Ozone ISA0
— Ozone ISA0
Ozone ISA0
— Ozone ISA2
(continued)
6-4
-------
Table 6-2. Quantified and Unquantified Welfare Co-Benefits (Cont.)
Benefits Category
Specific Effect
Effect Has
Been
Quantified
Effect Has
Been
Monetized
More
Information
Improved Environment (Cont.)
Reduced effects from acid
deposition
Recreational fishing
Tree mortality and decline
Commercial fishing and forestry
effects
Recreational demand in terrestrial
and aquatic ecosystems
Other non-use effects
Ecosystem functions (e.g.,
biogeochemical cycles)
Section 6.6.2,
NOX SOX ISAb
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Reduced effects from
nutrient enrichment
Species composition and
biodiversity in terrestrial and
estuarine ecosystems
Coastal eutrophication
Recreational demand in terrestrial
and estuarine ecosystems
Other non-use effects
Ecosystem functions (e.g.,
biogeochemical cycles, fire
regulation)
Section 6.6.2,
NOy SOy ISA0
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Reduced vegetation effects
from ambient exposure to
SO2and NOX
Injury to vegetation from SO2
exposure
Injury to vegetation from NOX
exposure
Section 6.6.2,
NOX SOX ISA0
Section 6.6.2,
NOX SOX ISA0
Reduced ecosystem effects
from exposure to
methylmercury (through the
role of sulfate in
methylation)
Effects on fish, birds, and
mammals (e.g., reproductive
effects)
Commercial, subsistence and
recreational fishing
Section 6.2 and
6.6.3, Mercury
Study RTC°'d
Section 6.2 and
6.6.3, Mercury
Study RTC°
We quantify these co-benefits in an illustrative analysis using the methods discussed in this chapter for the
specific modeled scenario. These results are provided in Appendix 6.B, but these results of that illustrative scenario
are not an estimate of the co-benefits for the revised primary standard.
b We assess these co-benefits qualitatively due to time and resource limitations for this RIA.
0 We assess these co-benefits qualitatively because we do not have sufficient confidence in available data or
methods.
d We assess these co-benefits qualitatively because current evidence is only suggestive of causality or there are
other significant concerns over the strength of the association.
6-5
-------
6.3 Visibility Co-Benefits Approach
6.3.1 Visibility and Light Extinction Background
The illustrative emission reduction strategies designed to attain the revised and
alternative annual standards would reduce emissions of directly emitted PM2.5 as well as
precursor emissions such as NOX and S02 for an alternative annual standard at 11 u.g/m3. These
emission reductions would improve the level of visibility because these suspended particles and
gases impair visibility by scattering and absorbing light (U.S. EPA, 2009b). Visibility is also
referred to as visual air quality (VAQ),2 and it directly affects people's enjoyment of a variety of
daily activities (U.S. EPA, 2009b). 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, 2009b). This section
discusses the economic co-benefits associated with improved visibility as a result of emission
reductions associated with the revised and alternative annual standards.
Air pollution affects light extinction, a measure of how much the components of the
atmosphere scatter and absorb light. More light extinction means that the clarity of visual
images and visual range is reduced, all else held constant. Light extinction is the optical
characteristic of the atmosphere that occurs when light is either scattered or absorbed, which
converts the light to heat. Particulate matter and gases can both scatter and absorb light. Fine
particles with significant light-extinction efficiencies include sulfates, nitrates, organic carbon,
elemental carbon, and soil (Sisler, 1996). The extent to which any amount of light extinction
affects a person's ability to view a scene depends on both scene and light characteristics. For
example, the appearance of a nearby object (e.g., a building) is generally less sensitive to a
change in light extinction than the appearance of a similar object at a greater distance. See
Figure 6-1 for an illustration of the important factors affecting visibility.
According to the PM ISA, there is strong and consistent evidence that PM is the
overwhelming source of visibility impairment in both urban and remote areas (U.S. EPA,
2009b). After reviewing all of the evidence, the PM ISA concluded that the evidence was
sufficient to conclude that a causal relationship exists between PM and visibility impairment.
2 We use the term VAQ to refer to the visibility effects caused solely by air quality conditions, excluding fog.
6-6
-------
Light from clouds
scattered Into
sight path v
Image-forming
light scattered
out of sight path
Sunlight X
scattered
X
reflected
from ground
scattered Into
sight path
Image-forming
light absorbed
Figure 6-1. Important Factors Involved in Seeing a Scenic Vista (Malm, 1999)
Visibility is commonly measured as either light extinction (3ext), which is defined as the
loss of light per unit of distance in terms of inverse megameters (Mm"1), or using the deciview
(dv) metric, which is a logarithmic function of extinction (Pitchford and Malm, 1994). Deciviews,
a unitless measure of visibility, are standardized for a reference distance in such a way that one
deciview corresponds to a change of about 10% in available light.3 Pitchford and Malm (1994)
characterize a change of one deciview as "a small but perceptible scenic change under many
circumstances."4 Extinction and deciviews are both physical measures of the amount of
visibility impairment (e.g., the amount of "haze"), with both extinction and deciview increasing
as the amount of haze increases. Using the relationships derived by Pitchford and Malm (1994),
Note that deciviews are inversely related to visual range, such that a decrease in deciviews implies an increase in
visual range (i.e., improved visibility). Conversely, an increase in deciviews implies a decrease in visual range (i.e.,
decreased visibility). Deciview, in effect, is a measure of the lack of visibility.
4 An instantaneous change of less than 1 deciview (i.e., less than 10% in the light extinction budget) represents a
measurable improvement in visibility but may not be perceptible to the eye. The visibility co-benefits approach
described in this chapter reflects annual average changes in visibility, which are likely made up of periods with
changes less than one deciview and periods with changes exceeding one deciview. Annual averages appear to
more closely correspond to the economic literature relied upon for valuation of visibility changes in this analysis.
6-7
-------
Deciviews = 10 * In (—) = 10 * In (^) (6.1)
V VR / V 10 /
where VR denotes visual range (in kilometers) and 3ext denotes light extinction (in Mm"1).5
Annual average visibility conditions (reflecting light extinction due to both
anthropogenic and non-anthropogenic sources) vary regionally across the U.S. and by season
(U.S. EPA, 2009b). Particulate sulfate is the dominant source of regional haze in the eastern U.S.
(>50% of the particulate light extinction) and an important contributor to haze elsewhere in the
country (>20% of particulate light extinction) (U.S. EPA, 2009b). Particulate nitrate is an
important contributor to light extinction in California and the upper Midwestern U.S.,
particularly during winter (U.S. EPA, 2009b). Smoke plumes from large wildfires dominate many
of the worst haze periods in the western U.S., while Asian dust only caused a few of the worst
haze episodes, primarily in the more northerly regions of the west (U.S. EPA, 2009b). Higher
visibility impairment levels in the East are due to generally higher concentrations of fine
particles, particularly sulfates, and higher average relative humidity levels (U.S. EPA, 2009b).
Humidity increases visibility impairment because some particles such as ammonium sulfate and
ammonium nitrate absorb water and form droplets that become larger when relative humidity
increases, thus resulting in increased light scattering (U.S. EPA, 2009b).
Reductions in air pollution from implementation of various programs associated with
the Clean Air Act Amendments of 1990 (CAAA) provisions have resulted in substantial
improvements in visibility, and will continue to do so in the future. Because trends in haze are
closely associated with trends in particulate sulfate and nitrate due to the simple relationship
between their concentration and light extinction, visibility trends have improved as emissions
of S02 and NOX have decreased over time due to air pollution regulations such as the Acid Rain
Program (U.S. EPA, 2009b). For example, Figure 6-2 shows that visual range increased nearly
50% in the eastern U.S. since 1992.6 The EPA's recent regulations such as the Cross-State Air
Pollution Rule (U.S. EPA, 2011c) and the Mercury and Air Toxics Standard (U.S. EPA, 2011d) are
anticipated to reduce S02 emissions down to 2 million tons nationally, which would lead to
substantial further improvement in visibility levels in the Eastern U.S. Calculated from light
extinction efficiencies from Trijonis et al. (1987, 1988), annual average visual range under
natural conditions in the East is estimated to be 150 km ± 45 km (i.e., 65 to 120 miles) and 230
km ± 35 km (i.e., 120 to 165 miles) in the West (Irving, 1991). Figure 6-2 reflects the average
5 It has been noted that, for a given deciview value, there can be many different visual ranges, depending on the
other factors that affect visual range—such as light angle and altitude. See Appendix 6a for more detail.
' In Figure 6-2, the "best days" are defined as the best 20% of days, the "mid-range days" are defined as tl
20%, and the "worst days" are defined as the worst 20% of days (IMPROVE, 2010).
6-8
-------
trends in visual ranges at select monitors in the eastern and western areas of the U.S. since
1992 using data from the IMPROVE monitoring network (U.S. EPA (2008) updated; IMPROVE
(2010)). As an illustration of the improvements in visibility attributable to the CAAA, Figure 6-3
depicts the modeled improvements in visibility associated with all the CAAA provisions in 2020
compared to a counterfactual scenario without the CAAA (U.S. EPA, 2011b). While visibility
trends have improved in most National Parks, the recent data show that these areas continue
to suffer from visibility impairment beyond natural background levels (U.S. EPA, 2009b).
A. Western U.S.
B. Eastern U.S.
Best visibility days.
Mid-range visibility days
Worst visibility days
'92
'94
'96
'98
'00
'02
'04
'06
Year
Coverage: 30 monitoring sites in the western U.S. and 11 monitoring sites in the eastern U.S. with sufficient •/
data to assess visibility trends from 1992 to 2008.
Visual ranges are calculated from the measured levels of different components within airborne particles and
these components' light extinction efficiencies.
Data source: IMPROVE, 2010
Monitoring sites
East
Figure 6-2. Visibility in Selected National Parks and Wilderness Areas in the U.S., 1992-
2008a'b
Source: U.S. EPA (2008) updated, IMPROVE (2010).
6-9
-------
Without CAAA
With CAAA
15 18 24 27 30
Visibility in Deciviews, 2020
0369
Best
Figure 6-3. Estimated Improvement in Annual Average Visibility Levels Associated with the
CAAA Provisions in 2020
Source: U.S. EPA, 2011b.7
7 It is important to note that visibility levels shown in these maps were modeled differently than the modeling
conducted for this analysis using an earlier method that we would currently use, including coarser grid resolution
(i.e., 36 km instead of 12 km). In addition, please note that these maps present annual average visibility levels,
which are different than the short-term averages being considered for the secondary standard.
6-10
-------
6.3.2 Quantifying Light Extinction for Assessing Visibility Co-benefits
For this RIA, we do not have air quality model runs for the regulatory baseline and the
alternative standard levels that would allow us to calculate the visibility co-benefits of attaining
the revised primary standard. However, we provide an illustrative analysis in Appendix 6.B
using the 2020 base case and 2020 control case simulations that were used to develop the air
quality ratios.8 In our approach, we generate light extinction estimates using the CMAQ model
in conjunction with the IMPROVE (Interagency Monitoring of Protected Visual Environments)
algorithm that estimates light extinction as a function of PM concentrations and relative
humidity levels (U.S. EPA, 2009b).9 The procedure for calculating light extinction associated with
the revised and alternative annual standards is described in detail in Chapter 3 of this RIA. In
addition, Appendix 6.A describes how the spatial resolution of the light extinction estimates
would be adjusted in our approach.
It is important to note that the light extinction estimates used in our approach represent
annual averages, which is different from the averaging times currently being considered for the
secondary PM NAAQS. While the annual averages are influenced by days with extremely
impaired visibility, the light extinction data is not sufficient to provide higher temporal
resolution than quarterly averages. While we suspect that the most impaired days would have
disproportionately improved visibility from the emission reductions associated with attaining
the revised or alternative primary standards, we are not able to quantify those impacts. These
data gaps result in an underestimate of visibility co-benefits associated with extreme days. We
recognize that recent advice from the Science Advisory Board's Advisory Council on Clean Air
Compliance Analysis (SAB-Council) recommends estimating visibility co-benefits considering
daytime visibility on days with severe impairment (U.S. EPA-SAB, 2010a), but the available data
and valuation studies do not allow such fine temporal resolution.
While our approach is a substantial improvement in the methods to estimate light
extinction nationally, we are still developing a method to estimate coarse particle
concentrations for the entire continental U.S. for estimating light extinction. As an interim
solution, our approach includes sensitivity analyses to show the potential impact of omitting
coarse particles from the light extinction estimates for recreational and residential visibility. For
these sensitivity analyses, we selected the levels of coarse particles to represent the full range
of possible annual concentrations from a recent report on the IMPROVE monitoring network
These simulations are described in Chapter 3.
9 According to the PM ISA, the IMPROVE algorithm performs reasonably well despite its simplicity (U.S. EPA,
2009b).
6-11
-------
(Debell et al., 2006). Specifically, for these sensitivity analyses, we assume four levels of coarse
particles: no coarse particles, 5 u.g/m3 nationwide, 15 u.g/m3 in the Southwest with 5 u.g/m3 in
the rest of the country, and 15 u.g/m3 in the Southwest with 8 u.g/m3 in the rest of the
country.10 In Table 6-3, we provide a qualitative assessment of how key assumptions in the
estimation of light extinction would affect the visibility co-benefits.
Table 6-3. Key Assumptions in the Light Extinction Estimates Affecting the Visibility Co-
Benefits Approach3
Key Assumption Direction of Bias Magnitude of Effect
The light extinction estimates are annual averages to
correspond with the valuation studies. People may value large
changes to the haziest days differently than small changes to Potential
many days. We assume that annual average light extinction is Underestimate
the most appropriate temporal scale for estimating visibility
benefits.
Coarse particles are a component of light extinction, but we
were unable to include coarse particles in the light extinction Potential
a Very Low
estimates. We provide sensitivity analyses with up to 15 ng/m Overestimate
in the Southwest and 8 ng/m3 in the rest of the country.
a A description of the classifications for magnitude of effects can be found in Appendix 5.B of this RIA.
6.3.3 Visibility Valuation Overview
In the Clean Air Act Amendments of 1977, the U.S. Government recognized visibility's
value to society by establishing a national goal to protect national parks and wilderness areas
from visibility impairment caused by manmade pollution.11 Air pollution impairs visibility in both
residential and recreational settings, and an individual's willingness to pay (WTP) to improve
visibility differs in these two settings. Benefits of residential visibility relate to the impact of
visibility changes on an individual's daily life (e.g., at home, at work, and while engaged in
routine recreational activities). Benefits of recreational visibility relate to the impact of visibility
changes manifested at parks and wilderness areas that are expected to be experienced by
recreational visitors.
Both recreational and residential visibility benefits consist of use values and nonuse
values. Use values include the aesthetic benefits of better visibility, improved road and air
safety, and enhanced recreation in activities like hunting and birdwatching. Nonuse values are
10 We define "Southwest" for this sensitivity analysis to be the states of California, Nevada, Utah, Arizona, New
Mexico, Colorado, and Texas.
11 See Section 169(a) of the Clean Air Act.
6-12
-------
based on a belief that the environment ought to exist free of human-induced haze. This
includes the value of better visibility for use by others now and in the future (bequest value).
Nonuse values may be more important for recreational areas, particularly national parks and
monuments.
The relationship between a household's WTP and changes in visibility can be derived
from a number of contingent valuation (CV) studies published in the peer-reviewed economics
literature. The studies used in the approach to estimate the residential and recreational
visibility co-benefits associated with the revised and alternative annual standards are described
in the following sections. In addition to CV studies, hedonic valuation studies (Beron et al.,
2001, 2004) also demonstrate that visibility has value, but we are unable to apply these
valuation estimates in the context of estimating the visibility co-benefits associated with
national regulations that reduce air pollution (Leggett and Neumann, 2004).
6.3.3.1 Visibility Valuation Approach
In our approach, we assume that individuals value visibility for aesthetic reasons rather
than viewing visibility as a proxy for other impacts associated with air pollution, such as health
or ecological improvements. Some studies in the literature indicate that individuals may have
difficulty distinguishing visibility from other aspects of air pollution (e.g., McClelland et al.,
1993; Chestnut and Rowe, 1990c; Carson, Mitchell, and Rudd, 1990). Because visual air quality
is inherently multi-attribute, it is a challenge for all visibility valuation studies to isolate the
value of visibility from the collection of intertwined benefits. Each study used in our approach
attempts to isolate visibility from other effect categories, but the different studies take
different approaches (U.S. EPA, 2009b).12 However, the degree to which the studies were
successful in convincing respondents to focus solely on visibility is unclear
Similarly, it is important to try to distinguish residential visibility from recreational
visibility co-benefits, specifically whether these can these can be treated as distinct and additive
benefit categories based on the available literature. In our approach, we assume that
residential and recreational visibility co-benefits are distinct and separable. It is conceivable
that respondents to the recreational visibility survey may have partially included values for their
own residential visibility when evaluating changes at national parks and wilderness areas
located in their region of the country. In our approach, we take care to minimize the number of
overlapping areas and their contributions. Specifically, we believe that the potential for double-
counting recreational and residential visibility is minimal for several reasons. First, in our
12 See Leggett and Neumann (2004) for a more detailed discussion of this issue.
6-13
-------
approach, we only include a subset of areas in the primary estimates of recreational and
residential visibility co-benefits, which overlap in only a few places.13 Second, a number of the
overlapping counties are wilderness areas, which would contribute little to the overall
monetized visibility co-benefits due to low visitation rates, rather than highly visited national
parks. For example, Los Angeles County is home to the San Gabriel Wilderness Area, which has
10 thousand annual visitors (NFS, 2008). If we were to exclude the residential visibility co-
benefits that accrue to 10 million residents in Los Angeles County and only include the very
small recreational visibility co-benefits for the wilderness area, we would be substantially
biasing the overall estimates downward. For these reasons, we believe that the potential for
double-counting is minimal.
In the next sections, we describe the methodology and limitations of the recreational
and residential visibility approach. Consistent with the health benefits analysis, the monetized
visibility co-benefits would be adjusted for inflation and income growth. These co-benefits
would be specific to the analysis year, and as population and income increase over time, these
co-benefits would be expected to increase each year for the same incremental change in light
extinction.
6.3.4 Recreational Visibility
6.3.4.1 Methodology
Our approach for estimating recreational visibility co-benefits is well-established and
has been used in numerous analyses by the EPA (U.S. EPA, 1999; 2005; 2006; 2010; 2011b). In
our approach, recreational visibility co-benefits apply to Class 1 areas, such as National Parks
and Wilderness Areas.14 Although other recreational settings such as National Forests, state
parks, or even hiking trails or roadside areas have important scenic vistas, a lack of suitable
economic valuation literature to identify these other areas and/or a lack of visitation data
prevents us from generating estimates for those recreational vista areas.
Under the 1999 Regional Haze Rule (64 FR 35714), states are required to set goals
develop long-term strategies to improve visibility in Class 1 areas, with the goal of achieving
13 As described in detail in Sections 6.3.3 and 6.3.4, our approach includes only a subset of visibility co-benefits in
the main benefits estimates, while providing the rest of the visibility co-benefits in sensitivity analyses.
14 Hereafter referred to as Class 1 areas, which are defined as areas of the country such as national parks, national
wilderness areas, and national monuments that have been set aside under Section 169(a) of the Clean Air Act to
receive the most stringent degree of air quality protection. Mandatory Class 1 federal lands fall under the
jurisdiction of three federal agencies, the National Park Service, the Fish and Wildlife Service, and the Forest
Service. EPA has designated 156 areas as mandatory Class 1 federal areas for visibility protection, including
national parks that exceed 6,000 acres and wilderness areas that exceed 5,000 acres (40 CFR §81.400).
6-14
-------
natural background visibility levels by 2064. In conjunction with the U.S. National Park Service
(NFS), the U.S. Forest Service (USFS), other Federal land managers, and State organizations in
the U.S., the EPA has supported visibility monitoring in national parks and wilderness areas
since 1988. The monitoring network known as IMPROVE includes 156 sites that represent the
Class 1 areas across the country (U.S. EPA, 2009b).15 The IMPROVE monitoring network
measures fine particles, coarse particles, and key PM2.s constituents that affect visibility, such
as sulfate, nitrate, organic and elemental carbon, soil dust, and several other elements.
Figure 6-4 identifies where each of these parks are located in the U.S.
Produced by NPS Air Resources Division
* Rainbow Lake, Wl and Bradwell Bay, FL are Class 1 Areas
where visibility is not an important air quality related value
••"-Tlafcakato
X/^WMH Vcfcanew
NPS Units
FWS Units
• FS Units
Figure 6-4. Mandatory Class 1 Areas in the U.S.
For recreational visibility, the EPA relies upon a contingent valuation (CV) survey
conducted in 1988 (Chestnut and Rowe, 1990a; 1990b) to estimate the recreational visibility co-
benefits. Although there are several other studies in the literature on recreational visibility
valuation, they are even older and use less robust methods. In the EPA's judgment, despite the
inherent limitations in the survey, the Chestnut and Rowe study served as the basis for
The formula used to estimate light extinction from concentrations of PM constituents and relative humidity is
referred to as the IMPROVE algorithm.
6-15
-------
monetary estimates of the co-benefits of visibility changes in recreational areas fora number of
previous EPA rulemakings. This study serves as an essential input to our approach for
estimating the co-benefits from improving recreational visibility.
In our approach, we assume that the household WTP is higher if the Class 1 recreational
area is located close to the person's home (i.e., in the same region of the country). People
appear to be willing to pay more for visibility improvements at parks and wilderness areas that
are in the same region as their household than at those that are not in the same region as their
household (Chestnut and Rowe, 1990a, 1990b). This is plausible, because people are more likely
to visit, be familiar with, and care about parks and wilderness areas in their own part of the
country. However, studies have also found many people who had never visited and never
planned to visit the parks still had positive values for visibility improvements in those locations
(Chestnut and Rowe, 1990b).
The Chestnut and Rowe survey measured the demand for visibility in Class 1 areas
managed by the NPS in three broad regions of the country: California, the Colorado Plateau
(Southwest), and the Southeast.16 Respondents in five states were asked about their WTP to
protect national parks or NPS-managed wilderness areas within a particular region. The survey
used photographs reflecting different visibility levels in the specified recreational areas. The
authors used the survey data to estimate household WTP values for improved visibility in each
region.
The separate regions were developed to capture differences in household WTP values
based on proximity to recreational areas. Chestnut (1997) also concluded that, fora given
region, a substantial proportion of the WTP is attributable to one specific park within the
region. This so called "indicator park" is the most well-known and frequently visited park within
a particular region. The indicator parks for the three studied park regions are Yosemite National
Park for the California region, the Grand Canyon National Park for the Southwest region, and
Shenandoah National Park for the Southeast region. In accordance with the methodology in
Chestnut (1997), our approach calculates the benefits from households for a particular region
for a given change in visibility at a particular Class 1 area. In theory, summing benefits from
households in all regions would yield the total monetary benefits associated with a given
visibility improvement at a particular park, which could then be summed with other parks and
16 The Colorado Plateau (Southwest) region is defined as the states of Colorado, New Mexico, Arizona, and Utah.
The Southeast region is defined as the states of West Virginia, Virginia, North Carolina, South Carolina, Georgia,
Florida, Alabama, Mississippi, Louisiana, Tennessee, and Kentucky. The California region includes the state of
California and one wilderness area in Nevada.
6-16
-------
regions to estimate national benefits. Because recreational visibility benefits may reflect the
value an individual places on visibility improvements regardless of whether the person plans to
visit the park, all households in the U.S. are assumed to derive some benefit from
improvements to Class 1 areas.
To value recreational visibility improvements associated with its rulemakings, the EPA
developed a valuation WTP equation function based on the baseline of visibility, the magnitude
of the visibility improvement, and household income. This function requires light extinction
estimates measured as visual range. The behavioral parameters of this equation were taken
from an analysis of the survey described in Chestnut and Rowe (1990a, 1990b). These
parameters were used to calibrate WTP for the visibility changes resulting from this rule.17 As an
example, household WTP for a visibility improvement at a park in its region takes the following
form:
WTP(A
-------
1997). Using the income elasticity calculated by Chestnut (1997), the recreational visibility
benefits assume a 1% increase in income is associated with a 0.9% increase in WTP for a given
change in visibility. WTP responses reported in Chestnut and Rowe (1990a, 1990b) were also
region-specific, rather than park-specific. As visibility improvements are not constant across all
parks in a region, we must infer park-specific visibility parameters in order to calculate WTP for
projected visibility changes. As the quantity and quality of parks differs between regions, we
apportion the regional WTP parameters based on relative visitation rates at the different parks,
because this statistic likely captures both park quality (more people visit parks with more
desirable attributes, so collective WTP is likely higher) and quantity (more people visit parks in a
region if the parks are more numerous, so collective WTP is likely higher).19 We also adjust the
co-benefits for inflation and growth in real income.
Recreational visibility co-benefits can be calculated as the sum of the household WTPs
for changes in light extinction. We assume that each household is valuing the first or only
visibility change that occurs in a particular area. The co-benefits at particular areas can be
calculated by assuming that the subset of visibility changes of interest is the first or the only set
of changes being valued by households. Estimating benefit components in this way will yield
slightly upwardly biased estimates of co-benefits, because disposable income is not reduced by
the WTPs for any prior visibility improvements. The upward bias should be extremely small,
however, because all of the WTPs for visibility changes are very small relative to income.
In our approach, the primary estimate for recreational visibility only includes co-benefits
for 86 Class 1 areas in the original study regions (i.e., California, the Southwest, and the
Southeast).20 These co-benefits reflect the value to households living in the same region as the
Class 1 area as well as values for all households in the United States living outside the state
containing the Class 1 area.
19 We use 2008 park visitation data from the National Park Service Statistical Abstracts (IMPS, 2008), as this is the
most current data available. Where the data for a particular park was not representative of normal visitation
rates at that park (for example due to fire damage that occurred during that year), we substitute data from the
prior year. We use 1997 visitation data for those wilderness areas not included in the National Park Service
Statistical Abstracts, as more current data is not readily available. As visitation rates for Wilderness Areas are
small compared to visitation rates in National Parks, the inaccuracies generated by using 1997 data are likely to
also be small.
20 The 86 Class 1 areas in the three studied park regions represented 68% of the total visitor days to Class 1 areas in
2008 (NPS, 2008).
6-18
-------
The Chestnut and Rowe study did not measure values for visibility improvement in Class
1 areas in the Northwest, Northern Rockies, and Rest of U.S. regions.21 In order to obtain
estimates of WTP for visibility changes for the 70 additional Class 1 areas in these non-studied
regions, we have to transfer the WTP values from the studied regions.22 This co-benefits
transfer approach introduces additional uncertainty into the estimates. However, we have
taken steps to adjust the WTP values to account for the possibility that a visibility improvement
in parks within one region may not necessarily represent the same visibility improvement at
parks within a different region in terms of environmental improvement. This may be due to
differences in the scenic vistas at different parks, uniqueness of the parks, or other factors, such
as public familiarity with the park resource. To account for this potential difference, we
adjusted the transferred WTP being transferred by the ratio of visitor days in the two regions.23
A complete description of the co-benefits transfer method used to infer values for visibility
changes in Class 1 areas outside the study regions is provided in Appendix 6a of this RIA.
Table 6-4 indicates which studied park regions we used to estimate the value in the non-
studied park regions in our approach. Figure 6-5 shows how the visitation rates vary across
Class 1 areas and regions and indicates whether each Class 1 area is located within one of the
studied regions.
Table 6-4. WTP for Visibility Improvements in Class 1 Areas in Non-Studied Park Regions
Park Region
Source of WTP Estimate
1. Northwest
2. Northern Rockies
3. Rest of U.S.
Benefits transfer from California
Benefits transfer from Colorado Plateau
Benefits transfer from Southeast
The Northwest region is defined as the states of Washington and Oregon. The Northern Rockies region includes
the states of Idaho, Montana, Wyoming, North Dakota, and South Dakota. The Rest of the U.S. region includes all
other states not included in the other 5 regions.
22 The 70 additional Class 1 areas represented 32% of the total visitor days to Class 1 areas in 2008 (NPS, 2008).
23 For example, if total park visitation in a transfer region was less than visitation in a study region, transferred WTP
would be adjusted downward by the ratio of the two.
6-19
-------
O
Visitor_daysv*
0 - 50,000
O 50,001-500,000
(_) 500,001-1,000.000
1,000.001 -3,000.000
3.000.001 -6.000.000
Figure 6-5. Visitation Rates and Park Regions for Class 1 Areas3
aThe colors in this map correspond to the park regions used in the valuation study and the extrapolation to parks
in other regions. Red = California, light red = Northwest (extrapolated from California), blue = Colorado Plateau,
light blue = Northern Rockies (extrapolated from Colorado Plateau), green = Southeast, light green = Rest of U.S.
(extrapolated from Southeast).
In a more recent study, Smith et al. (2005) conducted a contingent valuation survey that
updated and expanded a portion of the 1988 survey by Chestnut and Rowe (1990). Specifically,
the Smith et al. (2005) survey relied on a panel maintained by Knowledge Networks with 2,020
participants completing the survey. Similar to the Chestnut and Rowe survey, the Smith et al.
survey assessed WTP for changes in summertime visibility using the base photograph of
Shenandoah National Park. Unlike the Chestnut and Rowe survey, the Smith et al. survey only
assessed the Shenandoah National Park, did not estimate in-region estimates of WTP, and
evaluated several options for incorporating budgetary constraints into the survey. The authors
concluded that WTP for recreational visibility is skewed and sensitive to information about
budgetary constraints. We are still evaluating the potential error identified by Smith et al.
(2005) regarding the visibility levels in the photographs for Shenandoah National Park in the
Chestnut and Rowe survey (1990a,b).
Even though this survey represents several advantages over the older survey (e.g., more
recent, national, demographically representative, larger sample, etc.), we are unable
6-20
-------
incorporate the results generated by this survey into our existing method for calculating
recreational visibility co-benefits because the survey did not account for the differences
between WTP for in-region parks and out-of-region parks. This omission precludes us from
combining this new survey for only 1 region of the country with the WTP for the other regions
of the country from Chestnut and Rowe (1990). Furthermore, Smith et al. (2005) provide a
variety of WTP estimates reflecting different versions of the survey and different methods of
summarizing the typical response, which makes it difficult to select estimates to incorporate
into the recreational visibility benefits calculation.
6.3.4.2 Recreational Visibility Limitations, Caveats, and Uncertainties
Our approach relies upon several data sources as inputs, including emission inventories,
air quality data from models (with their associated parameters and inputs), relative humidity
measurements, park information, economic data and assumptions for monetizing co-benefits.
Each of these inputs may contain uncertainty that would affect the recreational visibility co-
benefits estimates. Though we are unable to quantify the cumulative effect of all of these
uncertainties in our approach, we do provide information on uncertainty based on the available
data, including model evaluation24 and sensitivity analyses to characterize major omissions (i.e.,
benefits from parks in non-studied park regions and inclusion of coarse particles). Although we
strive to incorporate as many quantitative assessments of uncertainty as possible, we are
severely limited by the available data, and there are several aspects that we are only able to
address qualitatively. A summary of the key assumptions including direction and magnitude of
bias is provided in Table 6-5.
One major source of uncertainty for the estimation of recreational visibility co-benefits
is the benefits transfer process. Choices regarding the functional form and key parameters of
the estimating equation for WTP for the affected population could have significant effects on
the magnitude of the estimates. Assumptions about how individuals respond to changes in
visibility that are either very small or outside the range covered in the Chestnut and Rowe study
could also affect the estimates.
24See Chapter 4 for more information on model evaluation.
6-21
-------
Table 6-5. Summary of Key Assumptions in Estimating Recreational Visibility Co-benefits3
Key Assumption
Potential Magnitude
Direction of Bias of Effect
Chestnut and Rowe study covers parks in three regions: California, Underestimate Medium
Southwest, and Southeast. Benefits to other regions in the U.S.
are not included in the primary benefits estimate.
Benefits to other recreational settings, such as National Forests Underestimate Medium-Low
and state parks, are not included in our approach.
Chestnut and Rowe study conducted on populations in five states. Unclear Unclear
These results are applied to the entire U.S. population.
Individuals have a greater WTP for visibility changes in parks Unclear Unclear
within their region.
WTP values reflect only visibility improvements and not overall air Potential Unclear
quality improvements. Overestimate
We assume that there are 2.68 people per household. Because Potential Medium-Low
this estimate has been decreasing over time, this may Underestimate
underestimate the number of households.
a A description of the classifications for magnitude of effects can be found in Appendix 5.B of this RIA.
Since the valuation of recreational visibility co-benefits relies upon one study (Chestnut
and Rowe, 1990a; 1990b), all of the uncertainties within that study also pertain to any analysis
that uses it. In general, the survey design and implementation reflect the period in which the
Chestnut and Rowe study was conducted. Since that time, many improvements to the design of
stated preference surveys have been developed (e.g., Arrow, 1993), but we are currently
unaware of newer studies that we could incorporate into our visibility co-benefits
methodology. Although Chestnut and Rowe still offers the best available WTP estimates, the
study has a number of limitations, including:
• The vintage of the survey (late 1980s) invites questions whether the values would
still be valid for current populations, or more importantly for our approach, future
populations in 2020.
• The survey focused on visibility improvements in and around specific national parks
and wilderness areas. Given that national parks and wilderness areas exhibit unique
characteristics, it is not clear whether the WTP estimate obtained from this survey
can be transferred to other national parks and wilderness areas, even other parks
within the studied park regions, without introducing additional uncertainty.
6-22
-------
• The survey focused only on populations in five states, so the application of the
estimated values to populations outside those states requires that preferences of
populations in the five surveyed states be similar to those of non-surveyed states.
• There is an inherent difficulty in separating values expressed for visibility
improvements from an overall value for improved air quality. The survey attempted
to control for this by informing respondents that "other households are being asked
about visibility, human health, and vegetation protections in urban areas and at
national parks in other regions." However, most of the respondents did not feel that
they were able to segregate recreational visibility at national parks entirely from
residential visibility and health effects.
• It is not clear exactly what visibility improvements the respondents to the survey
were valuing. The WTP question asked about changes in average visibility, but the
survey respondents were shown photographs of only daytime, summer conditions,
when visibility is generally at its worst. It is possible that the respondents believed
those visibility conditions held year-round, in which case they would have been
valuing much larger overall improvements in visibility than what otherwise would be
the case. In our approach, the EPA assumed that respondents provided values for
changes in annual average visibility. Because most policies would result in a shift in
the distribution of visibility (usually affecting the worst days more than the best
days), the annual average may not be the most relevant metric for policy analysis.
• The survey did not include reminders of possible substitutes (e.g., visibility at other
parks) or budget constraints. These reminders are considered to be best practice for
stated preference surveys.
6.3.5 Residential Visibility
6.3.5.1 Methodology
Residential visibility co-benefits are those that occur from visibility changes in urban,
suburban, and rural areas where people live. These co-benefits are important because some
people living in certain urban areas may place a high value on unique scenic resources in or
near these areas that are outside of Class 1 areas. For example, the State of Colorado
established a local visibility standard for the Denver metropolitan area in 1990 (Ely et al., 1991).
In our approach, residential visibility improvements are defined as those that occur specifically
in Metropolitan Statistical Areas (MSAs).
In the Urban-focused Visibility Assessment (U.S. EPA, 2010b) and the Policy Assessment
for the Review of the PM NAAQS (U.S. EPA, 2011a), several preference studies provide the
6-23
-------
foundation for the secondary PM NAAQS.25 The three completed survey studies (all in the west)
included Denver, Colorado (Ely et al., 1991), one in the lower Fraser River valley near
Vancouver, British Columbia (BC), Canada (Pryor, 1996), and one in Phoenix, Arizona (BBC
Research & Consulting, 2003). A pilot focus group study was conducted in Washington, DC on
behalf of the EPA to inform the 2006 PM NAAQS review (Abt Associates Inc., 2001). Although
these studies indicate that some individuals considered the visual air quality associated with
ambient levels of air pollution in urban areas to be unacceptable, these studies do not provide
sufficient information on which to develop monetized co-benefits estimates. Specifically, the
public perception studies do not provide preferences expressed in dollar values, even though
they do suggest that the co-benefits associated with improving residential visibility are positive.
Studies in the peer-reviewed literature support a non-zero value for residential visibility
(e.g., Brookshire et al., 1982; Loehman et al., 1994). Furthermore, Chestnut and Rowe (1990c)
conclude that residential visibility co-benefits are likely to be at least as high as recreational
visibility co-benefits because of the quantity of time most people spend in and near their
homes and the substantial number of people affected. In previous assessments, the EPA used a
study on residential visibility valuation conducted in 1990 (McClelland et al., 1993). Consistent
with advice from SAB-Council, the EPA designated the McClelland et al. study as significantly
less reliable for regulatory benefit-cost analysis, although it does provide useful estimates on
the order of magnitude of residential visibility co-benefits (U.S. EPA-SAB, 1999).26 In our
approach for estimating residential visibility co-benefits, we replaced the previous methodology
with a new benefits transfer approach and incorporated additional valuation studies. This new
approach was developed for The Benefits and Costs of the Clean Air Act 1990 to 2020: EPA
Report to Congress (U.S. EPA, 2011)27 and reviewed by the SAB-Council (U. S. EPA-SAB, 2004,
2010a, 2010b).
25 For more detail about these preference studies, including information about study designs and sampling
protocols, please see Section 2 of the Particulate Matter Urban-Focused Visibility Assessment (U.S. EPA, 2010b).
26 EPA's Advisory Council on Clean Air Compliance Analysis noted that the McClelland et al. (1993) study may not
incorporate two potentially important adjustments. First, their study does not account for the "warm glow"
effect, in which respondents may provide higher willingness to pay estimates simply because they favor "good
causes" such as environmental improvement. Second, while the study accounts for non-response bias, it may not
employ the best available methods. As a result of these concerns, the Council recommended that residential
visibility be omitted from the overall primary benefits estimate. (U.S. EPA-SAB, 1999)
27 This report is also known as the Second Prospective 812 analysis.
6-24
-------
To value residential visibility improvements, the new approach draws upon information
reported in the Brookshire et al. (1979), Loehman et al. (1985) and Tolley et al. (1986) studies.28
Each of the studies provides estimates of household WTP to improve visibility conditions. While
uncertainty exists regarding the precision of these older, stated-preference residential
valuation studies, we believe their results support the argument that individuals have a non-
zero value for residential visibility improvements. These studies provide primary visibility values
for Atlanta, Boston, Chicago, Denver, Los Angeles, Mobile, San Francisco, and Washington
D.C.29
In accordance with Chestnut and Rowe (1990c), we utilize the WTP estimates and the
associated change in visual range from each study to estimate the 3 parameter for the eight
study areas. The 3 parameter represents the WTP for a specific improvement in visibility in a
specific location. Where studies provide multiple estimates for visual range improvements, we
estimate 3 by regressing the natural log of the ratio of visual range following and prior to
improvement against WTP. To express these value estimates in comparable terms across study
locations, we express household WTP for a change in visual range in a specific MSA using the
following function:
WTP(AVR)=3*ln(^) (6.3)
where:
VR0 = mean annual visual range in miles before the improvement,
VRi = mean annual visual range in miles after the improvement, and
3 = parameter.
28 Loehman et al. (1994) and Brookshire et al. (1982) published results in peer-reviewed journals based on the
same underlying data we obtained from Loehman et al. (1985) and Brookshire et al. (1979). While the specific
details need to compute visibility benefits using Tolley et al. (1986) were not published in a peer-reviewed
journal, the overall work including study and survey design was subject to peer review during study development
(see Leggett and Neumann, 2004 and Patterson et al., 2005). In addition, Tolley subsequently published a book
(Tolley and Fabian, 1988) based on this research, which notes in the preface that the methods were critiqued
throughout by various external economists. The EPA does not claim that this external critique necessarily
constituted a formal peer review process, but we provide this information for transparency regarding the review
of this work. The use of these studies as the only available information to estimate residential visibility co-benefits
in the main estimate was supported by the SAB-Council (U.S. EPA-SAB, 1999a, 2010a).
29 Recognizing potential fundamental issues associated with data collected in Cincinnati and Miami (e.g., see
Chestnut et al. (1986) and Chestnut and Rowe (1990c), we do not include values for these cities in our analysis.
The 8 MSAs where the valuation studies were conducted represent 15% of the total US population in 2020 (U.S.
Census).
6-25
-------
Total residential visibility co-benefits within a particular MSA are driven by visibility
improvements, population density, and the WTP value applied. Only those people living within
in an MSA are assumed to receive co-benefits from improved residential visibility. In other
words, unlike recreational visibility, we do not assume a non-use value by people who live
outside the MSA for residential visibility. Table 6-6 provides a summary of these valuation
estimates for each study location, as well as an illustrative implied WTP value for a 10%
improvement in visual range. As shown, the implied annual per-household WTP estimates for a
hypothetical 10% improvement ranges from $21 to $220, depending on the study area. It is not
surprising that such a range of values exists, as these study areas all feature different
landscapes and vistas, populations and prevailing visibility conditions.
Table 6-6. Summary of Residential Visibility Valuation Estimates
City
Study
(3 Estimate
Implied WTP for 10%
Improvement in Visual
Range (1990$, 1990
income)
Implied WTP for 10%
Improvement in Visual
Range (2006$, 2020
income)
Atlanta
Boston
Chicago
Denver
Los Angeles
Mobile
San Francisco
Washington, DC
Tolleyetal. (1986)
Tolleyetal. (1986)
Tolley et al. (1986)
Tolley et al. (1986)
Brookshireetal. (1979)
Tolley et al. (1986)
Loehman etal. (1985)
Tolley et al. (1986)
321
398
310
696
94
313
989
614
$31
$38
$30
$66
$9
$30
$94
$59
$72
$89
$69
$155
$21
$70
$220
$137
Similar to recreational visibility co-benefits, we then incorporate preference calibration
using the method developed by Smith, Van Houtven, and Pattanayak (2002), which is discussed
in more detail in Appendix 6a of this RIA. This preference calibration is a change since The
Benefits and Costs of the Clean Air Act 1990 to 2020: EPA Report to Congress (U.S. EPA, 2011)
intended to address the SAB-Council's concern (U.S. EPA-SAB, 2010a) regarding the
inconsistency regarding household income in the estimation of recreational and residential
visibility. To express these "preference-calibrated" value estimates across study locations, we
express household WTP for a change in visual range in a specific MSA using the following
function:
6-26
-------
WTP(AVK) = m - [mP + 9 * (VRPQ - VRP^]\P (6-4)
where:
m = household income,
p = shape parameter (0.1),
0 = WTP parameter corresponding to the visibility at MSA,
VR0 = starting visibility, and
VRi = visibility after change.
While the primary estimate for residential visibility includes co-benefits in only the eight
MSAs included in the valuation studies, people living in other urban areas also have non-zero
values for residential visibility. For this reason, our approach includes a sensitivity analysis for
the extrapolated residential visibility in the 351 additional MSAs.30 Because there is
considerable uncertainty about the validity of this benefit transfer approach, these
extrapolated co-benefits are included in a sensitivity analysis only. This is an important
distinction between the approach used in The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress (U.S. EPA, 2011), where all cities were included in the total
benefits approach. We believe that it is appropriate to deviate from the previous approach in
order to be consistent with the approach used to estimate recreational visibility co-benefits and
to recognize the uncertainty associated with extrapolating beyond the studied cities. Figure 6-6
indicates the study cities as well as the assignment of the other MSAs to the study cities.
The degree to which the three studies were successful in convincing respondents to
focus solely on visibility is unclear, as none of the three studies includes follow-up questions
necessary to investigate the issue. Furthermore, no other residential visibility CV studies
provide evidence regarding the degree to which health effects are embedded in visibility
values. Although the McClelland et al. (1991) study has a follow-up question designed to
allocate WTP across several categories, the CV question in the McClelland et al. study was
focused on air pollution generally rather than visibility. As a result, we do not adjust the results
from these studies to account for potentially embedded health effects.
There are many factors that could influence WTP for residential visibility, and these
factors vary across urban areas. In our approach, we utilize the benefit transfer approach
30 The 351 additional MSAs plus the 8 study area MSAs represent 84% of the total US population in 2020 (U.S.
Census).
6-27
-------
developed for The Benefits and Costs of the Clean Air Act 1990 to 2020: EPA Report to Congress
(U.S. EPA, 2011) report, but we recognize that there are alternative methods that we could
have used. We assigned a valuation study area to each MSA based on two factors: geographic
proximity to one of the eight study cities and elevation. Any MSA with a county elevation above
1,500 meters was assigned the Denver valuation instead of the nearest study area.31 Because
residents of Denver have a dramatic view of the Rocky Mountains that is rarely obstructed by
trees, it is plausible that they might have a greater interest in protecting visibility than a city
without nearby mountains. The geographic proximity factor is constrained in two areas. The
San Francisco valuation study is only assigned to the six counties in the San Francisco Bay area
MSAs because the study is unique among the three regarding the temporal description of
visibility conditions, landscape/vistas, and prevailing weather conditions. In addition, the Los
Angeles valuation was assigned to the Riverside MSA despite exceeding the elevation
threshold.32
31 Elevation data represent the county-level maximum, which were calculated using the ArcGIS Spatial Analyst tool
"Zonal Statistics" using the geographic database HYDRO1K for North America (U.S. Geological Survey, 1997). This
dataset and associated documentation are available on the Internet at
DEMhttp://eros.usgs.gov/#/Find_Data/Products_and_Data_Available/gtopo30/hydro/namerica.
32 Riverside MSA is assigned to the Los Angeles study area because a significant portion of Riverside County itself is
located in the South Coast Air Quality Management District, which can be considered by to be part of the same
regulated airshed as Los Angeles. The geographic assignment is preserved despite exceeding the elevation
threshold because Riverside is adjacent to one of the study cities and this region has a particular set of location-
specific characteristics that set it apart from Denver.
6-28
-------
V
I Study_MSAs
MSA_Continental_US
Study
Atlanta
Boston
Chicago
Denver
Los Angeles
Mobile
_ San Francisco
| Washington DC
Figure 6-6. Residential Visibility Study City Assignment
6.3.5.2 Residential Visibility Limitations, Caveats, and Uncertainties
Similar to recreational visibility co-benefits, there are many data inputs into the
residential visibility co-benefits that contribute to overall uncertainty. Our approach includes
sensitivity analyses to characterize major omissions (i.e., co-benefits in other MSAs and coarse
particles). A summary of the key assumptions including direction and magnitude of bias in our
approach is provided in Table 6-7.
The valuation studies relied upon for the residential visibility co-benefits, although
representing the best available estimates, have a number of limitations. These include the
following:
• The survey design and implementation reflects the period in which the surveys were
conducted. Since that time, many improvements to the stated preference methods
have been developed.
• The vintage of the surveys (1970s and 1980s) invites questions whether the values
are still valid for current populations, or more importantly for our approach, future
populations in 2020.
6-29
-------
• The survey focused only on populations in eight cities, so the transfer of the WTP
estimates values to populations outside those cities requires that their preferences
be similar to those in non-surveyed cities, as well as the visibility attributes be
similar across study and transfer MSAs.
• There is an inherent difficulty in separating values expressed for visibility
improvements from an overall value for improved air quality. The studies attempted
to control for this, but most of the respondents did not feel that they were able to
segregate residential visibility entirely from recreational visibility and health effects.
Table 6-7. Summary of Key Assumptions in the Residential Visibility Co-benefits3
Key Assumption
Direction of Bias
Magnitude of
Effect
Residential and recreational visibility benefits are distinct and Potential Medium-Low
separable. Overestimate
Estimates residential visibility benefits are limited to populations Underestimate Low
within the boundaries of MSAs. Areas outside of an MSA are not
included in our approach.
WTP values reflect only visibility improvements and not overall air Potential Medium-Low
quality improvements. Overestimate
WTP values from studies in Atlanta, Boston, Chicago, Denver, Los Unclear Unclear
Angeles, Mobile, San Francisco, and Washington D.C. can be
accurately transferred to MSAs across the U.S. based on proximity
and elevation
We assume that there are 2.68 people per household. Because Potential Medium-Low
this estimate has been decreasing over time, this may Underestimate
underestimate the number of households.
a A description of the classifications for magnitude of effects can be found in Appendix 5.B of this RIA.
6.3.5.3 Using Hedonic Economic Literature to Estimate Visibility Co-benefits
The hedonic model assumes that consumers do not value the consumption of a good
directly, but rather value the characteristics contained within a good. In the context of property
values, the consumer values both the physical attributes of the property (i.e., number of rooms,
square footage, etc.) as well as geographic and environmental attributes (e.g., proximity to
parks, visibility, etc.). Following the technique developed by Rosen (1974), property
characteristics are regressed on the observed price of the properties within a given housing
market to estimate the contribution of each characteristic to the overall price.
Numerous studies have applied hedonic methods to estimate the willingness to pay
(WTP) for air quality changes (see Smith and Huang (1995) and Boyle and Kiel (2001) for
6-30
-------
literature reviews), but fewer researchers have focused specifically on visibility. Studies that
have estimated the WTP to improve visibility have focused on specific housing markets like Los
Angeles (Murdoch and Thayer, 1998; Beron et al., 2001) or San Francisco (Graves et al., 1998),
and it is unclear if these results would be more broadly applicable to the rest of the county.
While the literature demonstrates a link between pollutant concentrations and home-buying
behavior, it is difficult to partition the WTP for changes in pollution between health and
aesthetic concerns. Murdoch and Thayer (1998) use visibility as a surrogate for environmental
quality and Beron et al. (2001) acknowledge that their parameter estimates likely reflect a
combination of visual aesthetics and an absence of health effects. Delucchi et al. (2002) deals
with this issue by partitioning WTP estimates from a hedonic model into health and visibility
components using results from previous contingent valuation studies, and find that the
estimate of visibility co-benefits is similar to estimates based on contingent valuation alone.
In 2004, the Advisory Council on Clean Air Compliance Analysis (SAB-Council)
recommended that the EPA evaluate the available studies addressing residential visibility and
consider the possibility of using hedonic property models to estimate residential visibility co-
benefits (U.S. EPA-SAB, 2004). In response to this recommendation, the EPA evaluated the
existing economic literature, and determined that there were substantial limitations that
precluded the Agency from using these studies to make inferences regarding individuals' WTP
for improved visibility (Leggett and Neumann, 2004). Specifically, the literature did not provide
support for the assumption that market participants are aware of the spatial variation in
visibility, and consider this variation when purchasing a home, and can successfully separate
visibility effects from health effects (Leggett and Neumann, 2004). This conclusion is also
supported by Delucchi et al. (2004), which found that hedonic price analysis does not capture
all of the health effects of air pollution because homebuyers may not be fully informed about
these effects.
Research since 2004 has attempted to address limitations of the hedonic method
through the use of U.S. Census microdata (Bayer et al., 2009), spatial statistical methods
(Anselin and Le Gallo, 2006; Anselin and Lozana-Gracia, 2009; Beron et al., 2004; Kim et al.,
2010) and more complete air quality data and information about nonattainment status (Chay
and Greenstone, 2005). However, none of these studies specifically address visibility, and they
are therefore of limited use at this time. While the current state of the literature does not
provide a basis for using hedonics-based approaches, continued innovations in methodology
and the further development of national, micro-level housing and demographic datasets may
open possibilities for national-scale hedonics-based benefit analysis in the future.
6-31
-------
Regardless of whether we use hedonic models or stated preference surveys to estimate
co-benefits arising from improved visibility, it is important to emphasize that estimates of WTP
for residential or recreational visibility improvements are not substitutes for health benefits. As
previously mentioned, people often have difficulty separating their health concerns from their
aesthetic concerns when evaluating preferences for visibility, which could overestimate
visibility co-benefits if not properly controlled. However, because we use a damage-function
approach to estimate health benefits (see Chapter 5 of this RIA), the health benefits estimates
are unaffected by any potential confounding with visibility preferences.
6.3.6 Discussion of Visibility Co-benefits
As described in the previous sections of this chapter, the estimation of visibility co-
benefits is complex and suffers from unavoidable limitations. While we are confident that the
underlying scientific literature supports a non-zero estimate for visibility co-benefits
attributable to emission reductions, we are less confident in the magnitude of those co-benefits
outside of previously studied locations. While acknowledging these limitations, it is important
to note that this general approach was included in The Benefits and Costs of the Clean Air Act
1990 to 2020: EPA Report to Congress (U.S. EPA, 2011), which was reviewed by the SAB-Council
(U.S. EPA-SAB, 2010a, 2010b). Although the SAB-Council highlighted concerns with the visibility
approach used in the study, it did not recommend that visibility benefits be excluded. We have
addressed the SAB-Council's concern regarding inconsistency between estimation of residential
and recreational visibility in our approach. However, we do not have the data to address the
SAB-Council's concern regarding inclusion of night-time benefits of visibility improvements in
our annual average, which may lead to an underestimation of visibility benefits. To minimize
uncertainties related to extrapolation and geographical double counting, our approach only
includes a subset of monetized visibility co-benefits in the core monetized visibility co-benefits
estimate to correspond with our higher level of confidence in recreational co-benefits within
the study regions and residential co-benefits within the study cities. Although we are confident
that visibility co-benefits extend beyond these studied areas, we are less confident about the
magnitude of those co-benefits. However, it is unclear the degree to which the visibility
valuation surveys were successful in controlling for potential double counting embedded health
benefits.
Consistent with the approach described in the proposal RIA, we have described a
revised approach for estimating visibility co-benefits, including light extinction estimation
methods, visitation data for Class 1 areas (used in extrapolating co-benefits), valuation studies
for residential visibility co-benefits, and the benefit transfer technique for residential co-
6-32
-------
benefits. Including residential visibility co-benefits in the core visibility co-benefits estimates
reflects an evolution in our understanding of the nature and importance of the effect on public
welfare from visibility impairment to a more multifaceted approach that includes non-Class 1
areas, such as urban areas. This evolution has occurred in conjunction with the expansion of
available PM data and information from associated studies of public perception, valuation and
personal comfort and well-being. While visibility preference studies (Abt Associates Inc., 2001,
Ely et al., 1991, Pryor, 1996, BBC Research & Consulting, 2003) also provide support for a non-
zero benefits estimate, these surveys did not include questions that would enable monetization
of those preferences.
Despite these improvements, we are limited by the available peer-reviewed studies on
visibility co-benefits, which have not undergone a similar expansion as the health literature.
Each of these valuation studies has limitations, which are identified in the sections 6.3.4.2 and
6.3.5.2. When the SAB-Council reviewed the visibility benefits analysis for The Benefits and
Costs of the Clean Air Act 1990 to 2020: EPA Report to Congress (U.S. EPA, 2011), they also
lamented on the need for additional research to improve methods and estimates (U.S. EPA-
SAB, 2010a, 2010b). Because of time and resource constraints, performing original research for
regulatory analyses of specific policy actions is infeasible. Most importantly, we are interested
in recently published national-scale visibility valuation studies that incorporate current CV best
practices, as the existing studies are limited to specific subset of geographic areas. Other
important research questions that remain unresolved include identifying factors that affect
valuation preferences in order to facilitate benefits transfer from the original studies to transfer
sites across localities, disentangling health and ecosystem valuation from visibility valuation,
usefulness of preference calibration, and potential role of hedonic valuation approaches. Many
of these same research needs were identified by Cropper (2000), but they have yet to be
addressed by the research community.
In Appendix 6.B, we provide the results of an illustrative analysis of the visibility co-
benefits associated with the 2020 base case 2020 control case simulation described in Chapter
3 that were used to develop the air quality ratios; however, we do not have air quality model
runs for the regulatory baseline and the alternative standard levels that would allow us to
calculate the visibility co-benefits of attaining the revised primary standard.
6.4 Materials Damage Co-benefits
Building materials including metals, stones, cements, and paints undergo natural
weathering processes from exposure to environmental elements (e.g., wind, moisture,
6-33
-------
temperature fluctuations, sunlight, etc.). Pollution can worsen and accelerate these effects.
Deposition of PM is associated with both physical damage (materials damage effects) and
impaired aesthetic qualities (soiling effects). Wet and dry deposition of PM can physically affect
materials, adding to the effects of natural weathering processes, by potentially promoting or
accelerating the corrosion of metals, by degrading paints and by deteriorating building
materials such as stone, concrete and marble (U.S. EPA, 2009b). The effects of PM are
exacerbated by the presence of acidic gases and can be additive or synergistic due to the
complex mixture of pollutants in the air and surface characteristics of the material. Acidic
deposition has been shown to have an effect on materials including zinc/galvanized steel and
other metal, carbonate stone (as monuments and building facings), and surface coatings
(paints) (Irving, 1991). The effects on historic buildings and outdoor works of art are of
particular concern because of the uniqueness and irreplaceability of many of these objects.
The PM ISA concludes that evidence is sufficient to support a causal relationship
between PM and effects on materials (U.S. EPA, 2009b). Considerable research has been
conducted on the effects of air pollutants on metal surfaces due to the economic importance of
these materials, especially steel, zinc, aluminum, and copper. Moisture is the single greatest
factor promoting metal corrosion; however, deposited PM can have additive, antagonistic or
synergistic effects. In general, S02 is more corrosive than NOX although mixtures of NOX, S02
and other particulate matter corrode some metals at a faster rate than either pollutant alone
(U.S. EPA, 2008). Metal structures are usually coated by alkaline corrosion product layers and
thus are subject to increased corrosion by acidic deposition. In addition, research has
demonstrated that iron, copper, and aluminum-based products are subject to increased
corrosion due to pollution (Irving, 1991). Information from both the PM ISA (U.S. EPA, 2009b)
and NOx/SOx ISA (U.S. EPA, 2008) suggest that the extent of damage to metals due to ambient
PM is variable and dependent upon the type of metal, prevailing environmental conditions, rate
of natural weathering and presence or absence of other pollutants
In addition, the deposition of PM can cause soiling, which is the accumulation of dirt,
dust, and ash on exposed surfaces such as metal, glass, stone and paint. Particles consisting
primarily of carbonaceous compounds can cause soiling of commonly used building materials
and culturally important items such as statues and works of art. Soiling occurs when PM
accumulates on an object and alters the optical characteristics (appearance). The reflectivity of
a surface may be changed or presence of particulates may alter light transmission. These
effects can reduce the aesthetic value of a structure or result in reversible or irreversible
damage to statues, artwork and architecturally or culturally significant buildings. Due to soiling
6-34
-------
of building surfaces by PM, the frequency and duration of cleaning or repainting may be
increased. In addition to natural factors, exposure to PM may give painted surfaces a dirty
appearance. Pigments in works of art can be degraded or discolored by atmospheric pollutants,
especially sulfates (U.S. EPA, 2008). Previous assessments estimated household soiling co-
benefits based on the Manuel et al. (1982) study of consumer expenditures on cleaning and
household maintenance. However, the data used to estimate household soiling damages in the
Manuel et al. study is from a 1972 consumer expenditure survey and as such may not
accurately represent consumer preferences in the future. In light of this significant limitation,
we believe that this study cannot provide reliable estimates of the likely magnitude of the co-
benefits of reduced PM household soiling.
In order to estimate the monetized co-benefits associated with reducing materials
damage and household soiling, quantitative relationships are needed between particle size,
concentration, chemical concentrations and frequency of maintenance and repair. Such an
analysis would require three steps:
1. Develop a national inventory of sensitive materials;
2. Derive concentration-response functions that relate material damage to change in
pollution concentration or deposition; and,
3. Estimate the value of lost materials and/or repair of damage.
Due to data limitations and uncertainties inherent in each of these steps, we have
chosen not to include a monetized estimate of materials damage and household soiling in this
analysis. The PM ISA concluded that there is considerable uncertainty with regard to interaction
of co-pollutants in regards to materials damage and soiling processes (U.S. EPA, 2009b).
Previous benefits analyses by the EPA have provided quantitative estimates of materials
damage (U.S. EPA, 2011b) and household soiling damage (U.S. EPA, 1999). Consistent with SAB
advice (U.S. EPA, 1998), we determined that the existing data are not sufficient to calculate a
reliable estimate of future year household soiling damages (U.S. EPA, 1998). These previous
analyses have shown that materials damage co-benefits are significantly smaller than the
health benefits associated with reduced exposure to PM2.s and ozone, or even visibility co-
benefits. However, studies of materials damage to historic buildings and outdoor artwork in
Sweden (Grosclaude and Soguel, 1994) indicate that these co-benefits could be an order of
magnitude larger than household soiling co-benefits.
6-35
-------
In the absence of quantified co-benefits, we provide a qualitative description of the
avoided damage associated with reducing PM and PM precursor pollutants. Table 6-8 shows
the effect of various PM2.s precursor pollutants and other co-pollutants on various materials.
Table 6-8. Materials Damaged by Pollutants Affected by this Rule (U.S. EPA, 2011b)
Pollutant
Unquantified Effects/ Damage to:
Sulfur oxides
Hydrogen ion and
nitrogen oxides
Carbon dioxide
Formaldehyde
Particulate matter
Ozone
Infrastructural materials—galvanized and painted carbon steel
Commercial buildings—carbonate stone, metal, and painted wood surfaces
Residential buildings—carbonate stone, metal, and painted wood surfaces
Monuments—carbonate stone and metal
Structural aesthetics
Automotive finishes—painted metal
Infrastructural materials—galvanized and painted carbon steel
Zinc-based metal products, such as galvanized steel
Commercial and residential buildings—carbonate stone, metal, and wood surfaces
Monuments—carbonate stone and metal
Structural aesthetics
Automotive finishes—painted metal
Zinc-based metal products, such as galvanized steel
Zinc-based metal products, such as galvanized steel
Household cleanliness (i.e., household soiling)
Rubber products (e.g., tires)
6.5 Climate Co-benefits
Actions taken by state and local governments to implement the revised annual primary
standard are likely to have implications for climate change because emission reductions
ultimately implemented to meet the standard may have impacts on emissions of long-lived
greenhouse gas (GHG) such as carbon dioxide (C02), short-lived climate forcers such as black
carbon (BC), and cooling aerosols like organic carbon (OC). Our ability to quantify the climate
effects of these revised standard is limited due to lack of available information on the co-
controlled GHG emission reductions, the energy and associated climate gas implications of
control technologies assumed in the illustrative regulatory alternatives, and remaining
uncertainties regarding the impact of long-lived and short-lived climate forcer impacts on
climate change. For this RIA, we discuss qualitatively the implications of potential emission
reductions in warming and cooling aerosols and changes in long-lived GHG emissions such as
6-36
-------
C02 for the regulatory alternatives. Implementation strategies undertaken by state and local
governments to comply with the standards may differ from the illustrative emission reduction
strategies in this RIA. It is important to note that the net climate forcing depends on the specific
combinations of emission reductions chosen to meet the revised standard because of the
differences in warming and cooling potential of the difference pollutants.
6.5.1 Climate Effects of Short Lived Climate Forcers
Pollutants that affect the energy balance of the earth are referred to as climate forcers.
A pollutant that increases the amount of energy in the Earth's climate system is said to exert
"positive radiative forcing," which leads to warming and climate change. In contrast, a pollutant
that exerts negative radiative forcing reduces the amount of energy in the Earth's system and
leads to cooling.
Long-lived gases such as C02 differ from short-lived pollutants such as BC in the length
of time they remain in the atmosphere affecting the earth's energy balance. Long-lived gases
remain in the atmosphere for hundreds to thousands of years. Short-lived climate forcers
(SLCFs), in contrast, remain in the atmosphere for short periods of time ranging from days to
weeks. The potential to affect near-term climate change and the rate of climate change with
policies to address these emissions is gaining attention nationally and internationally (e.g.,
Black Carbon Report to Congress (U.S. EPA, 2012b), Arctic Council Task Force, Global Methane
Initiative, and Convention on Long-Range Trans-boundary Air Pollution of the United Nations
Economic Commission for Europe). A recent United Nations Environmental Programme (UNEP)
study provides the most comprehensive analysis to date of the co-benefits of measures to
reduce SLCFs including methane, ozone, and black carbon assessing the health, climate, and
agricultural co-benefits of a suite of mitigation technologies. The report concludes that the
climate is changing now, and these changes have the potential to "trigger abrupt transitions
such as the release of carbon from thawing permafrost and biodiversity loss." While reducing
long-lived GHGs such as C02 is necessary to protect against long-term climate change, reducing
SLCF gases including BC and ozone is beneficial and will slow the rate of climate change within
the first half of this century (UNEP, 2011).
6.5.1.1 Climate Effects of Black Carbon
Black carbon is the most strongly light-absorbing component of PM2.5, and is formed by
incomplete combustion of fossil fuels, biofuels, and biomass. The short atmospheric lifetime of
BC lasting from days to weeks and the mechanisms by which BC affects climate distinguish it
from long-lived GHGs like C02. This means that actions taken to reduce the BC constituents in
6-37
-------
direct PM2.5 will have almost immediate effects on climate change. Emissions sources and
ambient concentrations of BC vary geographically and temporally resulting in climate effects
that are more regionally and seasonally dependent than the effects of long-lived, well-mixed
GHGs. Likewise, mitigation actions for BC will produce different climate impacts depending on
the region, season, and emission source category affected.
BC influences climate in multiple ways: directly absorbing light, reducing the reflectivity
("albedo") of snow and ice through deposition, and interacting with clouds. BC affects climate
directly by absorbing both incoming and outgoing radiation of all wavelengths. In contrast,
GHGs mainly trap outgoing infrared radiation from the earth's surface. Per unit of mass in the
atmosphere, BC can absorb a million times more energy than CCh (Bond and Sun 2005). This
strong absorptive capacity is the property most relevant to its potential to affect the Earth's
climate. When BC is deposited on snow and ice, it darkens the surface and decreases albedo,
thereby increasing absorption and accelerating melting. Finally, BC also affects climate
indirectly by altering the properties of clouds, affecting cloud reflectivity, precipitation, and
surface dimming. These indirect impacts of BC are associated with all ambient particles and
may lead to cooling, but are not associated with long-lived well mixed GHGs.
Regional climate impacts of BC are highly variable, and sensitive regions such as the
Arctic and the Himalayas are particularly vulnerable to the warming and melting effects of BC.
Snow and ice cover in the Western U.S. has also been affected by BC. Specifically, deposition of
BC on mountain glaciers and snow packs produces a positive snow and ice albedo effect,
contributing to the melting of snowpack earlier in the spring and reducing the amount of
snowmelt that normally would occur later in the spring and summer (Hadley et al. 2010). This
has implications for freshwater resources in regions of the U.S. dependent on snow-fed or
glacier-fed water systems. In the Sierra Nevada mountain range, Hadley et al. (2010) found BC
at different depths in the snowpack, deposited over the winter months by snowfall. In the
spring, the continuous uncovering of the BC contributed to the early melt. A model capturing
the effects of soot on snow in the western U.S. shows significant decreases in snowpack
between December and May (Qian et al., 2009). Snow water equivalent (the amount of water
that would be produced by melting all the snow) is reduced 2-50 millimeters (mm) in
mountainous areas, particularly over the Central Rockies, Sierra Nevadas, and western Canada.
A study found that biomass burning emissions in Alaska and the Rocky Mountain region during
the summer can enhance snowmelt. Dust deposition on snow, at high concentrations, can have
similar effects to BC (Koch et al., 2007). Similarly, a study done by Painter et al. (2007) in the
6-38
-------
San Juan Mountains in Colorado indicated a decrease in snow cover duration of 18-35 days as a
result of dust transported from non-local desert sources.
The illustrative emission reduction strategies evaluated for this rule include reductions
in BC emissions that will tend to have a beneficial cooling effect on the atmosphere. BC and
elemental carbon (EC) (or particulate elemental carbon (PEC)) are used interchangeably in this
report because the EPA traditionally estimates EC emissions rather than BC and for the purpose
of this analysis these measures are essentially equivalent.
6.5.1.2 Climate Effects of Nitrates, Sulfate, and Organic Carbon (excluding BC)
The composition of the total emissions mixture is also relevant as to whether emissions
are warming or cooling to the atmosphere. Pollutants such as S02, NOX, and most OC particles
tend to produce a cooling influence on climate. Exceptions include OC deposition on snow and
ice, which leads to increased melting.
In addition, it is important to account for the indirect effects of all PM constituents on
climate: all aerosols (including BC) affect climate indirectly by changing the reflectivity and
lifetime of clouds. The net indirect effect of all aerosols is very uncertain but is thought to be a
net cooling influence.
6.5.1.3 Climate Effects of Ozone
Ozone changes due to this revised annual standard are not estimated for this analysis
but may occur due to the NOX reductions estimated. Ozone is a well-known SLCF (U.S. EPA,
2006). Stratospheric ozone (the upper ozone layer) is beneficial because it protects life on Earth
from the sun's harmful ultraviolet (UV) radiation. In contrast, tropospheric ozone (ozone in the
lower atmosphere) is a harmful air pollutant that adversely affects human health and the
environment and contributes significantly to regional and global climate change. Due to its
short atmospheric lifetime, tropospheric ozone concentrations exhibit large spatial and
temporal variability (U.S. EPA, 2009). The discernible influence of ground level ozone on climate
leads to increases in global surface temperature and changes in hydrological cycles. While
reducing long-lived GHGs such as C02 is necessary to protect against long-term climate change,
reducing SLCF gases including ozone is beneficial and will slow the rate of climate change within
the first half of this century (UNEP, 2011).
6.5.1.4 SLCFs Summary and Conclusions
Assessing the net climate impact of SLCFs for the illustrative emission reduction
strategies is outside the scope of this regulatory analysis and requires climate atmospheric
6-39
-------
modeling not undertaken due to time and resource constraints. Information about the amount
of BC relative to non-BC constituents emitted from a source is important. In general, these non-
BC constituents are emitted in greater volume than BC, counteracting the warming influence of
BC. Qualitatively, it seems likely that BC emission reductions associated with directly emitted
PM2.5 reductions will be beneficial for the climate in terms of reduced radiative forcing and
deposition on snow and ice. Reductions in OC, sulfates and nitrates are likely to produce
warming in the atmosphere. The indirect impacts of aerosols on clouds and precipitation
remain the subject of great uncertainty making it more difficult to estimate the quantitative
impact of aerosol reductions on climate.
6.5.2 Climate Effects of long-lived Greenhouse Gases
The EPA Administrator found in 2009 that elevated concentrations of the six major
GHGs, including C02, endanger the public health and public welfare of current and future
generations (FR 77 66496). While addressing short-lived climate forcers can result in near-term
(and sometimes regionally specific) co-benefits as well as reductions in the rate of warming,
reductions of long-term warming would require mitigation of long-lived GHGs. We are unable
to quantify the impact of the illustrative emission reduction strategies for this rulemaking on
long-lived climate gases due lack of available data. However, State and Local governments may
want to consider human health, welfare, and climate implications of regulatory strategies
undertaken to implement the promulgated PM standards.
6.6 Ecosystem Co-benefits and Services
The effects of air pollution on the health and stability of ecosystems are potentially very
important. At present, it is difficult to measure the impact of reducing air pollution in a national
scale analysis across different types of ecosystems and different pollutant effects. Previous
science assessments by the EPA (U.S. EPA, 2006a; 2008c; 2009b) have determined that air
pollution can be directly linked to aquatic and terrestrial acidification, nutrient enrichment,
vegetation injury, and metal bioaccumulation in animals. Ecosystem services are a useful
conceptual 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, 2006c). Figure 6-7 provides the Millennium Ecosystem
Assessment's schematic demonstrating the connections between the categories of ecosystem
6-40
-------
services and human well-being. The interrelatedness of these categories means that any one
ecosystem may provide multiple services. Changes in these services can affect human well-
being by affecting security, health, social relationships, and access to basic material goods
(MEA, 2005).
CONSTITUENTS OF WELL-BEING
ECOSYSTEM SERVICES
Provisioning
FOOD
FRESH WATER
WOOD AND FIBER
FUEL
Supporting
NUTRIENT CYCLING
SOIL FORMATION
PRIMARY PRODUCTION
Regulating
CLIMATE REQULATION
FLOOD REGULATION
DISEASE REGULATION
WATER PURIFICATION
Cultural
AESTHETIC
SPIRITUAL
EDUCATIONAL
RECREATIONAL
LIFE ON EARTH - BIODIVERSITY
Security
PERSONAL SAFETY
SECURE RESOURCE ACCESS
SECURITY FROM DISASTERS
Basic material
for good life
ADEQUATE LIVELIHOODS
SUFFICIENT NUTRITIOUS FOOD
SHELTER
ACCESS TO GOODS
Health
STRENGTH
FEELING WELL
ACCESS TO CLEAN AIR
AND WATER
Good social relations
SOCIAL COHESION
MUTUAL RESPECT
ABIUTY TO HELP OTHERS
Freedom
of choice
and action
OPPORTUNITY TO BE
ABLE TO ACHIEVE
WHAT AN INDIVIDUAL
VALUES DO ING
AND BEING
Source: Millennium Ecosystem Assessment
Figure 6-7. Linkages between Categories of Ecosystem Services and Components of Human
Weil-Being from Millennium Ecosystem Assessment (MEA, 2005)
In the Millennium Ecosystem Assessment (MEA, 2005), ecosystem services are classified
into four main categories:
1. Provisioning: Products obtained from ecosystems, such as the production of food
and water
2. Regulating: Benefits obtained from the regulation of ecosystem processes, such as
the control of climate and disease
3. Cultural: Nonmaterial benefits that people obtain from ecosystems through spiritual
enrichment, cognitive development, reflection, recreation, and aesthetic
experiences
4. Supporting: Services necessary for the production of all other ecosystem services,
such as nutrient cycles and crop pollination
The monetization of ecosystem services generally involves estimating the value of
ecological goods and services based on what people are willing to pay (WTP) to increase
ecological services or by what people are willing to accept (WTA) in compensation for
6-41
-------
reductions in them (U.S. EPA, 2006c). There are three primary approaches for estimating the
monetary value of ecosystem services: market-based approaches, revealed preference
methods, and stated preference methods (U.S. EPA, 2006c). Because economic valuation of
ecosystem services can be difficult, nonmonetary valuation using biophysical measurements
and concepts also can be used. An example of a nonmonetary valuation method is the use of
relative-value indicators (e.g., a flow chart indicating uses of a water body, such as beatable,
fishable, swimmable, etc.). It is necessary to recognize that in the analysis of the environmental
responses associated with any particular policy or environmental management action, only a
subset of the ecosystem services likely to be affected are readily identified. Of those ecosystem
services that are identified, only a subset of the changes can be quantified. Within those
services whose changes can be quantified, only a few will likely be monetized, and many will
remain nonmonetized. The stepwise concept leading up to the valuation of ecosystems services
is graphically depicted in Figure 6-8.
Ecosystems
Ecological goods and services
affected by the policy
Planning and problem formulation
Goods and
services not
identified
Identified
goods and
services not
quantified
Quantified
goods and
services not
monetized
Figure 6-8. Schematic of the Benefits Assessment Process (U.S. EPA, 2006c]
6.6.1 Ecosystem Co-benefits for Metallic and Organic Constituents of PM
Several significant ecological effects are associated with deposition of chemical
constituents of ambient PM such as metals and organics (U.S. EPA, 2009b). The trace metal
constituents of PM include cadmium, copper, chromium, mercury, nickel, zinc, and lead. The
organics include persistent organic pollutants (POPs), polyaromatic hydrocarbons (PAHs) and
polybromiated diphenyl ethers (PBDEs). Exposure to PM for direct effects occur via deposition
(e.g., wet, dry or occult) to vegetation surfaces, while indirect effects occur via deposition to
6-42
-------
ecosystem soils or surface waters where the deposited constituents of PM then interacts with
biological organisms. While both fine and coarse-mode particles may affect plants and other
organisms, more often the chemical constituents drive the ecosystem response to PM (Grantz
et al., 2003). Ecological effects of PM include direct effects to metabolic processes of plant
foliage; contribution to total metal loading resulting in alteration of soil biogeochemistry and
microbiology, plant and animal growth and reproduction; and contribution to total organics
loading resulting in bioaccumulation and biomagnification across trophic levels.
The PM ISA concludes that a causal relationship is likely to exist between deposition of
PM and a variety of effects on individual organisms and ecosystems (U.S. EPA 2009b). Most
direct ecosystem effects associated with particulate pollution occur in severely polluted areas
near industrial point sources (quarries, cement kilns, metal smelting) (U.S. EPA, 2009b).
However the PM ISA also finds, in many cases, it is difficult to characterize the nature and
magnitude of effects and to quantify relationships between ambient concentrations of PM and
ecosystem response due to significant data gaps and uncertainties as well as considerable
variability that exists in the components of PM and their various ecological effects (U.S. EPA,
2009b).
Particulate matter can adversely impact plants and ecosystem services provided by
plants by deposition to vegetative surfaces (U.S. EPA, 2009b). Particulates deposited on the
surfaces of leaves and needles can block light, altering the radiation received by the plant. PM
deposition near sources of heavy deposition can obstruct stomata limiting gas exchange,
damage leaf cuticles and increase plant temperatures (U.S. EPA, 2009b). Plants growing on
roadsides exhibit impact damage from near-road PM deposition, having higher levels of
organics and heavy metals, and accumulate salt from road de-icing during winter months (U.S.
EPA, 2009b). In addition, atmospheric PM can convert direct solar radiation to diffuse radiation,
which is more uniformly distributed in a tree canopy, allowing radiation to reach lower leaves
(U.S. EPA, 2009b). Decreases in crop yields (a provisioning service) due to reductions in solar
radiation have been attributed to regional scale air pollution in other counties with especially
severe regional haze (Chameides et al., 1999).
In addition to damage to plant surfaces, deposited PM can be taken up by plants from
soil or foliage. Copper, zinc, and nickel have been shown to be directly toxic to vegetation under
field conditions (U.S. EPA, 2009b).The ability of vegetation to take up heavy metals is
dependent upon the amount, solubility and chemical composition of the deposited PM. Uptake
of PM by plants from soils and vegetative surfaces can disrupt photosynthesis, alter pigments
and mineral content, reduce plant vigor, decrease frost hardiness and impair root development.
6-43
-------
Particulate mattercan also contain organic air toxic pollutants, including PAHs, which
are a class of polycyclic organic matter (POM). PAHs can accumulate in sediments and
bioaccumulate in freshwater, flora and fauna. The uptake of organics depends on the plant
species, site of deposition, physical and chemical properties of the organic compound and
prevailing environmental conditions (U.S. EPA, 2009b). Different species can have different
uptake rates of PAHs. For example, zucchini (Cucurbita pepo) accumulated significantly more
PAHs than related plant species (Parrish et al., 2006). PAHs can accumulate to high enough
concentrations in some coastal environments to pose an environmental health threat that
includes cancer in fish populations, toxicity to organisms living in the sediment and risks to
those (e.g., migratory birds) that consume these organisms (Simcik et al., 1996; Simcik et al.,
1999). Atmospheric deposition of particles is thought to be the major source of PAHs to the
sediments of Lake Michigan, Chesapeake Bay, Tampa Bay and other coastal areas of the U.S.
(Arzavus, Dickhut, and Canuel, 2001).
Contamination of plant leaves by heavy metals can lead to elevated concentrations in
the soil. Trace metals absorbed into the plant, frequently bind to the leaf tissue, and then are
lost when the leaf drops. As the fallen leaves decompose, the heavy metals are transferred into
the soil (Cotrufo et al., 1995; Niklinska et al., 1998). Many of the major indirect plant responses
to PM deposition are chiefly soil-mediated and depend on the chemical composition of
individual components of deposited PM. Upon entering the soil environment, PM pollutants
can alter ecological processes of energy flow and nutrient cycling, inhibit nutrient uptake to
plants, change microbial community structure and, affect biodiversity. Accumulation of heavy
metals in soils depends on factors such as local soil characteristics, geologic origin of parent
soils, and metal bioavailability. Heavy metals, such as zinc, copper, and cadmium, and some
pesticides can interfere with microorganisms that are responsible for decomposition of soil
litter, an important regulating ecosystem service that serves as a source of soil nutrients (U.S.
EPA, 2009b). Surface litter decomposition is reduced in soils having high metal concentrations.
Soil communities have associated bacteria, fungi, and invertebrates that are essential to soil
nutrient cycling processes. Changes to the relative species abundance and community
composition are associated with deposited PM to soil biota (U.S. EPA, 2009b).
Atmospheric deposition can be the primary source of some organics and metals to
watersheds. Deposition of PM to surfaces in urban settings increases the metal and organic
component of storm water runoff (U.S. EPA, 2009b). This atmospherically-associated pollutant
burden can then be toxic to aquatic biota. The contribution of atmospherically deposited PAHs
to aquatic food webs was demonstrated in high elevation mountain lakes with no other
6-44
-------
anthropogenic contaminant sources (U.S. EPA, 2009b). Metals associated with PM deposition
limit phytoplankton growth, affecting aquatic trophic structure. Long-range atmospheric
transport of 47 pesticides and degradation products to the snowpack in seven national parks in
the Western U.S. was recently quantified indicating PM-associated contaminant inputs to
receiving waters during spring snowmelt (Hageman et al., 2006).
The recently completed Western Airborne Contaminants Assessment Project (WACAP) is
the most comprehensive database on contaminant transport and PM depositional effects on
sensitive ecosystems in the Western U.S. (Landers et al., 2008). In this project, the transport,
fate, and ecological impacts of anthropogenic contaminants from atmospheric sources were
assessed from 2002 to 2007 in seven ecosystem components (air, snow, water, sediment,
lichen, conifer needles and fish) in eight core national parks. The study concluded that
bioaccumulation of semi-volatile organic compounds occurred throughout park ecosystems, an
elevational gradient in PM deposition exists with greater accumulation in higher altitude areas,
and contaminants accumulate in proximity to individual agriculture and industry sources, which
is counter to the original working hypothesis that most of the contaminants would originate
from Eastern Europe and Asia.
Although there is considerable data on impacts of PM on ecological receptors, few
studies link ambient PM levels to observed effect. This is due, in part, to the nature, deposition,
transport and fate of PM in ecosystems. Some of the difficulties in quantifying the ecosystem
co-benefits associated with reduced PM deposition include the following:
• PM is not a single pollutant, but a heterogeneous mixture of particles differing in
size, origin and chemical composition. Since vegetation and other ecosystem
components are affected more by particulate chemistry than size fraction, exposure
to a given mass concentration of airborne PM may lead to widely differing plant or
ecosystem responses, depending on the particular mix of deposited particles.
• Composition of ambient PM varies in time and space and the particulate mixture
may have synergistic, antagonistic or additive effects on ecological receptors
depending upon the chemical species present.
• Presence of co-pollutants makes it difficult to attribute observed effects to
ecological receptors to PM alone or one component of deposited PM.
• Ecosystem effects linked to PM are difficult to determine because the changes may
not be observed until pollutant deposition has occurred for many decades.
Furthermore, many PM components bioaccumulate over time in organisms or
plants, making correlations to ambient levels of PM difficult.
6-45
-------
• Multiple ecological stressors can confound attempts to link specific ecosystem
responses to PM deposition. These stressors can be anthropogenic (e.g., habitat
destruction, eutrophication, other pollutants) or natural (e.g., drought, fire, disease).
Deposited PM interacts with other stressors to affect ecosystem patterns and
processes.
• Each ecosystem has a unique topography, underlying bedrock, soils, climate,
meteorology, hydrologic regime, natural and land use history, and species
composition. Sensitivity of ecosystem response can be highly variable in space and
time. Because of this variety and lack of data for most ecosystems, extrapolating
these effects from one ecosystem to another is highly uncertain.
6.6.2 Ecosystem Co-benefits from Reductions in Nitrogen and Sulfur Emissions
Emissions of the PM precursors, such as nitrogen and sulfur oxides occur over large
regions of North America. Once these pollutants are lofted to the middle and upper
troposphere, they typically have a much longer lifetime and, with the generally stronger winds
at these altitudes, can be transported long distances from their source regions. The length scale
of this transport is highly variable owing to differing chemical and meteorological conditions
encountered along the transport path (U.S. EPA, 2008c). Secondary particles are formed from
NOX and S02 gaseous emissions and associated chemical reactions in the atmosphere.
Deposition can occur in either a wet (i.e., rain, snow, sleet, hail, clouds, or fog) or dry form (i.e.,
gases or particles). Together these emissions are deposited onto terrestrial and aquatic
ecosystems across the U.S., contributing to the problems of acidification, nutrient enrichment,
and methylmercury production as represented in Figures 6-9 and 6-10. Although there is some
evidence that nitrogen deposition may have positive effects on agricultural and forest output
through passive fertilization, it is likely that the overall value is very small relative to other
health and welfare effects. In addition to deposition effects, S02 can affect vegetation at
ambient levels near pollution sources.
6-46
-------
Atmospheric
Fate and
Transport
Deposition
Process
Acidification
Nutrient
Enrichment
-L
Aquatic
Ecosystems
J.
Terrestrial
Ecosystems
Aquatic
Ecosystems
_L
Terrestrial
Ecosystems
Atmospheric
Fate and
Transport
Deposition
Process
1
Acidification
Aquatic
Ecosystems
Terrestrial
Ecosystems
Methylmercury
Production
Figure 6-9. Schematics of Ecological Effects of Nitrogen and Sulfur Deposition
6-47
-------
Ambtent Air
Concentration
Dissolution
2H
Oxidation
S02 » H2S04
NO, *HN03
Wet Deposition
», NrV, NOj,
VOC
t
Foliar and
offocts
Dry deposition
NO,, NH., SO.
i rim
Acidification of water + Eutrophlcatlon
Figure 6-10. Nitrogen and Sulfur Cycling, and Interactions in the Environment
Source: U.S. EPA, 2008c.
The atmospheric lifetimes of particles vary with particle size. Accumulation-mode
particles such as sulfates are kept in suspension by normal air motions and have a lower
deposition velocity than coarse-mode particles; they can be transported thousands of
kilometers and remain in the atmosphere for a number of days. They are removed from the
atmosphere primarily by cloud processes. Particulates affect acid deposition by serving as cloud
condensation nuclei and contribute directly to the acidification of rain. In addition, the gas-
phase species that lead to the dry deposition of acidity are also precursors of particles.
Therefore, reductions in NOX and S02 emissions will decrease both acid deposition and PM
concentrations, but not necessarily in a linear fashion (U.S. EPA, 2008c). Sulfuric acid is also
deposited on surfaces by dry deposition and can contribute to environmental effects (U.S. EPA,
2008c).
6.6.2.1 Ecological Effects of Acidification
Deposition of nitrogen and sulfur can cause acidification, which alters biogeochemistry
and affects animal and plant life in terrestrial and aquatic ecosystems across the U.S. Soil
acidification is a natural process, but is often accelerated by acidifying deposition, which can
decrease concentrations of exchangeable base cations in soils (U.S. EPA, 2008c). Major
6-48
-------
terrestrial effects include a decline in sensitive tree species, such as red spruce (Picea rubens)
and sugar maple (Acer saccharum) (U.S. EPA, 2008c). 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, 2008c). Decreases in the acid neutralizing capacity and
increases in inorganic aluminum concentration contribute to declines in zooplankton, macro
invertebrates, and fish species richness in aquatic ecosystems (U.S. EPA, 2008c).
Geology (particularly surficial geology) is the principal factor governing the sensitivity of
terrestrial and aquatic ecosystems to acidification from nitrogen and sulfur deposition (U.S.
EPA, 2008c). Geologic formations having low base cation supply generally underlie the
watersheds of acid-sensitive lakes and streams. Other factors contribute to the sensitivity of
soils and surface waters to acidifying deposition, including topography, soil chemistry, land use,
and hydrologic flow path (U.S. EPA, 2008c).
Aquatic Acidification. Aquatic effects of acidification have been well studied in the U.S.
and elsewhere at various trophic levels. These studies indicate that aquatic biota have been
affected by acidification at virtually all levels of the food web in acid sensitive aquatic
ecosystems. The ISA for N0x/S0x—Ecological Criteria concluded that the evidence is sufficient
to infer a causal relationship between acidifying deposition and effects on biogeochemistry
related to aquatic ecosystems and biota in aquatic ecosystems (U.S. EPA, 2008c). Effects have
been most clearly documented for fish, aquatic insects, other invertebrates, and algae.
Biological effects are primarily attributable to a combination of low pH and high inorganic
aluminum concentrations. Such conditions occur more frequently during rainfall and snowmelt
that cause high flows of water and less commonly during low-flow conditions, except where
chronic acidity conditions are severe. Biological effects of episodes include reduced fish
condition factor33, changes in species composition and declines in aquatic species richness
across multiple taxa, ecosystems and regions. These conditions may also result in direct fish
mortality (Van Sickle et al., 1996). Biological effects in aquatic ecosystems can be divided into
two major categories: effects on health, vigor, and reproductive success; and effects on
biodiversity. Surface water with ANC values greater than 50 u.eq/L generally provides moderate
protection for most fish (i.e., brook trout, others) and other aquatic organisms (U.S. EPA,
2009c). Table 6-9 provides a summary of the biological effects experienced at various ANC
levels.
33 Condition factor is an index that describes the relationship between fish weight and length, and is one measure
of sublethal acidification stress that has been used to quantify effects of acidification on an individual fish (U.S.
EPA, 2008f).
6-49
-------
Table 6-9. Aquatic Status Categories
Category Label ANC Levels Expected Ecological Effects
Acute
Concern
Severe
Concern
Elevated
Concern
<0 micro
equivalent per
Liter (u.eq/L)
0-20 u.eq/L
20-50 u.eq/L
Near complete loss of fish populations is expected. Planktonic communities have
extremely low diversity and are dominated by acidophilic forms. The number of
individuals in plankton species that are present is greatly reduced.
Highly sensitive to episodic acidification. During episodes of high acidifying
deposition, brook trout populations may experience lethal effects. Diversity and
distribution of zooplankton communities decline sharply.
Fish species richness is greatly reduced (i.e., more than half of expected species
can be missing). On average, brook trout populations experience sublethal
effects, including loss of health, reproduction capacity, and fitness. Diversity and
distribution of zooplankton communities decline.
Moderate 50-100 u.eq/L
Concern
Low Concern >100 u.eq/L
Fish species richness begins to decline (i.e., sensitive species are lost from lakes).
Brook trout populations are sensitive and variable, with possible sublethal
effects. Diversity and distribution of zooplankton communities also begin to
decline as species that are sensitive to acidifying deposition are affected.
Fish species richness may be unaffected. Reproducing brook trout populations
are expected where habitat is suitable. Zooplankton communities are unaffected
and exhibit expected diversity and distribution.
A number of national and regional assessments have been conducted to estimate the
distribution and extent of surface water acidity in the U.S. (U.S. EPA, 2008c). As a result, several
regions of the U.S. have been identified as containing a large number of lakes and streams that
are seriously impacted by acidification. Figure 6-11 illustrates those areas of the U.S. where
aquatic ecosystems are at risk from acidification.
Because acidification primarily affects the diversity and abundance of aquatic biota, it
also affects the ecosystem services that are derived from the fish and other aquatic life found in
these surface waters.
While acidification is unlikely to have serious negative effects on, for example, water
supplies, it can limit the productivity of surface waters as a source of food (i.e., fish). In the
northeastern United States, the surface waters affected by acidification are not a major source
of commercially raised or caught fish; however, they are a source of food for some recreational
and subsistence fishermen and for other consumers. For example, there is evidence that certain
population subgroups in the northeastern United States, such as the Hmong and Chippewa
ethnic groups, have particularly high rates of self-caught fish consumption (Hutchison and Kraft,
1994; Peterson etal., 1994). However, it is not known if and how their consumption patterns
6-50
-------
are affected by the reductions in available fish populations caused by surface water
acidification.
H|SH PMenliai Sensitivity
Acid Sensitive Wa:ers :USG3 j
SUrt«i
1.OOO
1 -;rr
Figure 6-11. Areas Potentially Sensitive to Aquatic Acidification
Source: U.S. EPA, 2008c.
Inland surface waters support several cultural services, including aesthetic and
educational services and recreational fishing. Recreational fishing in lakes and streams is among
the most popular outdoor recreational activities in the northeastern United States. Based on
studies conducted in the northeastern United States, Kaval and Loomis (2003) estimated
average consumer surplus values per day of $36 for recreational fishing (in 2007 dollars);
therefore, the implied total annual value of freshwater fishing in the northeastern United States
was $5.1 billion in 2006.34 For recreation days, consumer surplus value is most commonly
measured using recreation demand, travel cost models.
Another estimate of the overarching ecological co-benefits associated with reducing
lake acidification levels in Adirondacks National Park can be derived from the contingent
valuation (CV) survey (Banzhaf et al., 2006), which elicited values for specific improvements in
These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
6-51
-------
acidification-related water quality and ecological conditions in Adirondack lakes. The survey
described a base version with minor improvements said to result from the program, and a
scope version with large improvements due to the program and a gradually worsening status
quo. After adapting and transferring the results of this study and converting the 10-year annual
payments to permanent annual payments using discount rates of 3% and 5%, the WTP
estimates ranged from $48 to $107 per year per household (in 2004 dollars) for the base
version and $54 to $154 for the scope version. Using these estimates, the aggregate annual
benefits of eliminating all anthropogenic sources of NOX and SOX emissions were estimated to
range from $291 million to $829 million (U.S. EPA, 2009c).35
In addition, inland surface waters provide a number of regulating services associated
with hydrological and climate regulation by providing environments that sustain aquatic food
webs. These services are disrupted by the toxic effects of acidification on fish and other aquatic
life. Although it is difficult to quantify these services and how they are affected by acidification,
some of these services may be captured through measures of provisioning and cultural services.
Terrestrial Acidification. Acidifying deposition has altered major biogeochemical
processes in the U.S. by increasing the nitrogen and sulfur content of soils, accelerating nitrate
and sulfate leaching from soil to drainage waters, depleting base cations (especially calcium and
magnesium) from soils, and increasing the mobility of aluminum. Inorganic aluminum is toxic to
some tree roots. Plants affected by high levels of aluminum from the soil often have reduced
root growth, which restricts the ability of the plant to take up water and nutrients, especially
calcium (U. S. EPA, 2008c). These direct effects can, in turn, influence the response of these
plants to climatic stresses such as droughts and cold temperatures. They can also influence the
sensitivity of plants to other stresses, including insect pests and disease (Joslin et al., 1992)
leading to increased mortality of canopy trees. In the U.S., terrestrial effects of acidification are
best described for forested ecosystems (especially red spruce and sugar maple ecosystems)
with additional information on other plant communities, including shrubs and lichen (U.S. EPA,
2008c). The ISA for N0x/S0x—Ecological Criteria concluded that the evidence is sufficient to
infer a causal relationship between acidifying deposition and effects on biogeochemistry
related to terrestrial ecosystems and biota in terrestrial ecosystems (U.S. EPA, 2008c).
Certain ecosystems in the continental U.S. are potentially sensitive to terrestrial
acidification, which is the greatest concern regarding nitrogen and sulfur deposition U.S. EPA
35 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
6-52
-------
(2008c). Figure 6-12 depicts the areas across the U.S. that are potentially sensitive to terrestrial
acidification.
Area of Higas! Potential Sensitivity
Top Ouartilu N
I Top Quartile S
Figure 6-12. Areas Potentially Sensitive to Terrestrial Acidification
Source: U.S. EPA, 2008c.
Both coniferous and deciduous forests throughout the eastern U.S. are experiencing
gradual losses of base cation nutrients from the soil due to accelerated leaching from acidifying
deposition. This change in nutrient availability may reduce the quality of forest nutrition over
the long term. Evidence suggests that red spruce and sugar maple in some areas in the eastern
U.S. have experienced declining health because of this deposition. For red spruce, (Picea
rubens) dieback or decline has been observed across high elevation landscapes of the
northeastern U.S., and to a lesser extent, the southeastern U.S., and acidifying deposition has
been implicated as a causal factor (DeHayes et al., 1999). Figure 6-13 shows the distribution of
red spruce (brown) and sugar maple (green) in the eastern U.S.
Terrestrial acidification affects several important ecological endpoints, including
declines in habitat for threatened and endangered species (cultural), declines in forest
6-53
-------
aesthetics (cultural), declines in forest productivity (provisioning), and increases in forest soil
erosion and reductions in water retention (cultural and regulating).
Figure 6-13. Distribution of Red Spruce (pink) and Sugar Maple (green) in the Eastern U.S.
Source: U.S. EPA, 2008c.
Forests in the northeastern United States provide several important and valuable
provisioning services in the form of tree products. Sugar maples are a particularly important
commercial hardwood tree species, providing timber and maple syrup. In the United States,
sugar maple saw timber was nearly 900 million board feet in 2006 (USFS, 2006), and annual
production of maple syrup was nearly 1.4 million gallons, accounting for approximately 19% of
worldwide production. The total annual value of U.S. production in these years was
approximately $160 million (NASS, 2008).36 Red spruce is also used in a variety of products
including lumber, pulpwood, poles, plywood, and musical instruments. The total removal of red
spruce saw timber from timberland in the United States was over 300 million board feet in
2006 (USFS, 2006).
These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
6-54
-------
Forests in the northeastern United States are also an important source of cultural
ecosystem services—nonuse (i.e., existence value for threatened and endangered species),
recreational, and aesthetic services. Red spruce forests are home to two federally listed species
and one delisted species:
1. Spruce-fir moss spider (Microhexura montivaga)—endangered
2. Rock gnome lichen (Gymnoderma lineare)—endangered
3. Virginia northern flying squirrel (Glaucomys sabrinusfuscus)—de\'\sled, but
important
Forestlands support a wide variety of outdoor recreational activities, including fishing,
hiking, camping, off-road driving, hunting, and wildlife viewing. Regional statistics on
recreational activities that are specifically forest based are not available; however, more
general data on outdoor recreation provide some insights into the overall level of recreational
services provided by forests. More than 30% of the U.S. adult population visited a wilderness or
primitive area during the previous year and engaged in day hiking (Cordell et al., 2008). From
1999 to 2004, 16% of adults in the northeastern United States participated in off-road vehicle
recreation, for an average of 27 days per year (Cordell et al., 2005). The average consumer
surplus value per day of off-road driving in the United States was $25 (in 2007 dollars), and the
implied total annual value of off-road driving recreation in the northeastern United States was
more than $9 billion (Kaval and Loomis, 2003). More than 5% of adults in the northeastern
United States participated in nearly 84 million hunting days (U.S. FWS and U.S. Census Bureau,
2007). Ten percent of adults in northeastern states participated in wildlife viewing away from
home on 122 million days in 2006. For these recreational activities in the northeastern United
States, Kaval and Loomis (2003) estimated average consumer surplus values per day of $52 for
hunting and $34 for wildlife viewing (in 2007 dollars). The implied total annual value of hunting
and wildlife viewing in the northeastern United States was, therefore, $4.4 billion and $4.2
billion, respectively, in 2006 (U.S. EPA, 2009c).37
As previously mentioned, it is difficult to estimate the portion of these recreational
services that are specifically attributable to forests and to the health of specific tree species.
However, one recreational activity that is directly dependent on forest conditions is fall color
viewing. Sugar maple trees, in particular, are known for their bright colors and are, therefore,
an essential aesthetic component of most fall color landscapes. A survey of residents in the
37 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
6-55
-------
Great Lakes area found that roughly 30% of residents reported at least one trip in the previous
year involving fall color viewing (Spencer and Holecek, 2007). In a separate study conducted in
Vermont, Brown (2002) reported that more than 22% of households visiting Vermont in 2001
made the trip primarily for viewing fall colors.
Two studies estimated values for protecting high-elevation spruce forests in the
southern Appalachian Mountains. Kramer et al. (2003) conducted a contingent valuation study
estimating households' WTP for programs to protect remaining high-elevation spruce forests
from damages associated with air pollution and insect infestation. Median household WTP was
estimated to be roughly $29 (in 2007 dollars) for a smaller program, and $44 for the more
extensive program. Jenkins et al. (2002) conducted a very similar study in seven Southern
Appalachian states on a potential program to maintain forest conditions at status quo levels.
The overall mean annual WTP for the forest protection programs was $208 (in 2007 dollars).
Multiplying the average WTP estimate from these studies by the total number of households in
the seven-state Appalachian region results in an aggregate annual range of $470 million to $3.4
billion for avoiding a significant decline in the health of high-elevation spruce forests in the
Southern Appalachian region (U.S. EPA, 2009c).38
Forests in the northeastern United States also support and provide a wide variety of
valuable regulating services, including soil stabilization and erosion control, water regulation,
and climate regulation. The total value of these ecosystem services is very difficult to quantify
in a meaningful way, as is the reduction in the value of these services associated with total
nitrogen and sulfur deposition. As terrestrial acidification contributes to root damages, reduced
biomass growth, and tree mortality, all of these services are likely to be affected; however, the
magnitude of these impacts is currently very uncertain.
6.6.2.2 Ecological Effects from Nitrogen Enrichment
Aquatic Enrichment. The ISA for N0x/S0x—Ecological Criteria concluded that the
evidence is sufficient to infer a causal relationship between nitrogen deposition and the
alteration of species richness, species composition, and biodiversity in wetland, freshwater
aquatic and coastal marine ecosystems (U.S. EPA, 2008c).
One of the main adverse ecological effects resulting from nitrogen deposition,
particularly in the Mid-Atlantic region of the United States, is the effect associated with nutrient
enrichment in estuarine waters. A recent assessment of 141 estuaries nationwide by the
38 These estimates reflect the marginal value of the service for the hypothetical program described in the survey,
not the marginal change in the value of the service as a result of the emission reductions achieved by this rule.
6-56
-------
National Oceanic and Atmospheric Administration (NOAA) concluded that 19 estuaries (13%)
suffered from moderately high or high levels of eutrophication due to excessive inputs of both
N and phosphorus, and a majority of these estuaries are located in the coastal area from North
Carolina to Massachusetts (NOAA, 2007). For estuaries in the Mid-Atlantic region, the
contribution of atmospheric distribution to total N loads is estimated to range between 10%
and 58% (Valigura et al., 2001).
Eutrophication in estuaries is associated with a range of adverse ecological effects. The
conceptual framework developed by NOAA emphasizes four main types of eutrophication
effects—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. Low DO also degrades the aesthetic
qualities of surface water. In addition to often being toxic to fish and shellfish, and leading to
fish kills and aesthetic impairments of estuaries, HABs can, in some instances, also 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; therefore, declines in
SAV due to nutrient enrichment are an important source of concern. Low water clarity is in part
the result of accumulations of both algae and sediments in estuarine waters. In addition to
contributing to declines in SAV, high levels of turbidity also degrade the aesthetic qualities of
the estuarine environment.
Estuaries in the eastern United States are an important source of food production, in
particular fish and shellfish production. The estuaries are capable of supporting large stocks of
resident commercial species, and they serve as the breeding grounds and interim habitat for
several migratory species. To provide an indication of the magnitude of provisioning services
associated with coastal fisheries, from 2005 to 2007, the average value of total catch was $1.5
billion per year. It is not known, however, what percentage of this value is directly attributable
to or dependent upon the estuaries in these states.
In addition to affecting provisioning services through commercial fish harvests,
eutrophication in estuaries may also affect the demand for seafood. For example, a well-
publicized toxic pfiesteria bloom in the Maryland Eastern Shore in 1997, which involved
thousands of dead and lesioned fish, led to an estimated $56 million (in 2007 dollars) in lost
seafood sales for 360 seafood firms in Maryland in the months following the outbreak (Lipton,
1999).
6-57
-------
Estuaries in the United States also provide an important and substantial variety of
cultural ecosystem services, including water-based recreational and aesthetic services. The
water quality in the estuary directly affects the quality of these experiences. For example, there
were 26 million days of saltwater fishing coastal states from North Carolina to Massachusetts in
2006 (FWA and Census, 2007). Assuming an average consumer surplus value for a fishing day at
$36 (in 2007 dollars) in the Northeast and $87 in the Southeast (Kaval and Loomis, 2003), the
aggregate value was approximately $1.3 billion (in 2007 dollars) (U.S. EPA, 2009c).39 In addition,
almost 6 million adults participated in motorboating in coastal states from North Carolina to
Massachusetts, for a total of nearly 63 million days annually during 1999-2000 (Leeworthy and
Wiley, 2001). Using a national daily value estimate of $32 (in 2007 dollars) for motorboating
(Kaval and Loomis (2003), the aggregate value of these coastal motorboating outings was $2
billion per year (U.S. EPA, 2009c).40 Almost 7 million participated in birdwatching for 175 million
days per year, and more than 3 million participated in visits to non-beach coastal waterside
areas.
Estuaries and marshes have the potential to support a wide range of regulating services,
including climate, biological, and water regulation; pollution detoxification; erosion prevention;
and protection against natural hazards from declines in SAV (MEA, 2005). SAV can help reduce
wave energy levels and thus protect shorelines against excessive erosion, which increases the
risks of episodic flooding and associated damages to near-shore properties or public
infrastructure or even contribute to shoreline retreat.
We are unable to provide an estimate of the aquatic enrichment co-benefits associated
with the revised or alternative annual standards due to data, time, and resource limitations.
Terrestrial Enrichment. Terrestrial enrichment occurs when terrestrial ecosystems
receive N loadings in excess of natural background levels, through either atmospheric
deposition or direct application. Evidence presented in the ISA for N0x/S0x (U.S. EPA, 2008c)
supports a causal relationship between atmospheric N deposition and biogeochemical cycling
and fluxes of N and carbon in terrestrial systems. Furthermore, evidence summarized in the
report supports a causal link between atmospheric N deposition and changes in the types and
number of species and biodiversity in terrestrial systems. 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
39 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
'° These estimates reflect the total value of the service, r
result of the emission reductions achieved by this rule.
40 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
6-58
-------
ecosystem conditions and indicators. This long time scale also affects the timing of the
ecosystem service changes. The ISA for N0x/S0x—Ecological Criteria concluded that the
evidence is sufficient to infer a causal relationship between nitrogen deposition and the
alteration of species richness, species composition, and biodiversity in terrestrial ecosystems
(U.S. EPA, 2008c).
One of the main provisioning services potentially affected by N deposition is grazing
opportunities offered by grasslands for livestock production in the Central U.S. Although N
deposition on these grasslands can offer supplementary nutritive value and promote overall
grass production, there are concerns that fertilization may favor invasive grasses and shift the
species composition away from native grasses. This process may ultimately reduce the
productivity of grasslands for livestock production. Losses due to invasive grasses can be
significant; for example, based on a bioeconomic model of cattle grazing in the upper Great
Plains, Leitch, Leistritz, and Bangsund (1996) and Leistritz, Bangsund, and Hodur (2004)
estimated $130 million in losses due to a leafy spurge infestation in the Dakotas, Montana, and
Wyoming.41 However, the contribution of N deposition to these losses is still uncertain.
Terrestrial nutrient enrichment also affects cultural and regulating services. For
example, in California, Coastal Sage Scrub (CSS) habitat concerns focus on a decline in CSS and
an increase in nonnative grasses and other species, impacts on the viability of threatened and
endangered species associated with CSS, and an increase in fire frequency. Changes in Mixed
Conifer Forest (MCF) include changes in habitat suitability and increased tree mortality,
increased fire intensity, and a change in the forest's nutrient cycling that may affect surface
water quality through nitrate leaching (U.S. EPA, 2008c). CSS and MCF are an integral part of
the California landscape, and together the ranges of these habitats include the densely
populated and valuable coastline and the mountain areas. Numerous threatened and
endangered species at both the state and federal levels reside in CSS and MCF. The value that
California residents and the U.S. population as a whole place on CSS and MCF habitats is
reflected in the various federal, state, and local government measures that have been put in
place to protect these habitats, including the Endangered Species Act, conservation planning
programs, and private and local land trusts. CSS and MCF habitat are showcased in many
popular recreation areas in California, including several national parks and monuments. In
addition, millions of individuals are involved in fishing, hunting, and wildlife viewing in California
every year (DOI, 2007). The quality of these trips depends in part on the health of the
41 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
6-59
-------
ecosystems and their ability to support the diversity of plants and animals found in important
habitats found in CSS or MCF ecosystems and the parks associated with those ecosystems.
Based on analyses in the NOX SOX REA average values of the total benefits in 2006 from fishing,
hunting, and wildlife viewing away from home in California were approximately $950 million,
$170 million, and $3.6 billion, respectively (U.S. EPA, 2009c).42 In addition, data from California
State Parks (2003) indicate that in 2002, 69% of adult residents participated in trail hiking for an
average of 24 days per year. The aggregate annual benefit for California residents from trail
hiking in 2007 was $11 billion (U.S. EPA, 2009c).43 It is not currently possible to quantify the loss
in value of services due to nitrogen deposition as those losses are already reflected in the
estimates of the contemporaneous total value of these recreational activities. Restoration of
services through decreases in nitrogen deposition would likely increase the total value of
recreational services.
Fire regulation is also an important regulating service that could be affected by nutrient
enrichment of the CSS and MCF ecosystems by encouraging growth of more flammable grasses,
increasing fuel loads, and altering the fire cycle. Over the 5-year period from 2004 to 2008,
Southern California experienced, on average, over 4,000 fires per year burning, on average,
over 400,000 acres per year (National Association of State Foresters [NASF], 2009). It is not
possible at this time to quantify the contribution of nitrogen deposition, among many other
factors, to increased fire risk.
We are unable to provide an estimate of the terrestrial nutrient enrichment co-benefits
associated with the revised or alternative annual standards due to data, time, and resource
limitations. Methods are not yet available to allow estimation of changes in ecosystem services
due to nitrogen deposition.
6.7.2.3 Vegetation Effects Associated with Gaseous Sulfur Dioxide
Uptake of gaseous sulfur dioxide in a plant canopy is a complex process involving
adsorption to surfaces (leaves, stems, and soil) and absorption into leaves. S02 penetrates into
leaves through to the stomata, although there is evidence for limited pathways via the cuticle
(U.S. EPA, 2008c). Pollutants must be transported from the bulk air to the leaf boundary layer in
order to get to the stomata. When the stomata are closed, as occurs under dark or drought
conditions, resistance to gas uptake is very high and the plant has a very low degree of
42 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
result of the emission reductions achieved by this rule.
'3 These estimates reflect the total value of the service, r
result of the emission reductions achieved by this rule.
43 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
6-60
-------
susceptibility to injury. In contrast, mosses and lichens do not have a protective cuticle barrier
to gaseous pollutants or stomates and are generally more sensitive to gaseous sulfur and
nitrogen than vascular plants (U.S. EPA, 2008c). Acute foliar injury usually happens within hours
of exposure, involves a rapid absorption of a toxic dose, and involves collapse or necrosis of
plant tissues. Another type of visible injury is termed chronic injury and is usually a result of
variable S02 exposures over the growing season. Besides foliar injury, chronic exposure to low
S02 concentrations can result in reduced photosynthesis, growth, and yield of plants (U.S. EPA,
2008c). These effects are cumulative over the season and are often not associated with visible
foliar injury. As with foliar injury, these effects vary among species and growing environment.
S02 is also considered the primary factor causing the death of lichens in many urban and
industrial areas (Hutchinson et al., 1996). The ISA for N0x/S0x—Ecological Criteria concluded
that the evidence is sufficient to infer a causal relationship between S02 injury to vegetation
(U.S. EPA, 2008c).
6.6.2.4 Mercury-Related Co-benefits Associated with the Role ofSulfate in Mercury Methylation
Mercury is a persistent, bioaccumulative toxic metal that is emitted from in three forms:
gaseous elemental Hg (Hg°), oxidized Hg compounds (Hg+2), and particle-bound Hg (HgP).
Methylmercury (MeHg) is formed by microbial action in the top layers of sediment and soils,
after Hg has precipitated from the air and deposited into waterbodies or land. Once formed,
MeHg is taken up by aquatic organisms and bioaccumulates up the aquatic food web. Larger
predatory fish may have MeHg concentrations many times, typically on the order of one million
times, that of the concentrations in the freshwater body in which they live.
The NOX SOX ISA—Ecological Criteria concluded that evidence is sufficient to infer a
causal relationship between sulfur deposition and increased mercury methylation in wetlands
and aquatic environments (U.S. EPA, 2008c). Specifically, there appears to be a relationship
between S042" deposition and mercury methylation; however, the rate of mercury methylation
varies according to several spatial and biogeochemical factors whose influence has not been
fully quantified (see Figure 6-14). Therefore, the correlation between S042" deposition and
MeHg could not be quantified for the purpose of interpolating the association across
waterbodies or regions. Nevertheless, because changes in MeHg in ecosystems represent
changes in significant human and ecological health risks, the association between sulfur and
mercury cannot be neglected (U.S. EPA, 2008c).
6-61
-------
Spatial Factors
Lake/Reservoir
Sediment Disturbance
Upstream Wetlands
Upstream Forested Land
Upstream Erosion
Upstream Urban Land
Mercury
+
SRB
+
Sulfate
s
<;
Bioqeochemical Factors
Organic Matter
Sulfide
Salinity
Anoxia
Temperature
Methylmercury
Does not promote meihylatiori
Promotes metliylation
Figure 6-14. Spatial and Biogeochemical Factors Influencing MeHg Production
As research evolves and the computational capacity of models expands to meet the
complexity of mercury methylation processes in ecosystems, the role of interacting factors may
be better parsed out to identify ecosystems or regions that are more likely to generate higher
concentrations of MeHg. Figure 6-15 illustrates the type of current and forward-looking
research being developed by the U.S. Geological Survey (USGS) to synthesize the contributing
factors of mercury and to develop a map of sensitive watersheds. The mercury score
referenced in Figure 6-15 is based on S042" concentrations, acid neutralizing capacity (ANC),
levels of dissolved organic carbon and pH, mercury species concentrations, and soil types to
gauge the methylation sensitivity (Myers et al., 2007).
6-62
-------
Mercury sensitivity
tore (unitless)
Figure 6-15. Preliminary USGS Map of Mercury Methylation-Sensitive Watersheds
Source: Myers et al., 2007.
Interdependent biogeochemical factors preclude the existence of simple sulfate-related
mercury methylation models. It is clear that decreasing sulfate deposition is likely to result in
decreased MeHg concentrations. Future research may allow for the characterization of a usable
sulfate-MeHg response curve; however, no regional or classification calculation scale can be
created at this time because of the number of confounding factors.
Decreases in S042" deposition have already shown promising reductions in MeHg.
Observed decreases in MeHg fish tissue concentrations have been linked to decreased
acidification and declining S042" and mercury deposition in Little Rock Lake, Wl (Hrabik and
Watras, 2002), and to decreased S042" deposition in Isle Royale in Lake Superior, Ml (Drevnick
et al., 2007). Although the possibility exists that reductions in S042" emissions could generate a
pulse in MeHg production because of decreased sulfide inhibition in sulfate-saturated waters,
this effect would likely involve a limited number of U.S. waters (Harmon et al., 2007). Also,
because of the diffusion and outward flow of both mercury-sulfide complexes and S042",
increased mercury methylation downstream may still occur in sulfate-enriched ecosystems with
increased organic matter and/or downstream transport capabilities.
6-63
-------
Remediation of sediments heavily contaminated with mercury has yielded significant
reductions of MeHg in biotic tissues. Establishing quantitative relations in biotic responses to
MeHg levels as a result of changes in atmospheric mercury deposition, however, presents
difficulties because direct associations can be confounded by all of the factors discussed in this
section. Current research does suggest that the levels of MeHg and total mercury in ecosystems
are positively correlated, so that reductions in mercury deposited into ecosystems would also
eventually lead to reductions in MeHg in biotic tissues. Ultimately, an integrated approach that
involves the reduction of both sulfur and mercury emissions may be most efficient because of
the variability in ecosystem responses. Reducing SOX emissions could have a beneficial effect on
levels of MeHg in many waters of the United States.
MeHg is the only form of mercury that biomagnifies in the food web. Concentrations of
MeHg in fish are generally on the order of a million times the MeHg concentration in water. In
addition to mercury deposition, key factors affecting MeHg production and accumulation in fish
include the amount and forms of sulfur and carbon species present in a given waterbody. Thus,
two adjoining water bodies receiving the same deposition can have significantly different fish
mercury concentrations.
Methylmercury builds up more in some types offish and shellfish than in others. The
levels of methylmercury in high and shellfish vary widely depending on what they eat, how long
they live, and how high they are in the food chain. Most fish, including ocean species and local
freshwater fish, contain some methylmercury. In general, higher mercury concentrations are
expected in top predators, which are often large fish relative to other species in a waterbody.
The ecosystem service most directly affected by sulfate-mediated mercury methylation
is the provision offish for consumption as a food source. This service is of particular importance
to groups engaged in subsistence fishing, pregnant women and young children.
6.6.3 Ecosystem Co-benefits from Reductions in Mercury Emissions
Deposition of mercury to waterbodies can also have an impact on ecosystems and
wildlife. Mercury contamination is present in all environmental media with aquatic systems
experiencing the greatest exposures due to bioaccumulation. Bioaccumulation refers to the net
uptake of a contaminant from all possible pathways and includes the accumulation that may
occur by direct exposure to contaminated media as well as uptake from food.
Atmospheric mercury enters freshwater ecosystems by direct deposition and through
runoff from terrestrial watersheds. Once mercury deposits, it may be converted to organic
6-64
-------
methylmercury mediated primarily by sulfate-reducing bacteria. Methylation is enhanced in
anaerobic and acidic environments, greatly increasing mercury toxicity and potential to
bioaccumulate in aquatic foodwebs. A number of key biogeochemical controls influence the
production of methylmercury in aquatic ecosystems. These include sulfur, pH, organic matter,
iron, mercury "aging," and bacteria type and activity (Munthe et al., 2007).
Wet and dry deposition of oxidized mercury is a dominant pathway for bringing mercury
to terrestrial surfaces. In forest ecosystems, elemental mercury may also be absorbed by plants
stomatally, incorporated by foliar tissues and released in litterfall (Ericksen etal., 2003).
Mercury in throughfall, direct deposition in precipitation, and uptake of dissolved mercury by
roots (Rea et al., 2002) are also important in mercury accumulation in terrestrial ecosystems.
Soils have significant capacity to store large quantities of atmospherically deposited
mercury where it can leach into groundwater and surface waters. The risk of mercury exposure
extends to insectivorous terrestrial species such as songbirds, bats, spiders, and amphibians
that receive mercury deposition or from aquatic systems near the forest areas they inhabit
(Bergeron et al., 2010a, b; Cristol et al., 2008; Rimmer et al., 2005; Wada et al., 2009 & 2010).
Numerous studies have generated field data on the levels of mercury in a variety of wild
species. Many of the data from these environmental studies are anecdotal in nature rather than
representative or statistically designed studies. The body of work examining the effects of these
exposures is growing but still incomplete given the complexities of the natural world. A large
portion of the adverse effect research conducted to date has been carried out in the laboratory
setting rather than in the wild; thus, conclusions about overarching ecosystem health and
population effects are difficult to make at this time. In the sections that follow numerous
effects have been identified at differing exposure levels.
6.6.3.1 Mercury Effects on Fish
A review of the literature on effects of mercury on fish (Crump and Trudeau, 2009)
reports results for numerous species including trout, bass (large and smallmouth), northern
pike, carp, walleye, salmon and others from laboratory and field studies. The effects studied are
reproductive and include deficits in sperm and egg formation, histopathological changes in
testes and ovaries, and disruption of reproductive hormone synthesis. These studies were
conducted in areas from New York to Washington and while many were conducted by adding
MeHg to water or diet many were conducted at current environmental levels. While we cannot
determine at this time whether these reproductive deficits are affecting fish populations across
the United States it should be noted that it is possible that overtime reproductive deficits could
6-65
-------
have an effect on populations. Lower fish populations would conceivably impact the ecosystem
services like recreational fishing derived from having healthy aquatic ecosystems quite apart
from the effects of consumption advisories due to the human health effects of mercury.
6.6.3.2 Mercury Effects on Birds
In addition to effects on fish, mercury also affects avian species. In previous reports (U.S.
EPA, 1997; 2005), much of the focus has been on large piscivorous species in particular the
common loon. The loon is most visible to the public during the summer breeding season on
northern lakes and they have become an important symbol of wilderness in these areas
(Mclntyre and Barr, 1997). A multitude of loon watch, preservation, and protection groups have
formed over the past few decades and have been instrumental in promoting conservation,
education, monitoring, and research of breeding loons (Mclntyre and Evers, 2000; Evers, 2006).
Significant adverse effects on breeding loons from mercury have been found to occur including
behavioral (reduced nest-sitting), physiological (flight feather asymmetry) and reproductive
(chicks fledged/territorial pair) effects (Evers, 2008). Additionally Evers, et al (2008) report that
they believe that the weight of evidence indicates that population-level effects occur in parts of
Maine and New Hampshire, and potentially in broad areas of the loon's range.
Recently attention has turned to other piscivorous species such as the white ibis, and
great snowy egret. While considered to be fish-eating generally these wading birds have a very
wide diet including crayfish, crabs, snails, insects and frogs. These species are experiencing a
range of adverse effects due to exposure to mercury. The white ibis has been observed to have
decreased foraging efficiency (Adams and Frederick, 2008). Additionally ibises have been shown
to exhibit decreased reproductive success and altered pair behavior (Frederick and Jayasena,
2010). These effects include significantly more unproductive nests, male/male pairing, reduced
courtship behavior (head bobbing and pair bowing) and lower nestling production by exposed
males. In this study, a worst-case scenario suggested by the results could involve up to a 50%
reduction in fledglings due to MeHg in diet. These estimates may be conservative if male/male
pairing in the wild it could result in a shortage of partners for females and the effect of
homosexual breeding would be magnified. In egrets, mercury has been implicated in the
decline of the species in south Florida (Sepulveda, et al., 1999) and Hoffman (2010) has shown
that egrets show liver and possibly kidney effects. While ibises and egrets are most abundant in
coastal areas and these studies were conducted in south Florida and Nevada the ranges of
ibises and egrets extend to a large portion of the United States. Ibis territory can range inland
to Oklahoma, Arkansas and Tennessee. Egret range covers virtually the entire United States
except the mountain west.
6-66
-------
Insectivorous birds have also been shown to suffer adverse effects due to mercury
exposure. These songbirds such as Bicknell's thrush, tree swallows and the great tit have shown
reduced reproduction, survival, and changes in singing behavior. Exposed tree swallows
produced fewer fledglings (Brasso, 2008), lower survival (Hallinger, 2010) and had
compromised immune competence (Hawley, 2009). The great tit has exhibited reduced singing
behavior and smaller song repertoire in an area of high contamination in the vicinity of a
metallurgic smelter in Flanders (Gorissen, 2005).
6.6.3.3 Mercury Effects on Mammals
In mammals, adverse effects have been observed in mink and river otter, both fish
eating species. For otter from Maine and Vermont maximum concentrations on Hg in fur nearly
equal or exceed a concentration associated with mortality and concentration in liver for mink in
Massachusetts/Connecticut and the levels in fur from mink in Maine exceed concentrations
associated with acute mortality (Yates, 2005). Adverse sublethal effects may be associated with
lower Hg concentrations and consequently be more widespread than potential acute effects.
These effects may include increased activity, poorer maze performance, abnormal startle reflex,
and impaired escape and avoidance behavior (Scheuhammer et al., 2007).
6.6.3.4 Mercury Ecological Conclusions
The studies cited here provide a glimpse of the scope of mercury effects on wildlife
particularly reproductive and survival effects. These effects range across species from fish to
mammals and spatially across a wide area of the United States. The literature is far from
complete however. Much more research is required to establish a link between the ecological
effects on wildlife and the effect on ecosystem services (services that the environment provides
to people) for example recreational fishing, bird watching and wildlife viewing. The EPA is not,
however, currently able to quantify or monetize the co-benefits of reducing mercury exposures
affecting provision of ecosystem services.
6.6.4 Vegetation Co-benefits from Reductions in Ambient Ozone
Illustrative emission reduction strategies that include NOX emission reductions would
affect ambient ozone concentrations. Ozone causes discernible injury to a wide array of
vegetation (U.S. EPA, 2006a; Fox and Mickler, 1996). Air pollution can affect the environment
and affect ecological systems, leading to changes in the ecological community and influencing
the diversity, health, and vigor of individual species (U.S. EPA, 2006a). In terms of forest
productivity and ecosystem diversity, ozone may be the pollutant with the greatest potential
for regional-scale forest impacts (U.S. EPA, 2006a). Studies have demonstrated repeatedly that
6-67
-------
ozone concentrations commonly observed in polluted areas can have substantial impacts on
plant function (De Steiguer et al., 1990; Pye, 1988).
When ozone is present in the air, it can enter the leaves of plants, where it can cause
significant cellular damage. Like carbon dioxide (C02) 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,
2006a; 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, 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 overtime may lead to premature plant death. Furthermore, there is evidence that ozone
can interfere with the formation of mycorrhiza, 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, 2006a).
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
6-68
-------
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,
2006a; Winner, 1994). After injuries have occurred, plants may be capable of repairing the
damage to a limited extent (U.S. EPA, 2006a). 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.
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, 2006a). 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, 2006a, 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, 2006a).
6.6.4.1 Ozone Effects on Forests
Air pollution can affect the environment and affect ecological systems, leading to
changes in the ecological community and influencing the diversity, health, and vigor of
individual species (U.S. EPA, 2006a). Ozone has been shown in numerous studies to have a
strong effect on the health of many plants, including a variety of commercial and ecologically
important forest tree species throughout the United States (U.S. EPA, 2007b).
In the U.S., this data comes from the U.S. Department of Agriculture (USDA) Forest
Service Forest Inventory and Analysis (FIA) program. As part of its Phase 3 program (formerly
known as Forest Health Monitoring), FIA looks for visible foliar injury of ozone-sensitive forest
plant species at each ground monitoring site across the country (excluding woodlots and urban
trees) that meets certain minimum criteria. Because ozone injury is cumulative over the course
of the growing season, examinations are conducted in July and August, when ozone
concentrations and associated injury are typically highest.
6-69
-------
Monitoring of ozone injury to plants by the U.S. Forest Service has expanded over the
last 15 years from monitoring sites in 10 states in 1994 to nearly 1,000 monitoring sites in 41
states in 2002. Since 2002, the monitoring program has further expanded to 1,130 monitoring
sites in 45 states. Figure 6-16 shows the results of this monitoring program for the year 2002
broken down by U.S. EPA Regions.44 Figure 6-17 identifies the counties that were included in
Figure 6-16, and provides the county-level data regarding the presence or absence of ozone-
related injury. As shown in Figure 6-16, large geographic areas of EPA Regions 6, 8, and 10 were
not included in the assessment. Ozone damage to forest plants is classified using a subjective
five-category biosite index based on expert opinion, but designed to be equivalent from site to
site. Ranges of biosite values translate to no injury, low or moderate foliar injury (visible foliar
injury to highly sensitive or moderately sensitive plants, respectively), and high or severe foliar
injury, which would be expected to result in tree-level or ecosystem-level responses,
respectively (U.S. EPA, 2006a; Coulston, 2004). The highest percentages of observed high and
severe foliar injury, which are most likely to be associated with tree or ecosystem-level
responses, are primarily found in the Mid-Atlantic and Southeast regions. While the assessment
showed considerable regional variation in ozone injury, this assessment targeted different
ozone-sensitive species in different parts of the country with varying ozone sensitivity, which
contributes to the apparent regional differences. It is important to note that ozone can have
other, more significant impacts on forest plants (e.g., reduced biomass growth in trees) prior to
showing signs of visible foliar injury (U.S. EPA, 2006a).
Assessing the impact of ground-level ozone on forests in the U.S. involves understanding
the risks to sensitive tree species from ambient ozone concentrations and accounting for the
prevalence of those species within the forest. As a way to quantify the risks to particular plants
from ground-level ozone, scientists have developed ozone-exposure/tree-response functions by
exposing tree seedlings to different ozone levels and measuring reductions in growth as
"biomass loss." Typically, seedlings are used because they are easy to manipulate and measure
their growth loss from ozone pollution. The mechanisms of susceptibility to ozone within the
leaves of seedlings and mature trees are identical, and the decreases predicted using the
seedlings should be related to the decrease in overall plant fitness for mature trees, but the
magnitude of the effect may be higher or lower depending on the tree species (Chappelka and
Samuelson, 1998). In areas where certain ozone-sensitive species dominate the forest
community, the biomass loss from ozone can be significant. Experts have identified 2% annual
The data are based on averages of all observations collected in 2002, which is the last year for which data are
publicly available. For more information, please consult EPA's 2008 Report on the Environment (U.S. EPA, 2008b).
6-70
-------
Degree of injury:
None
Low Moderate High Severe
Percent of monitoring sites in each category:
Region 1
(54 sites)
Region 2
(42 sites)
Region 3
(111 sites)
Region 4
(227 sites)
Region 5
(180 sites)
Region 6
(59 sites)
Region 7
(63 sites)
Region 8
(72 sites)
Region 9
(80 sites)
Region 10
(57 sites)
68.5
61.9
55.9
16.7
11.
1 -
21.4
M
18.0
14.4
7.1
7.2
75.3
10.1
7.0
"~
75.6
18,3
6.1
3.7
2.4
4.5
3.5
4.0
94.9
85.7
9.5
-
5.1
3.2
•1.6
100.0
76.3
12.5
8.8
H.3
±1.3
100.0
EPA Regions
aCoverage: 945 monitoring sites,
located in 41 states.
"Totals may not add to 100% due to
rounding.
Data source: USDA Forest Service,
2006
cDegree of Injury: These categories reflect a subjective index based on expert opinion. Ozone can have other,
more significant impacts on forest plants (e.g., reduced biomass growth in trees) prior to showing signs of visible
foliar injury.
Figure 6-16. Visible Foliar Injury to Forest Plants from Ozone in U.S. by EPA Regions
a,b,c
6-71
-------
Foliar Injury
Absent
Present
Figure 6-17. Presence and Absence of Visible Foliar Injury, as Measured by U.S. Forest
Service, 2002
Source: U.S. EPA, 2007b.
biomass loss as a level of concern, which would cause long term ecological harm as the short-
term negative effects on seedlings compound to affect long-term forest health (Heck and
Cowling, 1997).
Ozone damage to the plants including the trees and understory in a forest can affect the
ability of the forest to sustain suitable habitat for associated species particularly threatened and
endangered species that have existence value—a nonuse ecosystem service—for the public.
Similarly, damage to trees and the loss of biomass can affect the forest's provisioning services
in the form of timber for various commercial uses. In addition, ozone can cause discoloration of
leaves and more rapid senescence (early shedding of leaves), which could negatively affect fall-
color tourism because the fall foliage would be less available or less attractive. Beyond the
aesthetic damage to fall color vistas, forests provide the public with many other recreational
and educational services that may be affected by reduced forest health including hiking, wildlife
viewing (including bird watching), camping, picnicking, and hunting. Another potential effect of
biomass loss in forests is the subsequent loss of climate regulation service in the form of
reduced ability to sequester carbon and alteration of hydrologic cycles.
6-72
-------
Some of the common tree species in the United States that are sensitive to ozone are
black cherry (Prunus serotina), tulip-poplar (Liriodendron tulipifera), and eastern white pine
(Pinus strobus). Ozone-exposure/tree-response functions have been developed for each of
these tree species, as well as for aspen (Populus tremuliodes), and ponderosa pine (Pinus
ponderosa) (U.S. EPA, 2007b).
6.6.4.2 Ozone Effects on Crops
Laboratory and field experiments have shown reductions in yields for agronomic crops
exposed to ozone, including vegetables (e.g., lettuce) and field crops (e.g., cotton and wheat).
Damage to crops from ozone exposures includes yield losses (i.e., in terms of weight, number,
or size of the plant part that is harvested), as well as changes in crop quality (i.e., physical
appearance, chemical composition, or the ability to withstand storage) (U.S. EPA, 2007b). The
most extensive field experiments, conducted under the National Crop Loss Assessment
Network (NCLAN) examined 15 species and numerous cultivars. The NCLAN results show that
"several economically important crop species are sensitive to ozone levels typical of those
found in the United States" (U.S. EPA, 2006a). In addition, economic studies have shown
reduced economic co-benefits as a result of predicted reductions in crop yields, directly
affecting the amount and quality of the provisioning service provided by these crops, associated
with observed ozone levels (Kopp et al., 1985; Adams et al., 1986; Adams et al., 1989). In
addition, visible foliar injury by itself can reduce the market value of certain leafy crops (such as
spinach, lettuce). According to the Ozone Staff Paper, there has been no evidence that crops
are becoming more tolerant of ozone (U.S. EPA, 2007b). Using the Agriculture Simulation Model
(AGSIM) (Taylor, 1994) to calculate the agricultural benefits of reductions in ozone exposure,
the EPA estimated that attaining a W126 standard of 13 ppm-hr would produce monetized
benefits of approximately $400 million to $620 million in 2006 (inflated to 2006 dollars) (U.S.
EPA, 2007b).45
6.6.4.3 Ozone Effects on Ornamental Plants
Urban ornamental plants are an additional vegetation category likely to experience
some degree of negative effects associated with exposure to ambient ozone levels. Several
ornamental species have been listed as sensitive to ozone (Abt Associates, 1995). Because
ozone causes visible foliar injury, the aesthetic value of ornamental plants (such as petunia,
geranium, and poinsettia) in urban landscapes would be reduced (U.S. EPA, 2007b). Sensitive
45These estimates illustrate the value of vegetation effects from a substantial reduction of ozone concentrations,
not the marginal change in ozone concentrations anticipated a result of the emission reductions achieved by this
rule.
6-73
-------
ornamental species would require more frequent replacement and/or increased maintenance
(fertilizer or pesticide application) to maintain the desired appearance because of exposure to
ambient ozone (U.S. EPA, 2007b). In addition, many businesses rely on healthy-looking
vegetation for their livelihoods (e.g., horticulturalists, landscapers, Christmas tree growers,
farmers of leafy crops, etc.). The ornamental landscaping industry is a multi-billion dollar
industry that affects both private property owners/tenants and governmental units responsible
for public areas (Abt Associates, 1995). Preliminary data from the 2007 Economic Census
indicate that the landscaping services industry, which is primarily engaged in providing
landscape care and maintenance services and installing trees, shrubs, plants, lawns, or gardens,
was valued at $53 billion (U.S. Census Bureau, 2010). Therefore, urban ornamentals represent a
potentially large unquantified benefit category. This aesthetic damage may affect the
enjoyment of urban parks by the public and homeowners' enjoyment of their landscaping and
gardening activities. In addition, homeowners may experience a reduction in home value or a
home may linger on the market longer due to decreased aesthetic appeal. In the absence of
adequate exposure-response functions and economic damage functions for the potential range
of effects relevant to ornamental plants, we cannot conduct a quantitative analysis to estimate
these effects.
We are unable to provide an estimate of the ozone crop co-benefits associated with the
revised or alternative annual standards due to data, time, and resource limitations.
6.7 References
Abt Associates, Inc. 1995. Urban ornamental plants: sensitivity to ozone and potential economic
losses. U.S. EPA, Office of Air Quality Planning and Standards, Research Triangle Park.
Under contract to RADIAN Corporation, contract no. 68-D3-0033, WA no. 6. pp. 9-10.
Available on the Internet at
.
Abt Associates, Inc. 2001. Assessing Public Opinions on Visibility Impairment due to Air
Pollution: Summary Report. Prepared for EPA Office of Air Quality Planning and
Standards; funded under EPA Contract No. 68-D-98-001. Abt Associates Inc., Bethesda,
MD. January. Available on the Internet at
.
Adams, Evan M., and Frederick, Peter C. 2008. "Effects of methylmercury and spatial complexity
on foraging behavior and foraging efficiency in juvenile white ibises (Eudocimus albus)."
Environmental Toxicology and Chemistry. Vol 27, No. 8.
6-74
-------
Adams, R. M., Glyer, J. D., Johnson, S. L, McCarl, B. A. 1989. "A reassessment of the economic
effects of ozone on U.S. agriculture." Journal of the Air Pollution Control Association, 39,
960-968'
Adams, R. M., Hamilton, S. A., McCarl, B. A. 1986. "The benefits of pollution control: the case of
ozone and U.S. agriculture." American Journal of Agricultural Economics, 34, 3-19.
Anselin, L., Le Gallo, J. 2006. "Interpolation of air quality measures in hedonic house price
models: Spatial aspects." Spatial Economic Analysis 1: 31-52
Anselin, L., Lozano-Gracia, N. 2008. "Errors in variables and spatial effects in hedonic house
price models of ambient air quality." Empirical Economics 34: 5-34.
Arrow, K.J., et al. 1993. Report of the NOAA Panel on Contingent Valuation. Federal Register.
Volume 58, Number 10. Pages 4601 to 4614. Available on the Internet at
.
Arzavus K.M., Dickhut R.M., and Canuel E.A. 2001. "Fate of Atmospherically Deposited
Polycyclic Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science &
Technology, 35, 2178-2183.
Banzhaf, S., D. Burtraw, D. Evans, and A. Krupnick. 2006. "Valuation of Natural Resource
Improvements in the Adirondacks." Land Economics 82:445-464.
Bayer, P., Keohane, N.O. and Timmins, C. 2009. "Migration and hedonic valuation: the case of
air quality," Journal of Environmental Economics and Management, 58 (2009) 1-14.
BBC Research & Consulting. 2003. Phoenix Area Visibility Survey. Draft Report. Available on the
Internet at .
Bergeron, CM., Bodinof, CM., Unrine, JM., Hopkins, WA. 2010a. "Mercury accumulation along a
contamination gradient and nondestructive indices of bioaccumulation in amphibians."
Environmental Toxicology and Chemistry 29(4), 980-988.
Bergeron, CM., Bodinof, CM., Unrine, JM., Hopkins, WA. 2010b. "Bioaccumulation and maternal
transfer of mercury and selenium in amphibians." Environmental Toxicology and
Chemistry 29(4), 989-997.
Beron KJ, Murdoch J, Thayer M. 2001. "The benefits of visibility improvement: New evidence
from Los Angeles metropolitan area." J Real Estate Finance Econ 22(2-3):319-337.
Beron KJ, Hanson Y, Murdoch JC, Thayer MA. 2004. "Hedonic price functions and spatial
dependence: implications for the demand for urban air quality." In: Anselin L, Florax RJ,
Rey SJ (eds) Advances in spatial econometrics: methodology, tools and applications.
Springer-Verlag, Berlin, pp 267-281.
6-75
-------
Bond, T.C., and Sun H.L. (2005) Can reducing BC emissions counteract global warming?
Environmental Science & Technology 39 (16).
Brasso, Rebecka L, and Cristol, Daniel A. 2008. "Effects of mercury exposure in the reproductive
success of tree swallows (Tachycineta bicolor)." Ecotoxicology. 17:133-141.
Brookshire, D.S., R.C. d'Arge, W.D. Schulze and M.A. Thayer. 1979. Methods Development for
Assessing Tradeoffs in Environmental Management, Vol. II: Experiments in Valuing Non-
Market Goods: A Case Study of Alternative Benefit Measures of Air Pollution Control in
the South Coast Air Basin of Southern California. Prepared for the U.S. Environmental
Protection Agency, Office of Research and Development. Available on the Internet at
.
Carson, R.T., R.C. Mitchell, and P.A. Rudd. 1990. "Valuing Air Quality Improvements: Simulating
a Hedonic Pricing Equation in the Context of a Contingent Valuation Scenario." In
Visibility and Fine Particles, Transactions of an AWMA/EPA International Specialty
Conference, C.V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
Available on the Internet at
.
Chameides. W.L., Yu, H., Liu, S.C., Bergin, M., Zhou, X., Mearns, L., Wang, G., Kiang, C.S., Saylor,
R.D., Luo, C. Huang, Y., Steiner, A., and Giorgi, F. 1999. "Case study of the effects of
atmospheric aerosols and regional haze on agriculture: an opportunity to enhance crop
yields in China through emission controls?" PNAS, 96: 13626-13633.
Chappelka, A.H., Samuelson, L.J. 1998. "Ambient ozone effects on forest trees of the eastern
United States: a review." New Phytologist, 139, 91-108.
Chay KY, Greenstone M. 2005. "Does air quality matter? evidence form the housing market." J
Po//tfconll3(2):376-424.
6-76
-------
Chestnut, L.G. 1997. Methodology for Estimating Values for Changes in Visibility in National
Parks. Prepared for Office of Policy, Planning and Evaluation, May. Available on the
Internet at
.
Chestnut, L.G. and R. Dennis. 1997. "Economic Benefits of Improvements in Visibility: Acid Rain
Provisions of the 1990 Clean Air Act Amend ments." Journal of Air and Waste
Management Association 47: 395-402.
Chestnut, L.G., and R.D. Rowe. 1990a. "New National Park Visibility Value Estimates." In
Visibility and Fine Particles, Transactions of an AWMA/EPA International Specialty
Conference, C.V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
Chestnut, L.G. and R.D. Rowe. 1990b. Preservation Values for Visibility Protection at the
National Parks. Prepared for U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, and National Park Service, Air Quality Management Division.
Available on the Internet at
.
Chestnut, L.G. and R.D. Rowe. 1990c. "Economic Valuation of Changes in Visibility: A State of
the Science Assessment for NAPAP." Section B5 in Irving, Patricia M., e.d., 1991. Acid
Deposition: State of Science and Technology, Volume IV, Control Technologies, Future
Emissions, and Effects Valuation, The U.S. National Acid Precipitation Assessment
Program, Report 27, page 27-153.
Chestnut, L.G., R.D. Rowe and J. Murdoch. 1986. Review of 'Establishing and Valuing the Effects
of Improved Visibility in Eastern United States.' Prepared for the U.S. Environmental
Protection Agency. October. Available on the Internet at
.
Cordell, H.K., C.J. Betz, G. Green, and M. Owens. 2005. Off-Highway Vehicle Recreation in the
United States, Regions and States: A National Report from the National Survey on
Recreation and the Environment (NSRE). Prepared for the U.S. Department of
Agriculture Forest Service, Southern Research Station, National OHV Policy and
Implementation Teams, Athens, GA. Available on the Internet at
Cordell, K., B. Leeworthy, G.T. Green, C. Betz, B. Stephens. 2008. The National Survey on
Recreation & the Environment. Research Work Unit 4953. Pioneering Research on
Changing Forest Values in the South and Nation, U.S. Department of Agriculture Forest
Service, Southern Research Station, Athens, GA. Available on the Internet at
.
6-77
-------
Cotrufo, M.F., De Santo A.V., Alfani A., Bartoli G., De Cristofaro A. 1995. "Effects of urban heavy
metal pollution on organic matter decomposition in Quercus ilex L. Woods."
Environmental Pollution, 89(1), 81-87.
Coulston, J.W., Riitters, K.H., Smith, G.C. 2004. "A preliminary assessment of the Montreal
process indicators of air pollution for the United States." Environmental Monitoring and
Assessment, 95, 57-74.
Cristol D. A., Brasso R. L., Condon A. M., Fovargue R. E., Friedman S. L., Hallinger K. K., Monroe
A. P., White A. E. 2008. "The movement of aquatic mercury through terrestrial food
webs." Science 320, 335-335.
Cropper, Maureen. 2000. "Has Economic Research Answered the Needs of Environmental
Policy?" Journal of Environmental Economics and Management 39, 328-350.
Crump, Kate L., and Trudeau, Vance L. 2009. "Mercury-induced reproductive impairment in
fish." Environmental Toxicology and Chemistry. Vol. 28, No. 5.
De Steiguer, J., Pye, J., Love, C. 1990. "Air Pollution Damage to U.S. Forests." Journal of Forestry,
88(8), 17-22.
DeBell, L. J.; Gebhart, K. A.; Hand, J. L.; Malm, W. C; Pitchford, M. L.; Schichtel, B. A.; White, W.
H. 2006. Spatial and Seasonal Patterns and Temporal Variability of Haze and its
Constituents in the United States, Report IV; Cooperative Institute for Research in the
Atmosphere, Nov; ISSN 0737-5352-74. Available on the Internet at
.
DeHayes, D.H., P.G. Schaberg, G.J. Hawley, and G.R. Strimbeck. 1999. "Acid rain impacts on
calcium nutrition and forest health." e/osc/ence49(10):789-800.
DeLucchi, M.A., J.J. Murphy, D.R. McCubbin. 2002. "The health and visibility cost of air
pollution: A comparison of estimation methods," Journal of Environmental
Management, 64 (2), pp. 139-152.
Ely, D.W., Leary, J.T., Stewart, T.R., Ross, D.M. 1991. "The Establishment of the Denver Visibility
Standard." Presented at Air and Waste Management Association 84th Annual Meeting,
June 16-21, Paper 91-48.4, 17pp. Available on the Internet at
.
Ericksen, J. A., Gustin, M. S., Schorran, D. E., Johnson, D. W., Lindberg, S. E., & Coleman, J. S.
(2003). "Accumulation of atmospheric mercury in forest foliage." Atmospheric
Environment 37(12), 1613-1622.
Evers, D.C., 2006. Status assessment and conservation plan for the common loon (Gavia immer)
in North America. U.S. Fish and Wildlife Service, Hadley, MA, USA.
6-78
-------
Evens, David C, Savoy, Lucas J., DeSorbo, Christopher R., Yates, David E., Hanson, William,
Taylor, Kate M., Siegel, Lori S., Cooley, John H. Jr., Bank, Michael S., Major, Andrew,
Munney, Kenneth, Mower, Barry F., Vogel, Harry S., Schoch, Nina, Pokras, Mark,
Goodale, Morgan W., Fair, Jeff. 2008. "Adverse effects from environmental mercury
loads on breeding common loons." Ecotoxicology. 17:69-81.
Fox, S., Mickler, R. A. (Eds.). 1996. Impact of Air Pollutants on Southern Pine Forests, Ecological
Studies. (Vol. 118, 513 pp.) New York: Springer-Verlag.
Frederick, Peter, and Jayasena, Nilmini. 2010. "Altered pairing behavior and reproductive
success in white ibises exposed to environmentally relevant concentrations of
methylmercury." Proceedings of the Royal Society B. doi: 10-1098.
Gorissen, Leen, Snoeijs, Tinne, Van Duyse, Els, and Eens, Marcel. 2005. "Heavy metal pollution
affects dawn singing behavior in a small passerine bird." Oecologia. 145: 540-509.
Grantz DA; Garner JHB; Johnson DW. 2003. "Ecological effects of particulate matter." Environ
Int, 29: 213-239.
Graves, Phil, James C. Murdoch, Mark A. Thayer, and Don Waldman. 1988. "The Robustness of
Hedonic Price Estimation: Urban Air Quality." Land Economics, 64: 220-33.
Grosclaude P. and N. Soguel. 1994. "Valuing damage to historic buildings using a contingent
market: a case study of road traffic externalities," Journal of Environmental Planning and
Management 37 (1994) (3), pp. 279-287.
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.
Hadley O.L., Corrigan C.E., Kirchstetter T.W., Cliff S.S., and Ramanathan V. (2010) Measured
black carbon deposition on the Sierra Nevada snow pack and implication for snow pack
retreat. Atmos. Chem. Phys. Discuss. 10, 10463-10485
(doi:10.5194/acpd-10-10463-2010).
Hageman KJ; Simonich SL; Campbell DH; Wilson GR; Landers DH. 2006. "Atmospheric deposition
of current-use and historic-use pesticides in snow at national parks in the western
United States." Environ Sci Technol, 40: 3174-3180
Hallinger, Kelly K., Cornell, Kerri L., Brasso, Rebecka L., and Cristol, Daniel A. 2010. "Mercury
exposure and survival in free-living tree swallows (Tachycineta bicolor)." Ecotoxicology.
Doi: 10.1007/S10646-010-0554-4.
6-79
-------
Hawley, Dana M., Hallinger, Kelly K., Cristol, Daniel A. 2009. "Compromised immune
competence in free-living tree swallows exposed to mercury." Ecotoxicology. 18:499-
503.
Heck, W.W. &Cowling E.B. 1997. "The need for a long term cumulative secondary ozone
standard—an ecological perspective." Environmental Management, January, 23-33.
Hoffman, David J., Henny, Charles J., Hill, Elwood F., Grover, Robert A., Kaiser, James L,
Stebbins, Katherine R. 2009. "Mercury and drought along the lower Carson River,
Nevada: III. Effects on blood and organ biochemistry and histopathology of snowy egrets
and black-crowned night-herons on Lahontan Reservoir, 2002-2006." Journal of
Toxicology and Environmental Health, Part A. 72: 20, 1223-1241.
Hutchison, R., and C.E. Kraft. 1994. "Hmong Fishing Activity and Fish Consumption." Journal of
Great Lakes Research 20(2):471-487.
Intergovernmental Panel on Climate Change 4th Assessment Report: Climate Change 2007
Synthesis Report Summary for Policy Makers Pachauri, R.K. and Reisinger, A. (Eds.).
Available on the Internet at
.
IMPROVE (Interagency Monitoring of Protected Visual Environments). 2010.1992-2008 data
from the IMPROVE network based on the "New IMPROVE algorithm" (updated
November, 2008). Available on the Internet at
Irving, P.M., e.d. 1991. Acid Deposition: State of Science and Technology, Volume III, Terrestrial,
Materials, Health, and Visibility Effects, The U.S. National Acid Precipitation Assessment
Program, Chapter 24, page 24-76.
Joslin, J.D., Kelly, J.M., and van Miegroet, H. 1992. "Soil chemistry and nutrition of North
American spruce-fir stands: evidence for recent change. " Journal of Environmental
Quality, 21, 12-30.
Kaval, P., and J. Loomis. 2003. Updated Outdoor Recreation Use Values With Emphasis On
National Park Recreation. Final Report October 2003, under Cooperative Agreement CA
1200-99-009, Project number IMDE-02-0070. Available on the Internet at
.
Kim, CW, Phipps, T, and Anselin L. 2003. "Measuring the benefits of air quality improvement: a
spatial hedonic approach." J Environ Econ Manage 45:24-39.
6-80
-------
Koch, D., Bond, T.C., Streets D., and Unger N. (2007) Linking future aerosol radiative forcing to
shifts in source activities. Geophys. Res. Lett. 34 (L05821), AGU
(doi:10.1029/2006GL028360, 2007). Available on the Internet at
.
Koch D., Bond T.C., Streets D., Unger N., and van der Werf G.R. (2007) Global impacts of
aerosols from particular source regions and sectors. Journal of Geophysical
Research-Atmospheres 112 (D02205).
Kopp, R. J., Vaughn, W. J., Hazilla, M., Carson, R. 1985. "Implications of environmental policy for
U.S. agriculture: the case of ambient ozone standards." Journal of Environmental
Management, 20, 321-331.
Kramer, A., T. Holmes, and M. Haefel. 2003. Contingent valuation of forest ecosystem
protection. Pp. 303-320 in Forests in a Market Economy. Edited by E.G. Sills and K.L. Abt.
Dordrecht, The Netherlands: Kluwer Academic Publishers.
Landers DH; Simonich SL; Jaffe DA; Geiser LH; Campbell DH; Schwindt AR; Schreck CB; Kent ML;
Hafner WD; Taylor HE; Hageman KJ; Usenko S; Ackerman LK; Schrlau JE; Rose NL; Blett
TF; Erway MM. (2008). The Fate, Transport and Ecological Impacts of Airborne
Contaminants in Western National Parks (USA). EPA/600/R-07/138. U.S. Environmental
Protection Agency, Office of Research and Development, NHEERL, Western Ecology
Division. Corvallis, Oregon.
Leeworthy, V.R., and P.C. Wiley. 2001. Current Participation Patterns in Marine Recreation.
Silver Spring, MD: U.S. Department of Commerce, National Oceanic and Atmospheric
Administration, National Ocean Service, Special Projects. November. Available on the
Internet at < http://www.naturetourismplanning.com/pdfs/NSRE_2.pdf>.
Leggett, C.G. and J.E. Neumann. 2004. Responding to SAB Council Comments on the May 2003
Draft Analytical Plan for the Section 812 Second Prospective—Visibility Benefits.
Prepared for U.S. Environmental Protection Agency OPAR. Available on the Internet at
.
Leistritz, F. L., D. A. Bangsund, and N. M. Hodur. 2004. "Assessing the Economic Impact of
Invasive Weeds: The Case of Leafy Spurge (Euphorbia esula)." Weed Technology
18:1392-1395.
Leitch, J. A, F. L Leistritz, and D. A Bangsund. 1996. "Economic effect of leafy spurge in the
upper Great Plains: methods, models, and results." Impact Assessment 14: 419-434.
6-81
-------
Lipton, D. W. 1999. "Pfiesteria's economic impact on seafood industry sales and recreational
fishing, p. 35-38." In B. L. Gardner and L. Koch (ed.), Proceedings of the Conference,
Economics of Policy Options for Nutrient Management and Pfiesteria. Center for
Agricultural and Natural Resource Policy, University of Maryland, College Park.
Loehman, E.T., D. Boldt, K. Chaikin. 1985. Measuring the Benefits of Air Quality Improvements
in the San Francisco Bay Area. From Methods Development for Environmental Control
Benefits Assessment, Volume IV. Prepared for the U.S. Environmental Protection
Agency, Office of Policy, Planning and Evaluation, September. Grant #R805059-01-0
Available on the Internet at
.
Loehman, E.T., S. Park, and D. Boldt. 1994. "Willingness to Pay for Gains and Losses in Visibility
and Health." Land Economics 70(4): 478-498.
Malm, WC. 1999. Introduction to visibility. Colorado State University, Fort Collins, CO, USA,
1983, Revised edition, 1999. Available on the Internet at
.
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.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck and M. Thayer.
1993. Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method.
Prepared for U.S. Environmental Protection Agency, Office of Policy, Planning and
Evaluation. September. Available on the Internet at
.
Mclntyre, J.W., Barr, J.F. 1997. "Common Loon (Gavia immer)" in: Pool A, Gill F (eds.) The Birds
of North America. Academy of Natural Sciences, Philadelphia, PA, 313
Mclntrye, J.W., and Evers, D.C.,(eds). 2000. "Loons: old history and new finding." Proceedings of
a Symposium from the 1997 meeting, American Ornithologists' Union. North American
Loon Fund, 15 August 1997, Holderness, NH, USA.
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).
6-82
-------
Mitro, Matthew G., Evens, David C., Meyer, Michael W., and Piper, Walter H. 2008. "Common
loon survival rates and mercury in New England and Wisconsin." Journal of Wildlife
Management 72(3): 665-673.
Munthe, J., Bodaly, R. A., Branfireun, B. A., Driscoll, C. T., Gilmour, C. C., Harris, R., et al. (2007).
"Recovery of Mercury-Contaminated Fisheries." Environmental Science & Technology
36(1), 33-44.
Murdoch, J. C. and M. A. Thayer. 1988. "Hedonic price estimation of variable urban air quality,"
J. Environ. Econ. Management 15,143-146.
National Agricultural Statistics Service (NASS). 2008. Maple Syrup—June 12, 2008: Maple Syrup
Production Up 30 Percent Nationwide. U.S. Department of Agriculture, National
Agricultural Statistics Service, New England Agricultural Statistics, Concord, NH.
National Association of State Foresters (NASF). 2009. Quadrennial Fire Review
2009.Washington, DC: NASF. Quadrennial Fire and Fuel Review Final Report 2009.
National Wildfire Coordinating Group Executive Board. January.
National Oceanic and Atmospheric Administration (NOAA). 2007. Annual Commercial Landing
Statistics. August. Available on the Internet at .
National Park Service (NPS). 2008. Statistical Abstract: 2008. Department of Interior, National
Park Service Social Science Program, Public Use Statistics Office, Denver, Colorado.
Available on the Internet at
.
Neill, Helen R., David M. Hassenzahl and Djeto D. Assane. 2007. "Estimating the Effect of Air
Quality: Spatial versus Traditional Hedonic Price Models," Southern Economic Journal
73, No. 4, pp. 1088-1111.
Neumann, J.E. and R.W. Patterson. 2009. Alternative Approach to Estimating Monetized
Benefits of Residential Visibility for the Section 812 Second Prospective. Prepared for U.S.
Environmental Protection Agency OAR/OPAR. Available on the Internet at
-------
Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory Analysis. Washington,
DC. Available on the internet at .
Painter, et al. 2007. "Impact of disturbed desert soils on duration of mountain snow cover."
Geophys. Res. Lett. 34, L12502.
Parrish ZD; White JC; Isleyen M; Gent MPN; lannucci-Berger W; Eitzer BD; Kelsey JW; Mattina
Ml. 2006. "Accumulation of weathered polycyclic aromatic hydrocarbons (PAHs) by
plant and earthworm species." Chemosphere, 64: 609-618.
Patterson, R.W., J.E. Neumann, and C.G. Leggett. 2005. Recommended Residential Visibility
Values for the Section 812 Second Prospective Analysis. Prepared for U.S. Environmental
Protection Agency OPAR. Available on the Internet at
.
Pryor, S.C. 1996. "Assessing Public Perception of Visibility for Standard Setting Exercises."
Atmospheric Environment 30, no. 15, pp. 2705-2716.
Pye, J.M. 1988. "Impact of ozone on the growth and yield of trees: A review." Journal of
Environmental Quality, 17, 347-360.
Qian Y., Gustafson W.I., Leung L.R., and Ghan S.J. 2009. "Effects of soot-induced snow albedo
change on snowpack and hydrological cycle in western United States based on Weather
Research Forecasting Chemistry and Regional Climate Simulations." Journal of
Geophysical Research 114 (D02108).
Rae, D.A. 1983. Benefits of Visual Air Quality in Cincinnati Results of a Contingent Ranking
Survey. Prepared for Electric Power Research Institute. May.
Rea, A. W., Lindberg, S. E., Scherbatskoy, T., & Keeler, G. J. 2002. "Mercury Accumulation in
Foliage over Time in Two Northern Mixed-Hardwood Forests." Water, Air, & Soil
Pollution 133(1), 49-67.
6-84
-------
Rimmer, C. C, McFarland, K. P., Evens, D. C, Miller, E. K., Aubry, Y., Busby, D., et al. 2005.
"Mercury Concentrations in Bicknell's Thrush and Other Insectivorous Passerines in
Montane Forests of Northeastern North America." Ecotoxicology 14(1), 223-240.
Rosen, S. 1974. "Hedonic prices and implicit markets: product differentiation in pure
competition." Journal of Political Economy 82(1): 34-55.
Scheuhammer, Anton M., Meyer Michael W., Sandheinrich, Mark B., and Murray, Michael W.
2007. "Effects of environmental methylmercury on the health of wild birds, mammals,
and fish," Ambio Vol.36, No.l.
Sepulveda, Maria S., Frederick, Peter C., Spalding, Marilyn G., and Williams, Gary E. Jr. 1999.
"Mercury contamination in free-ranging great egret nestlings (Ardea albus) from
southern Florida, USA," Environmental Toxicology and Chemistry Vol. 18, No.5.
Simcik M.F., Eisenreich S.J., Lioy P.J. 1999. "Source apportionment and source/sink relationship
of PAHs in the coastal atmosphere of Chicago and Lake Michigan." Atmospheric
Environment, 33, 5071-5079.
Simcik M.F., Eisenreich, S.J., Golden K.A., et al. 1996. "Atmospheric Loading of Polycyclic
Aromatic Hydrocarbons to Lake Michigan as Recorded in the Sediments." Environmental
Science and Technology, 30, 3039-3046.
Sisler, J.F. 1996. Spatial and seasonal patterns and long-term variability of the composition of
the haze in the United States: an analysis of data from the IMPROVE network. CIRA
Report, ISSN 0737-5352-32, Colorado State University.
Smith, G., Coulston, J., Jepsen, E., Prichard, T. 2003. "A national ozone biomonitoring program-
results from field surveys of ozone sensitive plants in Northeastern forests (1994-
2000)." Environmental Monitoring and Assessment, 87, 271-291.
Smith, A.E.; Kemp, M.A.; Savage, T.H.; Taylor, C.L. 2005. "Methods and Results from a New
Survey of Values for Eastern Regional Haze Improvements,"/ Air & Waste Manage
AssocSS, 1767-1779.
Smith, V.K., G. Van Houtven, and S.K. Pattanayak. 2002. "Benefit Transfer via Preference
Calibration." Land Economics 78:132-152.
Smith, V.K., and J.C. Huang. 1995. "Can markets value air quality? A meta-analysis of hedonic
property value models/'/ Polit. Econ, 103, pp. 209-227.
Spencer, D.M., and D.F. Holecek. 2007. "Basic characteristics of the fall tourism market."
Tourism Management 28:491-504.
Taylor R. 1994. "Deterministic versus stochastic evaluation of the aggregate economic effects of
price support programs." Agricultural Systems 44: 461-473.
6-85
-------
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.
Tolley, G., A. Randall, G. Blomquist, M. Brien, R. Fabian, G. Fishelson, A. Frankel, M. Grenchik, J.
Hoehn, A. Kelly, R. Krumm, E. Mensah, and T. Smith. 1986. Establishing and Valuing the
Effects of Improved Visibility in Eastern United States. Prepared for U.S. Environmental
Protection Agency, Office of Policy, Planning and Evaluation. March. U.S. Environmental
Protection Agency Grant #807768-01-0. Available on the Internet at
.
Tolley, George and Robert Fabian (eds). 1988. The Economic Value of Visibility. Series : Studies
in urban and resource economics. Mount Pleasant, MO: Blackstone.
Trijonis, J.C. et al. 1987. "Preliminary extinction budget results from the RESOLVE program," pp.
872-883. In: P.J. Bhardwaja, et al. Visibility Protection Research and Policy Aspects. Air
Pollution Control Assoc., Pittsburgh, PA.
Trijonis, J.C. et al. 1988. RESOLVE Project Final Report: Visibility conditions and Causes of
Visibility Degradation in the Mojave Desert of California. NWC TP #6869. Naval Weapons
Center, China Lake, CA. Available on the Internet at
.
United Nations Environmental Programme, World Meteorological Organization and Stockholm
Environment Institute (UNEP). 2011. Integrated Assessment of Black Carbon and
Tropospheric Ozone Summary for Decision Makers. Available on the Internet at
.
U.S. Census Bureau. 2010. "Sector 56: EC075612: Administrative Support and Waste
Management and Remediation Services: Industry Series: Preliminary Comparative
Statistics for the United States (2002 NAICS Basis): 2007 and 2002." American FactFinder
Available on the Internet at http://factfinder.census.gov (accessed August 25, 2010)
U.S. Department of the Interior (DOI). 2007. 2006 National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation. U.S. Department of the Interior, Fish and Wildlife
Service, and U.S. Department of Commerce, U.S. Census Bureau. Washington, DC.
Available on the Internet at .
6-86
-------
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 1999a. 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). 1999b. Final
Advisory by the Advisory Council on Clean Air Compliance Analysis on the 1999
Prospective Study of Costs and Benefits (1999) of Implementation of the Clean Air Act
Amendments (CAAA). EPA-SAB-COUNCIL-ADV-00-003. November. Available on the
Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2004. 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). 2007. Clean Air
Scientific Advisory Committee's (CASAC) Review of the Agency's Final Ozone Staff Paper.
EPA-CASAC-07-002. Available on the Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2010a. Advisory
on a Preliminary Draft of the Second Section 812 Prospective Study of the Benefits and
Costs of the Clean Air Act (April 2010). EPA-COUNCIL-10-004. July. Available on the
Internet at
.
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2010b. 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. December.
Available on the Internet at
.
6-87
-------
U.S. Environmental Protection Agency (U.S. EPA). 1997. Mercury Study Report to Congress.
Volume V: Health Effects of Mercury and Mercury Compounds. EPA-452/R-97-007. U.S.
EPA Office of Air Quality Planning and Standards, and Office of Research and
Development. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 1998. Advisory Council on Clean Air
Compliance Analysis Advisory on the Clean Air Act Amendments (CAAA) of 1990 Section
812 Prospective Study: Overview of Air Quality and Emissions Estimates: Modeling,
Health and Ecological Valuation Issues Initial Studies. EPA-SAB-COUNCIL-ADV-98-003.
Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 1999a. Regional Haze Regulations, July 1,
1999. (64 FR 35714, July 1, 1999).
U.S. Environmental Protection Agency (U.S. EPA). 1999b. 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.
EPA410-R-99-001. Available on the Internet at .
U.S. Environmental Protection Agency (U.S. EPA). 2004. Air Quality Criteria for Particulate
Matter Volume II of II. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency, Research Triangle
Park, NC EPA/600/P-99/002bF. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2005. Regulatory Impact Analysis of the Final
Clean Air Mercury Rule. Office of Air Quality Planning and Standards, Research Triangle
Park, NC, March; EPA report no. EPA-452/R-05-003. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2006b. Regulatory Impact Analysis, 2006
National Ambient Air Quality Standards for Particulate Matter. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. October. Available on the Internet
at < http://www.epa.gov/ttn/ecas/regdata/RIAs/Appendix%20l-
Visibility%20Benefits%20Methodology.pdf>.
U.S. Environmental Protection Agency (U.S. EPA). 2006c. Ecological Benefits Assessment
Strategic Plan. EPA-240-R-06-001. Office of the Administrator. Washington, DC. October.
Available on the Internet at
.
6-88
-------
U.S. Environmental Protection Agency (U.S. EPA). 2007a. Guidance on the Use of Models and
Other Analyses for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.s, and
Regional Haze. Office of Air Quality Planning and Standards. EPA-454/B-07-002. April.
.
U.S. Environmental Protection Agency (U.S. EPA). 2007b. Review of the National Ambient Air
Quality Standards for Ozone: Policy assessment of scientific and technical information.
Staff Paper. Office of Air Quality Planning and Standards. EPA-452/R-07-007a. July.
Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2008b. U.S. EPA's 2008 Report on the
Environment (Final Report). Washington, D.C., EPA/600/R-07/045F (NTIS PB2008-
112484). May. Available on the Internet at
. Updated in
December 2010, available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2008c. 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). 2009a. The NOX Budget Trading Program:
2008 Environmental Results. Clean Air Markets Division. September. Available on the
Internet at .
6-89
-------
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—RTP Division. December. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2009c. Risk and Exposure Assessment for
Review of the Secondary National Ambient Air Quality Standards for Oxides of Nitrogen
and Oxides of Sulfur (Final Report). EPA-452/R-09-008a. 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). 2010b. Particulate Matter Urban-Focused
Visibility Assessment—Final Document (Corrected). Office of Air Quality Planning and
Standards, Research Triangle Park, NC. July. Available on the Internet 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, Washington, DC.
March. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011c. 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 on the
Internet at:
U.S. Environmental Protection Agency (U.S. EPA). 2011d. Regulatory Impact Analysis for the
Final Mercury and Air Toxics Standards. EPA-452/R-11-011. December. Available on the
Internet at .
6-90
-------
U.S. Environmental Protection Agency (U.S. EPA). 2012a. Integrated Science Assessment for
Particulate Ozone and Related Photochemical Oxidants (Third External Review Draft).
EPA/600/R-10/076C. National Center for Environmental Assessment—RTP Division.
June. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2012b. Black Carbon Report to Congress. EPA-
450/R-12-001. March. Available on the Internet at
.
U.S. Fish and Wildlife Service and U.S. Census Bureau (FWS and Census). 2007. 2006 National
Survey of Fishing, Hunting, and Wildlife-Associated Recreation (FHWAR). FHW/06-NAT.
U.S. Department of the Interior, U.S. Fish and Wildlife Service, Washington, DC, and U.S.
Department of Commerce, U.S. Census Bureau, Washington, DC.
U.S. Forest Service (USFS). 2006. Forest Inventory and Analysis National Program: Forest
Inventory Data Online. Online database. U.S. Department of Agriculture Forest Service,
Forest Inventory and Analysis, Arlington, VA. Available on the Internet at
.
Valigura, R.A., R.B. Alexander, M.S. Castro, T.P. Meyers, H.W. Paerl, P.E. Stacy, and R.E. Turner.
2001. Nitrogen Loading in Coastal Water Bodies: An Atmospheric Perspective.
Washington, DC: American Geophysical Union.
Van Sickle, J., Baker, J.P., Simonin, H.A., Baldigo, B.P., Kretser, W.A., Sharpe, W.E. 1996.
"Episodic acidification of small streams in the northeastern United States: Fish mortality
in field bioassays." Ecological Applications, 6, 408-421.
Wada, H. and Cristol, D.A. and McNabb, F.M.A. and Hopkins, W.A. 2009. "Suppressed
adrenocortical responses and thyroid hormone levels in birds near a mercury-
contaminated river." Environmental Science & Technology 43(15), 6031-6038.
Wada, H., Yates, DE., Evers, DC., Taylor, RJ., Hopkins, WA. 2010. "Tissue mercury concentrations
and adrenocortical responses of female big brown bats (Eptesicusfuscus) near a
contaminated river." Ecotoxicology 19(7), 1277-1284.
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.
Wiscombe W. and Warren S. 1980. "A model for the spectral albedo of snow. Snow containing
atmospheric aerosols."/ Atmos. Sci. 37, 2734-2274.
6-91
-------
Yates, David E., Mayack, David T., Munney, Kenneth, Evens David C, Major, Andrew, Kaur,
Taranjit, and Taylor, Robert J. 2005. "Mercury levels in mink (Mustela vison) and river
otter (Lonra canadensis) from northeastern North America," Ecotoxicology 14, 263-274.
6-92
-------
APPENDIX 6.A
ADDITIONAL DETAILS REGARDING THE VISIBILITY BENEFITS METHODOLOGY
6.A.1 Introduction
Economic benefits may result from two broad categories of changes in light extinction:
(1) changes in "residential" visibility—i.e., the visibility in and around the locations where
people live; and (2) changes in "recreational" visibility at Class I areas—i.e., visibility at Class I
national parks and wilderness areas.1 In this analysis, only those recreational and residential
benefits in areas that have been directly studied in the valuation literature are included in the
primary presentation of benefits; recreational benefits in other U.S. Class I regions and
residential benefits in other metropolitan areas are presented as sensitivity analyses of visibility
benefits.
In Chapter 6 of this RIA, we provide an overview of the visibility benefits methodology
and results. This appendix provides additional detail regarding specific aspects of the visibility
benefits methodology and is organized as follows. Section 6.A.2 describes the process we used
to convert the modeled light extinction data to match the spatial scale of the visibility benefits
assessment. We present the basic utility model in Section 6.A.3. In Section 6.A.4 we discuss the
measurement of visibility, and the mapping from environmental "bads" to environmental
"goods." In Sections 6.A.5 and 6.A.6 we summarize the methodology for estimating the
parameters of the model corresponding to visibility at in-region and out-of-region Class I areas,
and visibility in residential areas, respectively, and we describe the methods used to estimate
these parameters. Section 6.A.7 describes the process for aggregating the recreational and
residential visibility benefits. Section 6.A.8 describes the adjustment to reflect income growth
over time. Section 6.A.9 provides all the parameters used to calculate visibility benefits.
6.A.2 Converting Modeled Light Extinction Estimates
To calculate visibility benefits, we use light extinction estimates generated by the CMAQ
model.2 Modeled light extinction estimates are measured in units of inverse megameters
(Mm"1). Because the valuation studies measure visibility in terms of visual range, we convert the
light extinction units from Mm"1 to visual range (in km) for both recreational and residential
1 Hereafter referred to as Class I areas, which are defined as areas of the country such as national parks, national
wilderness areas, and national monuments that have been set aside under Section 169(a) of the Clean Air Act
to receive the most stringent degree of air quality protection. Class I federal lands fall under the jurisdiction of
three federal agencies, the National Park Service, the Fish and Wildlife Service, and the Forest Service.
2 For more information regarding the CMAQ modeling conducted for the PM NAAQS RIA, please see Chapter 3 of
this RIA.
6.A-1
-------
visibility benefits. Using the relationships derived by Pitchford and Malm (1994), the formulas
for this conversion are
Deciviews = 10 * In ( —) = 10 * In ( —
where VR denotes visual range (in kilometers) and 3ext denotes light extinction (in Mm"1).
Because we leverage the tools and data prepared for previous analyses (U.S. EPA, 2011), we use
a two-step process to convert from Mm"1 to VR using deciviews as an intermediate conversion
instead of converting directly. Therefore, the full formula incorporating the two-step
conversion is
VR = 391 * e-.*
The spatial scale of the modeled light extinction estimates must also be adjusted to
correspond with the design of the valuation studies and the underlying population and
economic data. For the residential visibility benefits analysis, we convert the spatial resolution
of the light extinction estimates from 12-km grid to county-level. We use county-level light
extinction to match the MSA boundaries, population data, and household income data. We
used the geographic centroids of each 12-km grid cell with the Veronoi Neighborhood
Averaging (VNA) interpolation method in the BenMAP model for this conversion (Abt
Associates, 2010).
For the recreational visibility benefits analysis, we use the light extinction estimates
from 12-km grid cell located at the geographic center of the Class I area. Although we
considered using the IMPROVE monitor location instead, we selected the park centroid for
three reasons:
1. Consistency with previous method for estimating recreational visibility benefits
2. Not all Class I areas have monitors, and shared monitors may be outside park
3. Siting criteria for IMPROVE monitors do not include iconic scenic vista location
6.A.3 Basic Utility Model
Within the category of recreational visibility, further distinctions have been made. There
is evidence (Chestnut and Rowe, 1990) that an individual's WTP for improvements in visibility at
a Class I area is influenced by whether it is in the region in which the individual lives, or whether
it is somewhere else. In general, people appear to be willing to pay more for visibility
6.A-2
-------
improvements at parks and wilderness areas that are "in-region" than at those that are "out-of-
region." This is plausible, because people are more likely to visit, be familiar with, and care
about parks and wilderness areas in their own part of the country.
To value estimated changes in visibility, we use an approach that is consistent with
economic theory. Below we discuss an application of the Constant Elasticity of Substitution
(CES) utility function approach3 to value both residential visibility improvements and visibility
improvements at Class I areas in the United States. This approach is based on the preference
calibration method developed by Smith, Van Houtven, and Pattanayak (2002).
We begin with a CES utility function in which a household derives utility from
1. "all consumption goods," X,
2. visibility in the residential area in which the household is located ("residential
visibility"),4
3. visibility at Class I areas in the same region as the household ("in-region recreational
visibility"), and
4. visibility at Class I areas outside the household's region ("out-of-region recreational
visibility").
We have specified a total of six recreational visibility regions,5 so there are five regions
for which any household is out of region. The utility function of a household in the nth
residential area and the ;th region of the country is:
u,,, -- ( x> + az; + r*Ql +
k=\ j*i k=\
3 The constant elasticity of substitution utility function has been chosen for use in this analysis because of its
flexibility when illustrating the degree of substitutability present in various economic relationships (in this case,
the trade-off between income and improvements in visibility).
4 We remind the reader that, although residential and recreational visibility benefits estimation is discussed
simultaneously in this section, benefits are calculated and presented separately for each visibility category.
5 See Section 6.3.4 of this RIA for a description of the different recreational visibility considered in this analysis.
6.A-3
-------
where
Zn = the level of visibility in the nth residential area;
Qik = the level of visibility at the kth in-region park (i.e., the kth park in the ith region);
Qjk = the level of visibility at the kth park in the jth region (for which the household is
out of region), jVi;
Nj = the number of Class I areas in the ith region;
Nj = the number of Class I areas in the jth region (for which the household is out of
region), jVi; and
0, the y's and 6's are parameters of the utility function corresponding to the visibility
levels at residential areas, and at in-region and out-of-region Class I areas,
respectively.
In particular, the yik's are the parameters corresponding to visibility at in-region Class I
areas; the Si's are the parameters corresponding to visibility at Class I areas in region 1
(California), if i^l; the 62's are the parameters corresponding to visibility at Class I areas in
region 2 (Colorado Plateau), if i^2, and so forth. Because the model assumes that the
relationship between residential visibility and utility is the same everywhere, there is only one
0. The parameter p in this CES utility function is an important determinant of the slope of the
marginal WTP curve associated with any of the environmental quality variables. When p=l, the
marginal WTP curve is horizontal. When p
-------
Given estimates of p, 0, the v's and the 6's, the household's utility function and the
corresponding WTP functions are fully specified. The household's WTP for any set of changes in
the levels of visibility at in-region Class I areas, out-of-region Class I areas, and the household's
residential area can be shown to be:
rlk(Q0Plk ~
The household's WTP for a single visibility improvement will depend on its order in the
series of visibility improvements the household is valuing. If it is the first visibility improvement
to be valued, the household's WTP for it follows directly from the previous equation. For
example, the household's WTP for an improvement in visibility at the first in-region park, from
Qn = Qoii to Qn = Qm, is
if this is the first (or only) visibility change the household values.
6.A.4 Measure of Visibility: Environmental "Goods" Versus "Bads"
In the above model, Qand Z are environmental "goods." As the level of visibility
increases, utility increases. The utility function and the corresponding WTP function both have
reasonable properties. The first derivative of the indirect utility function with respect to Q (orZ)
is positive; the second derivative is negative. WTP for a change from Qo to a higher (improved)
level of visibility, Qi, is therefore a concave function of Qi, with decreasing marginal WTP.
The measure of visibility that is currently preferred by air quality scientists is the
deciview, which increases as visibility decreases. Deciview, in effect, is a measure of the lack of
visibility. As deciviews increase, visibility, and therefore utility, decreases. The deciview, then, is
a measure of an environmental "bad." There are many examples of environmental "bads"— all
types of pollution are environmental "bads." Utility decreases, for example, as the
concentration of particulate matter in the atmosphere increases.
One way to value decreases in environmental bads is to consider the "goods" with
which they are associated, and to incorporate those goods into the utility function. In
particular, if B denotes an environmental "bad," such that:
6.A-5
-------
and the environmental "good," Q, is a function of B,
=
then the environmental "bad" can be related to utility via the corresponding environmental
"good":6
V= V(m,Q)= V(m,F(By) .
The relationship between Q and B, F(B), is an empirical relationship that must be estimated.
There is a potential problem with this approach, however. If the function relating B and
Q is not the same everywhere (i.e., if for a given value of B, the value of Q depends on other
factors as well), then there can be more than one value of the environmental good
corresponding to any given value of the environmental bad, and it is not clear which value to
use. This has been identified as a problem with translating deciviews (an environmental "bad")
into visual range (an environmental "good"). It has been noted that, for a given deciview value,
there can be many different visual ranges, depending on the other factors that affect visual
range— such as light angle and altitude. We note here, however, that this problem is not unique
to visibility, but is a general problem when trying to translate environmental "bads" into
"goods."7
In order to translate deciviews (a "bad") into visual range (a "good"), we use a
relationship derived by Pitchford and Malm (1994) in which
397
DV= 10 * ln( - ,) ,
' VR'
where DV denotes deciview and VR denotes visual range (in kilometers). Solving for VR as a
function of DV yields
6 There may be more than one "good" related to a given environmental "bad." To simplify the discussion, however,
we assume only a single "good."
7 Another example of an environmental "bad" is particulate matter air pollution (PM). The relationship between
survival probability (Q) and the ambient PM level is generally taken to be of the form
2=1- OK . where V denotes the mortality rate (or level) when there is no ambient PM (i.e., when
PM=0). However, a is implicitly a function of all the factors other than PM that affect mortality. As these factors
change (e.g., from one location to another), awill change (just as visual range changes as light angle changes). It
is therefore possible to have many values of Q corresponding to a given value of PM, as the values of V vary.
6.A-6
-------
VR= 391*e-°JDV .
This conversion is based on specific assumptions characterizing the "average" conditions
of those factors, such as light angle, that affect visual range. To the extent that specific
locations depart from the average conditions, the relationship will be an imperfect
approximation.8
6.A.5 Estimating the Parameters for Visibility at Class I Areas: the y's and 6's
As noted in Section 6.A.3, if we consider a particular visibility change as the first or the
only visibility change valued by the household, the household's WTP for that change in visibility
can be calculated, given income (m), the "shape" parameter, p, and the corresponding
recreational visibility parameter. For example, a Southeast household's WTP for a change in
visibility at in-region parks (collectively) from Qi = Qoi to Qi = Qn is:
WTP(DQJ = m- [mr + gl(Q'01 - Qrn)]1/r
if this is the first (or only) visibility change the household values.
Alternatively, if we have estimates of m as well as WTPi"1 and WTPi°ut of in-region and
out-of-region households, respectively, for a given change in visibility from Q0i to Qn in
Southeast parks, we can solve for YI and 61 as a function of our estimates of m, WTPi"1 and
WTPi°ut, for any given value of p. Generalizing, we can derive the values of y and 6 for the jth
region as follows:
(m- WTPY - mp
and
(m -WTP°ut)p-mp
Chestnut and Rowe (1990) and Chestnut (1997) estimated WTP (per household) for
specific visibility changes at national parks in three regions of the United States— both for
households that are in-region (in the same region as the park) and for households that are out-
' Ideally, we would want the location-, time-, and meteorological condition-specific relationships between
deciviews and visual range, which could be applied as appropriate. This is probably not feasible, however.
6.A-7
-------
of-region. The Chestnut and Rowe study asked study subjects what they would be willing to pay
for each of three visibility improvements in the national parks in a given region. Study subjects
were shown a map of the region, with dots indicating the locations of the parks in question. The
WTP questions referred to the three visibility improvements in all the parks collectively; the
survey did not ask subjects' WTP for these improvements in specific parks individually.
Responses were categorized according to whether the respondents lived in the same region as
the parks in question ("in-region" respondents) or in a different region ("out-of-region"
respondents). The areas for which in-region and out-of-region WTP estimates are available
from Chestnut and Rowe (1990), and the sources of benefits transfer-based estimates that we
employ in the absence of estimates, are summarized in Table 6.A-1. In all cases, WTP refers to
WTP per household.
Table 6.A-1. Available Information on WTP for Visibility Improvements in National Parks
Region of Household
Region of Park In Region3 Out of Regionb
1. California WTP estimate from study WTP estimate from study
2. Colorado Plateau WTP estimate from study WTP estimate from study
3. Southeast United States WTP estimate from study WTP estimate from study
4. Northwest United States (based on benefits transfer from California)
5. Northern Rockies (based on benefits transfer from Colorado Plateau)
6. Rest of United States (based on benefits transfer from Southeast U.S.)
3 In-region" WTP is WTP for a visibility improvement in a park in the same region as that in which the household is
located. For example, in-region WTP in the "Southeast" row is the estimate of the average Southeast
household's WTP for a visibility improvement in a Southeast park.
b Out-of-region" WTP is WTP for a visibility improvement in a park that is not in the same region in which the
household is located. For example, out-of-region WTP in the "Southeast" row is the estimate of WTP for a
visibility improvement in a park in the Southeast by a household outside of the Southeast.
In the primary calculation of visibility benefits for this analysis, only visibility changes at
parks within visibility regions for which a WTP estimate was available from Chestnut and Rowe
(1990) are considered (for both in- and out-of-region benefits). Primary estimates will not
include visibility benefits calculated by transferring WTP values to visibility changes at parks not
included in the Chestnut and Rowe study. Transferred benefits at parks located outside of the
Chestnut and Rowe visibility regions will, however, be included as an alternative calculation.
6.A-8
-------
The values of the parameters in a household's utility function will depend on where the
household is located. The region-specific parameters associated with visibility at Class I areas
(that is, all parameters except the residential visibility parameter) are arrayed in Table 6.A-2.
The parameters in columns 1 through 3 can be directly estimated using WTP estimates from
Chestnut and Rowe (1990) (the columns labeled "Region 1," "Region 2," and "Region 3").
Table 6.A-2. Summary of Region-Specific Recreational Visibility Parameters to be Estimated in
Household Utility Functions
Region of Household
Region 1
Region 2
Region 3
Region 4
Region 5
Region 6
Region 1
via
6!
5,
5,
6!
5,
Region
Region 2 Region 3
62 63
V2 63
62 V3
62 63
62 63
62 63
of Park
Region 4
64
64
64
V4
64
64
Region 5
65
65
65
65
Vs
65
Region 6
66
66
66
66
66
Ve
a The parameters arrayed in this table are region-specific rather than park-specific or wilderness area-specific. For
example, 6! is the parameter associated with visibility at "Class I areas in region 1" for a household in any region
other than region 1. The benefits analysis must derive Class I area-specific parameters (e.g., 6lk, for the k Class I
area in the first region).
For the three regions covered in Chestnut and Rowe (1990a) (California, the Colorado
Plateau, and the Southeast United States), we can directly use the in-region WTP estimates
from the study to estimate the parameters in the utility functions corresponding to visibility at
in-region parks (YI); similarly, we can directly use the out-of-region WTP estimates from the
study to estimate the parameters for out-of-region parks (61). For the other three regions not
covered in the study, however, we must rely on benefits transfer to estimate the necessary
parameters.
While Chestnut and Rowe (1990) provide useful information on households' WTP for
visibility improvements in national parks, there are several significant gaps remaining between
the information provided in that study and the information necessary for the benefits analysis.
First, as noted above, the WTP responses were not park specific, but only region specific.
Because visibility improvements vary from one park in a region to another, the benefits analysis
must value park-specific visibility changes. Second, not all Class I areas in each of the three
regions considered in the study were included on the maps shown to study subjects. Because
6.A-9
-------
the focus of the study was primarily national parks, most Class I wilderness areas were not
included. Third, only three regions of the United States were included, leaving the three
remaining regions without direct WTP estimates.
In addition, Chestnut and Rowe (1990) elicited WTP responses for three different
visibility changes, rather than a single change. In theory, if the CES utility function accurately
describes household preferences, and if all households in a region have the same preference
structure, then households' three WTP responses corresponding to the three different visibility
changes should all produce the same value of the associated recreational visibility parameter,
given a value of p and an income, m. In practice, of course, this is not the case.
In addressing these issues, we take a three-phase approach:
1. We estimate region-specific parameters for the region in the modeled domain
covered by Chestnut and Rowe (1990a) (California, the Colorado Plateau, and the
Southeast)—YI, y2, and y3 and 61, 62, and 63.
2. We infer region-specific parameters for those regions not covered by the Chestnut
and Rowe study (the Northwest United States, the Northern Rockies, and the rest of
the U.S.)—v4, Y5, and Ye and 64, 65, and 66.
3. We derive park- and wilderness area-specific parameters within each region (YIR and
6ik, for k=l, ..., NI; Y2k and 62i<, for k=l, ..., N2; and so forth).
The question that must be addressed in the first phase is how to estimate a single
region-specific in-region parameter and a single region-specific out-of-region parameter for
each of the three regions covered in Chestnut and Rowe (1990) from study respondents' WTPs
for three different visibility changes in each region. All parks in a region are treated collectively
as if they were a single "regional park" in this first phase. In the second phase, we infer region-
specific recreational visibility parameters for regions not covered in the Chestnut and Rowe
study (the Northwest United States, the Northern Rockies, and the rest of the United States). As
in the first phase, we ignore the necessity to derive park-specific parameters at this phase.
Finally, in the third phase, we derive park- and wilderness area-specific parameters for each
region.
6.A.5.1 Estimating Region-Specific Recreational Visibility Parameters for the Region Covered
in the Chestnut and Rowe Study (Regions 1, 2, and 3)
Given a value of p and estimates of m and in-region and out-of-region WTPs for a
change from Q0 to Qi in a given region, the in-region parameter, Y, and the out-of-region
parameter, 6, for that region can be solved for. Chestnut and Rowe (1990), however,
6.A-10
-------
considered not just one, but three visibility changes in each region, each of which results in a
different calibrated y and a different calibrated 6, even though in theory all the y's should be
the same and similarly, all the 6's should be the same. For each region, however, we must have
only a single y and a single 6.
Denoting Y j as our estimate of y for the jth region, based on all three visibility changes,
we chose y j to best predict the three WTPs observed in the study for the three visibility
improvements in the jth region. First, we calculated YJI, i=l, 2, 3, corresponding to each of the
three visibility improvements considered in the study. Then, using a grid search method
beginning at the average of the three's f ji, we chose to minimize the sum of the squared
differences between the WTPs we predict using YJ and the three region-specific WTPs
observed in the study. That is, we selected to minimize:
;=1
where WTPy and WTPjjQ are the observed and the predicted WTPs for a change in visibility in
the jth region from Q0 = Qoi to Qi= On, i=l, ..., 3. An analogous procedure was used to select an
optimal 6, for each of the three regions in the Chestnut and Rowe study.
6.A.5.2 Inferring Region-Specific Recreational Visibility Parameters for Regions Not Covered
in the Chestnut and Rowe Study (Regions 4, 5, and 6)
One possible approach to estimating region-specific parameters for regions not covered
by Chestnut and Rowe (1990a) (y4, y5, and y6 and 64, 65, and 66) is to simply assume that
households' utility functions are the same everywhere, and that the environmental goods being
valued are the same—e.g., that a change in visibility at national parks in California is the same
environmental good to a Californian as a change in visibility at national parks in Minnesota is to
a Minnesotan.
For example, to estimate 64 in the utility function of a California household,
corresponding to visibility at national parks in the Northwest United States, we might assume
that out-of-region WTP for a given visibility change at national parks in the Northwest United
States is the same as out-of-region WTP for the same visibility change at national parks in
California (income held constant). Suppose, for example, that we have an estimated mean WTP
of out-of-region households for a visibility change from Q0i to On at national parks in California
(region 1), denoted WTPi°ut. Suppose the mean income of the out-of-region subjects in the
6.A-11
-------
study was m. We might assume that, for the same change in visibility at national parks in the
Northwest United States, WTP4out = WTPi°ut among out-of-region individuals with income m.
We could then derive the value of 64, given a value of p as follows:
(m- WTP°utY -mp
64 ~ (?£ - Qu
where Q04 = Qoi and Qi4 = Qn, (i.e., where it is the same visibility change in parks in region 4
that was valued at parks in the region 1).
This benefits transfer method assumes that (1) all households have the same preference
structures and (2) what is being valued in the Northwest United States (by a California
household) is the same as what is being valued in the California (by all out-of-region
households). While we cannot know the extent to which the first assumption approximates
reality, the second assumption is clearly problematic. National parks in one region are likely to
differ from national parks in another region in both quality and quantity (i.e., number of parks).
One statistic that is likely to reflect both the quality and quantity of national parks in a
region is the average annual visitation rate to the parks in that region. A reasonable way to
gauge the extent to which out-of-region people would be willing to pay for visibility changes in
parks in the Northwest United States versus in California might be to compare visitation rates in
the two regions.9 Suppose, for example, that twice as many visitor-days are spent in California
parks per year as in parks in the Northwest United States per year. This could be an indication
that the parks in California are in some way more desirable than those in the Northwest United
States and/or that there are more of them—i.e., that the environmental goods being valued in
the two regions ("visibility at national parks") are not the same.
A preferable way to estimate 64, then, might be to assume the following relationship:
WTPout n
VY -L -L ^ /J^
°ut
WTP
(income held constant), where HI = the average annual number of visitor-days to California
parks and n4 = the average annual number of visitor-days to parks in the Northwest United
States. This implies that
9 We acknowledge that reliance on visitation rates does not get at nonuse value.
6.A-12
-------
for the same change in visibility in region 4 parks among out-of-region individuals with income
m. If, for example, ni = 2n4, WTP4out would be half of WTPi°ut. The interpretation would be the
following: California national parks have twice as many visitor-days per year as national parks in
the Northwest United States; therefore they must be twice as desirable/plentiful; therefore,
out-of-region people would be willing to pay twice as much for visibility changes in California
parks as in parks in the Northwest United States; therefore a Californian would be willing to pay
only half as much for a visibility change in national parks in the Northwest United States as an
out-of-region individual would be willing to pay for the same visibility change in national parks
in California. This adjustment, then, is based on the premise that the environmental goods
being valued (by people out of region) are not the same in all regions.
The parameter 64 is estimated as shown above, using this adjusted WTP4out. The same
procedure is used to estimate 65 and 66. We estimate y4; y, and y6 in an analogous way, using
the in-region WTP estimates from the transfer regions, e.g.,
WTP'" = ^-*WTP'" .
ni
6.A.5.3 Estimating Park- and Wilderness Area-Specific Parameters
As noted above, Chestnut and Rowe (1990) estimated WTP for a region's national parks
collectively, rather than providing park-specific WTP estimates. The [s]and [s]are therefore
the parameters that would be in household utility functions if there were only a single park in
each region, or if the many parks in a region were effectively indistinguishable from one
another. Also noted above is the fact that the Chestnut and Rowe study did not include all Class
I areas in the regions it covered, focusing primarily on national parks rather than wilderness
areas. Most Class I wilderness areas were not represented on the maps shown to study
subjects. In California, for example, there are 31 Class I areas, including 6 national parks and 25
wilderness areas. The Chestnut and Rowe study map of California included only 10 of these
Class I areas, including all 6 of the national parks. It is unclear whether subjects had in mind "all
parks and wilderness areas" when they offered their WTPs for visibility improvements, or
whether they had in mind the specific number of (mostly) parks that were shown on the maps.
The derivation of park- and wilderness area-specific parameters depends on this.
6.A-13
-------
6.A.5.4 Derivation of Region-Specific WTP for National Parks and Wilderness Areas
If study subjects were lumping all Class I areas together in their minds when giving their
WTP responses, then it would be reasonable to allocate that WTP among the specific parks and
wilderness areas in the region to derive park- and wilderness area-specific v's and 6's for the
region. If, on the other hand, study subjects were thinking only of the (mostly) parks shown on
the map when they gave their WTP response, then there are two possible approaches that
could be taken. One approach assumes that households would be willing to pay some
additional amount for the same visibility improvement in additional Class I areas that were not
shown, and that this additional amount can be estimated using the same benefits transfer
approach used to estimate region-specific WTPs in regions not covered by Chestnut and Rowe
(1990a).
However, even if we believe that households would be willing to pay some additional
amount for the same visibility improvement in additional Class I areas that were not shown, it is
open to question whether this additional amount can be estimated using benefits transfer
methods. A third possibility, then, is to simply omit wilderness areas from the benefits analysis.
For this analysis we calculate visibility benefits assuming that study subjects lumped all Class I
areas together when stating their WTP, even if these Class I areas were not present on the map.
6.A.5.5 Derivation of Park- and Wilderness Area-Specific WTPs, Given Region-Specific WTPs
for National Parks and Wilderness Areas
The first step in deriving park- and wilderness area-specific parameters is the estimation
of park- and wilderness area-specific WTPs. To derive park and wilderness area-specific WTPs,
we apportion the region-specific WTP to the specific Class I areas in the region according to
each area's share of the region's visitor-days. For example, if WTPi"1 and WTPi°ut denote the
mean household WTPs in the Chestnut and Rowe (1990) study among respondents who were
in-region-1 and out-of-region-1, respectively, nik denotes the annual average number of visitor-
days to the kth Class I area in California, and ni denotes the annual average number of visitor-
days to all Class I areas in California (that are included in the benefits analysis), then we assume
that
* WTP™ ,
and
6.A-14
-------
Yl
k
lk
Using WTPjin and WTPj°ut, either from the Chestnut and Rowe study (for j = 1, 2, and 3) or
derived by the benefits transfer method (for j = 4, 5, and 6), the same method is used to derive
Class I area-specific WTPs in each of the six regions.
While this is not a perfect allocation scheme, it is a reasonable scheme, given the
limitations of data. Visitors to national parks in the United States are not all from the United
States, and certainly not all from the region in which the park is located. A very large proportion
of the visitors to Yosemite National Park in California, for example, may come from outside the
United States. The above allocation scheme implicitly assumes that the relative frequencies of
visits to the parks in a region/rom everyone in the world is a reasonable index of the relative
WTP of an average household in that region (WTPj'n) or out of that region (but in the United
States) (WTPj°ut) for visibility improvements at these parks.10
A possible problem with this allocation scheme is that the relative frequency of visits is
an indicator of use value but not necessarily of nonuse value, which may be a substantial
component of the household's total WTP for a visibility improvement at Class I areas. If park A
is twice as popular (i.e., has twice as many visitors per year) as park B, this does not necessarily
imply that a household's WTP for an improvement in visibility at park A is twice its WTP for the
same improvement at park B. Although an allocation scheme based on relative visitation
frequencies has some obvious problems, however, it is still probably the best way to allocate a
collective WTP.
6.A.5.6 Derivation of Park- and Wilderness Area-Specific Parameters, Given Park- and
Wilderness-Specific WTP
Once the Class I area-specific WTPs have been estimated, we could derive the park- and
wilderness area-specific y's and 6's using the method used to derive region-specific y's and 6's.
Recall that that method involved (1) calibrating y and 6 to each of the three visibility
improvements in the Chestnut and Rowe study (producing three y's and three 6's),
(2) averaging the three y's and averaging the three 6's, and finally, (3) using these average y and
6 as starting points for a grid search to find the optimal y and the optimal 6— i.e., the y and 6
10 This might be thought of as two assumptions: (1) that the relative frequencies of visits to the parks in a region
from everyone in the world is a reasonable representation of the relative frequency of visits from people in the
United States—i.e., that the parks that are most popular (receive the most visitors per year) in general are also
the most popular among Americans; and (2) that the relative frequency with which Americans visit each of their
parks is a good index of their relative WTPs for visibility improvements at these parks.
6.A-15
-------
that would allow us to reproduce, as closely as possible, the three in-region and three out-of-
region WTPs in the study for the three visibility changes being valued.
Going through this procedure for each national park and each wilderness area
separately would be very time consuming, however. We therefore used a simpler approach,
which produces very close approximations to the y's and 6's produced using the above
approach. If:
WTPj"1 = the in-region WTP for the change in visibility from Qo to Qi in the jth region;
WTPjk"1 = the in-region WTP for the same visibility change (from Q0 to Qi) in the kth
Class I area in the jth region (= Sjk*WTPj'n, where Sjk is the kth area's share of
visitor-days in the jth region);
m = income;
Yj* = the optimal value of v for the jth region; and
Yjk = the value of Yjk calibrated to WTPjk"1 and the change from Q0 to Qi;
then11:
(m- WTPfY - mp
7j *
and
(m- WTPmkY - m
r,k =
(<2op-0n
which implies that:
ajk *
where:
(m- WTPmkY - mp
a, =
(m- WTPmy - m
11 Vj* is only approximately equal to the right-hand side because, although it is the optimal value designed to
reproduce as closely as possible all three of the WTPs corresponding to the three visibility changes in the
Chestnut and Rowe study, Vj* will n°t exactly reproduce any of these WTPs.
6.A-16
-------
We use the adjustment factor, ajk, to derive YJR from YJ*, for the kth Class I area in the jth
region. We use an analogous procedure to derive 6jk from 6j* for the kth Class I area in the jth
region (where, in this case, we use WTPj°ut and WTPjkout instead of WTPjin and WTPjkin).12
6.A.6 Estimating the Parameter for Visibility in Residential Areas: 6
In previous assessments, EPA used a study on residential visibility valuation conducted
in 1990 (McClelland et al., 1993). Consistent with advice from EPA's Science Advisory Board
(SAB), EPA designated the McClelland et al. study as significantly less reliable for regulatory
benefit-cost analysis, although it does provide useful estimates on the order of magnitude of
residential visibility benefits (U.S. EPA-SAB, 1999). 13 In order to estimate residential visibility
benefits in this analysis, we have replaced the previous methodology with a new benefits
transfer approach and incorporated additional valuation studies. This new approach was
developed for The Benefits and Costs of the Clean Air Act 1990 to 2020: EPA Report to Congress
(U.S. EPA, 2011) and reviewed by the SAB (U. S. EPA-SAB, 2010). To value residential visibility
improvements, the new approach draws upon information from the Brookshire et al. (1979),
Loehman et al. (1985) and Tolley et al. (1984) studies.14 These studies provide primary visibility
values for Atlanta, Boston, Chicago, Denver, Los Angeles, Mobile, San Francisco, and
Washington D.C.15
The estimation of 0 is a simpler procedure for residential visibility benefits, involving a
straightforward calibration using the study income and WTP for each study city:
p
(m - WTPy - m
^ '
12 This method uses a single in-region WTP and a single out-of-region WTP per region. Although the choice of WTP
will affect the resulting adjustment factors (the ajk's) and therefore the resulting vjk's and 6jk's, the effect is
negligible. We confirmed this by using each of the three in-region WTPs in California and comparing the
resulting three sets of Vjk's ar|d 6jk's, which were different from each other by about one one-hundredth of a
percent.
13 EPA's Advisory Council on Clean Air Compliance Analysis noted that the McClelland et al. (1993) study may not
incorporate two potentially important adjustments. First, their study does not account for the "warm glow"
effect, in which respondents may provide higher willingness to pay estimates simply because they favor "good
causes" such as environmental improvement. Second, while the study accounts for non-response bias, it may
not employ the best available methods. As a result of these concerns, the Council recommended that
residential visibility be omitted from the overall primary benefits estimate. (U.S. EPA-SAB, 1999)
14 Loehman et al. (1985) and Brookshire et al. (1979) were subsequently published in peer-reviewed journals (see
Loehman et al. (1994) and Brookshire et al. (1982). The Tolley et al. (1984) work was not published, but was
subject to peer review during study development.
15 Recognizing potential fundamental issues associated with data collected in Cincinnati and Miami (e.g., see
Chestnut et al. (1986) and Chestnut and Rowe (1990c), we do not include values for these cities in our analysis.
6.A-17
-------
where:
m = household income,
p = shape parameter (0.1),
0 = WTP parameter corresponding to the visibility at MSA,
Z0 = starting visibility, and
Zi = visibility after change.
Where studies provide multiple estimates for visual range improvements for a single
study city, we estimate one 0 as the simple average of the 0 calculated for each set of visual
range improvements.
6.A.7 Putting It All Together: The Household Utility and WTP Functions
Given an estimate of 0, derived as shown in Section 6.A.4, and estimates of the y's and
6's, derived as shown in Section 6.A.3, based on an assumed or estimated value of p, the utility
and WTP functions for a household in any region are fully specified. We could therefore
estimate the value to that household of visibility changes from any baseline level to any
alternative level in the household's residential area and/or at any or all of the Class I areas in
the United States, in a way that is consistent with economic theory. In particular, the WTP of a
household in the ith region and the nth residential area for any set of changes in the levels of
visibility at in-region Class I areas, out-of-region Class I areas, and the household's residential
area is:
-QfJ
The national benefits associated with any suite of visibility changes would be calculated
as the sum of these household WTPs for those changes. The benefit of any subset of visibility
changes (e.g., changes in visibility only at Class I areas in California) can be calculated by setting
all the other components of the WTP function to zero (that is, by assuming that all other
visibility changes that are not of interest are zero). This is effectively the same as assuming that
the subset of visibility changes of interest is the first or the only set of changes being valued by
households. Estimating benefit components in this way will yield slightly upward biased
estimates of benefits, because disposable income, m, is not being reduced by the WTPs for any
prior visibility improvements. That is, each visibility improvement (e.g., visibility at Class I areas
in the California) is assumed to be the first, and they cannot all be the first. The upward bias
6.A-18
-------
should be extremely small, however, because all of the WTPs for visibility changes are very
small relative to income.
Although we recognize that the approach described above is most consistent with
economic theory, we have chosen to not use this function with income constraints on overall
WTP. Instead, we simply add the total preference calibrated recreational visibility benefits to
the preference-calibrated residential visibility benefits. Again, because all of the WTPs for
visibility changes are very small relative to income, the upward bias should be extremely small.
6.A.8 Income Elasticity and Income Growth Adjustment for Visibility Benefits
Growth in real income over time is an important component of benefits analysis.
Economic theory argues that WTP for most goods (such as environmental protection) will
increase if real incomes increase. There is substantial empirical evidence that the income
elasticity16 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 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. Empirical
evidence of the effect of real income on WTP gathered to date is based on studies examining
the former. The Environmental Economics Advisory Committee (EEAC) of the Science Advisory
Board (SAB) advised 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,
2000a). A recent advisory by another committee associated with the SAB, the Advisory Council
on Clean Air Compliance Analysis, 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 (U.S. EPA-SAB, 2004)" and that "The same increase should be assumed
for the WTP for serious nonfatal health effects (U.S. EPA-SAB, 2004)," they note that "given the
16 Income elasticity is a common economic measure equal to the percentage change in WTP for a 1% change in
income.
6.A-19
-------
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, 2004)." Until these conflicting advisories have been reconciled, 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.
We assume that the WTP for improved visibility would increase with growth in real
income. 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.
Details of the general procedure to account for projected growth in real U.S. income
between 1990 and 2020 can be found in Kleckner and Neumann (1999). Specifically, we use the
elasticity for visibility benefits provided in Chestnut (1997).
In addition to elasticity estimates, projections of real gross domestic product (GDP) and
populations from 1990 to 2020 are needed to adjust benefits to reflect real per capita income
growth. We used projections of real GDP provided in Kleckner and Neumann (1999) for the
years 1990 to 2010.17 We used projections of real GDP provided by Standard and Poor's (2000)
for the years 2010 to 2020.18 Visibility benefits are adjusted by multiplying the unadjusted
benefits by the appropriate adjustment factor.
6.A.9 Summary of Parameters
In Tables 6.A-3 through 6.A-6, we provide the parameters used to calculate recreational
and residential visibility benefits.
17 U.S. Bureau of Economic Analysis, Table 2A (available at http://www.bea.doc.gov/bea/dn/0897nip2/ tab2a.htm.
and U.S. Bureau of Economic Analysis, Economics and Budget Outlook. Note that projections for 2007 to 2010
are based on average GDP growth rates between 1999 and 2007.
18 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.
6.A-20
-------
Table 6.A-3. Mean Annual Household WTP for Changes in Visual Range for Recreational
Visibility (1990$)a
Region
California
Southwest
Southeast
WTP
In-region
$66.41
$80.19
$71.42
$50.12
$72.67
$61.40
$66.41
$82.70
$75.18
WTP
Out-of-region
$43.85
$53.88
$51.37
$45.11
$55.13
$48.87
$35.08
$53.88
$47.61
Starting Visual
Range (miles)
90
90
45
155
155
115
25
25
10
Ending Visual
Range (miles)
125
150
90
200
250
155
50
75
25
Study Household
Income
$48,759
$48,759
$48,759
a Based on Chestnut (1997) and adjusted for study sample income and currency year
Table 6.A-4. Region-Specific Parameters for Recreational Visibility Benefits3
Region Optimal y Optimal 6
California 0.00517633 0.003629603
Southwest 0.006402706 0.005092572
Southeast 0.003552379 0.002163346
Northwest 0.001172669 0.000823398
Northern Rockies 0.005263445 0.004176339
Rest of U.S. 0.001211215 0.000738149
a Calculated using methodology described in sections 6.A.3 through 6.A.4
6.A-21
-------
Table 6.A-5. Mean Annual Household WTP for Changes in Visual Range for Residential
Visibility
City
Atlanta
(Tolley et al., 1984)
Boston
(Tolley et al., 1984)
Chicago
(Tolley et al., 1984)
Denver
(Tolley et al., 1984)
Los Angeles
(Brookshire et al.,
1979)
Mobile
(Tolley et al., 1984)
San Francisco
(Loehman et al., 1985)
Washington, DC
(Tolley et al., 1984)
WTP in
Original
Year's $
$188
$281
$82
$119
$139
$171
$202
$269
$121
$144
$115
$154
$43
$116
$71
$168
$197
$71
$238
$303
Starting
Visual
Range
(miles)
12
12
12
12
18
18
9
9
10
10
50
50
2
2
12
10
10
16.3
15
15
Ending
Visual
Range
(miles)
22
32
22
32
28
38
18
30
20
30
60
70
12
28
28
20
30
18.6
25
35
Study
Household
Income
$19,900a
$19,900a
$27,600d
$27,600d
$25,000a
$25,000a
$30,000b
$30,000b
$29,400d
$29,400d
$32,000°
$32,000°
$15,200d
$15,200d
$15,200d
$20,200a
$20,200a
$26,100°
$27,500a
$27,500a
Year of
Original
Data
1982
1982
1984
1984
1982
1982
1981
1981
1984
1984
1984
1984
1978
1978
1978
1982
1982
1980
1982
1982
0 if p = 0.1
(1990$,
1990
income)
0.033446
0.031661
0.010738
0.009417
0.026636
0.019049
0.022313
0.016696
0.013180
0.009732
0.038558
0.027803
0.003866
0.006716
0.011702
0.026078
0.018882
0.045307
0.036866
0.027804
0ifp =
0.1
(Simple
Average)
0021316
0 022843
0 015480
0033181
0.007428
0 022480
0 045307
0 032335
a See Chestnut et al. (1986), pages 5-5 through 5-10.
b See Tolley et al., (1984), page 127.
° See Loehman et al. (1985), page 38.
d Historical median income data by MSA from U.S. Census (1990).
6.A-22
-------
Table 6.A-6. Parameters for Income Growth Adjustment for Visibility Benefits
Adjustment Step Parameter Estimate
Central Estimate of Elasticitya 0.90
Adjustment Factor Used to Account for Projected Real Income Growth in 2020b 1.517
a Derivation of estimates can be found in Kleckner and Neumann (1999) and Chestnut (1997).
b Based on elasticity values reported in Table 5-3, U.S. Census population projections, and projections of real GDP
per capita.
6.A.10 References
Abt Associates, Inc. 2010. Environmental Benefits and Mapping Program (Version 4.0).
Bethesda, MD. Prepared for U.S. Environmental Protection Agency Office of Air Quality
Planning and Standards. Research Triangle Park, NC. Available on the Internet at
.
Brookshire, D.S., R.C. d'Arge, W.D. Schulze and M.A. Thayer. 1979. Methods Development for
Assessing Tradeoffs in Environmental Management, Vol. II: Experiments in Valuing Non-
Market Goods: A Case Study of Alternative Benefit Measures of Air Pollution Control in
the South Coast Air Basin of Southern California. Prepared for the U.S. Environmental
Protection Agency, Office of Research and Development. Available on the Internet at
.
Chestnut, L.G., and R.D. Rowe. 1990. "A New National Park Visibility Value Estimates." In
Visibility and Fine Particles, Transactions of an AWMA/EPA International Specialty
Conference, C.V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
Chestnut, L.G., R.D. Rowe and J. Murdoch. 1986. Review of 'Establishing and Valuing the Effects
of Improved Visibility in Eastern United States.' Prepared for the U.S. Environmental
Protection Agency. October. Available on the Internet at
.
6.A-23
-------
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.
Available on the Internet at .
Loehman, E.T., D. Boldt, K. Chaikin. 1985. Measuring the Benefits of Air Quality Improvements
in the San Francisco Bay Area. From Methods Development for Environmental Control
Benefits Assessment, Volume IV. Prepared for the U.S. Environmental Protection
Agency, Office of Policy, Planning and Evaluation, September. Grant #R805059-01-0
Available on the Internet at .
Loehman, E.T., S. Park, and D. Boldt. 1994. "Willingness to Pay for Gains and Losses in Visibility
and Health." Land Economics 70(4): 478-498.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck and M. Thayer.
1993. Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method.
Prepared for U.S. Environmental Protection Agency, Office of Policy, Planning and
Evaluation. September. Available on the Internet at
.
Pitchford, M.L., and W.C. Malm. 1994. "Development and Applications of a Standard Visual
Index." Atmospheric Environment 28(5):1049-1054.
Sisler, J.F. 1996. Spatial and Seasonal Patterns and Long Term Variability of the Composition of
the Haze in the United States: An Analysis of Data from the IMPROVE Network. Colorado
State University, Cooperative Institute for Research in the Atmosphere (CIRA), ISSN
0737-5352-32. Fort Collins, CO. July. Available on the Internet at .
Smith, V.K., G. Van Houtven, and S.K. Pattanayak. 2002. "Benefit Transfer via Preference
Calibration." Land Economics 78:132-152.
Tolley, G., A. Randall, G. Blomquist, M. Brien, R. Fabian, G. Fishelson, A. Frankel, M. Grenchik, J.
Hoehn, A. Kelly, R. Krumm, E. Mensah, and T. Smith. 1984. Establishing and Valuing the
Effects of Improved Visibility in Eastern United States. Prepared for U.S. Environmental
Protection Agency, Office of Policy, Planning and Evaluation. March. U.S. Environmental
Protection Agency Grant #807768-01-0. Available on the Internet at
.
6.A-24
-------
U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 1999. 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 (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, Washington, DC.
March. Available on the Internet 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 on the Internet
at .
6.A-25
-------
APPENDIX 6.B
ILLUSTRATIVE SCENARIO OF RECREATIONAL AND RESIDENTIAL VISIBILITY BENEFITS
6.B.I Synopsis
In this Appendix, we provide an illustrative example analysis of visibility benefits that
applies the visibility benefits methodology described in Chapter 6 and Appendix 6A to a specific
modeled scenario. For this illustrative example, we use the 2020 base case simulation and the
2020 control case simulation from the CMAQ model that were used to develop the air quality
ratios.1 In this Appendix, we refer to this specific scenario as the "illustrative scenario," which is
not a surrogate of the revised annual primary standard. Because we do not have air quality
model runs for the regulatory baseline and the revised or alternative annual standards, we
cannot that calculate the visibility co-benefits of attaining the revised annual primary standard.
It is important to emphasize that this illustrative scenario does not reflect an emissions control
strategy for any specific annual standard level, which is important because light extinction can
vary substantially depending on the specific combination of S02, NOX, or directly emitted
particles reduced and the magnitude and location of those emission reductions. In addition, the
visibility benefits in this chapter cannot be added to the health benefits of the revised or
alternative standards. We provide this illustrative scenario to demonstrate the results of
applying the methodology for estimating benefits for scenarios with light extinction estimates.
6.B.2 Recreational Benefits Results
The modeling results indicate that visibility would improve in several Class I areas as a
result of emission reductions in the illustrative scenario. While we are unable to calculate the
specific contribution in this analysis, the emission reductions associated with these emission
reductions would help Class 1 areas to meet the goals of the Regional Haze rule. Table 6.B-1
identifies the visibility improvements in the 10 most visited parks for the illustrative scenario
using two light extinction metrics: visual range and deciview. The monetized benefits of
recreational visibility improvements are provided in Table 6.B-2 for the illustrative scenario.
Because there is considerable uncertainty about the accuracy of the benefit transfer to
other regions, we include the estimated recreational visibility benefits for parks in other regions
as a sensitivity analysis only. The sensitivity analysis results are not considered part of the total
monetized benefits. Table 6.B-3 provides the results of this sensitivity analysis. Figure 6.B-1
1 See Chapter 4 of this RIA for more information regarding the specific scenario in these modeling simulations
including the magnitude, location, and type of emission reductions. See Chapter 3 for more information regarding
how these modeling simulations were used to calculate the air quality ratios.
6.B-1
-------
shows how the monetized benefits for recreational visibility are distributed across Class I areas
for the illustrative scenario. This sensitivity analysis shows that the benefits in non-studied park
regions could be substantial. In addition, in Table 6.B.4, we provide an indication of the
potential impact of omitting coarse particles from the light extinction calculation using
surrogate coarse particle concentrations from Debell et al. (2006).2
Table 6.B-1. Annual Average Visibility Improvements in the Top 10 Most Visited Class I Areas
for the Illustrative Scenario in 2020a'b
Class 1 Area
Grand Canyon NP
Great Smoky Mountains NP
Yellowstone NPC
Yosemite NP
Sequoia-Kings NP
Glacier NPC
Rocky Mountain NP
Zion NP
Grand Teton NPC
Kings Canyon NP
State
AZ
NC/TN
WY
CA
CA
MT
CO
UT
WY
CA
Illustrative Scenario
Visual Range
(m) Deciviews
-
-
-
400 0.1
990 0.3
-
-
-
-
470 0.1
aBecause these benefits occur within the analysis year, the monetized benefits are the same for all discount rates.
Although the light extinction estimates do not reflect coarse particles, the rounded incremental visibility benefits
are unaffected.
bVisibility measured at the county of the geographic center of park. Top 10 most visited parks ranked by visitor
days in 2008 (NPS, 2008).
c Not included in the primary benefits because the parks are not located in the studied regions. Benefits for these
parks are included in the sensitivity analysis.
2 See Chapter 6 (Section 6.3.1) for more information regarding the purpose of this sensitivity analysis.
6.B-2
-------
Table 6.B-2. Recreational Visibility Benefits in Studied Regions for the Illustrative Scenario in
2020 (in millions of 2010$)a
Studied Park Region Illustrative Scenario Benefits
Southeast $4.3
Southwest
California $350
Total $350
aBecause these benefits occur within the analysis year, the monetized benefits are the same for all discount rates.
These benefits reflect the WTP for all U.S. households for parks in these regions. Although the light extinction
estimates do not reflect coarse particles, the rounded incremental visibility benefits are unaffected.
Table 6.B-3. Sensitivity Analysis for Recreational Visibility Benefits outside Studied Park
Regions for the Illustrative Scenario in 2020 (in millions of 2010$)a
Non-Studied Park Region Illustrative Scenario Benefits
Northwest
Northern Rockies
Rest of U.S.
Total for Non-Studied Parks Regions
Total including Studied Park Regions $350
aBecause these benefits occur within the analysis year, the monetized benefits are the same for all discount rates.
These benefits reflect the WTP for all U.S. households for parks in these regions. These sensitivity analysis results
are not considered part of the total monetized benefits. Although the light extinction estimates do not reflect
coarse particles, the rounded incremental visibility benefits are unaffected.
6.B-3
-------
Table 6.B-4. Sensitivity Analysis for Incorporating Coarse Particles into Recreational Visibility
Benefits for the Illustrative Scenario (in millions of 2010$)a
Illustrative Scenario Benefits
Primary
Recreational
Benefits
(Studied Park
Regions)
Sensitivity
Analysis
(Other Park
Regions)
No coarse particles
+ 5|jg/m3 coarse
+ 5 ng/m3 coarse with 15 ng/m3 in Southwest
+8 |Jg/m3 coarse with 15 ng/m3 in Southwest
No coarse particles
+ 5|jg/m3 coarse
+ 5 u.g/m3 coarse with 15 u.g/m3 in Southwest
+8 |Jg/m3 coarse with 15 u.g/m3 in Southwest
$350
$330
$300
$300
-
-
-
a Because these benefits occur within the analysis year, the monetized benefits are the same for all discount rates.
These benefits reflect the WTP for U.S. households who live in these regions for the parks in the study regions
(primary benefits) or parks in other regions (sensitivity analysis). The levels of coarse particles represent the full
range of possible annual concentrations from a recent report on the IMPROVE monitoring network (Debell et al.,
2006). We define Southwest to be the states of California, Nevada, Utah, Arizona, New Mexico, Colorado, and
Texas.
6.B-4
-------
Monetized Recreational
Visibility Benefits (2010$)
• No visibility improvement
• < $1 million
^P S1 million to $5 million
'> $5 million
Figure 6.B-1. Recreational Visibility Benefits in Class I Areas for the Illustrative Scenario in
2020a
aThe size of the circle in this map indicates the magnitude of the recreational benefits for each Class 1 Area. The
colors in this map indicate whether the park benefits are included in the primary benefits or in the sensitivity
analysis (i.e., non-studied park regions). Blue = primary benefits (studied park regions), Green = sensitivity analysis
(non-studied park regions).
6.B.3 Residential Benefits Results
The modeling results indicate that visibility would improve in many of the study areas as
a result of the emission reductions associated with emission reductions in the illustrative
scenario. Table 6.B-5 shows the monetized residential visibility benefits in the eight study areas.
These benefits reflect the value to households living within those MSAs, accounts for inflation,
and account for growth in real income since the original WTP estimates were developed.
The results of the sensitivity analysis for the monetized residential visibility benefits in
all MSAs for the illustrative scenario are provided in Table 6.B-6. Figure 6.B-2 shows how the
sensitivity analysis results are distributed geographically for the illustrative scenario. In
addition, Table 6.B-7 provides an indication of the potential impact of omitting coarse particles
in the calculation of light extinction.
6.B-5
-------
Table 6.B-5. Monetized Residential Visibility Benefits in Studied Areas in 2020 for the
Illustrative Scenario (millions of 2010$, 2020 income)3
Study Area
Atlanta
Boston
Chicago
Denver
Los Angeles
Mobile
San Francisco
Washington, DC
Total
Illustrative Scenario
-
-
$43
-
$110
-
$39
-
$190
aBecause these benefits occur within the analysis year, the monetized benefits are the same for all discount rates.
Although the light extinction estimates do not reflect coarse particles, the rounded incremental visibility benefits
are unaffected.
Table 6.B-6. Sensitivity Analysis for Monetized Residential Visibility Benefits in Other Areas
for the Illustrative Scenario in 2020 (in millions of 2010$)a
Extent of Benefit Transfer Illustrative Scenario Benefits
Additional MSAs in East $140
Additional MSAs in West $24
Additional MSAs in California $160
Total $330
Total including Study Areas $520
aBecause these benefits occur within the analysis year, the monetized benefits are the same for all discount rates.
These sensitivity analysis results are not considered part of the total monetized benefits. Although the light
extinction estimates do not reflect coarse particles, the rounded incremental visibility benefits are unaffected.
6.B-6
-------
Monetized Residential
Visibility Benefits (2010$)
| | Studied MSAs
No visibility improvement
^] <$1 million
^B S1 million lo $5 mlllion
| > $5 million
Figure 6.B-2. Residential Visibility Benefits for the Illustrative Scenario in 2020 (2010$)
Table 6.B-7. Sensitivity Analysis for Incorporating Coarse Particles into Residential Visibility
Benefits (in millions of 2010$)a
Illustrative Scenario Benefits
Primary
Residential
Benefits
Sensitivity Analysis
No coarse particles
+ 5 ng/m3 coarse
+ 5 ng/m3 coarse with 15 ng/m3 in Southwest
+8 u.g/m3 coarse with 15 u.g/m3 in Southwest
No coarse particles
+ 5 ng/m3 coarse
+ 5 ng/m3 coarse with 15 ng/m3 in Southwest
+8 ng/m3 coarse with 15 ng/m3 in Southwest
$190
$180
$150
$130
$330
$310
$250
$270
aBecause these benefits occur within the analysis year, the monetized benefits are the same for all discount rates.
These benefits reflect the WTP for U.S. households who live in these regions for the parks in the study regions
(primary benefits) or parks in other regions (sensitivity analysis). The levels of coarse particles represent the full
range of possible annual concentrations from a recent report on the IMPROVE monitoring network (Debell et al.,
2006). We define Southwest to be the states of California, Nevada, Utah, Arizona, New Mexico, Colorado, and
Texas.
6.B-7
-------
6.B.4 References
DeBell, L. J.; Gebhart, K. A.; Hand, J. L; Malm, W. C; Pitchford, M. L; Schichtel, B. A.; White, W.
H. 2006. Spatial and Seasonal Patterns and Temporal Variability of Haze and its
Constituents in the United States, Report IV; Cooperative Institute for Research in the
Atmosphere, Nov; ISSN 0737-5352-74. Available on the Internet at
.
National Park Service (NPS). 2008. Statistical Abstract: 2008. Department of Interior, National
Park Service Social Science Program, Public Use Statistics Office, Denver, Colorado.
Available on the Internet at
.
6.B-8
-------
CHAPTER 7
ENGINEERING COST ANALYSIS
7.1 Synopsis
This chapter summarizes the data sources and methodology used to estimate the
engineering costs of attaining the revised annual standard and two alternative annual standards
for the PM2.5 primary standard analyzed in this RIA. This chapter provides the estimates of the
annual engineering costs for illustrative control strategies designed to demonstrate attainment
of the revised annual standard of 12 u.g/m3 in conjunction with retaining the 24-hour standard
of 35 u.g/m3' as well as control strategies designed to demonstrate attainment of the alternative
annual standards of 13 and 11 u.g/m3 in conjunction with retaining the 24-hour standard of 35
u.g/m3 (referred to as 12, 13, and 11). These illustrative control strategies are outlined in
Chapter 4. The cost discussion for known controls in Section 7.2.2 is followed by a presentation
of estimates for the engineering costs of the additional emissions reductions that are needed
beyond the application of known controls to reach full attainment of the alternative standards
analyzed; the cost estimates derived from this approach, discussed in Section 7.2.3, are
referred to as "extrapolated" costs. By definition, no cost data currently exists for the additional
emissions reductions needed beyond known controls. We employ two methodologies for
estimating the costs of unidentified future controls, and both approaches assume either that
existing technologies can be applied in particular combinations or to specific sources that we
currently can't predict or that innovative strategies and new control options make possible the
emissions reductions needed for attainment by 2020.
The engineering costs described in this chapter generally include the costs of
purchasing, installing, operating, and maintaining the referenced control technologies. For a
variety of reasons, actual control costs may vary from the estimates 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 the revised and alternative standards, and EPA anticipates that local and
state governments will consider programs that are best suited for local and regional conditions.
Furthermore, based on past experience, EPA believes that it is reasonable to anticipate that the
marginal cost of control will decline over time due to technological improvements and more
widespread adoption of previously considered niche control technologies. Also, EPA recognizes
the extrapolated portion of the engineering cost estimates are uncertain because extrapolated
7-1
-------
costs do not contain information about which sectors may be affected or which control
measures may be employed in the future.
The engineering cost estimates are limited in their scope. This analysis focuses on the
emissions reductions needed for attainment of the revised and alternative standards that are
described earlier in this RIA. EPA understands that some states will incur costs designing State
Implementation Plans (SIPs) and implementing new control strategies to meet the revised
standard. However, EPA does not know what specific actions states will take to design their SIPs
to meet the 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. EPA
generally estimates state-level administrative costs in an information collection request (ICR)
that accompanies the implementation rule or guidance for each NAAQS (as opposed to
accompanying the issuance of the NAAQS)
7.2 Estimation of Engineering Control Costs
7.2.1 Data and Methods—Identified Control Costs (non-EGU Point and Area Sources)
After designing the hypothetical control strategy using the methodology discussed in
Chapter 4, EPA used the Control Strategy Tool1 (CoST) to estimate engineering control costs for
mobile, non-EGU point and area sources.2 CoST calculates engineering costs using either: (1) an
average annualized cost-per-ton estimate multiplied by the total tons of a pollutant reduced to
derive a total cost estimate, or (2) an equation that incorporates key emission source
information (e.g., unit capacity and stack flow information).3 Most control cost information
within CoST was developed based on the cost-per-ton approach because estimating
engineering costs using an equation requires more data, and these data are sometimes, but not
always, available.
The cost equations located in CoST require either unit capacity or stack flow to
determine annualized, capital and/or operating and maintenance (O&M) costs. Capital costs
1 The Control Strategy Tool recently underwent peer review by an ad hoc panel of experts. Responses to the peer
review are currently being developed and will be available by final promulgation of this rule.
2 Area sources are not necessarily non-urban sources.
3 Annualized costs represent an equal stream of yearly costs over the period the control technology is expected to
operate.
7-2
-------
are converted to annualized costs using the capital recovery factor (CRF).4 When the cost
equations and input data are available in CoST, the equations 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.
Engineering costs will differ between different emissions sources based upon quantity of
emissions reduced, plant capacity, or stack flow. Engineering costs will also differ in nominal
(i.e., not adjusted for inflation) terms by the year the costs are calculated for (i.e., 1999$ versus
2010$).5 For capital investment, in order to attain standards in 2020 we assume capital
investment occurs at the beginning of 2020. We make this simplifying assumption because we
do not know what all firms making capital investments will do and when they will do it. For
2020, our estimate of annualized costs includes annualized capital and O&M costs for those
controls included in our identified 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 along with the interest rate of 7%
incorporated into the CRF. We make no presumption of additional capital investment in years
beyond 2020. The EUAC method is discussed in detail in the EPA Air Pollution Control Cost
Manual.6 Applied controls and their respective engineering costs are provided in the PM NAAQS
docket.
7.2.2 Identified Control Costs
In this section, we provide engineering cost estimates for the control strategies
identified in Chapter 4 that include control technologies on non-EGU point sources and area
sources. Engineering costs generally refer to the capital equipment expense, the site
preparation costs for the application, and annual operating and maintenance costs. For this
analysis, we included known controls for all of the geographic areas likely to exceed the revised
4 The capital recovery factor formula is expressed as [r*(l+r)An/(l+r)An -1]*CC. Where r is the real rate of interest,
n is the number of time periods, and CC is the capital cost. For more information on this cost methodology and
CoST, please refer to the documentation at http://www.epa.gov/ttn/ecas/cost.htm, the EPA Air Pollution
Control Cost Manual found at http://epa.gov/ttn/catc/products.htmltfcccinfo, and EPA's Guidelines for
Preparing Economic Analyses, Chapter 6 found at
http://yosemite.epa.gov/ee/epa/eed.nsf/webpages/Guidelines.htmlftdownload.
5 The engineering costs will not be any different in real (inflation-adjusted) terms if calculated in 2010 versus other-
year dollars, if the other-year dollars are properly adjusted. For this analysis, all costs are reported in real 2010
dollars.
6 http://epa.gov/ttn/catc/products.htmltfcccinfo
7-3
-------
and/or alternative standards. We included all known controls at an annual cost of $20,000 per
ton or less, which included approximately 85% of known controls in the geographic areas likely
to exceed the revised and/or alternative standards.7 We did not include the small number of
known controls that had an annual cost of more than $20,000 per ton because either (i) the
remaining emissions sources were relatively small sources, and we believe they are already
controlled, or (ii) the equations in CoST were not applicable to these remaining emissions
sources.
Because we obtain control cost data from many sources, we are not always able to
obtain consistent data across original data sources.8 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% discount rate
because the majority of the available disaggregated control cost data is calculated using 7%.
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% discount rate. The use of these two
discount rates for cost estimation reflects the guidance for RIA preparation found in Circular A-
4, issued by OMB in September 2003.9 In general, we have some disaggregated data available
for non-EGU point source controls; we do not have any disaggregated control cost data for area
source controls.10 In this analysis, for the revised annual standard of 12 u.g/m3 and the
alternative standard of 13 u.g/m3 we did not have any disaggregated known control cost data,
and as such we were not able to recalculate known control costs using a 3% discount rate.
Because we were not able to recalculate known controls costs using a 3% discount rate, we are
not presenting known controls costs for the revised or alternative standards using that discount
rate. See Table 7-1 for a summary of sectors and known control costs.
7 For the known controls, for all of the geographic areas likely to exceed the revised and/or alternative standards
we include controls at an annual cost of $20,000 per ton or less. To estimate the costs associated with
unidentified future controls, or unknown controls, we employ a fixed-cost and hybrid methodology. The fixed-
cost methodology employs a primary cost estimate of $15,000/ton (2010 dollars), and the hybrid methodology
employs an initial, annual cost-per-ton estimate of $15,000/ton (2010 dollars). We explain the choices of these
parameters in this Section and Section 7.2.3.
8 Data sources can include states, as well as technical studies, which do not typically include references to the
original data source.
9 U.S. Office of Management and Budget. Circular A-4, September 17, 2003. Available at
http://www.whitehouse.gov/omb/circulars a004 a-4/.
10 For area source controls, total annualized costs are assumed to be calculated using a 7% discount rate.
7-4
-------
Table 7-1. Summary of Sectors, Emissions Reductions, and Known Annualized Control Costs
(millions of 2010$)a'b
Revised and
Alternative Standard Emissions Sector
13 u.g/m3 Non-EGU Point Sources
Area Sources
Total
12 u.g/m3 Non-EGU Point Sources
Area Sources
Total
11 u.g/m3 Non-EGU Point Sources
Area Sources
Total
Emissions Reductions0
-
53
53
380
430
800
23,000
2,300
25,400
7%
Discount Rate
—
$0.63
$0.63
$0.87
$4.3
$5.1
$88
$13
$100
All estimates rounded to two significant figures. Estimates may not sum due to rounding convention.
The emissions reductions and total costs are associated with partial attainment.
c The emissions reductions for the alternative standard of 11 include PM2.5 and SO2 emissions reductions.
The total annualized cost of control in each sector in the control scenario is summarized
by region in Table 7-2. Table 7-2 includes annualized control costs to allow for comparison
across regions and between costs and benefits. These numbers reflect the engineering costs
annualized at a discount rate of 7%. Engineering cost estimates presented throughout this and
subsequent chapters are based on a 7% discount rate.
Table 7-2. Partial Attainment Known Annualized Control Costs in 2020
for Revised and Alternative Standards Analyzed (millions of 2010$)a'b
Revised & Alternative Standard Region
13 |Jg/m3 East
West
California
Total
12 u.g/m3 East
West
California
Known Controls
$0.63
$0.63
$5.1
7-5
-------
Total $5.1
11 |jg/m3 East $96
West $0.45
California $5.3
Total $100
Estimates are rounded to two significant figures. As such, numbers may not sum down columns.
b Note that the estimates provided reflect incremental emissions reductions from an analytical baseline that gives
"credit" to the San Joaquin Valley and South Coast areas for emissions reductions expected to occur between
2020 and 2025 (when those areas are expected to demonstrate attainment with the revised and/or alternative
standards).
The total annualized engineering costs associated with the application of known
controls, incremental to the baseline and using a 7% discount rate, are approximately $5.1
million for the revised annual standard of 12, $630,000 for the less stringent alternative annual
standard of 13 u.g/m3, and $100 million for a more stringent alternative annual standard of 11
l-ig/m3.
7.2.3 Extrapolated Costs
This section presents the methodology and results of the extrapolated engineering cost
calculations for attainment of the revised annual PM2.5 standard of 12 u.g/m3, as well as
alternative annual standards of 13 u.g/m3and 11 u.g/m3. All costs presented for the illustrative
control strategies are calculated incrementally from the current PM2.5 standard of 15/35 u.g/m3,
therefore, any additional emission reductions needed to attain the current 24-hour standard of
35 u.g/m3are part of the baseline analysis and not presented here.
As mentioned earlier in this chapter, and as described in more detail in Chapter 4, the
application of the known control strategy was not successful in reaching attainment for all
areas for these alternative PM2.5 standards. Because some areas remained in nonattainment,
the engineering costs detailed in Section 7.2.2 represent the costs of partial attainment for the
revised annual standard of 12 u.g/m3 and alternative annual PM2.5 standards of 13 u.g/m3 and 11
u.g/m3. For the revised standard and each alternative standard and geographic area that cannot
reach attainment with known controls, we estimated the additional emissions reductions
needed for PM2.5 to attain the standard. See Chapter 3, Section 3.3.2 and Tables 3-7, 3-8, and 3-
9 for a detailed discussion of how the additional, needed emissions reductions were estimated
and for a summary of the needed emissions reductions for the revised annual standard and the
alternative annual standards.
7-6
-------
To generate estimates of the costs and benefits of meeting the revised and alternative
standards, in addition to the application of known controls EPA assumes the application of
unidentified future controls that make possible the additional emission reductions needed for
attainment in 2020. By definition, no cost data currently exists for unidentified future
technologies or innovative strategies. EPA used two methodologies for estimating the costs of
unidentified future controls: a fixed-cost methodology and a hybrid methodology. Both
approaches assume either that existing technologies can be applied in particular combinations
or to specific sources that we currently can't predict or that innovative strategies and new
control options make possible the emissions reductions needed for attainment by 2020.
However, the two approaches reflect different assumptions about technological progress and
innovation in emissions reductions strategies.
7.2.3.1 Fixed-Cost Methodology
The fixed-cost methodology uses a $15,000/ton estimate for each ton of PM2.s reduced;
the hybrid methodology is similar to the hybrid methodology used for the 2008 Ozone NAAQS
RIA cost analysis and is presented in more detail below. The fixed-cost methodology was
preferred by EPA's Science Advisory Board over two other options, including a marginal-cost-
based methodology. When reviewing the Office of Air and Radiation's Direct Cost Report and
Uncertainty Analysis Plan, the Science Advisory Board noted:
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.
Note that the choice of $15,000/ton for the fixed-cost methodology was based on the
precedent set in the March 2011 final report The Benefits and Costs of the Clean Air Act from
1990 to 2020.11'12 We also chose $15,000/ton for the national, initial cost-per-ton for use in the
11 We considered adjusting the $15,000/ton value and reviewed data on inflation between 2006 and 2010. We
found that during that period inflation was sufficiently low to not warrant a $/ton value adjustment; any such
adjustment would be considered well within the bounds of uncertainty in this analysis. To assess the sensitivity
of the results to the value of $15,000/ton, we also include sensitivity analyses at $10,000/ton and $20,000/ton
in Appendix 7A. In addition, the extrapolated costs do not rely on specific underlying data sources (e.g., Census
data) that periodically change and that would require updating based on those changes. As such, we currently
do not have either specific data or a particular rationale to change the $15,000/ton value.
12 The final report The Benefits and Costs of the Clean Air Act from 1990 to 2020 includes the following discussion
for the rationale for the $15,000 per ton threshold. Controls that are more costly than $15,000 per ton may
not be cost effective, and local air quality agencies would likely seek reductions from other unidentified control
measures. This is consistent with the practice of the South Coast Air Quality Management District, which
7-7
-------
hybrid methodology to facilitate direct comparison with the fixed-cost methodology.13 In
addition, we do not have reason to conclude that the initial cost-per-ton used in the hybrid
methodology should be different than the value used in the fixed-cost methodology. In the
proposal RIA, we requested comments or suggestions on methodologies for estimating the
costs of unspecified future controls to provide illustrative estimates of NAAQS costs. We did not
receive any direct comments on methodologies, but we did receive comments from the San
Joaquin Valley Air Pollution Control District (SJV APCD) on the magnitude of their potential
investments needed to meet the revised annual standard of 12 u.g/m3 relative to our total cost
estimates. The SJV APCD commented that expenditures in their jurisdiction alone could likely
be more than our total cost estimate. Their comment provides additional context for the need
to improve existing or identify new methodologies for estimating the costs of unspecified
future controls.
7.2.3.2 Hybrid Cost Methodology
The hybrid methodology generates a total annual cost curve for PlV^.sfor unknown
future controls that might be applied in order to move toward 2020 attainment. The hybrid
methodology has the advantage of incorporating information about how significant the needed
reductions from unspecified control technologies are relative to the known control measures
and matching that information with expected increasing per-ton cost for applying unknown
controls.14 For PM2.5 reductions needed in each area, the cost begins with a national constant
cost-per-ton for PM2.5 and increases as emissions reduction for PM2.5 are needed, reflecting the
expectation that average per-ton control costs are likely to be higher in areas needing a higher
ratio of emission reductions from unknown to known controls. For example, to attain a revised
annual standard of 12 u.g/m3, all of the needed emissions reductions for Los Angeles County
were from known controls at an average cost of $6,000 per ton; whereas for Riverside County
approximately 95 percent of the needed emissions reductions were from unknown future
controls at an average cost of $290,000 per ton. For other geographical areas, the average cost
attempts to identify viable alternatives for any control requirements with an estimated cost exceeding $16,500
per ton. When costs are above this threshold, the South Coast Air Quality Management District conducts more
detailed cost-effectiveness and economic impact analyses of the controls.
13 For the known controls, for all of the geographic areas likely to exceed the revised and/or alternative standards
we include controls at an annual cost of $20,000 per ton or less. To estimate the costs associated with
unidentified future controls, or unknown controls, we employ a fixed-cost and hybrid methodology. The fixed-
cost methodology employs a primary cost estimate of $15,000/ton (2010 dollars), and the hybrid methodology
employs an initial, annual cost-per-ton estimate of $15,000/ton (2010 dollars). We explain the choices of these
parameters in this Section and Section 7.2.2.
14 In applying the hybrid methodology, EPA reviewed the data to ensure that the estimated, additional emissions
reductions selected for each geographic area were not greater than the remaining uncontrolled emissions in
that geographic area. The highest percent selected was 90%.
7-8
-------
per ton for unknown controls ranged from $19,000 to $28,000 per ton. The incremental
improvement in air quality for an unknown control is determined using an area-by-area ratio of
air quality improvement to air quality change, which is discussed in more detail in Chapter 3.
EPA developed a model of increasing total annualized costs for controlling PM2.5
emissions. The simplest form of ax2 + bx + c was used where x is the tons of a particular
pollutant to be reduced in a particular area and a, b, and c are constants. For the hybrid
methodology b is set to be a national, initial cost-per-ton (N) for unknown controls for PM2.5,
and c is set to zero because there is no cost to imposing no control. The hybrid methodology
has a different a for PM2.5for each geographic area. For a particular geographic area a is N/E
where
N = national, initial annualized cost/ton (b from above) of $15,000 per ton.
E = by geographic area, is the denominator and represents all particulate
emission reductions achieved (from applying known and unknown
controls to obtain the 15/35 baseline, as well as known controls to
achieve the alternative standard) prior to estimating needed emission
reductions from unknown controls to achieve the alternative standard.
U = unknown emissions reductions by geographic area and standard.
T = cost by geographic area and standard, or —U2 + NU (i.e., ox2-/-bx).
An example of the hybrid methodology is provided below. In this example, in Area B the
percentage of total PM2.5 reductions needed from unknown controls relative to total emissions
reductions needed (e.g., 100/150, or 67%) is larger than the percentage of total PM2.5
reductions needed from unknown controls relative to total emissions reductions needed in
Area A (e.g., 100/200, or 50%). Because Area B needs a higher portion of emissions reductions
from unknown controls, total cost using the hybrid methodology is higher in Area B. This
illustration shows that the relative costs of unknown controls reflect the expectation that
average per-ton control costs for the same number of unknown tons are likely to be higher in
Area B, which needs a higher ratio of emission reductions from unknown to known controls.
7-9
-------
Example of Applying Hybrid Methodology
Area A
Area B
PM2.s Emissions
Reductions Achieved
With Known Controls
100
50
PM2.s Emissions
Reductions Needed From
Unknown Controls
100
100
Area A—Cost Using Hybrid Methodology
$15,000
^100 tons reduced
Average cost/ton = $30,000
* 100 tons needed2 + ($15,000 * 100 tons needed) = $3,000,000
Area B—Cost Using Hybrid Methodology
( ( $15,000 \ ,\
* 100 tons needed2 + ($15,000 * 100 tons needed) = $4, 500,000
\\50tonsreducedJ I
Average cost/ton = $45,000
7.2.3.3 Fixed-Cost and Hybrid Methodology Extrapolated Cost Estimates
Extrapolated cost estimates are provided using a 7% discount rate because known
control measure information is available at 7% for all measures applied in this analysis.
Table 7-3 provides the extrapolated cost estimates using both the fixed-cost and hybrid
methodologies described above. The extrapolated cost estimates range from $48 million
(2010$) to $340 million (2010$) for the revised standard of 12 u.g/m3. We included sensitivity
analyses using both the alternative fixed cost-per-ton and the hybrid methodologies in
Appendix 7.A.
Table 7-3. Extrapolated Costs by Revised and Alternative Standard Analyzed3'15 (millions of
2010$)
Extrapolated Cost
Fixed-Cost Methodology Hybrid
Standard Region 7%
13 |Jg/m3 East —
West -
California $10
Total $10
Methodologyc
7%
—
—
$100
$100
7-10
-------
12 |Jg/m3 East
West
California
Total
11 u.g/m3 East
West
California
Total
—
—
$48
$48
$71
$1.3
$150
$220
—
—
$340
$340
$650
$3.3
$940
$1,600
Estimates are rounded to two significant figures.
Note that the estimates provided reflect incremental emissions reductions from an analytical baseline that gives
"credit" to the San Joaquin Valley and South Coast areas for emissions reductions expected to occur between
2020 and 2025 (when those areas are expected to demonstrate attainment with the revised and/or alternative
standards).
In applying the hybrid methodology, Plumas County, CA and Shoshone County, ID did not have any known PM
controls. We took the following approach to estimate prior emissions reductions for these two counties for use
in the hybrid methodology cost calculations: for the remaining counties, by county we summed (i) emissions
reductions from known controls and (ii) extrapolated emissions reductions to meet the 15/35 baseline and
divided each county's sum by that county's base case PM emissions. We selected the overall minimum
percentage and for each of the two counties without any known PM controls, we multiplied that overall
minimum percentage by the specific county's base case PM emissions.
Of note is the geographic distribution of extrapolated costs. For the revised and
alternative standards, the above costs indicate that control measures applied in California
represent a significant portion of the extrapolated costs. Using the fixed-cost methodology, for
the revised annual standard of 12 u.g/m3 and the alternative annual standards of 13 u.g/m3 and
11 u.g/m3, the California component of the extrapolated cost estimates represents 100%, 100%,
and 67%, respectively, of the nationwide extrapolated cost estimates. Using the hybrid
methodology, for the revised annual standard of 12 u.g/m3and the alternative annual standards
of 13 u.g/m3 and 11 u.g/m3, the California component of the extrapolated cost estimates
represents 100%, 100%, and 59%, respectively, of the nationwide extrapolated cost estimates.
Because no cost data exists for unknown future controls, it is unclear whether approaches using
hypothetical cost curves will be more accurate or less accurate in forecasting total national
costs of unknown controls than a fixed-cost methodology that uses a range of national cost-
per-ton values.
Estimating engineering costs for emission reductions needed beyond those from known
controls to reach attainment in 2020 is inherently a challenging exercise. As described later in
this chapter, our experience with Clean Air Act implementation shows that technological
7-11
-------
advances and development of innovative strategies can reduce emissions and reduce the costs
of emerging technologies over time. Technological change may provide new possibilities for
controlling emissions as well as reducing the cost and effectiveness of known controls through
technological improvements or higher control efficiencies.
7.2.3.4 Interpreting Extrapolated Cost Estimates
The two estimates do not represent lower and upper bound estimates, but simply
represent estimates generated by two different methodologies. The fixed-cost methodology
assumes that technological change and innovation will result in the availability of additional
controls by 2020 that are similar in cost to the higher end of the cost range for current known
controls. The hybrid methodology assumes that while additional controls may become
available by 2020, they become available at an increasing cost and the increasing cost varies by
geographic area and by degree of difficulty associated with obtaining the needed emissions
reductions. Without an initial parameter estimate, i.e., $15,000/ton, we are not able to predict
which methodology will generate a higher cost estimate; however, with the same initial
parameter estimate of $15,000/ton, the hybrid methodology will generate a higher cost
estimate.
7.2.4 Total Cost Estimates
In the supporting statement for the Information Collection Request Revision for
Particulate Matter 2.5 Ambient Air Monitoring, 40 CFR Part 58 we estimate the incremental
cost of relocating 21 existing near-roadway monitors, and those costs are included in the total
national cost estimates presented below. The amendments to the ambient air monitoring
regulations will revise the network design requirements for PM2.s monitoring sites, resulting in
moving 21 monitors to established near-road monitoring stations by January 1, 2015. The
incremental cost associated with moving these 21 monitors is a one-time cost of $28,570.15
Tables 7-4 and 7-5 present a summary of the total national costs of attaining the revised
annual standard of 12 u.g/m3and the alternative annual standards of 13 u.g/m3 and 11 u.g/m3in
2020. This summary includes the known and extrapolated costs. As discussed in Section 7.2.2,
we were not able to recalculate any known control costs using a 3% discount rate. As such, both
15 EPA is not increasing the size of the national PM2.5 monitoring network; the Agency anticipates that states would
be able to relocate existing monitors to meet the near-roadway requirement. For purposes of estimating cost,
only 21 monitors will be moved by 2015. Data from these monitors, along with other monitors in the area,
could be used to determine whether the area is meeting both the annual and 24-hour standards. However,
data from these monitors would not be available in time for use in making initial attainment and
nonattainment designations.
7-12
-------
known and extrapolated costs were calculated at a 7% discount rate only. The total cost
estimates are $53 million (2010$) and $350 million (2010$) for the revised annual standard of
12 u.g/m3; $11 million and $100 million for the alternative annual standard of 13 u.g/m3; and
$320 million and $1,700 million for the alternative annual standard of 11 u.g/m3. To further
evaluate potential costs ranges, we included sensitivity analyses using both the alternative fixed
cost-per-ton and the hybrid methodologies in Appendix 7.A. In addition, Appendix 7.A includes
costs and information needed to calculate those costs, by county, to meet 12 u,g/m3.
For the revised annual standard of 12 u.g/m3, the total cost estimates are comprised of
between 90 and 97 percent extrapolated cost estimates, and the estimated total cost using the
hybrid methodology is roughly 6.5 times more than the estimated total cost using the fixed-cost
methodology.16 Because the hybrid methodology reflects increasing marginal costs in areas
needing a higher ratio of emissions reductions from unknown to known controls, it could be
more representative of total costs. In an effort to consider the potential fitness of the
extrapolated cost estimates, we reviewed the South Coast Air Quality Management District's
(SCAQMD) 2012 Air Quality Management Plan (AQMP)17, and we located data on recent
emission reduction credit (ERC) transactions in both the SCAQMD and SJV APCD. While this
information provides context for the extrapolated cost estimates, the current relationship
between available controls and costs to reduce emissions may or may not be applicable in 2020
because of changes in innovation and advances in technology.
The SCAQMD's 2012 AQMP includes information on control measures to meet the
current 24-hour standard of 35. This list of control measures includes further PM2.s controls for
under-fired charbroilers at a cost per ton reduced of $15,000. This control cost matches the
parameter used in the fixed-cost methodology, as well as the initial value used for the hybrid
methodology and is supportive of our selection of that value.
To provide context for the hybrid methodology's increasing per-ton cost format we
obtained the California Air Resources Board's 2009 and 2010 Emission Reduction Offset
Transaction Costs, Summary Report and reviewed the PMi0 ERC prices in both the SCAQMD and
the SJV APCD. To some degree, ERC transaction prices reflect a choice between installing a
more stringent control or purchasing ERCs. Between 2009 and 2010 PMi0 ERC_prices in SJV
APCD ranged from $40,000 per ton per year (tpy) to $70,000/tpy, and PMi0 ERC prices in the
SCAQMD ranged from $575,000/tpy to more than $1.9 million/tpy. These prices reflect both
16 Note that the extrapolated cost estimates do not represent lower and upper bound estimates, but simply
represent estimates generated by the fixed-cost and hybrid methodologies.
17 Available at http://www.aqmd.gov/aqmp/2012aqmp/draft/index.html.
7-13
-------
marginal costs that are higher than the fixed-cost estimates and marginal costs that are not
inconsistent with the higher cost estimates generated using the hybrid methodology.
Table 7-4. Total Costs by Revised and Alternative Standard Analyzed (millions of 2010$)a,
Fixed-Cost Methodology15'0
Revised and Alternative
Standard Region
13 u.g/m3 East
West
California
Total
12 u.g/m3 East
West
California
Total
11 u.g/m3 East
West
California
Total
Known Control
Costs
—
—
$0.63
$0.63
—
—
$5.1
$5.1
$96
$0.45
$5.3
$100
Unknown Control
Costs— Fixed-Cost
Methodology
—
—
$10
$10
—
—
$48
$48
$71
$1.3
$150
$220
Total Costs
—
—
$11
$11
—
—
$53
$53
$170
$1.8
$160
$320
Estimates are rounded to two significant figures. As such, numbers may not sum down columns.
b All control costs are presented at a 7% discount rate only.
Note that the estimates provided reflect incremental emissions reductions from an analytical baseline that gives
"credit" to the San Joaquin Valley and South Coast areas for emissions reductions expected to occur between
2020 and 2025 (when those areas are expected to demonstrate attainment with the revised and/or alternative
standards).
7-14
-------
Table 7-5 Total Costs by Revised and Alternative Standard Analyzed (millions of 2010$)a,
Hybrid Methodology15'0
Revised and Alternative
Standard Region
13 u.g/m3 East
West
California
Total
12 u.g/m3 East
West
California
Total
11 u.g/m3 East
West
California
Total
Known Control
Costs
—
—
$0.63
$0.63
—
—
$5.1
$5.1
$96
$0.45
$5.3
$100
Unknown Control
Costs— Hybrid
Methodology
—
—
$100
$100
—
—
$340
$340
$650
$3.3
$940
$1,600
Total Costs
—
—
$100
$100
—
—
$350
$350
$750
$3.8
$950
$1,700
a Estimates are rounded to two significant figures. As such, numbers may not sum down columns.
All control costs are presented at a 7% discount rate only.
Note that the estimates provided reflect incremental emissions reductions from an analytical baseline that gives
"credit" to the San Joaquin Valley and South Coast areas for emissions reductions expected to occur between
2020 and 2025 (when those areas are expected to demonstrate attainment with the revised and/or alternative
standards).
7.3 Changes in Regulatory Cost Estimates over Time
There are many examples in which technological innovation and "learning by doing"
have made it possible to achieve greater emission reductions than had been feasible earlier, or
have reduced the costs of emissions control in relation to original estimates. Studies have
concluded that costs of some EPA programs have been less than originally estimated, due in
part to EPA's inability to predict and account for future technological innovation in regulatory
impact analyses.18 Additionally, technological change will affect baseline conditions for our
analysis. Technical change may lead to potential improvements in the efficiency with which
' Harrington et al. (2000) and previous studies cited by Harrington. Harrington, W., R.D. Morgenstern, and P.
Nelson. 2000. "On the Accuracy of Regulatory Cost Estimates." Journal of Policy Analysis and Management
19(2):297-322.
7-15
-------
firms produce goods and services; for example, firms may use less energy to produce the same
quantities of output.
Increasing marginal abatement costs could possibly induce the type of innovation that
would result in lower costs than estimated in this chapter. By 2020, breakthrough technologies
in control equipment could result in a downward shift in the marginal abatement cost curve for
such equipment (Figure 7-1)19 as well as a decrease in its slope, reducing marginal costs per unit
of abatement. In addition, elevated abatement costs may result in significant increases in the
cost of production and would likely induce production efficiencies, in particular those related to
energy inputs, which would lower emissions from the production side.
Cost/
Ton
MCo
MCi
9o
Slope = / Slope =>
MC
LONG
Induced Technology Shift
S02 Reductions
Figure 7-1. Technological Innovation Reflected by Marginal Cost Shift
7.3.1 Examples of Technological Advances in Pollution Control
There are a number of examples of low-emissions technologies and pollution control
equipment developed and/or commercialized over the past 15 to 20 years, such as
• Selective catalytic reduction (SCR) and ultra-low NOX burners for NOX emissions
• Scrubbers, which achieve 95% and potentially greater S02 control on boilers
• Sophisticated new valve seals and leak detection equipment for refineries and
chemical plants
Figure 7-1 shows a linear marginal abatement cost curve. It is possible that the shape of the marginal abatement
cost curve is non-linear.
7-16
-------
• 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 are much cleaner 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
• Market penetration of gas-electric hybrid vehicles, and clean fuels
• The development of retrofit technology to reduce emissions from in-use vehicles
and non-road equipment
These technologies were not commercially available two decades ago, and some did not
even exist. Yet today, all of these technologies are on the market, and many are widely
employed. Several are key components of major pollution control programs.
"Learning by doing" or "learning curve impacts," a distinct concept from technological
innovation, has also made it possible to achieve greater emissions reductions than had been
feasible earlier or has reduced the costs of emissions control compared 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.
Impacts such as these would manifest themselves as a lower expected cost to operate
technologies in the future compared to what costs may have been.
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 for the
Clean Air Act Second Section 812 Prospective Analysis.20 In that report, learning curve
adjustments were included for those sectors and technologies for which learning curve data
were available. A typical learning curve adjustment example is to reduce either capital or O&M
costs by a certain percentage given a doubling of output from that sector or for that
technology. In 1936, T.P. Wright was the first to characterize the relationship between
increased productivity and cumulative production. He analyzed man-hours required to
20 E.H. Pechan and Associates, Inc. and Industrial Economics, Incorporated. 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/oar/sect812/febll/costfullreport.pdf.
7-17
-------
assemble successive airplane bodies. He suggested the relationship is a log linear function,
since he observed a constant linear reduction in man-hours every time the total number of
airplanes assembled was doubled. The relationship he devised between number assembled and
assembly time is called Wright's Equation (Gumerman and Marnay, 2004).21 This equation,
shown below, has been shown to be widely applicable in manufacturing:
Wright's Equation: CN = C0 * Nb, (7.2)
where:
N = cumulative production
CN = cost to produce Nth unit of capacity
Co = cost to produce the first unit
b = learning parameter = In (l-LR)/ln(2), where
LR = learning by doing rate, or cost reduction per doubling of capacity or output.
The percentage adjustments to costs can range from 5 to 20%, depending on the sector
and technology. Learning curve adjustments were prepared in a memo by lEc supplied to U.S.
EPA and applied for the mobile source sector (both onroad and nonroad) and for application of
various ECU control technologies within the Draft Direct Cost Report.22 Advice received from
the SAB Advisory Council on Clean Air Compliance Analysis in June 2007 indicated an interest in
expanding the treatment of learning curves to those portions of the cost analysis for which no
learning curve impact data are currently available. Examples of these sectors are non-EGU point
sources and area sources. The memo by lEc outlined various approaches by which learning
curve impacts can be addressed for those sectors. The recommended learning curve impact
adjustment for virtually every sector considered in the Draft Direct Cost Report is a 10%
reduction in O&M costs for two doublings of cumulative output, with proxies such as
cumulative fuel sales or cumulative emissions reductions being used when output data was
unavailable.
21 Gumerman, Etan and Marnay, Chris. Learning and Cost Reductions for Generating Technologies in the National
Energy Modeling System (NEMS), Ernest Orlando Lawrence Berkeley National Laboratory, University of
California at Berkeley, Berkeley, CA. January 2004, LBNL-52559.
22 Industrial Economics, Inc. Proposed Approach for Expanding the Treatment of Learning Curve Impacts for the
Second Section 812 Prospective Analysis: Memorandum, prepared for U.S. EPA, Office of Air and Radiation,
August 13, 2007.
7-18
-------
For this RIA, we do not have the necessary data for cumulative output, fuel sales, or
emission reductions for all sectors included in our analysis in order to properly generate control
costs that reflect learning curve impacts. Clearly, the effect of including these impacts would be
to lower our estimates of costs for our control strategies in 2020, but we are not able to include
such an analysis in this RIA.
7.3.2 Influence on Regulatory Cost Estimates
Studies indicate that it is not uncommon for pre-regulatory cost estimates to be higher
than later estimates, in part because of an inability to predict technological advances. Over
longer time horizons, the opportunity for technical advances is greater.
7.3.2.1 Multi-Rule Study
Harrington et al. of Resources for the Future (RFF)23 conducted an analysis of the
predicted and actual costs of 28 federal and state rules, including 21 issued by U.S. EPA and the
Occupational Safety and Health Administration (OSHA), and found a tendency for predicted
costs to overstate actual implementation costs. Costs were considered accurate if they fell
within the analysis error bounds or if they fell within 25% (greater or less than) of the predicted
amount. They found that predicted total costs were overestimated for 14 of the 28 rules, while
total costs were underestimated for only three rules. Differences can result because of quantity
differences (e.g., overestimate of pollution reductions) or differences in per-unit costs (e.g.,
cost per unit of pollution reduction). Per-unit costs of regulations were overestimated in 14
cases, while they were underestimated in six cases. In the case of U.S. EPA rules, the Agency
overestimated per-unit costs for five regulations, underestimated them for four regulations
(three of these were relatively small pesticide rules), and accurately estimated them for four.
Based on an examination of eight rules, "for those rules that employed economic incentive
mechanisms, overestimation of per-unit costs seems to be the norm," the study said. In
addition, Harrington et al. also states that overestimation of total costs can be due to error in
the quantity of emissions reductions achieved, which would also cause the benefits to be
overestimated.
It should be noted that many (though not all) of the U.S. EPA rules examined by
Harrington et al. had compliance dates of several years, which allowed a limited period for
technical innovation. In contrast, the progress demonstration and compliance dates for a
23 Harrington, W., R.D. Morgenstern, and P. Nelson. 2000. "On the Accuracy of Regulatory Cost Estimates." Journal
of Policy Analysis and Management 19(2):297-322.
7-19
-------
attaining a NAAQS occur over a longer time horizon and could allow for possible technical
innovation.
7.3.2.2 Acid Rain S02 Trading Program
Recent cost estimates of the Acid Rain S02 trading program by RFF and MIT have been
as much as 83% lower than originally projected by EPA (see Table 7-6).24 As noted in the RIA for
the Clean Air Interstate Rule, the 1989 ex ante numbers associated with the Acid Rain Program
were an overestimate in part because of the limitation of economic modeling to predict
technological improvement of pollution controls and other compliance options, such as fuel
switching. In part, the fuel switching from high-sulfur to low-sulfur coal was spurred by a
reduction in rail transportation costs due to deregulation of rail rates during the 1990s.
Harrington et al. report that scrubbing turned out to be more efficient (95% removal vs. 80-
85% removal) and more reliable (95% vs. 85% reliability) than expected, and that unanticipated
opportunities arose to blend low- and high-sulfur coal in older boilers up to a 40/60 mixture,
compared with the 5/95 mixture originally estimated.
Table 7-6. Phase 2 Cost Estimates
Phase 2 Cost Estimates
Ex ante estimates $2.7 to $6.2 billion3
Ex post estimates $1.0 to $1.4 billion
3 2010 Phase II cost estimate in 1995$.
7.3.2.3 Chlorofluorocarbon Phase-Out
EPA used a combination of regulatory, market-based (i.e., a cap-and-trade system
among manufacturers), and voluntary approaches to phase out the most harmful ozone
depleting substances. The phase out was done more efficiently than either EPA or industry
originally anticipated. The phase out for Class I substances was implemented
4-6 years faster, included 13 more chemicals, and cost 30% less than was predicted at the time
the 1990 Clean Air Act Amendments were enacted.25
24 Carlson, Curtis, Dallas R. Burtraw, Maureen, Cropper, and Karen L Palmer. 2000. "Sulfur Dioxide Control by
Electric Utilities: What Are the Gains from Trade?" Journal of Political Economy 108(#6):1292-1326.
Ellerman, Denny. January 2003. Ex Post Evaluation of Tradable Permits: The U.S. SO2 Cap-and-Trade Program.
Massachusetts Institute of Technology Center for Energy and Environmental Policy Research.
25 Holmstead, Jeffrey, 2002. 'Testimony of Jeffrey Holmstead, Assistant Administrator, Office of Air and Radiation,
U.S. Environmental Protection Agency, Before the Subcommittee on Energy and air Quality of the committee
on Energy and Commerce, U.S. House of Representatives, May 1, 2002, p. 10.
7-20
-------
The Harrington et al. study states, "When the original cost analysis was performed for
the CFC phase-out it was not anticipated that the hydrofluorocarbon HFC-134a could be
substituted for CFC-12 in refrigeration." However, as Hammit26 notes "since 1991 most new U.S.
automobile air conditioners have contained HFC-134a (a compound for which no commercial
production technology was available in 1986) instead of CFC-12" (p. 13). Hammit cites a similar
story for HCFRC-141b and 142b, which are currently substituting for CFC-11 in important foam-
blowing applications.
7.3.3 Influence of Regulation on Technological Change
We cannot estimate the interplay between EPA regulation and technology improvement
but have reason to believe it may be significant. There is emerging research on technology-
forcing polices (i.e., where a regulator specifies a policy standard that cannot be met with
existing technology or cannot be met with existing technology at an acceptable cost, and over
time market demand will provide incentives for industry to develop the appropriate
technology). This is illustrated by Gerard and Lave (2005). They demonstrate through a careful
review of policy history that the 1970 CAA legislated dramatic improvements in the reduction of
emissions for 1975 and 1976 automobiles. Those mandated improvements went beyond the
capabilities of existing technologies. But the regulatory pressure "pulled" forth or "forced"
catalytic converting technology in 1975.
Popp (2003) and Keohane (2002) have both provided empirical evidence that Title IV led
to induced technological change. Popp provides evidence that since Title IV there has been
technological innovations that have improved the removal efficiency of scrubbers. Keohane
provides evidence that fossil-fuel fired electric utilities that were subject to Title IV were, for a
given increase in the cost of switching to low sulfur coal, more likely to install a scrubber.
7.4 Uncertainties and Limitations
EPA based its estimates of emissions control costs on the best available information
from available engineering studies of air pollution controls and developed a reliable modeling
framework for analyzing the cost, emissions changes, and other impacts of regulatory controls.
However, our cost analysis is subject to uncertainties and limitations, which we document on a
qualitative basis in Table 7-7 below. For additional discussion of how we assess uncertainty, see
Section 5.5.7.
26 Hammit, J.K. (2000). "Are the costs of proposed environmental regulations overestimated? Evidence from the
CFC phase out." Environmental and Resource Economics, 16(#3): 281-302.
7-21
-------
Table 7-7. Summary of Qualitative Uncertainty for Modeling Elements of PM Engineering
Costs
Potential Source of Uncertainty
Direction
of
Potential
Bias
Magnitude of
Impact on
Monetized
Costs3
Degree of
Confidence
in Our
Analytical
Approach13
Ability to
Assess
Uncertaintyc
Uncertainties Associated with Engineering Costs
Engineering Cost Estimates
• Capital recovery factor estimates (7% and 3%)
• Estimates of private compliance cost
• Increased advancement in control
technologies as well as reduction in costs over
time
• Cost estimates for PM10
Both Medium-high Medium
Tier 2
Unquantified Costs
• Costs of federal and state administration of SIP
program
• Transactional costs
Low
Medium
Medium
Tierl
Extrapolated Costs
Both
High
Low
Tierl
Magnitude of Impact
High—If error could influence the total costs by more than 25%
Medium—If error could influence the total costs by 5%-25%
Low—If error could influence the total costs by less than 5%
Degree of Confidence in Our Analytic Approach
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
Low—Limited data exists to support the selected approach
Ability to Assess Uncertainty (using WHO Uncertainty Framework)
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
7.5 References
Carlson, Curtis, Dallas R. Burtraw, Maureen, Cropper, and Karen L. Palmer. 2000. "Sulfur Dioxide
Control by Electric Utilities: What Are the Gains from Trade?" Journal of Political
Economy 108(#6):1292-1326.
Ellerman, Denny. January 2003. Ex Post Evaluation of Tradable Permits: The U.S. S02 Cap-and-
Trade Program. Massachusetts Institute of Technology Center for Energy and
Environmental Policy Research.
7-22
-------
Gerard, D. and L.B. Lave (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.
Gumerman, Etan and Marnay, Chris. Learning and Cost Reductions for Generating Technologies
in the National Energy Modeling System (NEMS), Ernest Orlando Lawrence Berkeley
National Laboratory, University of California at Berkeley, Berkeley, CA. January 2004,
LBNL-52559.
Hammit, J.K. (2000). "Are the costs of proposed environmental regulations overestimated?
Evidence from the CFC phase out." Environmental and Resource Economics, 16(#3): 281-
302.
Harrington, W., R.D. Morgenstern, and P. Nelson. 2000. "On the Accuracy of Regulatory Cost
Estimates." Journal of Policy Analysis and Management 19(2):297-322.
Holmstead, Jeffrey, 2002. "Testimony of Jeffrey Holmstead, Assistant Administrator, Office of
Air and Radiation, U.S. Environmental Protection Agency, Before the Subcommittee on
Energy and air Quality of the committee on Energy and Commerce, U.S. House of
Representatives, May 1, 2002, p. 10.
ICF Consulting. 2005. The Clean Air Act Amendment: Spurring Innovation and Growth While
Cleaning the Air. Washington, DC. Available at
http://www.icfi.com/Markets/Environment/doc files/caaa-success.pdf.
Industrial Economics, Inc. Proposed Approach for Expanding the Treatment of Learning Curve
Impacts for the Second Section 812 Prospective Analysis: Memorandum, prepared for
U.S. EPA, Office of Air and Radiation, August 13, 2007.
Keohane, Nathaniel. 2002. "Environmental Policy and the Choice of Abatement Technique"
Evidence from Coal-Fired Power Plants." February. Yale School of Management.
Office of Management and Budget (OMB). 2003. Circular A-4. Available at
http://www. whitehouse.gov/omb/circu la rs_a004_a-4/
Popp, David. 2003. "Pollution Control Innovations and the Clean Air Act of 1990." Autumn.
Journal of Policy Analysis and Management, 22(4), 641-60.
E.H. Pechan and Associates and Industrial Economics, Direct Cost Estimates for the Clean Air Act
Second Section 812 Prospective Analysis: Draft Report, prepared for U.S. EPA, Office of
Air and Radiation, February 2007. Available at
http://www.epa.gov/oar/sect812/mar07/direct cost draft.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2002. EPA Air Pollution Control Cost
Manual—Sixth Edition (EPA452/B-02-001). Office of Air Quality Planning and Standards,
Research Triangle Park, NC. Also Available at http://epa.gOV/ttn/catc/dirl/c allchs.pdf.
7-23
-------
U.S. Environmental Protection Agency (U.S. EPA). 2006. AirControlNET4.1 Control Measures
Documentation Report. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. Also Available at http://www.epa.gov/ttnecasl/models/
DocumentationReport.pdf.
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.
U.S. Environmental Protection Agency (U.S. EPA). 2010. Guidelines for Preparing Economic
Analysis. Office of Policy, National Center for Environmental Economics. Also Available
at http://yosemite.epa.gov/ee/epa/eed.nsf/Webpages/Guidelines.html/
$file/Guidelines.pdf
U.S. Environmental Protection Agency (U.S. EPA). 2011. The Benefits and Costs of the Clean Air
Act from 1990 to 2020. Final Report for the U.S. Environmental Protection Agency,
Office of Air and Radiation. Also available at
http://www.epa.gov/air/sect812/febll/fullreport.pdf
7-24
-------
APPENDIX 7.A
DATA TO CALCULATE COSTS TO MEET 12 U.G/M3 AND SENSITIVITY ANALYSES OF
EXTRAPOLATED COST ESTIMATES
7.A.1 PM2.5 Emission Reductions and Costs to Meet 12 (ig/m3
Table?.A.I below includes costs and information needed to calculate those costs,
by county, to meet 12 u,g/m3. The Table includes the PM2.5 emissions reductions needed to
reach the 15/35 u,g/m3level because the hybrid methodology includes a parameter that uses
the quantity of prior emissions reductions.1
Table 7.A.1 PM2.5 Emission Reductions and Costs to Meet 12 u,g/m3
FIPS
Code
06037
06065
06025
06029
06107
06047
06071
County
Los Angeles
Riverside
Imperial
Kern
Tulare
Merced
San
Bernardino
Known
Controls
PM2.5
Emissions
Reductions
to Reach Emissions
15/35 ug/m3 Reductions Costs
743 $4.5
53 $0.63
404
1,769
726
76
988
12 u.g/m3
Unknown
Controls
Emissions
Reductions
--
980
294
418
635
19
844
Hybrid
Methodology
Costs
--
$290
$7.6
$7.8
$18
$0.36
$23
1 EPA developed a model of increasing total annualized costs for controlling PM2.5 emissions-- ax2 +bx + c where x
is the tons of a particular pollutant to be reduced in a particular area and a, b, and c are constants. For a
particular geographic area o is N/E where (i) N is a national, initial annualized cost/ton of $15,000 per ton and
(ii) E is, by geographic area, the denominator and represents all particulate emission reductions achieved (from
applying known and unknown controls to obtain the 15/35 baseline, as well as known controls to achieve the
alternative standard) prior to estimating needed emission reductions from unknown controls to achieve the
alternative standard.
7.A-1
-------
7.A.2 Sensitivity Analyses of Extrapolated Cost Estimates
Because of the uncertainties associated with estimating costs for the PM2.5 NAAQS and
because a significant portion of the estimated emissions reductions and related costs for
attaining the NAAQS come from unknown controls, it is important to test the sensitivity of the
assumptions applied to estimate unknown controls. The sensitivity analyses below are included
to help characterize the uncertainty for the cost estimates from unknown controls and the
responsiveness of the cost estimates to varying parameter estimates and assumptions. Note
that the tables below include cost estimates associated with unknown controls and not total
cost estimates.
While there are many approaches to sensitivity analysis, we selected analyses below,
keeping emissions estimates constant, to show variability in the cost estimates and remain
consistent with the benefits analysis. Note that the extrapolated cost estimates are provided
using a 7 percent discount rate because known control measure information is available at 7
percent for all measures applied in this analysis.
7.A.2.1 Sensitivity Analysis of Fixed-Cost Methodology
Table 7.A.2 below presents the sensitivity analysis of the fixed-cost methodology and
includes, by region and revised and alternative standard, the primary cost estimate of
$15,000/ton. The Table also includes, by region and revised and alternative standard, cost
estimates using $10,000/ton and $20,000/ton. For the revised standard of 12/35, the total cost
estimate associated with unknown control costs ranges from $32 million to $64 million,
depending on the fixed-cost-per-ton assumed.
Table 7.A.2 Sensitivity Analysis of Fixed-Cost Methodology for Unknown Controls by Revised
and Alternative Standard Analyzed (millions of 2010$)a
Extrapolated Costs
Revised &
Alternative
Standard
13 ng/m3
12 ng/m3
Region
East
West
California
Total
East
West
$10,000/ton
7%
—
—
$6.7
$6.7
—
—
$15,000/ton
7%
—
—
$10
$10
—
—
$20,000/ton
7%
—
—
$13
$13
—
—
7.A-2
-------
California
Total
11 ug/m3 East
West
California
Total
$32
$32
$48
$0.86
$97
$150
$48
$48
$71
$1.3
$150
$220
$64
$64
$95
$1.7
$190
$290
Estimates are rounded to two significant figures.
7.A.2.2 Sensitivity Analysis of Alternative Hybrid Methodology
Table 7.A.3 below presents the sensitivity analysis of the alternative hybrid
methodology. To be consistent with the sensitivity analysis of the fixed-cost methodology, the
Table also includes, by region and revised and alternative standard, cost estimates using
alternate parameter estimates for the initial cost per ton. For the revised standard of 12/35,
the total cost estimate associated with unknown control costs ranges from $230 million to $460
million.
Table 7.A.3 Sensitivity Analysis of Hybrid Methodology for Unknown Controls by Revised and
Alternative Standard Analyzed (millions of 2010$)a
Revised &
Alternative
Standard
13 ug/m3
12 ug/m3
11 ug/m3
Region
East
West
California
Total
East
West
California
Total
East
West
California
Total
Extrapolated
$10,000/ton
7%
—
—
$69
$69
—
—
$230
$230
$430
$2.2
$630
$1,100
Costs
$15,000/ton
7%
—
—
$100
$100
—
—
$340
$340
$650
$3.3
$940
$1,600
$20,000/ton
7%
—
—
$140
$140
—
—
$460
$460
$870
$4.4
$1,300
$2,100
Estimates are rounded to two significant figures.
7.A-3
-------
CHAPTER 8
COMPARISON OF BENEFITS AND COSTS
8.1 Synopsis
This chapter compares estimates of the benefits with costs and summarizes the net
benefits of the revised annual standard of 12 u.g/m3 and the alternative annual standards of 13
u.g/m3 and 11 u.g/m3 relative to the analytical baseline that includes recently promulgated
national regulations and additional emissions reductions needed to attain the existing 15/35
u.g/m3 standards, as well as adjustments to NOX emissions in the San Joaquin and South Coast
areas.
8.2 Comparison of Benefits and Costs
The EPA's illustrative analysis has estimated the health and welfare benefits and costs
associated with the revised annual PM NAAQS. The results in Table 8-1 for 2020 suggest there
will be significant health and welfare benefits and these benefits will outweigh the costs
associated with the illustrative control strategies in 2020. In the analysis, we estimate the net
benefits of the revised annual PM2.5 standard of 12 u.g/m3 and alternative annual standards of
13 u.g/m3 and 11 u.g/m3, incremental to the 2020 analytical baseline. For the revised annual
standard of 12 u.g/m3, net benefits are estimated to be $3.7 billion to $9 billion at a 3% discount
rate and $3.3 billion to $8.1 billion at a 7% discount rate in 2020 (2010 dollars). For an
alternative annual standard of 13 u.g/m3, net benefits are estimated to be $1.2 billion to $2.9
billion at the 3% discount rate and $1.1 billion to $2.6 billion at the 7% discount rate. Net
benefits of an alternative annual PM2.5 standard of 11 u.g/m3are estimated to be $11 billion to
$29 billion at a 3% discount rate and $10 billion to $26 billion at a 7% discount rate in 2020.
For the revised annual standard of 12 u.g/m3, the EPA estimates that the benefits of full
attainment exceed the costs of full attainment by 11 to 154 times at a 7% discount rate. For the
alternative annual standard of 13 u.g/m3, the EPA estimates that the benefits of full attainment
exceed the costs of full attainment by 11 to 246 times at a 7% discount rate. For the alternative
annual standards of 11 u.g/m3, the EPA estimates that the benefits of full attainment exceed the
costs of full attainment by 7 to 81 times at a 7% discount rate.
8-1
-------
Table 8-1. Total Monetized Benefits, Total Costs, and Net Benefits in 2020 (millions of
2010$)—Full Attainment3
Revised
Annual
Standard
12
Total
3%
Discount
Ratec
$53 to
$350
Costs"
7% Discount
Rate
$53 to 350
Monetized
3%
Discount
Rate
$4,000 to
$9,100
Benefits d
7% Discount
Rate
$3,600 to
$8,200
Net
3% Discount
Rateb
$3,700 to
$9,000
Benefits
7% Discount
Rate
$3,300 to
$8,100
Alternative
Standards
13
11
$11 to $100
$320 to
$1,700
$11 to $100
$320 to
$1,700
$1,300 to
$2,900
$13,000 to
$29,000
$1,200 to
$2,600
$12,000 to
$26,000
$1,200 to
$2,900
$11,000 to
$29,000
$1,100 to
$2,600
$10,000 to
$26,000
a These estimates reflect incremental emissions reductions from an analytical baseline that gives an "adjustment"
to the San Joaquin and South Coast areas in California for NOX emissions reductions expected to occur between
2020 and 2025, when those areas are expected to demonstrate attainment with the revised standards. Full
benefits of the revised standards in those two areas will not be realized until 2025.
b The two cost estimates do not represent lower- and upper-bound estimates but represent estimates generated
by two different methodologies. The lower estimate is generated using the fixed-cost methodology, which
assumes that technological change and innovation will result in the availability of additional controls by 2020 that
are similar in cost to the higher end of the cost range for current, known controls. The higher estimate is generated
using the hybrid methodology, which assumes that while additional controls may become available by 2020, they
become available at an increasing cost and the increasing cost varies by geographic area and by degree of difficulty
associated with obtaining the needed emissions reductions.
c Due to data limitations, we were unable to discount compliance costs for all sectors at 3%. See Chapter 7, Section
7.2.2 for additional details on the data limitations. As a result, the net benefit calculations at 3% were computed by
subtracting the costs at 7% from the monetized benefits at 3%.
d The reduction in premature deaths each year accounts for over 90% of total monetized benefits. Mortality risk
valuation assumes discounting over the SAB-recommended 20-year segmented lag structure. Not all possible
benefits or disbenefits are quantified and monetized in this analysis. B is the sum of all unquantified 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. The range of benefits reflects the range of the central
estimates from two mortality cohort studies (i.e., Krewski et al. [2009] to Lepeule et al. [2012]).
Figure 8-1 demonstrates the size of the benefits relative to costs for the revised annual
standards of 12 u.g/m3 at the 7% discount rate. This figure shows benefits for two different
mortality studies and costs using two methods for extrapolating costs to emissions reductions
associated with unknown controls.
8-2
-------
Benefits using Lepeule etal. (2012)
$8.2 billion*
Benefits using Krewskietal. (2009)
S3.6 billion*
Costs using Hybrid Approach
$350 million
z
Costs using Fixed Costs
$53 million
Mote: Relative size of benefits and costs are to scale.
Figure 8-1. Monetized Benefit to Cost Comparison for the Revised Annual Standard of 12
u.g/m3 in 2020 (7% Discount Rate)
Note: Relative size of benefits and costs are to scale.
Figure 8-2 displays the range of net benefits for the selected standards using the two
epidemiology functions and 12 expert elicitation functions for PM-related premature mortality
that the EPA employs in its analysis of benefits. As shown in the figure, the benefits exceed
costs in every combination analyzed.
8-3
-------
$7.0 n
$6.0 •
$5.0 •
S$4.0
5 $3.0
$2.0 •
$1.0 •
$0.0
1
Lepeule etal.
Krewskiet al.
II
2 Cost estimates combined with total monetized benefits estimates derived from 2
epidemiology functions and 12 expert functions
Figure 8-2. Net Benefits for Revised Annual Standard of 12 u.g/m3 in 2020 at a 7% Discount
Rate
Due to data and methodology limitations, the EPA was unable to quantify some health
benefits associated with exposure to PM2.5, as well as the additional co-benefits from
improvements in welfare effects associated with emission reductions to attain the primary
standard, such as visibility. Tables 8-2 and 8-3 summarize the human health and welfare
categories contained within the core benefits estimate as well as those categories that are
unquantified in the core benefits estimate. Because the illustrative emission reduction strategy
for the revised annual standard at 12 u.g/m3 consisted of only directly emitted PM2.5, these
tables are limited to only those categories associated with emission reductions of directly
emitted PM2.5.1 It is important to emphasize that the list of unquantified benefit categories is
not exhaustive, nor is quantification of each effect complete.
1 For the unquantified benefits associated with emission reductions of NOX and SO2, please see Chapters 5 and 6.
8-4
-------
Table 8-2. Human Health Effects of Ambient PM2.5
Benefits Category
Specific Effect
Effect Has
Been
Quantified
Effect Has
Been
Monetized
More
Information
Improved Human Health
Reduced
incidence of
premature
mortality from
exposure to PM2.5
Adult premature mortality based on cohort
study estimates and expert elicitation
estimates (age >25 or age >30)
Infant mortality (age <1)
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)
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Section 5.6
Emergency department visits for
cardiovascular effects (all ages)
Strokes and cerebrovascular disease (age 50-
79)
Other cardiovascular effects (e.g., other 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
Section 5.6
Section 5.6
PM ISA
PM ISAb
PM ISA
b,c
PM ISA
b,c
aWe quantify these benefits in a sensitivity analysis, but not in the core analysis.
bWe assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
°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.
8-5
-------
Table 8-3. Welfare Co-Benefits of PM2.5
Benefits Category
Effect Has
Been
Specific Effect Quantified
Effect Has
Been More
Monetized Information
Improved Environment
Reduced visibility impairment
Reduced climate effects
Reduced effects on materials
Reduced effects from PM
deposition (metals and
organics)
Visibility in Class 1 areas in SE, SW, a
and CA regions
Visibility in Class 1 areas in other —
regions
Visibility in 8 cities —
Visibility in other residential areas —
Climate impacts from PM —
Household soiling —
Materials damage (e.g., corrosion, —
increased wear)
Effects on Individual organisms —
and ecosystems
Section 6.3,
Appendix 6b
a Section 6.3,
Appendix 6b
Section 6.3,
Appendix 6b
Section 6.3,
Appendix 6b
— Section 6.5, PM
ISAb
— Section 6.4, PM
ISAb
— Section 6.4, PM
ISA0
— Section 6.6.1,
PM ISAb
a We quantify these co-benefits in an illustrative analysis, but these results of that illustrative scenario are not an
estimate of the co-benefits for the revised primary standard.
b We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
8.3 Discussion and Conclusions
An extensive body of scientific evidence documented in PM ISA indicates that PM2.5 can
penetrate deep into the lungs and cause serious health effects, including premature death and
other non-fatal illnesses (U.S. EPA, 2009). As described in the preamble to the rule, the
revisions to the standards are based on an integrative assessment of an extensive body of new
scientific evidence (U.S. EPA, 2009). Health studies published since the PM ISA (e.g., Pope et al.
[2009]) confirm that recent levels of PM2.s have had a significant impact on public health. Based
on the air quality analysis in this RIA, the EPA projects that nearly all counties with PM2.5
monitors in the United States would meet an annual standard of 12 u.g/m3without additional
Federal, State, or local PM control programs. This demonstrates the substantial progress that
the United States has made in reducing air pollution emissions over the last several decades.
8-6
-------
Regulations such as the EPA's recent Mercury and Air Toxics Standards (MATS) and other
Federal programs such as diesel standards will provide substantial improvements in regional
concentrations of PM2.s. Our analysis shows a few areas would still need additional emissions
reductions to address local sources of air pollution, including ports and uncontrolled industrial
emissions. For this reason, we have designed the RIA analysis to focus on local controls in these
few areas. We estimate that these additional local controls would yield benefits well in excess
of costs.
The setting of a NAAQS does not compel specific pollution reductions and as such does
not directly result in costs or benefits. For this reason, NAAQS RIAs are merely illustrative. The
NAAQS RIAs illustrate the potential costs and benefits of additional steps States could take to
attain a revised air quality standard nationwide beyond rules already on the books. We base
our illustrative estimates on an array of emission control strategies for different sources. The
costs and benefits identified in this RIA will not be realized until specific controls are mandated
by SIPs or other Federal regulations. In short, NAAQS RIAs hypothesize, but do not prescribe,
the control strategies that States may choose to enact when implementing a revised NAAQS.
It is important to emphasize that the EPA does not "double count" the costs or the
benefits of our rules. Emission reductions achieved under rules that require specific actions
from sources—such as MATS—are in the baseline of this NAAQS analysis, as are emission
reductions needed to meet the current NAAQS. For this reason, the cost and benefits estimates
provided in this RIA and all other NAAQS RIAs should not be added to the estimates for
implementation rules.
In calculating the costs, the EPA assumed the application of a significant number of
unidentified future controls that would make possible the additional emissions reductions
needed for attainment in 2020. EPA used two methodologies—the fixed-cost and hybrid
methodologies—for estimating the costs of unidentified future controls, and both approaches
assume either that existing technologies can be applied in particular combinations or to specific
sources that we currently can't predict or that innovative strategies and new control options
make possible the emissions reductions needed for attainment by 2020. Estimates generated
by the two approaches do not represent lower- and upper-bound estimates but simply
represent estimates generated by two different methodologies. The fixed-cost methodology
implicitly assumes that technological change and innovation will result in the availability of
additional controls by 2020 that are similar in cost to the higher end of the cost range for
current controls. The hybrid methodology implicitly assumes that while additional controls
become available by 2020, they become available at an increasing cost and the increasing cost
8-7
-------
varies by geographic area and by degree of difficulty associated with obtaining the needed
emissions reductions.
For the revised annual standard of 12 u.g/m3, the total cost estimates comprise between
90 and 97% extrapolated cost estimates, and the estimated total cost using the hybrid
methodology is roughly 6.5 times more than the estimated total cost using the fixed-cost
methodology. Because the hybrid methodology reflects increasing marginal costs in areas
needing a higher ratio of emissions reductions from unknown to known controls, it could be
more representative of total costs. In an effort to consider the potential fitness of the
extrapolated cost estimates, we reviewed the South Coast Air Quality Management District's
(SCAQMD) 2012 Air Quality Management Plan (AQMP), and we located data on recent emission
reduction credit (ERC) transactions in both the SCAQMD and San Joaquin Valley Air Pollution
Control District (SJV APCD). While this information provides context for the extrapolated cost
estimates, the current relationship between available controls and costs to reduce emissions
may or may not be applicable in 2020 because of changes in innovation and advances in
technology.
The SCAQMD's 2012 AQMP includes information on control measures to meet the
current 24-hour standard of 35 u.g/m3, including further PM2.s controls for under-fired
charbroilers at a cost per ton reduced of $15,000. This control cost matches the parameter used
in the fixed-cost methodology, as well as the initial value used for the hybrid methodology and
is supportive of our selection of that value. In addition, the California Air Resources Board's
2009 and 2010 Emission Reduction Offset Transaction Costs, Summary Report included PM-m
ERC prices in both the SCAQMD and the SJV APCD. To some degree, ERC transaction prices
reflect a choice between installing a more stringent control and purchasing ERCs. Between 2009
and 2010 PMin ERC prices in SJV APCD ranged from $40,000 per ton per year (tpy) to
$70,000/tpy, and PMin ERC prices in the SCAQMD ranged from $575,000/tpy to more than $1.9
million/tpy. These prices reflect both marginal costs that are higher than the fixed-cost
estimates and marginal costs that are not inconsistent with the higher cost estimates generated
using the hybrid methodology. For further discussion of the total cost estimates, refer to
Section 7.2.4 in Chapter 7 of this RIA.
Furthermore, the monetized benefits estimates presented in this RIA are not intended
to capture the full burden of PM to public health but rather represent the incremental benefits
expected upon attaining the revised annual primary standard of 12 u.g/m3. In comparison,
modeling by Fann et al. (2012) estimated that 2005 levels of air pollution were responsible for
between 130,000 and 320,000 PM2.5-related deaths, or between 6.1% and 15% of total deaths
8-8
-------
from all causes in the continental United States. The monetized benefits associated with
attaining the proposed range of standards appear modest when viewed within the context of
the potential overall public health burden of PM2.5and ozone air pollution estimated by Fann et
al. (2012), but this is primarily because regulations already on the books will make great strides
toward reducing future levels of PM. One important distinction between the total public health
burden estimated for 2005 air pollution levels and the estimated benefits in this RIA is that
ambient levels of PM2.5 will have improved substantially by 2020, due to major emissions
reductions resulting from implementation of Federal regulations. For example, we estimate
that S02 emissions (an important PM2.5 precursor) in the United States would fall from 14
million tons in 2005 to less than 5 million tons by 2020 (a reduction of 66%). For this reason,
States will only need to achieve small air quality improvements to reach the proposed PM
standards. As shown in the recent RIA for MATS (U.S. EPA, 2011b), implementing other Federal
and State air quality actions will address a substantial fraction of the total public health burden
of PM2.5 and ozone air pollution.
The NAAQS are not set at levels that eliminate the risk of air pollution. 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 PM 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, 2010c). While benefits occurring below the standard
are assumed to be more uncertain than those occurring above the standard, the EPA considers
these to be legitimate components of the total benefits estimate. Although there are greater
uncertainties at lower PM2.5 concentrations, there is no evidence of a population-level
threshold in PM2.5-related health effects in the epidemiology literature.
Lastly, the EPA recognizes that there are uncertainties in both the cost and benefit
estimates provided in this RIA. The EPA was unable to monetize fully all of the benefits
associated with reaching these standards in this RIA, including other health effects of PM,
visibility effects, ecosystem effects, and climate effects. If the EPA were able to monetize all of
the benefits, the benefits would exceed the estimated costs by an even greater margin.
8.4 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.
8-9
-------
Farm, N., A. Lamson, K. Wesson, D. Risley, S.C. Anenberg, and B.J. Hubbell. 2012. "Estimating
the National Public Health Burden Associated with Exposure to Ambient PM2.5 and
Ozone. Risk Analysis." Risk Analysis 32(1): 81-95.
Krewski, D., R.T. Burnett, M.S. Goldbert, K. Hoover, J. Siemiatycki, M. Jerrett, M. Abrahamowicz,
and W.H. White. 2009. "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.
Lepeule, J., F. Laden, D. Dockery, and J. Schwartz. 2012. "Chronic Exposure to Fine Particles and
Mortality: An Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009."
Environ Health Perspect. In press. Available at: http://dx.doi.org/10.1289/ehp.1104660.
Pope, C.A. Ill, E. Majid, and D. Dockery. 2009. "Fine Particle Air Pollution and Life Expectancy in
the United States." New England Journal of Medicine 360: 376-386.
U.S. Environmental Protection Agency (U.S. EPA). 2006. Regulatory Impact Analysis, 2006
National Ambient Air Quality Standards for Particulate Matter, Chapter 5. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. October. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%205-Benefits.pdf.
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—RTP Division. December. Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546.
U.S. Environmental Protection Agency (U.S. EPA). 2010. 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 on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011a. 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:
http://www.epa.gov/airtransport/pdfs/FinalRIA.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011b. Regulatory Impact Analysis for the
Final Mercury and Air Toxics Standards. EPA-452/R-11-011. December. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/matsriafinal.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2011c. 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:
http://www.epa.gov/oar/sect812/febll/fullreport.pdf.
8-10
-------
U.S. Environmental Protection Agency (U.S. EPA). 2011d. Policy Assessment for the Review of
the Particulate Matter National Ambient Air Quality Standards. EPA-452/D-11-003. April.
Available at: http://www.epa.gov/ttnnaaqs/standards/pm/s_pm_2007_pa.html.
8-11
-------
CHAPTER 9
STATUTORY AND EXECUTIVE ORDER REVIEWS
9.1 Synopsis
This chapter summarizes the Statutory and Executive Order (EO) impact analyses
relevant for the PM NAAQS Regulatory Impact Analysis. For each EO and Statutory requirement
we describe both the requirements and the way in which our analysis addresses these
requirements.
9.2 Executive Order 12866: Regulatory Planning and Review
Under section 3(f)(l) of Executive Order 12866 (58 FR 51735, October 4, 1993), this
action is an "economically significant regulatory action" because it is likely to have an annual
effect on the economy of $100 million or more. The $100 million threshold can be triggered by
either costs or benefits, or a combination of them. Accordingly, the EPA submitted this action to
the Office of Management and Budget (OMB) for review under Executive Orders 12866 and
13563 (76 FR 3821, January 21, 2011), and any changes made in response to OMB
recommendations have been documented in the docket for this action.
9.3 Paperwork Reduction Act
This action does not impose an information collection burden under the provisions of
the Paperwork Reduction Act, 44 U.S.S. 3501 et seq. Burden is defined at 5 CFR 1320.3(b).
There are no information collection requirements directly associated with revisions to a NAAQS
under section 109 of the CAA.
9.4 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 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 this 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.
9-1
-------
After considering the economic impacts of this final rule on small entities, I certify that
this action will not have a significant economic impact on a substantial number of small entities.
This final rule will not impose any requirements on small entities. Rather, this rule establishes
national standards for allowable concentrations of particulate matter 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).
9.5 Unfunded Mandates Reform Act
This action contains no Federal mandates under the provisions of Title II of the
Unfunded Mandates Reform Act of 1995 (UMRA), 2 U.S.C. 1531-1538 for state, local, or tribal
governments or the private sector. The action imposes no enforceable duty on any state, local
or tribal governments or the private sector. Therefore, this action is not subject to the
requirements of sections 202 or 205 of the UMRA.
This action is also not subject to the requirements section 205 of the UMRA because it
contains no regulatory requirements that might significantly or uniquely affect small
governments. This action 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 rule contains
no regulatory requirements that might significantly or uniquely affect small governments.
Furthermore, in setting a NAAQS, 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 pursuant to the Unfunded Mandates Reform Act would not furnish
any information which the court could consider in reviewing the NAAQS). 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, the EPA will address, as appropriate, unfunded mandates if
and when it proposes any revisions to 40 CFR parts 51 or 58.
9.6 Executive Order 13132: Federalism
This action 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
9-2
-------
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 and review NAAQS; however, CAA section 116 preserves the rights of
states to establish more stringent requirements if deemed necessary by a state. Furthermore,
this final rule does not impact CAA section 107 which establishes that the states have primary
responsibility for implementation of the NAAQS. Finally, as noted in section D (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 action.
However, as also noted in section D (above) on UMRA, the EPA recognizes that states
will have a substantial interest in this rule and any corresponding revisions to associated air
quality surveillance requirements, 40 CFR part 58.
9.7 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
national standards to address the health and welfare effects of particulate matter. Historically,
the EPA's definition of "tribal implications" has been limited to situations in which it can be
shown that a rule has impacts on the tribes' ability to govern or implications for tribal
sovereignty. Based on this historic definition, this action does not have Tribal implications, as
specified in Executive Order 13175 (65 FR 67249, November 9, 2000), i.e. because 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. Nevertheless, we were aware that many tribes would be
interest in this rule and we undertook a number of outreach activities to inform tribes about
the PM NAAQS review and offered to two consultations with tribes.
Although Executive Order 13175 does not apply to this rule, the EPA undertook a
consultation process including: prior to proposal on March 29,2012 we sent letters to tribal
leadership inviting consultation on the rule and then sent a second round of letters offering
consultation after the proposal was issued on June 29, 2012. We conducted outreach and
information calls to tribal environmental staff on May 9, 2012; June 15, 2012; and August 1,
2012. We also participated on the National Tribal Air Association call on June 28, 2012.
9-3
-------
As a result we received comments from the National Tribal Air Association, the Southern
Ute Mountain Ute Tribe, and the Navajo Nation EPA.
9.8 Executive Order 13045: Protection of Children from Environmental Health and Safety
Risks
This action is subject to Executive Order 13045 (62 FR 19885, April 23, 1997) because it
is an economically significant regulatory action as defined by Executive Order 12866, and the
EPA believes that the environmental health or safety risk addressed by this action may have a
disproportionate effect on children. Accordingly, we have evaluated the environmental health
or safety effects of PM exposures on children. The protection offered by these standards may
be especially important for children because childhood represents a lifestage associated with
increased susceptibility to PM-related health effects. Because children have been identified as
an at-risk population, we have carefully evaluated the environmental health effects of exposure
to PM pollution among children. Discussions of the results of the evaluation of the scientific
evidence and policy considerations pertaining to children are contained in sections III.B, III.D,
III.E, IV.B, and IV.C of the rule's preamble.
9.9 Executive Order 13211: Actions that Significantly Affect Energy Supply, Distribution or
Use
This action is not a "significant energy action" as defined in Executive Order 13211, (66
FR 28355, May 22, 2001) because it is not likely to have a significant adverse effect on the
supply, distribution, or use of energy. The purpose of this action concerns the review of the
NAAQS for PM. The action does not prescribe specific pollution control strategies by which
these ambient standards will be met. Such strategies are 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.
9.10 National Technology Transfer and Advancement Act
Section 12(d) of the National Technology Transfer and Advancement Act of 1995
(NTTAA), Public Law 104-113, section 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.
9-4
-------
This final rulemaking involves technical standards for environmental monitoring and
measurement. Specifically, the EPA proposes to retain the indicators for fine (PM2.5) and
coarse (PM10) particles. The indicator for fine particles is measured using the Reference
Method for the Determination of Fine Particulate Matter as PM2.5 in the Atmosphere
(appendix L to 40 CFR part 50), which is known as the PM2.5 FRM, and the indicator for coarse
particles is measured using the Reference Method for the Determination of Particulate Matter
as PM10 in the Atmosphere (appendix J to 40 CFR part 50), which is known as the PM10 FRM.
To the extent feasible, the EPA employs a Performance-Based Measurement System
(PBMS), which does not require the use of specific, prescribed analytic methods. The PBMS is
defined as a set of processes wherein the data quality needs, mandates or limitations of a
program or project are specified, and serve as criteria for selecting appropriate methods to
meet those needs in a cost-effective manner. It is intended to be more flexible and cost
effective for the regulated community; it is also intended to encourage innovation in analytical
technology and improved data quality. Though the FRM defines the particular specifications for
ambient monitors, there is some variability with regard to how monitors measure PM,
depending on the type and size of PM and environmental conditions. Therefore, it is not
practically possible to fully define the FRM in performance terms to account for this variability.
Nevertheless, our approach in the past has resulted in multiple brands of monitors being
approved as FRM for PM, and we expect this to continue. Also, the FRMs described in 40 CFR
part 50 and the equivalency criteria described in 40 CFR part 53, constitute a performance-
based measurement system for PM, since methods that meet the field testing and performance
criteria can be approved as FEMs. Since finalized in 2006 (71 FR, 61236, October 17, 2006) the
new field and performance criteria for approval of PM2.5 continuous FEMs has resulted in the
approval of six approved FEMs. In summary, for measurement of PM2.5 and PM10, the EPA
relies on both FRMs and FEMs, with FEMs relying on a PBMS approach for their approval. The
EPA is not precluding the use of any other method, whether it constitutes a voluntary
consensus standard or not, as long as it meets the specified performance criteria.
9.11 Executive Order 12898: Federal Actions to Address Environmental Justice in Minority
Populations and Low-Income Populations
Executive Order 12898 (59 FR 7629, February 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
9-5
-------
or environmental effects of their programs, policies, and activities on minority populations and
low-income populations in the United States.
The EPA maintains an ongoing commitment to ensure environmental justice for all
people, regardless of race, color, national origin, or income. Ensuring environmental justice
means not only protecting human health and the environment for everyone, but also ensuring
that all people are treated fairly and are given the opportunity to participate meaningfully in
the development, implementation, and enforcement of environmental laws, regulations, and
policies.
The EPA has identified potential disproportionately high and adverse effects on minority
and/or low-income populations related to PM2.5 exposures. In addition, the EPA has identified
persons from lower socioeconomic strata as an at-risk population for PM-related health effects.
As a result, the EPA has carefully evaluated the potential impacts on low-income and minority
populations as discussed in section III.E.3.a of the rule's preamble. Based on this evaluation and
consideration of public comments on the proposal, the EPA is eliminating the spatial averaging
provisions as part of the form of the annual standard to avoid potential disproportionate
impacts on at-risk populations. The Agency expects this final rule will lead to the establishment
of uniform NAAQS for PM. The Integrated Science Assessment and Policy Assessment contain
the evaluation of the scientific evidence and policy considerations that pertain to these
populations. These documents are available as described in the Supplementary Information
section of the rule's preamble and copies of all documents have been placed in the public
docket for this action.
9-6
-------
CHAPTER 10
QUALITATIVE DISCUSSION OF EMPLOYMENT IMPACTS OF AIR QUALITY REGULATIONS
10.1 Introduction
Executive Order 13563 states that federal agencies should consider the effect of
regulations on employment. According to the Executive Order, "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"
(Executive Order 13563, 2011). Although a stand-alone analysis of employment impacts is not
typically included in a standard cost-benefit analysis,1 employment impacts are currently of
particular concern due to recent economic conditions reflecting relatively high levels of
unemployment. This chapter provides a context for considering the potential influence of
environmental regulation on growth and job shifts in the U.S. economy. Section 10.2 addresses
the particular influence of this proposed rule on employment. Section 10.3 presents a
descriptive overview of the peer-reviewed literature relevant to evaluating the effect of air
quality regulation on employment. Finally, in Section 10.4, we offer several conclusions.
10.2 Influence of NAAQS Controls on Employment
Peer-reviewed econometric studies that estimate the impact of air quality regulation on
net overall employment and within the regulated sector converge on the finding that any net
employment effects, whether positive or negative, have been small. This finding holds for even
major nationwide environmental regulations. Therefore, given the overall small effect
environmental regulations have been shown to have on net employment in the regulated
sectors, we do not expect them to have a significant impact on the overall economy.
Estimating specific employment impacts from a new NAAQS standard is particularly
challenging for two reasons. First, the NAAQS targets a level of public health protection that
individual areas have flexibility to meet in a variety of ways, and the primary regulatory activity
and implementation occur at the state or local level. Under these circumstances, states and
localities are given considerable flexibility in choosing which strategies to adopt to meet the
NAAQS target. State and local officials can consider a variety of economic impacts including
employment impacts of various control strategies, as well as other factors, when designing
their state implementation plans (SIPs). This makes it challenging to predict how specific sectors
will be impacted and how those impacts vary across regions of the country. Analyses in the RIA
are based on a particular NAAQS compliance scenario that reflects assumptions about control
1 This is the case except to the extent that labor costs are part of total costs in a cost-benefit analysis.
10-1
-------
measures applied across all sectors and locations, specific control strategies adopted by the
states, and associated extrapolated costs. EPA believes this compliance scenario supports
reasonably illustrative quantitative estimates of the potential overall economic effects of the
revised NAAQS. However, EPA does not consider this illustrative, aggregate compliance
scenario to be sufficiently certain and precise to support quantitative projections of outcomes
in particular locations, sectors, or markets, including labor markets, in light of the scarcity of
applicable studies that can be used to generate such estimates. Therefore, this RIA does not
include quantitative projections of aggregate shifts in employment.
Second, we anticipate that national employment levels will be changing during the
period that the NAAQS is being implemented, a period that may be greater than 10 years for
some areas, following designations of nonattainment. Although current unemployment rates
remain high relative to historical averages largely due to the sharp increase in unemployment
that began in early 2008 (U.S. Department of Labor, Bureau of Labor Statistics, 2012a), current
data suggest unemployment rates have been declining in recent months (U.S. Department of
Labor, Bureau of Labor Statistics, 2012b). Policies to meet the NAAQS in all areas will not go
into effect for several years. By this time, we anticipate the economy will have had a chance to
recover toward higher employment levels that more closely approximate full employment. In
addition, over a period of 10 years or longer, potentially significant changes in technology,
growth and distribution of economic activities, and other key determinants of local and national
labor market conditions further complicate projections of future employment and the potential
incremental effect of regulatory programs.
Although a quantitative assessment of employment consequences of today's proposed
revision to the national ambient PM standards remains beyond the reach of available data and
modeling tools, EPA is in the process of supporting the development of tools and research that
could assist in the future. In the interim, some insights on the potentially relevant
consequences of revising ambient air pollution standards can be gained by considering
currently available literature, including its limitations. In light of these challenges, Section 10.3
focuses on qualitative insights from currently available peer-reviewed literature on the impact
of air quality regulations in general.
10.3 The Current State of Knowledge Based on the Peer-Reviewed Literature
There is limited peer-reviewed econometric literature estimating employment effects of
environmental regulations. We present an overview here, highlighting studies with particular
relevance for NAAQS. Determining the direction of employment effects in the regulated
10-2
-------
industries is challenging because of competing effects. Complying with the new or more
stringent regulation requires additional inputs, including labor, and may alter the relative
proportions of labor and capital used by regulated firms in their production processes.
When the economy is at full employment, an environmental regulation is unlikely to
have a considerable impact on net employment in the long run. Instead, labor would primarily
be reallocated from one productive use to another (e.g., from producing electricity or steel to
producing pollution abatement equipment). Theory supports the argument that, in the case of
full employment, the net national employment effects from environmental regulation are likely
to be small and transitory (e.g., as workers move from one job to another). There is reason to
believe that when the economy is operating at less than full employment environmental
regulation could result in a short-run net increase in employment.2 Several empirical studies
suggest that net employment impacts may be positive but small even in the regulated sector.
Taken together, the peer-reviewed literature does not contain evidence that environmental
regulation would have a notable impact on net employment across the whole economy.
This discussion focuses on both short- and long-term employment impacts in the
regulated industries, as well as on the environmental protection sector for construction of
needed pollution control equipment prior to the compliance date of the regulation. EPA is
committed to using the best available science and the relevant theoretical and empirical
literature in this assessment and is pursuing efforts to support new research in this field.
10.3.1 Immediate and Short-Run Employment Impacts
Environmental regulations are typically phased in to allow firms time to invest in the
necessary technology and process changes to meet the new standards. Whatever effects a
regulation will have on employment in the regulated sector will typically occur only after a
regulation takes effect or in the long term, as new technologies are introduced. However, the
environmental protection sector (pollution control equipment) often sees immediate
employment effects. When a regulation is promulgated, the first response of industry is to
order pollution control equipment and services to comply with the regulation when it becomes
effective. This can produce a short-term increase in labor demand for specialized workers
within the environmental protection sector related to design, construction, installation, and
operation of the new pollution control equipment required by the regulation (see Schmalansee
and Stavins, 2011; Bezdek, Wendling, and Diperna, 2008).
2 See Schmalansee and Stavin (2011).
10-3
-------
As the NAAQS are implemented, it is possible that the regulated sector will experience
short-run changes in employment. Because it is the states' responsibility to design their SIPs
over the next few years, we cannot assess the short-term effects of those SIPs on the regulated
sector with sufficient precision to quantify the resulting incremental effects on employment.
However, as previously noted, even in a full employment case, there may be transitory effects
as workers change jobs. Some workers may need to retrain or relocate in anticipation of the
new requirements or require time to search for new jobs, while shortages in some sectors or
regions could bid up wages to attract workers.
It is important to recognize that these adjustment costs can entail local labor
disruptions, and, although the net change in the national workforce might be small, gross
reductions in employment can still have negative impacts on individuals and communities. The
peer-reviewed literature that is currently available is focused on medium- and long-term
employment impacts and does not offer much insight into the short-term balance between
increased employment in the environmental protection sector and possible decreased
employment in some regulated sectors.
10.3.2 Long-Term Employment Impacts on the Regulated Industry
Determining the direction of net employment effects in regulated industries is
challenging because of competing effects. Morgenstern, Pizer, and Shih (2002) discuss how
environmental regulations can be understood as requiring regulated firms to add a new output
(environmental quality) to their product mix. Although legally compelled to produce this new
output, regulated firms have to finance this additional production input with the proceeds of
sales of their other (market) products. The current literature on employment impacts of air
quality regulations can be disaggregated into two types of approaches or models: 1) structural
and 2) reduced-form models. Two papers that present a formal structural model of the
underlying profit-maximizing/cost-minimizing problem of the firm are Berman and Bui (2001)
and Morgenstern, Pizer, and Shih (2002). Berman and Bui (2001) developed an innovative
approach to estimating the effect of environmental regulations designed to meet a NAAQS
(e.g., ozone and NOX) requirement in California on employment. Berman and Bui's model
allows environmental regulation to operate via two separate mechanisms: 1) the output
elasticity of labor demand and 2) the effect of pollution abatement activities on demand for
variable factors, combined with the marginal rates of technical substitution between
abatement activity and variable factors, including labor. Berman and Bui show how Neoclassical
economic theory predicts that the output effect is, in most cases, negative, while the direction
of the second, composite effect is indeterminate, making the overall net effect ambiguous.
10-4
-------
Morgenstern, Pizer, and Shin (2002) developed a similar structural model to Berman and
Bui's (2001) model. Their model focuses on three mechanisms whereby environmental
regulation may impact employment in regulated industries. The first mechanism is the demand,
or output, effect, where new compliance costs increase the cost of production, raising prices
and thereby reducing consumer demand, which, in turn, reduces labor demand. The second
mechanism is the cost effect, which increases the demand for inputs, including labor, because
more inputs are now required to produce the same amount of output. Finally, the factor-shift
effect notes how regulated firms' production technologies may be more or less labor intensive
after complying with the regulation (i.e., more/less labor is required relative to capital per
dollar of output), implying an ambiguous overall net effect on labor demand. Conceptually, this
theoretical approach, which is very similar to Berman and Bui's approach, could be applied to
NAAQS. However, Morgenstern et al.'s empirical approach uses pollution abatement
expenditures for only four highly polluting/regulated sectors (pulp and paper, plastics, steel,
and petroleum refining) to estimate effects on net employment; therefore, their empirical
results are not directly applicable to the full range of manufacturing and nonmanufacturing
industries affected by NAAQS. Regardless, their work represents one of the most rigorous
attempts to quantify the net employment impacts of regulation on the regulated sector.
Morgenstern et al. conclude from their empirical results that increased pollution abatement
expenditures generally have not caused a significant change in net employment in those four
sectors. More specifically, their results suggest that, on average across the industries studied,
each additional $1 million ($1987) spent on pollution abatement resulted in a (statistically
insignificant) net increase of 1.5 jobs.
Berman and Bui (2001) use their model to empirically examine how an increase in local
air quality regulation that reduces NOX emissions as a precursor to ozone and PMi0 affects
manufacturing employment in the South Coast Air Quality Management District (SCAQMD),
which incorporates Los Angeles and its suburbs. During the time frame of their study, 1979 to
1992, the SCAQMD enacted some of the country's most stringent air quality regulations. Using
SCAQMD's local air quality regulations, which are more stringent than federal and state
regulations, Berman and Bui identify the effect of environmental regulations on net
employment in the regulated sectors.3 They compare changes in employment in affected plants
to those in other plants in the same industries but in regions not subject to the local
regulations. The authors find that "while regulations do impose large costs, they have a limited
effect on employment"—even when exit and dissuaded entry effects are considered (Berman
3 Note, like Morgenstern, Pizer, and Shih (2002), this study does not estimate the number of jobs created in the
environmental protection sector.
10-5
-------
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 [emphasis added], even when exit and dissuaded
entry effects are included" (Berman and Bui, 2001, p. 269). In their view, the limited effects
likely arose because 1) the regulations applied disproportionately to capital-intensive plants
with relatively little employment, 2) the plants sold to local markets where competitors were
subject to the same regulations (so that sales were relatively unaffected), and 3) abatement
inputs served as complements to employment. Although Berman and Bui focus on more sectors
than Morgenstern et al. and focus specifically on air regulations, the study only examined
impacts in Southern California and impacts may differ in other nonattainment areas.
Other studies, including Henderson (1996), Becker and Henderson (2000), Greenstone
(2002), and List et al. (2003), have taken a reduced-form approach to ask a related but quite
different question regarding the impact of environmental regulation on economic activity. All of
these studies examined the effect of attainment status, with respect to NAAQS, on various
forms of economic activity (e.g., employment growth, plant openings and closings, investment).
Polluting plants already located in and new polluting plants wanting to open in nonattainment
counties (counties not in compliance with one or more NAAQSs) are likely to face more
stringent air pollution regulations to help bring them into compliance. Thus, the stringency in
environmental regulations may vary spatially, which may affect the spatial distribution of
economic activity but not necessarily the overall level of economic activity. These studies find
limited evidence that employment grows more slowly, investment is lower, or fewer new
polluting plants open in nonattainment areas relative to attainment areas. However, this
evidence does not mean that there is less aggregate economic activity as a result of
environmental regulation nor does it provide evidence regarding absolute growth rates; it
simply suggests that the relative growth rate of some sectors may differ between attainment
and nonattainment areas. The approach used in all of these other studies is not capable of
estimating net employment effects as would be necessary for a national rulemaking, only
certain aspects of gross labor flows in selected areas.
10.4 Conclusion
The long-term effects of a regulation on the environmental protection sector (which
provides goods and services to the regulated sector) are difficult to assess. Employment in the
industry supplying pollution control equipment is likely to increase with the increased demand
from the regulated industry for the equipment.4 According to U.S. Department of Commerce
See Bezdek, Wendling, and Diperna (2008), for example, and U.S. Department of Commerce (2010).
10-6
-------
(2010) data, by 2008, there were 119,000 environmental technology (ET) firms generating
approximately $300 billion in revenues domestically (2% of national gross domestic product),
producing $43.8 billion in exports (2% of total exports), and supporting nearly 1.7 million jobs
(0.93% of total jobs). Air pollution control accounted for 18% of the domestic ET market and
16% of exports. Small and medium-size companies represent 99% of private ET firms, producing
20% of total revenue. The remaining 1% of companies are large companies supplying 49% of ET
revenue (OEEI, 2010).5
As described above, deriving estimates of how regulations will impact economy-wide
net employment is a difficult task, especially in the case of setting a new NAAQS, given that
economic theory predicts that the net effect of an environmental regulation on regulated
sectors and the overall economy is indeterminate (not necessarily positive or negative). Peer-
reviewed econometric studies that use a structural approach, applicable to overall net effects in
the regulated sectors, converge on the finding that any net employment effects of
environmental regulation in general, whether positive or negative, have been small and have
not affected employment in the national economy in a significant way.
10.5 References
Becker, R. and V. Henderson (2000). "Effects of Air Quality Regulations on Polluting Industries."
Journal of Political Economy 108(2): 379-421.
Berman, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand: Evidence
from the South Coast Air Basin." Journal of Public Economics 79(2): 265-295.
Bezdek, R. H., R. M. Wendling, and P. Diperna (2008). "Environmental Protection, the Economy,
and Jobs: National and Regional Analyses." Journal of Environmental Management
86(1): 63-79.
Executive Order 13563 (January 21, 2011). "Improving Regulation and Regulatory Review.
Section 1. General Principles of Regulation." Federal Register 76(14): 3821-3823.
Greenstone, M. (2002). "The Impacts of Environmental Regulations on Industrial Activity:
Evidence from the 1970 and 1977 Clean Air Act Amendments and the Census of
Manufactures." Journal of Political Economy 110(6): 1175-1219.
Henderson, V. (1996). "Effects of Air Quality Regulation." The American Economic Review 86(4):
789-813.
5 To calculate the percentages, total national 2008 GDP ($14,369.1 billion), exports ($1,842.68 billion), and
employment (181.75 million employees) were obtained from Bureau of Economic Analysis, U.S. Census Bureau,
and Woods & Poole, respectively.
10-7
-------
List, J. A., D. L. Millimet, P. G. Fredriksson, and W. W. McHone (2003). "Effects of Environmental
Regulations on Manufacturing Plant Births: Evidence from a Propensity Score Matching
Estimator." The Review of Economics and Statistics 85(4): 944-952.
Morgenstern, R. D., W. A. Pizer, and J-S. Shih (2002). "Jobs Versus the Environment: An
Industry-Level Perspective." Journal of Environmental Economics and Management
43(3): 412-436.
Schmalansee, R. and R. Stavins (2011). "A Guide to Economic and Policy Analysis for the
Transport Rule." White Paper. Boston, MA: Exelon Corp.
U.S. Department of Commerce, International Trade Administration (2010, April). Environmental
Industries Fact Sheet, using 2008 data from Environmental Business International, Inc.
http://web.ita.doc.gov/ete/eteinfo.nsf/068f3801d047f26e85256883006ffa54/4878b7e2
fc08ac6d85256883006c452c?OpenDocument.
U.S. Department of Labor, Bureau of Labor Statistics (2012a). Databases, Tables & Calculators
by Subject. http://data.bls.gov/timeseries/LNS14000000. Accessed April 16, 2012.
U.S. Department of Labor, Bureau of Labor Statistics (2012b, April 6). "The Employment
Situation—March 2012." News Release.
http://www.bls.gov/news.release/pdf/empsit.pdf. Accessed April 16, 2012.
10-8
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-452/R-12-005
Environmental Protection Health and Environmental Impacts Division December 2012
Agency Research Triangle Park, NC
------- |