&EPA
United States
Environmental Protection
Agency
Regulatory Impact Analysis for the Final
Mercury and Air Toxics Standards
-------
EPA-452/R-11-011
December 2011
Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, NC
-------
CONTENTS
Section Page
Executive Summary ES-1
ES.l Key Findings ES-1
ES.1.1 Health Co-Benefits ES-4
ES.1.2 Welfare Co-Benefits ES-5
ES.2 Not All Benefits Quantified ES-10
ES.3 Costs and Employment Impacts ES-15
ES.4 Small Entity and Unfunded Mandates Impacts ES-16
ES.5 Limitations and Uncertainties ES-17
ES.6 References ES-21
1 Introduction and Background 1-1
1.1 Introduction 1-1
1.2 Background for Final Mercury and Air Toxics Standards 1-2
1.2.1 NESHAP 1-2
1.2.2 NSPS 1-4
1.3 Appropriate & Necessary Analyses 1-4
1.4 Provisions of the Final Mercury and AirToxics Standards 1-5
1.4.1 What Is the Source Category Regulated by the Final Rule? 1-5
1.4.2 What Are the Pollutants Regulated by the Rule? 1-6
1.4.3 What Are the Emissions Limits? 1-6
1.4.4 What are the Startup, Shutdown, and Malfunction Requirements? .... 1-11
1.5 Baseline and Years of Analysis 1-12
1.6 Benefits of Emission Controls 1-13
-------
1.7 Cost of Emission Controls 1-14
1.8 Organization of the Regulatory Impact Analysis 1-14
1.9 References 1-15
2 Electric Power Sector Profile 2-1
2.1 Introduction 2-1
2.2 Power Sector Overview 2-1
2.2.1 Generation 2-2
2.2.2 Transmission 2-5
2.2.3 Distribution 2-6
2.3 Deregulation and Restructuring 2-6
2.4 Emissions of Mercury and Other Hazardous Air Pollutants from Electric
Utilities 2-7
2.5 Pollution Control Technologies 2-9
2.6 HAP Regulation in the Power Sector 2-11
2.6.1 Programs Targeting HAP 2-11
2.6.2 Programs Targeting S02 and NOx 2-12
2.7 Revenues, Expenses, and Prices 2-14
2.7.1 Natural Gas Market 2-18
2.8 Electricity Demand and Demand Response 2-19
2.9 References 2-20
3 Cost, Economic, and Energy Impacts 3-1
3.1 Background 3-2
3.2 Projected Emissions 3-10
3.3 Projected Compliance Costs 3-14
3.4 Projected Compliance Actions for Emissions Reductions 3-15
3.5 Projected Generation Mix 3-17
-------
3.6 Projected Withdrawals from Service 3-18
3.7 Projected Capacity Additions 3-20
3.8 Projected Coal Production for the Electric Power Sector 3-21
3.9 Projected Retail Electricity Prices 3-23
3.10 Projected Fuel Price Impacts 3-25
3.11 Key Differences in EPA Model Runs for MATS Modeling 3-27
3.12 Projected Primary PM Emissions from Power Plants 3-28
3.13 Illustrative Dry Sorbent Injection Sensitivity 3-28
3.14 Additional Compliance Costs Analyzed for Covered Units 3-29
3.14.1 Compliance Cost for Oil-Fired Units 3-29
3.14.2 Monitoring, Reporting and Record-keeping Costs 3-30
3.14.3 Total Costs Projected for Covered Units under MATS 3-30
3.15 Limitations of Analysis 3-31
3.16 Significant Energy Impact 3-35
3.17 References 3-35
Appendix 3A Compliance costs for oil-fired electric generating units 3A-1
3A.1 Methodology and Assumptions 3A-1
3A.1.1 Base Case 3A-1
3A.1.2 Policy Case 3A-3
3A.1.3 Cost Sensitivities Related to Mandatory Natural Gas Curtailment 3A-4
3A.2 Results 3A-5
3A.3 References 3A-7
4 Mercury and other HAP Benefits Analysis 4-1
4.1 Introduction 4-2
4.2 Impact of Mercury on Human Health 4-5
IV
-------
4.2.1 Introduction 4-5
4.2.2 Neurologic Effects 4-6
4.2.3 Cardiovascular Impacts 4-6
4.2.4 Genotoxic Effects 4-6
4.2.5 Immunotoxic Effects 4-7
4.2.6 Other Human Toxicity Data 4-7
4.3 Impact of Mercury on Ecosystems and Wildlife 4-7
4.3.1 Introduction 4-7
4.3.2 Effects on Fish 4-8
4.3.3 Effects on Birds 4-9
4.3.4 Effects on Mammals 4-10
4.4 Mercury Risk and Exposure Analyses—Data Inputs and Assumptions 4-11
4.4.1 Introduction 4-11
4.4.2 Data Inputs 4-11
4.4.3 Mercury Concentrations in Freshwater Fish 4-15
4.5 Linking Changes in Modeled Mercury Deposition to Changes in Fish Tissue
Concentrations 4-18
4.5.1 Introduction 4-18
4.5.2 Use of Mercury Maps to Project Changes in Fish Tissue
Concentrations 4-18
4.5.3 The Science of Mercury Processes and Variability in Aquatic
Ecosystems 4-23
4.5.4 Summary 4-31
4.6 Analysis of the Dose-Response Relationship Between Maternal Mercury
Body Burden and Childhood IQ 4-32
4.6.1 Introduction 4-32
4.6.2 Epidemiological Studies of Mercury and Neurodevelopmental
Effects 4-34
4.6.3 Statistical Analysis 4-35
4.6.4 Strengths and Limitations of the IQ Dose-Response Analysis 4-36
4.6.5 Possible Confounding from Long-Chained Polyunsaturated Fatty
Acids 4-39
4.7 Mercury Benefits Analysis Modeling Methodology 4-40
4.7.1 Introduction 4-40
-------
4.7.2 Estimation of Exposed Populations and Fishing Behaviors 4-41
4.7.3 Estimation of Lost Future Earnings 4-47
4.8 Mercury Benefits and Risk Analysis Results 4-50
4.8.1 Baseline Incidence 4-50
4.8.2 IQ Loss and Economic Valuation Estimates 4-56
4.8.3 Primary Results for National Analysis of Exposures from
Recreational Freshwater Fish Consumption 4-57
4.8.5 Discussion of Assumptions, Limitations, and Uncertainties 4-59
4.8.6 Overall Conclusions 4-68
4.9 Benefits Associated with Reductions in Other HAP than Mercury 4-70
4.9.1 Hazards 4-75
4.10 References 81
4A Analysis of Trip Travel Distance for Recreational Freshwater Anglers 4A-1
4A.1 Data 4A-1
4A.2 Analysis of Travel Distance Data 4A-1
4A.3 Summary Results Applied in the Population Centroid Approach 4A-5
5 Health and Welfare Co-Benefits 5-1
Synopsis 5-1
5.1 Overview 5-3
5.2 Benefits Analysis Methods 5-11
5.2.1 Health Impact Assessment 5-12
5.2.2 Economic Valuation of Health Impacts 5-13
5.2.3 Adjusting the Results of the PM2.s co-benefits Analysis to Account
for the Emission Reductions in the Final Mercury and Air Toxics
Standards 5-15
5.3 Uncertainty Characterization 5-18
5.4 Benefits Analysis Data Inputs 5-22
5.4.1 Demographic Data 5-22
VI
-------
5.4.2 Effect Coefficients 5-23
5.4.3 Baseline Incidence Estimates 5-38
5.4.4 Economic Valuation Estimates 5-41
5.4.5 Hospital Admissions Valuation 5-51
5.5 Unquantified Health and Welfare Benefits 5-61
5.5.1 Visibility Valuation 5-61
5.5.2 Ecosystem Services 5-68
5.5.3 Ecosystem Benefits of Reduced Nitrogen and Sulfur Deposition 5-70
5.5.4 Ecological Effects Associated with Gaseous Sulfur Dioxide 5-79
5.5.5 Nitrogen Enrichment 5-80
5.5.6 Benefits of Reducing Ozone Effects on Vegetation and Ecosystems .... 5-83
5.5.7 Unquantified S02 and N02-Related Human Health Benefits 5-89
5.6 Social Cost of Carbon and Greenhouse Gas Co-Benefits 5-90
5.7 Co-Benefits Results 5-94
5.8 Discussion 5-105
5.9 References 5-106
5A Impact of the Interim Policy Scenario on Emissions 5A-1
5A.1 Introduction 5A-1
5A.2 Overview of Modeling Platform and Emissions Processing Performed 5A-1
5A.3 Development of 2005 Base Year Emissions 5A-2
5A.4 Development of Future year baseline Emissions 5A-9
5A.5 Development of Future Year Control Case Emissions for Air Quality
Modeling 5A-21
5B Impact of the Interim Policy Scenario on Air Quality 5B-1
5C Health and Welfare Co-Benefits of the Modeled Interim Policy Scenario 5C-1
5D PM2.5 Co-Benefits of the Final Rule by State 5D-1
VII
-------
5D.1 Introduction 1
5D.2 Methods 1
5D.3 Limitations and uncertainties 2
5D.4 Results 2
5E Summary of Expert Opinions on the Existence of a Threshold in the
Concentration-Response Function for PM2.5-related Mortality 5E-1
6 Employment and Economic Impact Analysis 1
6.1 Employment Impacts for the MATS 2
6.2 Employment Impacts Primarily on the Regulated Industry: Morgenstern,
Pizer, and Shih (2002) 3
6.3 Employment Impacts of the MATS-Pollution Control Sector Approach by
2015 8
6.3.1 Overall Approach and Methodology for Pollution Control Sector
Approach 10
6.3.2 Summary of Employment Estimates from Pollution Control Sector
Approach 11
6.3.3 Other Employment Impacts of MATS 11
Note: See Appendix 6A for more detail 12
6.4 Summary of Employment Impacts 12
6.5 Potential Effect of Electricity Price Increase on Economy-Wide
Production Costs 13
6.6 Estimating Social Cost and Economic Impacts 16
6.7 References 17
6A.1 Overall Approach 2
6A.1.2 Employment Changes due to New Pollution Control Equipment 3
6A.1.3 Results 6
6A.2 Results Summary 11
6A.3 Detailed Methodology 11
VIM
-------
6A.3.1 Pollution Control Equipment Labor 11
6A.3.2 Retirement Labor 13
6A.3.3 Fuel Use Labor 14
6A.4 References 16
7 Statutory and Executive Order Analyses 7-1
7.1 Introduction 7-2
7.2 Executive Order 12866: Regulatory Planning and Review and Executive
Order 13563, Improving Regulation and Regulatory Review 7-2
7.3 Paperwork Reduction Act 7-4
7.4 Final Regulatory Flexibility Analysis 7-5
7.4.1 Reasons Why Action Is Being Taken 7-5
7.4.2 Statement of Objectives and Legal Basis for Final Rules 7-6
7.4.3 Summary of Issues Raised during the Public Comment Process on
thelRFA 7-6
7.4.4 Description and Estimate of the Affected Small Entities 7-13
7.4.5 Compliance Cost Impacts 7-14
7.4.6 Description of Steps to Minimize Impacts on Small Entities 7-18
7.5 Unfunded Mandates Reform Act (UMRA) Analysis 7-21
7.5.1 Identification of Affected Government Entities 7-22
7.5.2 Compliance Cost Impacts 7-22
7.6 Executive Order 13132, Federalism 7-26
7.7 Executive Order 13175, Consultation and Coordination with Indian Tribal
Governments 7-27
7.8 Protection of Children from Environmental Health and Safety Risks 7-28
7.9 Statement of Energy Effects 7-28
7.10 National Technology Transfer and Advancement Act 7-29
7.11 Environmental Justice 7-35
7.11.1 Environmental Justice Impacts 7-35
IX
-------
7.11.2 Analysis of High Risk Sub-Populations 7-42
7.11.3 Characterizing the Distribution of Health Impacts across
Populations 7-51
7.12 Congressional Review Act 7-56
7.13 References 7-57
8 Comparison of Benefits and Costs 8-1
8.1 Comparison of Benefits and Costs 8-1
8.2 References 8-3
-------
LIST OF FIGURES
Number Page
ES-1. Economic Value of Estimated PM2.5-Related Health Co-Benefits According to
Epidemiology or Expert-Derived PM Mortality Risk Estimate ES-8
ES-2. Net Benefits of the MATS Rule According to PM2.s Epidemiology or Expert-
Derived Mortality Risk Estimate ES-9
2-1. Fossil Fuel-Fired Electricity Generating Facilities, by Size 2-4
2-2. Status of State Electricity Industry Restructuring Activities 2-6
2-3. National Average Retail Electricity Price (1960-2009) 2-16
2-4. Average Retail Electricity Price by State (cents/kWh), 2009 2-16
2-5. Natural Gas Spot Price, Annual Average (Henry Hub) 2-17
2-6. Electricity Growth Rate (3 Year Rolling Average) and Projections from the Annual
Energy Outlook 2011 2-18
3-1. Geographic Distribution of Affected Units, by Facility, Size and Fuel Source in
2012 3-8
3-2. S02 Emissions from the Power Sector in 2015 with and without MATS 3-11
3-3. NOX Emissions from the Power Sector in 2015 with and without MATS 3-11
3-4. Mercury Emissions from the Power Sector in 2015 with and without MATS 3-12
3-5. Hydrogen Chloride Emissions from the Power Sector in 2015 with and without
MATS 3-13
3-6. Operating Pollution Control Capacity on Coal-fired Capacity (by Technology) with
the Base Case and with MATS, 2015 (GW) 3-15
3-7. Generation Mix with the Base Case and with MATS, 2015-2030 3-17
3-8. Total Coal Production by Coal-Producing Region, 2007 (Million Short Tons) 3-22
3-9. Retail Price Model Regions 3-25
3A-1. 2006-2010 Heat Input Apportioned by Fuel for Oil-Fired Units Subject to
Mandatory Natural Gas Curtailment 3A-2
4-1. Spatial and Biogeochemical Factors Influencing MeHg Production 4-26
4-2. Preliminary USGS Map of Mercury Methylation-Sensitive Watersheds Derived
from More Than 55,000 Water Quality Sites aqnd 2,500 Watersheds (Myers
etal., 2007) 4-27
XI
-------
4-3. Methodology for Estimating and Linking Exposed Populations and Levels of
Mercury Exposure 4-40
4-4. Linking Census Tracts to Demographic Data and Mercury Fish Tissue Samples 4-42
4-5. Estimated Chronic Census Tract Carcinogenic Risk from HAP Exposure from
Outdoor Sources (2005 NATA) 4-70
4-6. Estimated Chronic Census Tract Noncancer (Respiratory) Risk from HAP Exposure
from Outdoor Sources (2005 NATA) 4-71
5-1. Illustration of BenMAP Approach 5-11
5-2. Data Inputsand Outputs for the BenMAP Model 5-12
5-3. Important Factors Involved in Seeing a Scenic Vista (Malm, 1999) 5-61
5-4. Mandatory Class I Areas in the U.S 5-62
5-5. Linkages Between Categories of Ecosystem Services and Components of Human
Well-Being from Millennium Ecosystem Assessment (MEA, 2005) 5-67
5-6. Schematic of the Benefits Assessment Process (U.S. EPA, 2006b) 5-68
5-7. Schematics of Ecological Effects of Nitrogen and Sulfur Deposition 5-69
5-8. Areas Potentially Sensitive to Aquatic Acidification (U.S. EPA, 2008b) 5-72
5-9. Areas Potentially Sensitive to Terrestrial Acidification (U.S. EPA, 2008b) 5-74
5-10. Distribution of Red Spruce (Pink) and Sugar Maple (Green) in the Eastern U.S.
(U.S. EPA, 2008b) 5-75
5-11. Ozone Injury to Forest Plants in U.S. by EPA Regions, 2002 5-84
5-12. Estimated Black Cherry, Yellow Poplar, Sugar Maple, Eastern White Pine, Virginia
Pine, Red Maple, and Quaking Aspen Biomass Loss due to Current Ozone
Exposure, 2006-2008 (U.S. EPA, 2009b) 5-85
5-13. Economic Value of Estimated PM2.5-Related Health co-benefits of the Mercury
and Air Toxics Standards in 2016 According to Epidemiology or Expert-Derived
PM Mortality Risk Estimate 5-99
5-14. Percentage of Total PM-Related Mortalities of the Mercury and Air Toxics
Standards in 2016 Avoided by Baseline AirQuality Level 5-100
5-15. Cumulative Percentage of Total PM-Related Mortalities of the Mercury and Air
Toxics Standards in 2016 Avoided by Baseline Air Quality Level 5-102
5B-1. Map of the Photochemical Modeling Domains. The black outer box denotes the
36 km national modeling domain; the red inner box is the 12 km western U.S.
grid; and the blue inner box is the 12 km eastern U.S. grid 5B-2
5B-2. Change in Design Values Between the 2017 Baseline and 2017 Control
Simulations. Negative numbers indicate lower (improved) design values in the
control case compared to the baseline 5B-5
5B-3. Change in Design Values Between the 2017 Base Case and 2017 Control
Simulations. Negative numbers indicate lower (improved) design values in the
control case compared to the baseline 5B-6
XII
-------
5B-4. Change in 20% Worst Days Between the 2017 Baseline and 2017 Control
Simulations. Negative numbers indicate lower (improved) visibility expressed in
deciviews in the control case compared to the baseline 5B-7
5B-5. Change in Design Values Between the 2017 Baseline and 2017 Control
Simulations. Negative numbers indicate lower (improved) design values in the
control case compared to the baseline 5B-8
5C-1. Comparison of state-level S02 emission changes between the interim modeled
scenario and the final policy 5C-3
5C-2. Estimated Reduction in Excess PM2.5-Related Premature Deaths Estimated to
Occur in Each County in 2016 as a Result of the Interim Modeled Mercury and Air
Toxics Standards 5C-6
7-1. Modeled African-American Population Below the Poverty Level by Census Tract
in the Southeast for 2016 7-45
7-2. Modeled White Population Below the Poverty Level by Census Tract in the
Southeast for 2016 7-46
7-3. Modeled Female Population Below the Poverty Level by Census Tract for 2016 7-47
7-4. Modeled Hispanic Population by Census Tract for 2016 7-48
7-5. Modeled Laotian Population by Census Tract for 2016 7-49
7-6. Modeled Chippewa Population by Census Tract in the Great Lakes Area for 2016 7-50
XIII
-------
LIST OF TABLES
Number Page
ES-1. Summary of EPA's Estimates of Annualized Benefits, Costs, and Net Benefits of
the Final MATS in 2016 (billions of 2007$) ES-2
ES-2: Projected Electricity Generating Unit (ECU) Emissions of S02, NOX, Mercury,
Hydrogen Chloride, PM, and C02 with the Base Case and with MATS, 2015 ES-2
ES-3. Estimated Reduction in Incidence of Adverse Health Effects of the Mercury and
AirToxics Standards (95% confidence intervals) ES-5
ES-4. Estimated Economic Value of Health and Welfare Co-Benefits of the Mercury and
AirToxics Standards (95% confidence intervals, billions of 2007$) ES-6
ES-5. Human Health Effects of Pollutants Affected by the Mercury and Air Toxics
Standards ES-10
ES-6. Environmental Effects of Pollutants Affected by the Mercury and Air Toxics
Standards ES-12
ES-7. Estimated Employment Impact Table ES-15
1-1. Emission Limitations for Coal-Fired and Solid Oil-Derived Fuel-Fired EGUs 1-6
1-2. Emission Limitations for Liquid Oil-Fired EGUs 1-7
1-3. Alternate Emission Limitations for Existing Coal- and Oil-Fired EGUs 1-8
1-4. Alternate Emission Limitations for New Coal- and Oil-Fired EGUs 1-9
2-1. Existing Electricity Generating Capacity by Energy Source, 2009 2-1
2-2. Total U.S. Electric Power Industry Retail Sales in 2009 (Billion kWh) 2-2
2-3. Electricity Net Generation in 2009 (Billion kWh) 2-2
2-4. Coal Steam Electricity Generating Units, by Size, Age, Capacity, and Efficiency
(Heat Rate) 2-3
2-5. U.S. Anthropogenic Mercury Emissions, 1990-2010 2-7
2-6. U.S. Hydrogen Chloride Emissions, 2005 and 2010 2-7
2-7. Revenue and Expense Statistics for Major U.S. Investor-Owned Electric Utilities
for 2009 ($millions) 2-15
2-8. Projected Revenues by Service Category in 2015 for Public Power and Investor-
Owned Utilities (billions) 2-15
3-1. Emissions Limitations for Coal-Fired and Solid Oil-Derived Fuel-Fired Electric
Utility Steam Generating Units 3-4
3-2. Emissions Limitations for Liquid Oil-Fired Electric Utility Steam Generating Units 3-5
XIV
-------
3-3. 2009 U.S. Electricity Net Generation and EPA Base Case Projections for 2015-
2030 (Billion kWh) 3-7
3-4. Projected Emissions of S02, NOX, Mercury, Hydrogen Chloride, PM, and C02 with
the Base Case and with MATS, 2015 3-10
3-5. Annualized Compliance Cost for MATS Requirements on Coal-fired Generation 3-14
3-6. Generation Mix with the Base Case and the MATS, 2015 (Thousand GWh) 3-16
3-7. Characteristics of Covered Operational Coal Units and Additional Coal Units
Projected to Withdraw as Uneconomic under MATS, 2015 3-17
3-8. Total Generation Capacity by 2015 (GW) 3-19
3-9. Total Generation Capacity by 2030 (GW) 3-20
3-10. 2015 Coal Production for the Electric Power Sector with the Base Case and MATS
(Million Tons) 3-21
3-11. 2015 Power Sector Coal Use with the Base Case and the MATS, by Coal Rank
(TBtu) 3-22
3-12. Projected Contiguous U.S. and Regional Retail Electricity Prices with the Base
Case and with the MATS (2007 cents/kWh) 3-24
3-13. Average Minemouth and Delivered Coal Prices with the Base Case and with
MATS(2007$/MMBtu) 3-26
3-14. 2015-2030 Weighted Average Henry Hub (spot) and Delivered Natural Gas Prices
with the Base Case and with MATS (2007$/MMBtu) 3-26
3-15. Cost Impacts of Compliance Actions for Oil-Fired Units 3-30
3-16. Total Costs Projected for Covered Units under MATS, 2015 (billions of 2007$) 3-31
3A-1. Oil-fired EGUs by Fuel Type 3A-2
3A-2. Least Cost NEEDS Modeled Fuels for Oil-fired EGUs 3A-3
3A-3. Percentage of Total Heat Input Derived from Oil for Oil-Fired Units Subject to
Mandatory Natural Gas Curtailment (2008-2010) 3A-5
3A-4. Costs to Achieve the MATS Emission Limitations for Oil-Fired Units 3A-6
4-1. Summary of FWHAR State-Level Recreational Fishing Characteristics 4-11
4-2. Summary of HUC-level Average Mercury Fish Tissue Concentration Estimates 4-16
4-4. Summary of Baseline Mercury Fish Tissue Concentrations 4-50
4-5. Baseline Levels of Mercury Exposure and IQ Impacts Due to Freshwater Self-
Caught Fish Consumption 4-51
4-6. Summary Estimates of the Aggregate Size and Present Value of IQ Losses Under
Alternative Base Case and Emissions Control Scenarios 4-56
4-7. Aggregate Benefit Estimates for Reductions IQ Losses Associated with Alternative
Emissions Reduction Scenarios 4-56
4-8. Unquantified Health and Ecosystem Effects Associated with Exposure to Mercury....4-65
xv
-------
4A-1. Reported Trip Travel Distance for Freshwater Anglers (Miles) 4A-2
4A-2. Demographic Characteristics of Freshwater Anglers 4A-2
4A-3. Demographic Characteristics of Freshwater Anglers 4A-3
4A-4. OLS Regression Results for Determinants of Reported Trip Travel Distance (Miles)....4A-4
4A-5. Travel Distance Frequencies by Demographic Group (Percentage in Each Distance
Category) 4A-6
5-1. Estimated Monetized Co-benefits of the Mercury and Air Toxics Standards in
2016 (billions of 2007$) 5-3
5-2. Human Health Effects of Pollutants Affected by the Mercury and Air Toxics
Standards 5-4
5-3. Environmental Effects of Pollutants Affected by the Mercury and Air Toxics
Standards 5-6
5-4. Primary Sources of Uncertainty in the Benefits Analysis 5-18
5-5. Criteria Used When Selecting C-R Functions 5-22
5-6. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts 5-24
5-7. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population 5-37
5-8. Asthma Prevalence Rates Used for this Analysis 5-39
5-9. Expected Impact on Estimated Benefits of Premature Mortality Reductions of
Differences Between Factors Used in Developing Applied VSL and Theoretically
Appropriate VSL 5-43
5-10. Alternative Direct Medical Cost of Illness Estimates for Non-fatal Heart Attacks 5-48
5-11. Estimated Costs Over a 5-Year Period (in 2006$) of a Non-fatal Myocardial
Infarction 5-49
5-12. Unit Values for Economic Valuation of Health Endpoints (2006$) 5-51
5-13. Elasticity Values Used to Account for Projected Real Income Growth 5-57
5-14. Adjustment Factors Used to Account for Projected Real Income Growth 5-59
5-15. Aquatic Status Categories 5-71
5-16. Social Cost of Carbon (SCC) Estimates (per tonne of C02) for 2016 (in 2007$) 5-90
5-17. Monetized Co-Benefits of C02 Emissions Reductions in 2016 (in millions of
2007$) 5-91
5-18. Estimated Reduction in Incidence of Adverse Health Effects of the Mercury and
Air Toxics Standards in 2016 (95% confidence intervals) 5-95
5-19. Estimated Economic Value of Health and Welfare co-benefits of the Mercury and
Air Toxics Standards in 2016 (95% confidence intervals, billions of 2007$) 5-96
5-20. Estimated Reduction in Incidence of Premature Adult Mortality due to the Mercury and
Air Toxics Standards in 2016 Occurring Above and Below the Lowest Measured Levels in
the Underlying Epidemiology Studies 5-101
XVI
-------
5A-1. 2005 US Emissions by Sector 5A-5
5A-2. 2005 Base Year S02 Emissions (tons/year) for States by Sector 5A-5
5A-3. 2005 Base Year PM2.5 Emissions (tons/year) for States by Sector 5A-7
5A-4. Summary of Mobile Source Control Programs Included in the Future Year
Baseline 5A-11
5A-5. Control Strategies and/or Growth Assumptions Included in the Future Year
Baseline for Non-EGU Stationary Sources 5A-13
5A-6. Summary of Modeled Base Case Annual Emissions (tons/year) for 48 States by
Sector: S02 and PM2.5 5A-15
5A-7. Future Year Baseline S02 Emissions (tons/year) for States by Sector 5A-15
5A-8. Future Year Baseline PM2.5 Emissions (tons/year) for States by Sector 5A-17
5A-9. Future Year Baseline ECU CAP Emissions (tons/year) by State 5A-19
5A-10. Summary of Emissions Changes for the MATS AQ Modeling in the Lower 48
States 5A-22
5A-11. ECU Emissions Totals for the Modeled MATS Control Case in the Lower 48 States ..5A-22
5A-12. State Specific Changes in Annual ECU S02 for the Lower 48 States 5A-24
5A-13. State Specific Changes in Annual ECU PM2.5forthe Lower48 States 5A-26
5B-1. Geographic Elements of Domains Used in Photochemical Modeling 5B-3
5C-1. Estimated Reduction in Incidence of Adverse Health Effects of the Interim
Modeled Mercury and Air Toxics Standards in 2016 (95% confidence intervals) 5C-4
5C-2. Estimated Economic Value of Health and Welfare Benefits of the Interim
Modeled Mercury and Air Toxics Standards in 2016 (95% confidence intervals,
billions of 2007$) 5C-5
5C-3. Estimated Economic Value of Adult Mortality Benefits by Pollutant, in Total and
Per Ton of Emissions Reduced Interim Modeled Mercury and Air Toxics Standard
in 2016 (95% confidence intervals, 2007$) 5C-8
5D-1. Estimated Reduction in Incidence of Premature Adult Mortality for the Mercury
and Air Toxics Standards in 2016 by State 5D-3
5D-2. Estimated Economic Value of Health Benefits of the Mercury and Air Toxics
Standard in 2016 by State (billions of 2007$, 3% discount rate) 5D-5
6-1. Percent of Abatement Expenditures in Different PACE Studies from Add-On or
End-of-Line Control Measures 6-3
6-2. Employment Impacts Within the Regulated Industry Using Peer-Reviewed Study
Estimates using Morgenstern et al. (2002) 6-7
6-3. Increased Pollution Control Installations due to MATS, by 2015 (GW) 6-10
XVII
-------
6-4. Employment Effects Using the Environmental Protection Sector Approach for the
MATS (in Job-Years) 6-10
6-5. Other Employment Impacts of MATS (in Job-Years) 6-11
6-6. Estimated Employment Impact for the MATS 6-12
6A-1. Net Employment Changes for 2015 (job-years) 6A-2
6A-2. Increased Pollution Control Demand due to the final MATS, 2015 (GW) 6A-4
6A-3. Estimated Pollution Control Resource Needs (Quantity and Prices Used) 6A-6
6A-4. Jobs Due to Pollution Control Equipment under the final MATS (Job-years in
2015) 6A-7
6A-5. Annual Job Losses due to Coal Capacity Retirements for 2015 6A-8
6A-6. Annual Employment Impacts Due To Changes in Coal Use for 2015 6A-10
6A-7. Annual Employment Impact due to Changes in Fuel Use (2015) 6A-10
6A-8. Installation Labor Requirement 6A-11
6A-9. Resources Needed for Operation 6A-12
6A-10. Operating Labor Assumptions 6A-13
6A-11. Inputs to Labor from Retirements 6A-13
6A-12. Inputs to Labor for Fuel Use 6A-14
7-1. Projected Impact of MATS on Small Entities in 2015 7-15
7-2. Incremental Annualized Costs under MATS Summarized by Ownership Group and
Cost Category in 2015 (2007$ millions) 7-17
7-3. Incremental Annualized Costs under MATS Summarized by Ownership Group and
Cost Category (2007$ millions) in 2015 7-23
7-4. Summary of Potential Impacts on Government Entities under MATS in 2015 7-24
7-5. Comparative Summary of the Demographics within 5 Kilometers (3 Miles) of the
Affected Sources (population in millions) 7-39
7-6. Reported Distributions of Self-Caught Freshwater Fish Consumption Rates
Among Selected Potentially High-Risk Subpopulations 7-44
7-7. Change in the Percentage of All Deaths Attributable to PM2.s Before and After
Implementation of MATS by 2016 for Each Populations, Stratified by Race 7-52
7-8. Change in the Percentage of All Deaths Attributable to PM2.5 Before and After
Implementation of MATS by 2016 for Each Population, Stratified by Race and
Poverty Level 7-53
7-9. Change in the Percentage of All Deaths Attributable to PM2.s Before and After
the Implementation of MATS by 2016 for Each Population, Stratified by
Educational Attainment 7-53
8-1. Summary of EPA's Estimates of Annualized Benefits, Costs, and Net Benefits of
the Final MATS in 2016 (billions of 2007$) 8-2
XVIII
-------
EXECUTIVE SUMMARY
This Regulatory Impact Analysis (RIA) presents the health and welfare benefits, costs,
and other impacts of the final Mercury and Air Toxics Standards (MATS) in 2016.
ES.l Key Findings
This rule will reduce emissions of Hazardous Air Pollutants (HAP), including mercury,
from the electric power industry. As a co-benefit, the emissions of certain PM2.s precursors such
as S02 will also decline. EPA estimates that this final rule will yield annual monetized benefits
(in 2007$) of between $37 to $90 billion using a 3% discount rate and $33 to $81 billion using a
7% discount rate. The great majority of the estimates are attributable to co-benefits from 4,200
to 11,000 fewer PM2.5-related premature mortalities. The monetized benefits from reductions
in mercury emissions, calculated only for children exposed to recreationally caught freshwater
fish, are expected to be $0.004 to $0.006 billion in 2016 using a 3% discount rate and $0.0005
to $0.001 billion using a 7% discount rate. The annual social costs, approximated by the
compliance costs, are $9.6 billion (2007$) and the annual monetized net benefits are $27 to $80
billion using 3% discount rate or $24 to $71 billion using a 7% discount rate.1 The benefits
outweigh costs by between 3 to 1 or 9 to 1 depending on the benefit estimate and discount
rate used. There are some costs and important benefits that EPA could not monetize, such as
other mercury reduction benefits and those for the HAP other than mercury being reduced by
this final rule. Upon considering these limitations and uncertainties, it remains clear that the
benefits of the MATS are substantial and far outweigh the costs. Employment impacts
associated with the final rule are estimated to be small.
The benefits and costs in 2016 of the final rule are in Table ES-1. The emission
reductions from the electricity sector that are expected to result from the rule are reported in
Table ES-2.
1 As discussed in Chapter 3, costs were annualized using a 6.15% discount rate.
ES-1
-------
Table ES-1. Summary of EPA's Estimates of Annualized3 Benefits, Costs, and Net Benefits of
the Final MATS in 2016b (billions of 2007$)
Estimate Estimate
Description (3% Discount Rate) (7% Discount Rate)
Costs0 $9.6 $9.6
Benefitsd'e'f $37 to $90 + B $33 to $81 + B
Net benefits (benefits-costs)8 $27 to $80 + B $24 to $71 + B
a All estimates presented in this report represent annualized estimates of the benefits and costs of the final MATS
in 2016 rather than the net present value of a stream of benefits and costs in these particular years of analysis.
b Estimates rounded to two significant figures and represent annualized benefits and costs anticipated for the
year 2016.
°Total social costs are approximated by the compliance costs. Compliance costs consist of IPM projections,
monitoring/reporting/recordkeeping costs, and oil-fired fleet analysis costs. For a complete discussion of these
costs refer to Chapter 3. Costs were annualized using a 6.15% discount rate.
Total benefits are composed primarily of monetized PM-related health benefits. The reduction in premature
fatalities each year accounts for over 90% of total monetized benefits. Benefits in this table are nationwide and
are associated with directly emitted PM2.5 and SO2 reductions. The estimate of social benefits also includes CO2-
related benefits calculated using the social cost of carbon, discussed further in Chapter 5.
e Not all possible benefits or disbenefits are quantified and monetized in this analysis. B is the sum of all
unquantified benefits and disbenefits. 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. Estimates
here are subject to uncertainties discussed further in the body of the document. Potential benefit categories
that have not been quantified and monetized are listed in Table ES-5.
f Mortality risk valuation assumes discounting over the SAB-recommended 20-year segmented lag structure.
Results reflect the use of 3% and 7% discount rates consistent with EPA and OMB guidelines for preparing
economic analyses (EPA, 2000; OMB, 2003).
8 Net benefits are rounded to two significant figures. Columnar totals may not sum due to rounding.
Table ES-2: Projected Electricity Generating Unit (ECU) Emissions of SO2, NOX, Mercury,
Hydrogen Chloride, PM, and CO2 with the Base Case and with MATS, 2015 a'b
Million Tons Thousand Tons CO,
Base
MATS
All EGUs
Covered EGUs
All EGUs
Covered EGUs
S02
3.4
3.3
2.1
1.9
NOX
1.9
1.7
1.9
1.7
Mercury
(Tons)
28.7
26.6
8.8
6.6
HCI
48.7
45.3
9.0
5.5
PM2.5
277
270
227
218
(Million Metric
Tonnes)
2,230
1,906
2,215
1,883
Source: Integrated Planning Model run by EPA, 2011
The year 2016 is the compliance year for MATS, thouj
for compliance in 2016 for IPM emissions and costs due to availability of modeling impacts in that year.
The year 2016 is the compliance year for MATS, though as we explain in later chapters, we use 2015 as a proxy
ES-2
-------
ES.1.1 Health Co-Benefits
The final MATS Rule is expected to yield significant health co-benefits by reducing
emissions not only of HAP such as mercury, but also significant co-benefits by reducing to direct
fine particles (PM2.5) and sulfur dioxide, which contributes to the formation of PM2.5.
Our analyses suggest this rule would yield co-benefits in 2016 of $37 to $90 billion
(based on a 3% discount rate) and $33 to $81 billion (based on a 7% discount rate). This
estimate reflects the economic value of a range of avoided health outcomes including 510
fewer mercury-related IQ points lost as well as avoided PM2.5-related impacts, including 4,200
to 11,000 premature deaths, 4,700 nonfatal heart attacks, 2,600 hospitalizationsfor respiratory
and cardiovascular diseases, 540,000 lost work days, and 3.2 million days when adults restrict
normal activities because of respiratory symptoms exacerbated by PM2.5. We also estimate
substantial additional health improvements for children from reductions in upper and lower
respiratory illnesses, acute bronchitis, and asthma attacks. See Table ES-3 for a list of the annual
reduction in health effects expected in 2016 and Table ES -4 for the estimated value of those
reductions. In addition, we include in our monetized co-benefits estimates the effect from the
reduction in C02 emissions resulting from this rule. We calculate the co-benefits associated
with these emission reductions using the interagency estimates of the social cost of carbon
(sec)1.
It is important to note that the health co-benefits from reduced PM2.5 exposure reported
here contain uncertainty, including from the following key assumptions:
1. The PM2.5-related co-benefits of the regulatory alternatives were derived
through a benefit per-ton approach, which does not fully reflect local variability in
population density, meteorology, exposure, baseline health incidence rates, or other
local factors that might lead to an over-estimate or under-estimate of the actual co-
benefits of controlling PM precursors. In addition, differences in the distribution of
emissions reductions across states between the modeled scenario and the final rule
scenario add uncertainty to the final benefits estimates.
1 Docket ID EPA-HQ-OAR-2009-0472-114577, Technical Support Document: Social Cost of Carbon for Regulatory
Impact Analysis Under Executive Order 12866, Interagency Working Group on Social Cost of Carbon, with
participation by Council of Economic Advisers, Council on Environmental Quality, Department of Agriculture,
Department of Commerce, Department of Energy, Department of Transportation, Environmental Protection
Agency, National Economic Council, Office of Energy and Climate Change, Office of Management and Budget,
Office of Science and Technology Policy, and Department of Treasury (February 2010). Also available at
http://www.epa.gov/otaq/climate/regulations.htm
ES-3
-------
2. 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 via transported precursors emitted from EGUs may differ
significantly from direct PM2.5 released from diesel engines and other industrial
sources, but the scientific evidence is not yet sufficient to allow differential effects
estimates by particle type.
3. We assume that the health impact function for fine particles is linear within the
range of ambient concentrations under consideration. Thus, the estimates include
health co-benefits from reducing fine particles in areas with varied concentrations of
PM2.5, including both regions that are in attainment with fine particle standard and
those that do not meet the standard down to the lowest modeled concentrations.
A large fraction of the PM2.5-related benefits associated with this rule occur below the
level of the National Ambient Air Quality Standard (NAAQS) for annual PM2.5 at 15 u.g/m3, which
was set in 2006. It is important to emphasize that NAAQS are not set at a level of zero risk.
Instead, the NAAQS reflect the level determined by the Administrator to be protective of public
health within an adequate margin of safety, taking into consideration effects on susceptible
populations. While benefits occurring below the standard may be less certain than those
occurring above the standard, EPA considers them to be legitimate components of the total
benefits estimate.
Based on the modeled interim baseline which is approximately equivalent to the final
baseline (see Appendix 5A), 11% and 73% of the estimated avoided premature deaths occur at
or above an annual mean PM2.5 level of 10 u.g/m3 (the LML of the Laden et al. 2006 study) and
7.5 u.g/m3 (the LML of the Pope et al. 2002 study), respectively. These are the source studies for
the concentration-response functions used to estimate mortality benefits. As we model
avoided premature deaths among populations exposed to levels of PM2.5; we have lower
confidence in levels below the LML for each study. However, studies using data from more
recent years, during which time PM concentrations have fallen, continue to report strong
associations with mortality. EPA briefly describes these uncertainties below and in more detail
in the benefits chapter of this RIA.
£5.1.2 Welfare Co-Benefits
The term welfare co-benefits covers both environmental and societal benefits of
reducing pollution, such as reductions in damage to ecosystems, improved visibility and
improvements in recreational and commercial fishing, agricultural yields, and forest
ES-4
-------
productivity. EPA did not quantify any of the important welfare co-benefits expected from the
final MATS, but these are discussed in detail in Chapter 5.
Table ES-3. Estimated Reduction in Incidence of Adverse Health Effects of the Mercury and
Air Toxics Standards (95% confidence intervals)3'15
Impact
Eastern U.S.0
Western U.S.
Total
Mercury-Related Endpoints
IQ Points Lost
PM-Related Endpoints
Premature death
510.8
Pope et al. (2002) (age
>30)
Laden et al. (2006) (age
>25)
Infant (< lyear)
Chronic bronchitis
Non-fatal heart attacks (age >
18)
Hospital admissions-
respiratory (all ages)
Hospital admissions-
cardiovascular (age > 18)
Emergency room visits for
asthma (age < 18)
Acute bronchitis (age 8-12)
Lower respiratory symptoms
(age 7-14)
Upper respiratory symptoms
(asthmatics age 9-18)
Asthma exacerbation
(asthmatics age 6-18)
Lost work days (ages 18-65)
Minor restricted-activity days
(ages 18-65)
4,100
(1,100-7,000)
10,000
(4,800 - 16,000)
19
(-21-59)
2,700
(89 - 5,400)
4,600
(1,200-8,100)
820
(320-1,300)
1,800
(1,200-2,100)
3,000
(1,500-4,500)
6,000
(-1,400 - 13,000)
77,000
(30,000 - 120,000)
58,000
(11,000 - 110,000)
130,000
(4,500 - 430,000)
520,000
(440,000 - 600,000)
3,100,000
(2,500,000 - 3,700,000)
130
(30 - 220)
320
(140-510)
1
(-1-2)
100
(-1-210)
120
(25-210)
17
(6-27)
42
(27-50)
110
(52-160)
250
(-69 - 560)
3,100
(1,100-5,200)
2,400
(360 - 4,400)
5,200
(-6 - 18,000)
21,000
(18,000-24,000)
120,000
(99,000 - 150,000)
4,200
(1,200-7,200)
11,000
(5,000 - 17,000)
20
(-22-61)
2,800
(88-5,600)
4,700
(1,200 - 8,300)
830
(330 - 1,300)
1,800
(1,200-2,200)
3,100
(1,600-4,700)
6,300
(-1,400-14,000)
80,000
(31,000-130,000)
60,000
(11,000-110,000)
130,000
(4,500 - 450,000)
540,000
(460,000 - 620,000)
3,200,000
(2,600,000 - 3,800,000)
Estimates rounded to two significant figures; column values will not sum to total value.
The negative estimates for certain endpoints are the result of the weak statistical power of the study used to
calculate these health impacts and do not suggest that increases in air pollution exposure result in decreased
health impacts.
c Includes Texas and those states to the north and east.
ES-5
-------
Table ES-4. Estimated Economic Value of Health and Welfare Co-Benefits of the Mercury and
Air Toxics Standards (95% confidence intervals, billions of 2007$)a
Impact
Pollutant Eastern U.S.b
Western U.S.
Avoided IQ loss associated with methylmercury exposure from self-caught fish consumption
anglers
3% discount rate
7% discount rate
Adult premature death (Pope et al.,
3% discount rate
7% discount rate
Adult premature death (Laden et al.
3% discount rate
7% discount rate
Infant premature death
Chronic bronchitis
Non-fatal heart attacks
3% discount rate
7% discount rate
Hospital admissions— respiratory
Hospital admissions-
cardiovascular
Emergency room visits for asthma
Acute bronchitis
Lower respiratory symptoms
Upper respiratory symptoms
Asthma exacerbation
Lost work days
Hg
Hg
2002 PM
PM
PM
mortality estimate)
$33
($2.6 -$99)
$30
($2.3 - $90)
$1.0
(<$0.01-$3.1)
$0.9
(<$0.01-$2.8)
Total
among recreational
$0.004 - $0.006
$0.0005 -$0.001
$34
($2.6 - $100)
$30
($2.4 -$92)
, 2006 PM mortality estimate)
PM
PM
PM
PM
PM
PM
PM
PM
PM
PM
PM
PM
PM
PM
$84
($7.4 -$240)
$76
($6.7 -$220)
$0.2
($-0.2 -$0.8)
$1.3
($0.1 -$6.1)
$0.5
($0.1 -$1.3)
$0.4
($0.1 -$1.0)
25 $0.01
(<$0.01 - $0.02)
25 $0.03
(<$0.01 - $0.05)
<$0.01
25 <$0.01
<$0.01
25 <$0.01
<$0.01
$0.1
($0.1 -$0.1)
$2.6
($0.1 -$7.6)
$2.3
($0.1 -$6.9)
<$0.01
$0.1
(<$0.01-$0.2)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$87
($7.5 - $250)
$78
($6.8 -$230)
$0.2
($-0.2 - $0.8)
$1.4
($0.1 -$6.4)
$0.5
($0.1 -$1.3)
$0.4
($0.1 -$1.0)
$0.01
($0.01 - $0.02)
$0.03
(<$0.01 - $0.05)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$0.1
($0.1 -$0.1)
(continued)
ES-6
-------
Table ES-4. Estimated Economic Value of Health and Welfare Co-Benefits of the Mercury and
Air Toxics Standards (95% confidence intervals, billions of 2007$)a (continued)
Impact
Minor restricted-activity days
Pollutant
PM2.5
Eastern U.S.b
$0.2
($0.1 -$0.3)
Western U.S.
<$0.01
Total
$0.2
($0.1 -$0.3)
CO2-related benefits
(3% discount rate) CO2 $0.36
Monetized total Benefits (Pope et al., 2002 PM2.5 mortality estimate)
3% discount rate
7% discount rate
Monetized total Benefits (Laden
3% discount rate
7% discount rate
$35+B
($2.8 -$110)
$32+B
($2.5 - $98)
et al., 2006 PM2.5 mortality estimate)
$87+B
($7.5 -$250)
$78+B
($6.8 -$230)
$1.1+B
($0.03 - $3.4)
$1.0+B
($0.03 -$3.1)
$2.7+B
($0.1 -$7.9)
$2.4+B
($0.1 -$7.2)
$37+B
($3.2 -$110)
$33+B
($2.9 - $100)
$90+ B
($8.0 - $260)
$81+B
($7.3 - $240)
a Economic value adjusted to 2007$ using GDP deflator. Estimates rounded to two significant figures. The negative
estimates for certain endpoints are the result of the weak statistical power of the study used to calculate these
health impacts and do not suggest that increases in air pollution exposure result in decreased health impacts.
Confidence intervals reflect random sampling error and not the additional uncertainty associated with
accounting for differences in air quality baseline forecasts described in Chapter 5. The net present value of
reduced CO2 emissions are calculated differently than other benefits. The same discount rate used to discount
the value of damages from future emissions (SCC at 5, 3, 2.5 percent) is used to calculate net present value of
SCC for internal consistency. This table shows monetized CO2 co-benefits at discount rates at 3 and 7 percent
that were calculated using the global average SCC estimate at a 3% discount rate because the interagency
workgroup on this topic deemed this marginal value to be the central value. In section 5.6 we also report CO2 co-
benefits using discount rates of 5 percent (average), 2.5 percent (average), and 3 percent (95th percentile).
b Includes Texas and those states to the north and east.
Figure ES-1 summarizes an array of PM2.5-related monetized benefits estimates based
on alternative epidemiology and expert-derived PM-mortality estimate.
Figure ES-2 summarizes the estimated net benefits for the final rule by displaying all
possible combinations of health and climate co-benefits and costs. Each of the 14 bars in each
graph represents a separate point estimate of net benefits under a certain combination of cost
and benefit estimation methods. Because it is not a distribution, it is not possible to infer the
likelihood of any single net benefit estimate.
ES-7
-------
$140
$120
$100
c
O
PM2.5 Benefits estimates derived from 2 epidemiology functions and 12 expert
functions
Figure ES-1. Economic Value of Estimated PM2.5-Related Health Co-Benefits According to
Epidemiology or Expert-Derived PM Mortality Risk Estimate3'6
Based on the modeled interim baseline, which is approximately equivalent to the final baseline (see Appendix
5A)
Column total equals sum of PM2.s-related mortality and morbidity benefits.
ES-8
-------
$120
PM2 5 Benefits estimates derived from 2 epidemiology functions and 12 expert
functions
Figure ES-2. Net Benefits of the MATS Rule According to PM2.5 Epidemiology or Expert-
Derived Mortality Risk Estimate3'*1
3 Based on the modeled interim baseline, which is approximately equivalent to the final baseline (see Appendix
5A)
b Column total equals sum of PM2.5-related mortality and morbidity benefits.
ES.2 Not All Benefits Quantified
EPA was unable to quantify or monetize all of the health and environmental benefits
associated with the final MATS Rule. EPA believes these unquantified benefits could be
substantial, including the overall value associated with HAP reductions, value of increased
agricultural crop and commercial forest yields, visibility improvements, and reductions in
nitrogen and acid deposition and the resulting changes in ecosystem functions. Tables ES-5 and
ES-6 provide a list of these benefits.
ES-9
-------
Table ES-5. Human Health Effects of Pollutants Affected by the Mercury and Air Toxics
Standards
Benefits Category
Effect Has
Been
Specific Effect Quantified
Effect Has
Been More
Monetized Information3
Improved Human Health
Reduced incidence of
premature mortality
from exposure to PM2.5
Reduced incidence of
morbidity from
exposure to PM2.5
Reduced incidence of
mortality from
exposure to ozone
Reduced incidence of
morbidity from
exposure to ozone
Adult premature mortality based on cohort S
study estimates and expert elicitation
estimates (age >25 or age >30)
Infant mortality (age <1) S
Non-fatal heart attacks (age > 18) S
Hospital admissions— respiratory (all ages) S
Hospital admissions— cardiovascular (age S
Emergency room visits for asthma (age <18) S
Acute bronchitis (age 8-12) S
Lower respiratory symptoms (age 7-14) S
Upper respiratory symptoms (asthmatics S
age 9-11)
Asthma exacerbation (asthmatics age 6-18) S
Lost work days (age 18-65) S
Minor restricted-activity days (age 18-65) S
Chronic bronchitis (age >26) S
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
Premature mortality based on short-term —
study estimates (all ages)
Premature mortality based on long-term —
study estimates (age 30-99)
Hospital admissions— respiratory causes —
(age > 65)
Hospital admissions— respiratory causes —
(age <2)
Emergency room visits for asthma (all ages) —
Minor restricted-activity days (age 18-65) —
S Section 5.4
•/ Section 5.4
•/ Section 5.4
S Section 5.4
•/ Section 5.4
•/ Section 5.4
•/ Section 5.4
S Section 5.4
•/ Section 5.4
•/ Section 5.4
•/ Section 5.4
•/ Section 5.4
S Section 5.4
- PM ISA0
- PM ISA0
- PMISA°'d
- PMISA°'d
— Ozone CD, Draft
Ozone ISAb
— Ozone CD, Draft
Ozone ISAb
— Ozone CD, Draft
Ozone ISAb
— Ozone CD, Draft
Ozone ISAb
— Ozone CD, Draft
Ozone ISAb
— Ozone CD, Draft
Ozone ISAb
(continued)
ES-10
-------
Table ES-5. Human Health Effects of Pollutants Affected by the Mercury and Air Toxics
Standards (continued)
Benefits Category
Reduced incidence of
morbidity from
exposure to NO2
Reduced incidence of
morbidity from
exposure to SO2
Reduced incidence of
morbidity from
exposure to methyl
mercury (through
reduced mercury
deposition as well as
the role of sulfate in
methylation )
Specific Effect
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
Asthma hospital admissions (all ages)
Chronic lung disease hospital admissions (age
>65)
Respiratory emergency department visits (all
ages)
Asthma exacerbation (asthmatics age 4-18)
Acute respiratory symptoms (age 7-14)
Premature mortality
Other respiratory effects (e.g., airway
hyperresponsiveness and inflammation, lung
function, other ages and populations)
Respiratory hospital admissions (age > 65)
Asthma emergency room visits (all ages)
Asthma exacerbation (asthmatics age 4-12)
Acute respiratory symptoms (age 7-14)
Premature mortality
Other respiratory effects (e.g., airway
hyperresponsiveness and inflammation, lung
function, other ages and populations)
Neurologic effects— IQ loss
Other neurologic effects (e.g., developmental
delays, memory, behavior)
Cardiovascular effects
Genotoxic, immunologic, and other toxic
effects
Effect Has Effect Has
Been Been
Quantified Monetized More Information
— — Ozone CD, Draft
Ozone ISAb
— — Ozone CD, Draft
Ozone ISAb
— — Ozone CD, Draft
Ozone ISA0
— — Ozone CD, Draft
Ozone ISAd
— — Ozone CD, Draft
Ozone ISAd
- - NO2 ISAb
- - NO2 ISAb
- - NO2 ISAb
- - NO2 ISAb
- - NO2 ISAb
- - NO2 ISAc'd
- - NO2 ISAc'd
- - SO2 ISAb
- - SO2 ISAb
- - SO2 ISA
- - SO2 ISAb
S02 ISAc'd
- - SO2 ISAc'd
S S IRIS; NRC, 2000b
- - IRIS; NRC, 2000c
- - IRIS; NRC, 2000c'd
- - IRIS; NRC, 2000c'd
For a complete list of references see Chapter 5.
We assess these benefits qualitatively due to time and resource limitations for this analysis.
We 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.
ES-11
-------
Table ES-6. Environmental Effects of Pollutants Affected by the Mercury and Air Toxics
Standards
Benefits Category Specific Effect
Effect Has Effect Has
Been Been
Quantified Monetized
More
Information3
Improved Environment
Reduced visibility
impairment
Visibility in Class I areas in SE, SW, and
CA regions
Visibility in Class I areas in other regions
Visibility in residential areas
Reduced effects from
PM deposition (metals
and organics)
Effects on Individual organisms and
ecosystems
PM ISA
PM ISA
PM ISAb
Reduced climate
effects
Reduced effects on
materials
Global climate impacts from
Climate impacts from ozone
Other climate impacts (e.g.,
other impacts)
C02
and PM -
other GHGs, —
Household soiling —
Materials damage (e.g., corrosion, —
increased wear)
S Section 5.6
— Section 5.6
- IPCCC
- PM ISA0
- PM ISA0
PM ISA0
Reduced vegetation
and ecosystem effects
from exposure to
ozone
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)
Ozone CD, Draft
Ozone ISA0
Ozone CD, Draft
Ozone ISAb
Ozone CD, Draft
Ozone ISAb'd
Ozone CD, Draft
Ozone ISA0
Ozone CD, Draft
Ozone ISA0
Ozone CD, Draft
Ozone ISA0
Ozone CD, Draft
Ozone ISA0
Ozone CD, Draft
Ozone ISA0
(continued)
ES-12
-------
Table ES-6. Environmental Effects of Pollutants Affected by the Mercury and Air Toxics
Standards (continued)
Benefits Category
Specific Effect
Effect Has Effect Has
Been Been
Quantified Monetized
More
Information
Reduced effects from
acid deposition
Recreational fishing
Tree mortality and decline
Commercial fishing and forestry effects
Recreational demand in terrestrial and
aquatic ecosystems
Other nonuse effects
Ecosystem functions (e.g.,
biogeochemical cycles)
NOX SOX ISA
NOX SOX ISA0
NOX SOX ISA0
NOX SOX ISA0
NOX SOX ISA0
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)
NOX SOX ISA
NOX SOX ISA0
NOX SOX ISA0
NOX SOX ISA0
NOX SOX ISA0
Reduced vegetation
effects from ambient
exposure to SO2 and
NOX
Injury to vegetation from SO2 exposure
Injury to vegetation from NOX exposure
NOXSOXISA°
NOX SOX ISA0
Reduced incidence of
morbidity from
exposure to methyl
mercury (through
reduced mercury
deposition as well as
the role of sulfate in
methylation )
Effects on fish, birds, and mammals (e.g.,
reproductive effects)
Commercial, subsistence and
recreational fishing
Mercury Study
RTC'
,c,d
Mercury Study
RTC°
For a complete list of references see Chapter 5.
We assess these benefits qualitatively due to time and resource limitations for this analysis.
We 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.
ES-13
-------
ES.3 Costs and Employment Impacts
The projected annual incremental private costs of the final MATS Rule to the electric
power industry are $9.6 billion in 2015.1 These costs represent the total cost to the electricity-
generating industry of reducing HAP emissions to meet the emissions limits set out in the rule.
Estimates are in 2007 dollars. These total costs of the rule are estimated using the Integrated
Planning Model (IPM), as well as additional analyses for oil-fired units and monitoring/record-
keeping costs.
There are several national changes in energy prices that result from the final MATS Rule.
Retail electricity prices are projected to increase in the contiguous US by an average of 3.1% in
2015 with the final MATS Rule. On a weighted average basis between 2015 and 2030,
consumer natural gas price anticipated to increase from 0.3% to 0.6% depending on consumer
class in response to the final MATS Rule.
There are several other types of energy impacts associated with the final MATS Rule. A
small amount of coal-fired capacity, about 4.7 GW (less than 2 percent of all coal-fired capacity
in 2015), is projected to become uneconomic to maintain by 2015. These units are
predominantly smaller and less frequently-used generating units dispersed throughout the
contiguous US. If current forecasts of either natural gas prices or electricity demand were
revised in the future to be higher, that would create a greater incentive to keep these units
operational. Coal production for use in the power sector is projected to decrease by 1 percent
by 2015, and we expect slightly reduced coal demand in Appalachia and the West with the final
MATS Rule.
In addition to addressing the costs and benefits of the final MATS Rule, EPA has
estimated a portion of the employment impacts of this rulemaking. We have estimated two
types of impacts. One provides an estimate of the employment impacts on the regulated
industry over time. The second covers the short-term employment impacts associated with the
construction of needed pollution control equipment until the compliance date of the
regulation. We expect that the rule's impact on employment will be small, but will (on net)
result in an expected increase in employment.
1 The year 2016 is the compliance year for MATS, though as we explain in later chapters, we use 2015 as a proxy
for compliance in 2016 for IPM emissions, costs and economic impact analysis due to availability of modeling
impacts in that year.
ES-14
-------
The approaches to estimate employment impacts use different analytical techniques,
are applied to different industries during different time periods, and use different units of
analysis. No overlapping estimates are summed. Estimates of employment changes per dollar
of expenditure on pollution control from Morgenstern et al. (2002) are used to estimate the
ongoing annual employment impacts for the regulated entities (the electric power sector) as a
result of this rule. The short term estimates for employment needed to design, construct, and
install the control equipment in the three year period before the compliance date are also
provided using an approach that estimates employment impacts for the environmental
protection sector based on forecast changes from IPM on the number and scale of pollution
controls and labor intensities in relevant sectors. Finally, some of the other types of
employment impacts that will be ongoing are estimated using IPM outputs and labor
intensities, as reported in Chapter 6, but not included in this table because they omit some
potentially important categories.
In Table ES-7, we show the employment impacts of the MATS Rule as estimated by the
environmental protection sector approach and by the Morgenstern approach.
Table ES-7. Estimated Employment Impact Table
Annual (Reoccurring)
One Time (Construction During
Compliance Period)
Environmental protection sector
approach3
Not applicable
46,000
Net effect on electric utility sector
employment from Morgenstern
et al., approach0
8,000
-15,000 to 30,000d
Not Applicable
3 These one-time impacts on employment are estimated in terms of job-years.
This estimate is not statistically different from zero.
These annual or reoccurring employment impacts are estimated in terms of production workers as defined by
the US Census Bureau's Annual Survey of Manufacturers (ASM).
95% confidence interval
ES.4 Small Entity and Unfunded Mandates Impacts
After preparing an analysis of small entity impacts, EPA cannot certify that there will be
no SISNOSE (significant economic impacts on a substantial number of small entities) for this
rule. Of the 82 small entities affected, 40 are projected to have costs greater than 1 percent of
their revenues. The exclusion of units smaller than 25 Megawatt capacity (MW) as per the
requirements of the Clean Air Act has already significantly reduced the burden on small entities,
ES-15
-------
and EPA participated in a Small Business Regulatory Enforcement Fairness Act (SBREFA) Panel
to examine ways to mitigate the impact of the proposed Toxics Rule on affected small entities
EPA examined the potential economic impacts on state and municipality-owned entities
associated with this rulemaking based on assumptions of how the affected states will
implement control measures to meet their emissions. These impacts have been calculated to
provide additional understanding of the nature of potential impacts and additional information.
According to EPA's analysis, of the 96 government entities considered in this, EPA
projects that 42 government entities will have compliance costs greater than 1 percent of base
generation revenue in 2015, based on our assumptions of how the affected states implement
control measures to meet their emissions budgets as set forth in this rulemaking.
Government entities projected to experience compliance costs in excess of 1 percent of
revenues may have some potential for significant impact resulting from implementation of
MATS.
ES.5 Limitations and Uncertainties
Every analysis examining the potential benefits and costs of a change in environmental
protection requirements is limited to some extent by data gaps, limitations in model
capabilities (such as geographic coverage), and variability or uncertainties in the underlying
scientific and economic studies used to configure the benefit and cost models. Despite the
uncertainties, we believe this benefit-cost analysis provides a reasonable indication of the
expected economic benefits and costs of the final MATS Rule.
For this analysis, such uncertainties include possible errors in measurement and
projection for variables such as population growth and baseline incidence rates; uncertainties
associated with estimates of future-year emissions inventories and air quality; variability in the
estimated relationships between changes in pollutant concentrations and the resulting changes
in health and welfare effects; and uncertainties in exposure estimation.
Below is a summary of the key uncertainties of the analysis:
Costs
• Compliance costs are used to approximate the social costs of this rule. Social costs
may be higher or lower than compliance costs and differ because of preexisting
distortions in the economy, and because certain compliance costs may represent
shifts in rents.
ES-16
-------
• Analysis does not capture employment shifts as workers are retrained at the same
company or re-employed elsewhere in the economy.
• We do not include the costs of certain relatively small permitting costs associated
with updating Title V permits.
• Technological innovation is not incorporated into these cost estimates. Thus, these
cost estimates may be potentially higher than what may occur in the future, all other
things being the same.
Benefits
* The mercury concentration estimates for the analysis come from several different
sources.
• The mercury concentration estimates used in the model were based on simple
temporal and spatial averages of reported fish tissue samples. This approach
assumes that the mercury samples are representative of "local" conditions (i.e.,
within the same HUC 12) in similar waterbodies (i.e., rivers or lakes).
• State-level averages for fishing behavior of recreational anglers are applied to each
modeled census tract in the state; which does not reflect within-state variation in
these factors.
• Application of state-level fertility rates to specific census tracts (and specifically to
women in angler households.
• Applying the state-level individual level fishing participation rates to approximate
the household fishing rates conditions at a block level.
• Populations are only included in the model if they are within a reasonable distance
of a waterbody with fish tissue MeHg samples. This approach undercounts the
exposed population (by roughly 40 to 45%) and leads to underestimates of national
aggregate baseline exposures and risks and underestimates of the risk reductions
and benefits resulting from mercury emission reductions.
• Assumption of 8 g/day fish consumption rate for the general population in
freshwater angler households.
• The dose-response model used to estimate neurological effects on children because
of maternal mercury body burden has several important uncertainties, including
selection of IQ as a primary endpoint when there may be other more sensitive
endpoints, selection of the blood-to-hair ratio for mercury, and the dose-response
estimates from the epidemiological literature. Control for confounding from the
ES-17
-------
potentially positive cognitive effects offish consumption and, more specifically,
omega-3 fatty acids.
Valuation of IQ losses using a lost earning approach has several uncertainties,
including (1) there is a linear relationship between IQ changes and net earnings
losses, (2) the unit value applies to even very small changes in IQ, and (3) the unit
value will remain constant (in real present value terms) for several years into the
future. Each unit value for IQ losses has two main sources of uncertainty (1). The
statistical error in the average percentage change in earnings as a result of IQ
changes and (2) estimates of average lifetime earnings and costs of schooling.
Based on the modeled interim baseline which is approximately equivalent to the
final baseline (see Appendix 5A), 11% and 73% of the estimated avoided premature
deaths occur at or above an annual mean PM2.5 level of 10 u.g/m3 (the LML of the
Laden etal. 2006 study) and 7.5 u.g/m3(the LML of the Popeetal. 2002 study),
respectively. These are the source studies for the concentration-response functions
used to estimate mortality benefits. As we model avoided premature deaths among
populations exposed to levels of PM2.5 that are successively lower than the LML of
each study our confidence in the results diminishes. However, studies using data
from more recent years, during which time PM concentrations have fallen, continue
to report strong associations with mortality.
There are uncertainties related to the health impact functions used in the analysis.
These include: within study variability; across study variation; the application of
concentration-response (C-R) functions nationwide; extrapolation of impact
functions across population; and various uncertainties in the C-R function, including
causality and thresholds. Therefore, benefits may be under- or over-estimates.
Analysis is for 2016, 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.
This analysis omits certain unquantified effects due to lack of data, time and
resources. These unquantified endpoints include other health and ecosystem
effects. EPA will continue to evaluate new methods and models and select those
most appropriate for estimating the benefits of reductions in air pollution. Enhanced
collaboration between air quality modelers, epidemiologists, toxicologists,
ecologists, and economists should result in a more tightly integrated analytical
framework for measuring benefits of air pollution policies.
PM2.s mortality co-benefits represent a substantial proportion of total monetized
benefits (over 90%), and these estimates have following key assumptions and
uncertainties.
ES-18
-------
o The PM2.5 -related co-benefits of the alternative scenarios were derived
through a benefit per-ton approach, which does not fully reflect local
variability in population density, meteorology, exposure, baseline health
incidence rates, or other local factors that might lead to an over-estimate
or under-estimate of the actual benefits of this rule.
o 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 via transported
precursors emitted from EGUs may differ significantly from direct PM2.5
released from diesel engines and other industrial sources, but no clear
scientific grounds exist for supporting differential effects estimates by
particle type.
o We assume that the health impact function for fine particles is linear
within the range of ambient concentrations under consideration. Thus,
the estimates include health benefits from reducing fine particles in areas
with varied concentrations of PM2.5, including both regions that are in
attainment with fine particle standard and those that do not meet the
standard down to the lowest modeled concentrations.
o To characterize the uncertainty in the relationship between PM2.5and
premature mortality, 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.
ES-19
-------
ES.6 References
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. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press.
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, doi: 10.1111/J.1539-
6924.2009.01227.x [Online 9 Apr 2009].
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.
U.S. Environmental Protection Agency (EPA). December 2010. Guidelines for Preparing
Economic Analyses. EPA 240-R-10-001.
U.S. Office of Management and Budget (OMB). 2003. Circular A-4 Guidance to Federal Agencies
on Preparation of Regulatory Analysis.
Woodruff, T.J., J. Grillo, and K.C. Schoendorf. 1997. "The Relationship Between Selected Causes
of Postneonatal Infant Mortality and Particulate Infant Mortality and Particulate Air
Pollution in the United States." Environmental Health Perspectives 105(6):608-612.
ES-20
-------
CHAPTER 1
INTRODUCTION AND BACKGROUND
1.1 Introduction
In this action, EPA is addressing the emissions of mercury and other hazardous air
pollutants by coal- and oil-fired electricity generating units. This document presents the health
and welfare benefits of the final Mercury and Air Toxics Standards and compares the benefits of
this rule to the estimated costs of implementing the rule in 2016. This chapter contains
background information on the rule and an outline of the chapters of this Regulatory Impact
Analysis (RIA).
1.2 Background for Final Mercury and Air Toxics Standards
1.2.1 NESHAP
This action finalizes National Emission Standards for Hazardous Air Pollutants (NESHAP)
for new and existing coal- and oil-fired electricity generating units (EGUs) meeting the definition
found in Clean Air Act (CAA) section 112(a)(8). EPA is promulgating these standards to meet its
statutory obligation to address HAP emissions from these sources under CAA section 112(d).
The final NESHAP for new and existing coal- and oil-fired EGUs will be promulgated under 40
CFRpart63, subpart UUUUU.
On December 20, 2000 (65 FR 79825), EPA determined that regulation of coal- and oil-
fired EGUs under CAA section 112 was appropriate and necessary, in accordance with CAA
section 112(n)(l)(A). EPA at the same time added coal- and oil-fired EGUs to the list of source
categories requiring regulation under CAA section 112(d). The December 2000 listing triggered
the deadline established by Congress in CAA section 112(c)(5) under which EPA has two years
from the date of listing in which to promulgate "emissions standards under section (d) of this
section."
In 2002, EPA initiated a CAA section 112(d) standard setting process for coal- and oil-
fired EGUs, and on January 30, 2004, proposed CAA section 112(d) standards for mercury (Hg)
emissions from coal-fired EGUs and nickel (Ni) emissions from oil-fired EGUs, and, in the
alternative, proposed to remove EGUs from the CAA section 112(c) list based on a finding that it
was neither appropriate nor necessary to regulate EGUs pursuant to CAA section 112. EPA
never finalized the proposed CAA section 112(d) standard. The Agency finalized the CAA section
111 alternative, after taking and responding to extensive public comments on both sets of
regulatory options, by issuing a de-listing rule (Section 112(n) Revision Rule; 70 FR 15994;
1-1
-------
March 29, 2005) and a final rule, the Clean Air Mercury Rule (CAMR), establishing Hg emissions
standards for coal-fired EGUs under CAA section 111 on May 18, 2005 (70 FR 28606). The
removal of EGUs from the CAA section 112 list was challenged in the United States (U.S.) Court
of Appeals for the District of Columbia Circuit (D.C. Circuit Court).
Petitions for reconsideration were filed by a number of parties in summer 2005. EPA
responded to the petitions with a final notice of reconsideration on June 9, 2006 (71 FR 33388).
Petitions for judicial review were filed on November 29, 2006, by a number of parties1 (State of
New Jersey, etal., v. EPA, 517 F.3d 574).
On Februarys, 2008, the D.C. Circuit Court vacated the Section 112(n) Revision Rule
(State of New Jersey, et al., v. EPA, 517 F.3d 574), and subsequently denied rehearing and
rehearing en bane of that decision. As a part of the decision, the D.C. Circuit Court also vacated
CAMR, reverting to the December 2000 regulatory determination and requiring the
development of emission standards under CAA section 112(d) (MACT standards) for coal- and
oil-fired EGUs. The litigation process continued until, on January 29, 2009, EPA requested of the
Department of Justice (DOJ) that the Government's appeals be withdrawn.
On December 18, 2008, several environmental and public health organizations
("Plaintiffs")2 filed a complaint in the D.C. District Court (Civ. No. l:08-cv-02198 (RMC)) alleging
that the Agency had failed to perform a nondiscretionary duty under CAA section 304(a)(2) by
failing to promulgate final section 112(d) standards for HAP from coal- and oil-fired EGUs by the
statutorily-mandated deadline, December 20, 2002, 2 years after such sources were listed
under section 112(c). EPA settled that litigation. A Consent Decree was issued on April 15, 2010,
that calls for EPA to, no later than March 16, 2011, sign for publication in the Federal Register a
notice of proposed rulemaking setting forth EPA's proposed emission standards for coal- and
oil-fired EGUs and, no later than November 16, 2011, sign for publication in the Federal Register
a notice of final rulemaking. EPA and the litigants agreed to a 30-day extension in order to fully
respond to the 960,000 comments received on the proposed rule. This agreement extended the
signing deadline to December 16, 2011.
1 Environmental Petitioners; the National Congress of American Indians and Treaty Tribes; ARIPPA; American Coal
for Balanced Mercury Regulations, et al.; United Mine Workers of America; Alaska Industrial Development and
Export Authority; the States of New Jersey, California, Connecticut, Delaware, Illinois, Maine, Maryland,
Massachusetts, Michigan, Minnesota, New Hampshire, New Mexico, New York, Pennsylvania, Rhode Island,
Vermont, and Wisconsin; and the City of Baltimore, MD.
2 American Nurses Association, Chesapeake Bay Foundation, Inc., Conservation Law Foundation, Environment
America, Environmental Defense Fund, Izaak Walton League of America, Natural Resources Council of Maine,
Natural Resources Defense Council, Physicians for Social Responsibility, Sierra Club, The Ohio Environmental
Council, and Waterkeeper Alliance, Inc.
1-2
-------
On March 16, 2011, in response to the D.C. Circuit Court's vacatur, EPA proposed CAA
section 112(d) NESHAP for all coal- and oil-fired EGUs that reflect the application of the
maximum achievable control technology (MACT) consistent with the requirements of CAA
sections 112(d)(2) and (3). This action finalizes that proposed rule. This final rule is intended to
protect air quality and promote public health by reducing emissions of the hazardous air
pollutants (HAP) listed in CAA section 112(b).
1.2.2 NSPS
Section lll(b)(l)(b) of the CAA requires EPA to periodically review and revise the New
Source Performance Standards (NSPS) as necessary to reflect improvements in methods for
reducing emissions. The NSPS for EGUs (40 CFR part 60, subpart Da) were originally
promulgated on June 11,1979 (44 FR 33580). On February 27, 2006, EPA promulgated
amendments to the NSPS for particulate matter (PM), sulfur dioxide (S02), and nitrogen oxides
(NOX) contained in the standards of performance for EGUs (71 FR 9866). EPA was subsequently
sued by the offices of multiple states attorneys general and environmental organizations on the
amendments. The Petitioners alleged that EPA failed to correctly identify the best system of
emission reductions for the amended S02 and NOX standards. The Petitioners also claimed that
it is appropriate to establish emission limits for fine particulate matter and condensable
particulate matter. Based upon further examination of the record, EPA has determined that
certain issues in the rule warrant further consideration. On September 4, 2009, EPA was
granted a voluntary remand without vacatur of the 2006 amendments. EPA considers it
appropriate to respond to the NSPS voluntary remand in conjunction with the ECU NESHAP
since it allows EPA to more comprehensively consider the impact on the utility sector.
Therefore, even though there was no judicial timetable to complete the NSPS remand, EPA
proposed it in conjunction with the NESHAP. We also proposed several minor amendments,
technical clarifications, and corrections to existing provisions of the fossil fuel-fired ECU and
large and small industrial-commercial-institutional steam generating units NSPS, 40 CFR part 60,
subparts D, Db, and DC. The NSPS and amendments are being finalized along with the NESHAP
in this action.
The title Mercury and Air Toxics Standards (MATS) used in the remainder of this RIA
refers to the combination of the ECU NESHAP and NSPS.
1.3 Appropriate & Necessary Analyses
In the preamble to the proposed rule, EPA confirmed the December 2000 finding that it
is appropriate to regulate emissions of Hg and other HAP from EGUs because emissions of
1-3
-------
those pollutants pose hazards to public health and the environment and EGUs are the largest or
among the largest contributors of many of those HAP. We also confirmed that it is necessary to
regulate EGUs under section 112 for a variety of reasons, including that hazards to public health
and the environment posed by HAP emissions from EGUs remain after imposition of the
requirements of the Clean Air Act. This confirmation was supported in part by several new
analyses of the hazards to public health posed by both mercury and non-mercury HAP. For
more information on the finding and the analyses to support them, please refer to the
preamble of the final rule.
1.4 Provisions of the Final Mercury and Air Toxics Standards
1.4.1 What Is the Source Category Regulated by the Final Rule?
The final MATS addresses emissions from new and existing coal- and oil-fired EGUs. In
general, if an ECU burns coal (either as a primary fuel or as a supplementary fuel) or any
combination of coal with another fuel where the coal accounts for more than 10 percent of the
average annual heat input during any 3 calendar years or for more than 15 percent of the
annual heat input during any one calendar year, the unit is considered to be coal-fired under
this final rule. If a unit is not a coal-fired unit and burns only oil or burns oil in combination with
a fuel other than coal where the oil accounts for more than 10 percent of the average annual
heat input during any 3 calendar years or for more than 15 percent of the annual heat input
during any one calendar year, the unit is considered to be oil-fired under this final rule.
CAA section 112(a)(8) defines an ECU as:
a fossil fuel-fired combustion unit of more than 25 megawatts electric (MWe) that
serves a generator that produces electricity for sale. A unit that cogenerates steam and
electricity and supplies more than one-third of its potential electric output capacity and
more than 25 MWe output to any utility power distribution system for sale is also an
electric utility steam generating unit.
This action established 40 CFR part 63, subpart UUUUU, to address HAP emissions from
new and existing coal- and oil-fired EGUs. EPA must determine what is the appropriate
maximum achievable control technology (MACT) for those units under sections 112(d)(2) and
(d)(3)oftheCAA.
EPA has divided coal- and oil-fired EGUs into the following subcategories:
• Units designed for not low rank virgin coal;
1-4
-------
• units designed for low rank virgin coal;
• IGCC units;
• Solid oil-derived fuel-fired units;
• Continental liquid oil-fired units; and
• Non-continental liquid oil-fired units.
1.4.2 What Are the Pollutants Regulated by the Rule?
The final NESHAP regulates emissions of HAP. Available emissions data show that
several HAP that are formed during the combustion process or which are contained within the
fuel burned are emitted from coal- and oil-fired EGUs. The individual HAP include mercury,
arsenic, cadmium, lead, and nickel, among others. EPA describes the health effects of these and
other HAP emitted from the operation of coal- and oil-fired EGUs in Chapter 4 of this RIA. These
HAP emissions are known to cause or contribute significantly to air pollution, which may
reasonably be anticipated to endanger public health or welfare.
In addition to reducing HAP, the emission control technologies that will be installed on
coal- and oil-fired EGUs to reduce HAP will also reduce sulfur dioxide (S02) and particulate
matter (PM). A wide range of human health and welfare effects are linked to the emissions of
PM and S02. These human health and welfare effects are discussed extensively in Chapter 5 of
this RIA.
1.4.3 What Are the Emissions Limits?
Under section 112(d), EPA must establish emission standards for major sources that
"require the maximum degree of reduction in emissions of the HAP subject to this section" that
EPA determines is achievable taking into account certain statutory factors. These are referred
to as maximum achievable control technology or MACT standards. The MACT standards for
existing sources must be at least as stringent as the average emissions limitation achieved by
the best performing 12 percent of existing sources in the category (for which the Administrator
has emissions information) or the best performing 5 sources for source categories with less
than 30 sources. This level of minimum stringency is referred to as the MACT floor, and EPA
cannot consider cost in setting the floor. For new sources, MACT standards must be at least as
stringent as the control level achieved in practice by the best controlled similar source.
The numerical emission standards that are being finalized for new and existing coal- and
oil-fired EGUs units are shown in Tables 1-1 and 1-2. In some cases, affected sources have the
1-5
-------
choice of complying with an emissions standard per unit of input or an output based standard,
which are provided in parentheses below the input-based standard. These standards must be
complied with on a 30-day rolling average basis if using continuous monitoring. If
demonstrating compliance on the basis of a stack test, units must demonstrate compliance by
conducting periodic stack tests on a quarterly basis.
Table 1-1. Emission Limitations for Coal-Fired and Solid Oil-Derived Fuel-Fired EGUs
Filterable Part icu late
Subcategory Matter
Existing Unit designed for not
low rank virgin coal
Existing Unit designed for low
rank virgin coal
Existing - IGCC
Existing -Solid oil-derived
New unit designed for not low
rank virgin coal
New unit designed for coal low
rank virgin coal
New -IGCC
New -Solid oil-derived
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.040 Ib/MMBtu
(0.40 Ib/MWh)
0.0080 Ib/MMBtu
(0.090 Ib/MWh)
0.0070 Ib/MWh
0.0070 Ib/MWh
0.070 lb/MWhb
0.090 lb/MWhc
0.020 Ib/MWh
Hydrogen Chloride
0.0020 Ib/MMBtu
(0.020 Ib/MWh)
0.0020 Ib/MMBtu
(0.020 Ib/MWh)
0.00050 Ib/MMBtu
(0.0050 Ib/MWh)
0.0050 Ib/MMBtu
(0.080 Ib/MWh)
0.00040 Ib/MWh
0.00040 Ib/MWh
0.0020 lb/MWhd
0.00040 Ib/MWh
Mercury
1.2 Ib/TBtu
(0.020 Ib/GWh)
4.0 lb/TBtua
(0.040 lb/GWha)
2.5 Ib/TBtu
(0.030 Ib/GWh)
0.20 Ib/TBtu
(0.0020 Ib/GWh)
0.00020 Ib/GWh
0.040 Ib/GWh
0.0030 lb/GWhe
0.0020 Ib/GWh
Note: In some cases, affected units may comply with either an input-based standard or an output-based standard,
shown in parentheses below the input-based standard.
Ib/MMBtu = pounds pollutant per million British thermal units fuel input
Ib/TBtu = pounds pollutant per trillion British thermal units fuel input
Ib/MWh = pounds pollutant per megawatt-hour electric output (gross)
Ib/GWh = pounds pollutant per gigawatt-hour electric output (gross)
a Beyond-the-floor limit. The MACT floor for this subcategory is 11.0 Ib/TBtu (0.20 Ib/GWh)
b Duct burners on syngas; based on permit levels in comments received
c Duct burners on natural gas; based on permit levels in comments received
d Based on best-performing similar source
e Based on permit levels in comments received
1-6
-------
Table 1-2. Emission Limitations for Liquid Oil-Fired EGUs
Subcategory
Existing- Liquid
Existing- Liquid
New- Liquid oil
New- Liquid oil
oil-continental
oil-non-continental
- continental
- non-continental
Filterable PM
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.070 Ib/MWh
0.20 Ib/MWh
Hydrogen Chloride
0.0020 Ib/MMBtu
(0.010 Ib/MWh)
0.00020 Ib/MMBtu
(0.0020 Ib/MWh)
0.00040 Ib/MWh
0.0020 Ib/MWh
Hydrogen Fluoride
0.00040 Ib/MMBtu
(0.0040 Ib/MWh)
0.000060 Ib/MMBtu
(0.00050 Ib/MWh)
0.00040 Ib/MWh
0.00050 Ib/MWh
Note: In some cases, affected units may comply with either an input-based standard or an output-based standard,
shown in parentheses below the input-based standard.
We are also finalizing alternate equivalent emission standards for certain subcategories
in three areas: S02 (for HCI), individual non-Hg metals, and total non-Hg metals (for filterable
PM) from coal- and solid oil-derived fuel-fired EGUs, and individual and total metals (for
filterable PM) from oil-fired EGUs. These alternate emission limitations are provided in Tables
1-3 and 1-4. We are finalizing an alternate limitation of 1 percent moisture in the liquid oil as
an alternate to the HCI and HF emission limits for both liquid oil subcategories (i.e., continental
and non-continental).
1-7
-------
Table 1-3. Alternate Emission Limitations for Existing Coal- and Oil-Fired EGUs
00
Liquid Oil
S02
Total non-Hg metals
Antimony, Sb
Arsenic, As
Beryllium, Be
Cadmium, Cd
Chromium, Cr
Cobalt, Co
Lead, Pb
Manganese, Mn
Mercury, Hg
Nickel, Ni
Selenium, Se
Coal-fired EGUs
0.20 Ib/MMBtu
(1.5 Ib/MWh)
0.000050 Ib/MMBtu
(0.50 Ib/GWh)
0.80 Ib/TBtu
(0.0080 Ib/GWh)
1.1 Ib/TBtu
(0.020 Ib/GWh)
0.20 Ib/TBtu
(0.0020 Ib/GWh)
0.30 Ib/TBtu
(0.0030 Ib/GWh)
2.8 Ib/TBtu
(0.030 Ib/GWh)
0.80 Ib/TBtu
(0.0080 Ib/GWh)
1.2 Ib/TBtu
(0.020 Ib/GWh)
4.0 Ib/TBtu
(0.050 Ib/GWh)
NA
3.5 Ib/TBtu
(0.040 Ib/GWh)
5.0 Ib/TBtu
(0.060 Ib/GWh)
IGCC
NA
0.000060 Ib/MMBtu
(0.50 Ib/GWh)
1.4 Ib/TBtu
(0.020 Ib/GWh)
1.5 Ib/TBtu
(0.020 Ib/GWh)
0.10 Ib/TBtu
(0.0010 Ib/GWh)
0.15 Ib/TBtu
(0.0020 Ib/GWh)
2.9 Ib/TBtu
(0.030 Ib/GWh)
1.2 Ib/TBtu
(0.020 Ib/GWh)
190 Ib/MMBtu
(1.8 Ib/MWh)
2.5 Ib/TBtu
(0.030 Ib/GWh)
NA
6.5 Ib/TBtu
(0.070 Ib/GWh)
22 Ib/TBtu
(0.30 Ib/GWh)
Continental
NA
0.00080 Ib/MMBtu
(0.0080 lb/MWh)a
13 Ib/TBtu
(0.20 Ib/GWh)
2.8 Ib/TBtu
(0.030 Ib/GWh)
0.20 Ib/TBtu
(0.0020 Ib/GWh)
0.30 Ib/TBtu
(0.0020 Ib/GWh)
5.5 Ib/TBtu
(0.060 Ib/GWh)
21 Ib/TBtu
(0.30 Ib/GWh)
8.1 Ib/TBtu
(0.080 Ib/GWh)
22 Ib/TBtu
(0.30 Ib/GWh)
0.20 Ib/TBtu
(0.0020 Ib/GWh)
110 Ib/TBtu
(1.1 Ib/GWh)
3.3 Ib/TBtu
(0.040 Ib/GWh)
Non-continental
NA
0.00060 Ib/MMBtu
(0.0070 lb/MWh)a
2.2 Ib/TBtu
(0.020 Ib/GWh)
4.3 Ib/TBtu
(0.080 Ib/GWh)
0.60 Ib/TBtu
(0.0030 Ib/GWh)
0.30 Ib/TBtu
(0.0030 Ib/GWh)
31 Ib/TBtu
(0.30 Ib/GWh)
110 Ib/TBtu
(1.40 Ib/GWh)
4.9 Ib/TBtu
(0.080 Ib/GWh)
20 Ib/TBtu
(0.30 Ib/GWh)
0.040 Ib/TBtu
(0.00040 Ib/GWh)
470 Ib/TBtu
(4.1 Ib/GWh)
9.8 Ib/TBtu
(0.20 Ib/GWh)
Solid Oil-derived
0.30 Ib/MMBtu
(2.0 Ib/MWh)
0.000040 Ib/MMBtu
(0.6 Ib/GWh)
0.80 Ib/TBtu
(0.0080 Ib/GWh)
0.30 Ib/TBtu
(0.0050 Ib/GWh)
0.060 Ib/TBtu
(0.00060 Ib/GWh)
0.30 Ib/TBtu
(0.0040 Ib/GWh)
0.8 Ib/TBtu
(0.020 Ib/GWh)
1.1 Ib/TBtu
(0.020 Ib/GWh)
0.80 Ib/TBtu
(0.020 Ib/GWh)
2.3 Ib/TBtu
(0.040 Ib/GWh)
NA
9.0 Ib/TBtu
(0.2 Ib/GWh)
1.2 Ib/TBtu
(0.020 Ib/GWh)
NA = Not applicable
a Includes Hg
-------
Table 1-4. Alternate Emission Limitations for New Coal- and Oil-Fired EGUs
Liquid Oil, Ib/GWh
S02
Total metals
Antimony, Sb
Arsenic, As
Beryllium, Be
Cadmium, Cd
Chromium, Cr
Cobalt, Co
Lead, Pb
Mercury, Hg
Manganese, Mn
Nickel, Ni
Selenium, Se
Coal-fired EGUs
0.40 Ib/MWh
0.060 Ib/GWh
0.0080 Ib/GWh
0.0030 Ib/GWh
0.00060 Ib/GWh
0.00040 Ib/GWh
0.0070 Ib/GWh
0.0020 Ib/GWh
0.0020 Ib/GWh
NA
0.0040 Ib/GWh
0.040 Ib/GWh
0.0060 Ib/GWh
IGCCa
0.40 Ib/MWh
0.40 Ib/GWh
0.020 Ib/GWh
0.020 Ib/GWh
0.0010 Ib/GWh
0.0020 Ib/GWh
0.040 Ib/GWh
0.0040 Ib/GWh
0.0090 Ib/GWh
NA
0.020 Ib/GWh
0.070 Ib/GWh
0.30 Ib/GWh
Continental
NA
0.00020
lb/MWhb
0.010
0.0030
0.00050
0.00020
0.020
0.030
0.0080
0.00010
0.020
0.090
0.020
Non-continental
NA
0.0070
lb/MWhb
0.0080
0.060
0.0020
0.0020
0.020
0.30
0.030
0.00040
0.10
4.1
0.020
Solid Oil-Derived
0.40 Ib/MWh
0.60 Ib/GWh
0.0080 Ib/GWh
0.0030 Ib/GWh
0.00060 Ib/GWh
0.00070 Ib/GWh
0.0060 Ib/GWh
0.0020 Ib/GWh
0.020 Ib/GWh
NA
0.0070 Ib/GWh
0.040 Ib/GWh
0.0060 Ib/GWh
NA = Not applicable
a Based on best-performing similar source
b Includes Hg
EPA is finalizing a beyond-the-floor standard for Hg only of 4.0 Ibs/trillion BTU for all
existing and new units designed to burn low BTU virgin coal based on the availability of
activated carbon injection (ACI) for cost-effective Hg control. When considering beyond-the-
floor options, EPA must consider not only the maximum degree of reduction in emissions of
HAP, but must take into account costs, energy, and non-air quality health and environmental
impacts when doing so. We are finalizing a beyond-the-floor standard for these units
because the Agency considers the cost of incremental reductions beyond the MACT floor
standard of 11 Ibs/trillion BTUs to be reasonable. While the primary IPM analysis discussed in
Chapter 3 requires compliance with the beyond-the-floor limit, EPA performed a supplemental
analysis at proposal that estimates the difference in impacts between regulating coal-fired units
designed for lignite at the floor limit and at the beyond-the-floor limit modeled. This analysis
(the IPM Beyond the Floor Cost TSD) shows that if the units were only required to meet a
standard of 11 Ibs/trillion BTUs, the units would emit approximately an additional 3,854 Ibs at a
1-9
-------
reduced annualized cost of $86.7 million. EPA also performed an analysis of beyond-the-floor
alternatives which can be found in the Beyond the MACT Floor Analysis TSD. Based on these
analyses, EPA concluded that the beyond-the-floor standard achieved significant additional
benefits when compared to the costs of the standard.
Pursuant to CAA section 112(h), we are finalizing a work practice standard for organic
HAP, including emissions of dioxins and furans, from all subcategories of ECU. The work
practice standard being finalized for these EGUs would require the implementation of an
annual performance test program as described the preamble. We are finalizing work practice
standards because the data confirm that the significant majority of the measured organic HAP
emissions from EGUs are below the detection levels of the EPA test methods, and, as such, EPA
considers it impracticable to reliably measure emissions from these units.
The ECU NESHAP PM and S02 standards for new and modified facilities are as stringent
or more stringent than the NSPS amendments. Thus, the only impacts unique to the NSPS
amendments are those for the NOx emissions limits for new and modified facilities. In the
baseline for this analysis and in compliance with MATS, no source is expected to trigger the
NSPS limitations for new or modified sources. Therefore, we have concluded that there are no
costs or benefits associated with the NSPS amendments that are unique to these amendments.3
The NSPS requirements are described in detail in the preamble.
1.4.4 What are the Startup, Shutdown, and Malfunction Requirements?
Consistent with Sierra Club v. EPA (551 F.3d 1019 (DC Cir. 2008), cert, denied, 130 S. Ct.
1735 (U.S. 2010)), EPA proposed numerical emission standards that would apply at all times,
including during periods of startup, shutdown, and malfunction. In this final rule, EPA has
evaluated comments and other data concerning startup and shutdown periods and, for the
reasons explained below, is establishing work practice standards for startup and shutdown
periods as the terms are defined in the final rule.
EPA has revised this final rule to require sources to meet a work practice standard,
which requires following the manufacturer's recommended procedures for minimizing periods
of startup and shutdown, for all subcategories of new and existing coal- and oil-fired EGUs (that
would otherwise be subject to numeric emission limits) during periods of startup and
shutdown. As discussed elsewhere in the preamble, we considered whether performance
testing, and therefore, enforcement of numeric emission limits, would be practicable during
3 If the NESHAP requirements were not simultaneously analyzed with the NSPS amendments, then we would
expect that the cost and benefits of the NSPS would be small.
1-10
-------
periods of startup and shutdown. After reviewing comments and other data regarding the
nature of these periods of operation, the EPA is finalizing a work practice standard for periods
of start up and shut down. EPA will revisit this decision during the mandatory 8-year review
cycle.
Periods of startup, normal operations, and shutdown are all predictable and routine
aspects of a source's operations. However, by contrast, malfunction is defined as a "sudden,
infrequent, and not reasonably preventable failure of air pollution control and monitoring
equipment, process equipment or a process to operate in a normal or usual manner..." (40 CFR
63.2) EPA has determined that malfunctions should not be viewed as a distinct operating mode
and, therefore, any emissions that occur at such times do not need to be factored into
development of CAA section 112(d) standards, which, once promulgated, apply at all times.
In the event that a source fails to comply with the applicable CAA section 112(d)
standards as a result of a malfunction event, EPA would determine an appropriate response
based on, among other things, the good faith efforts of the source to minimize emissions during
malfunction periods, including preventative and corrective actions, as well as root cause
analyses to ascertain and rectify excess emissions. EPA would also consider whether the
source's failure to comply with the CAA section 112(d) standard was, in fact, "sudden,
infrequent, not reasonably preventable" and was not instead "caused in part by poor
maintenance or careless operation" (40 CFR 63.2).
1.5 Baseline and Years of Analysis
The emissions scenarios for the RIA reflect the Cross-State Air Pollution Rule (CSAPR) as
finalized in July 2011 and the emissions reductions of SOx, NOx, directly emitted PM, and C02
are consistent with application of federal rules, state rules and statutes, and other binding,
enforceable commitments in place by December 2010 for the analysis timeframe. Consistent
with the mercury risk deposition modeling for MATS, EPA did not model non-federally
enforceable mercury-specific emissions reduction rules in the base case or MATS policy case
(see preamble Section III.A for further detail). This approach does not significantly affect the
projections underlying the cost and benefit results presented in this RIA. The baseline
specifications used for these analyses are described in more detail in Chapter 3, Chapter 4, and
Chapter 5 of this RIA. The ECU and non-EGU regulatory and air quality baseline used for the co-
benefits analysis is described in Appendix 5A.
The costs and co-benefits from reductions in S02 and direct PM emissions are calculated
using a baseline that includes the Cross State Air Pollution Rule (CSAPR; 76 FR 48208) finalized
1-11
-------
July 6, 2011. EPA has subsequently proposed minor modifications to the state-level S02 budgets
in CSAPR. These modifications are expected to result in small changes in the levels of S02
emission reductions in a number of states. These changes in the baseline levels of S02
emissions may result in slightly larger reductions in emissions and, consequently, slightly higher
benefits being attributed to MATS. The impact on control costs is uncertain, but likely to be
minimal given that only 1% of units are potentially affected. These modifications have not yet
been finalized, but EPA expects the overall impact on MATS to be low.
Mercury reductions were not remodeled between the proposal and final rule for either
the appropriate and necessary analysis or the RIA. As a result, the analysis presented in Chapter
4 uses the MATS proposal baseline that includes proposed, but not final, CSAPR, as well as the
mercury standards as proposed rather than as finalized. Furthermore, there were some
differences in the treatment of the baseline at proposal relative to the baseline included here in
that it included non-federally enforceable state rules. These differences do not have a
significant impact on total mercury emissions. Mercury benefits are magnitudes smaller than
the co-benefits presented here and do not impact the final rounded benefits estimates.
The year 2016 is the compliance year for MATS, though as we explain in later chapters,
we use 2015 as a proxy for compliance in 2016 for our cost analysis due to availability of
modeling impacts in that year. All estimates presented in this report represent annualized
estimates of the benefits and costs of the final MATS in 2016 rather than the net present value
of a stream of benefits and costs in these particular years of analysis.
1.6 Benefits of Emission Controls
The benefits of the final MATS are discussed in Chapters 4 and 5 of this report. Annual
monetized benefits of $37 to 90 billion (3 percent discount rate, 2007$) or $33 to 81 billion
(7 percent discount rate, 2007$) are expected for the final rule in 2016.
Since the final rule requirements were finalized after the completion of the air quality
modeling for this rule, EPA used benefit-per-ton (BPT) factors to quantify the changes in PM2.5-
related health impacts and monetized benefits based on changes in S02 and direct PM2.5
emissions. These BPT factors were based on an interim baseline and policy scenario for which
full-scale ambient air quality modeling and air quality-based human health benefits
assessments were performed. These BPT estimates were then multiplied by the amount of
emission reductions expected from MATS as finalized to estimate the benefits of the rule. The
BPT approach is methodologically consistent with the technique reported in Fann, Fulcher, &
Hubbell (2009), and has been used in previous RIAs, including the recent Ozone NAAQS RIA
1-12
-------
(U.S. EPA, 2008), the N02 NAAQS RIA (U.S. EPA, 2010), the proposed Mercury and Air Toxics
Standards RIA (U.S. EPA 2011a), and the Cross-State Air Pollution Rule (U.S. EPA, 2011b).
1.7 Cost of Emission Controls
EPA analyzed the costs of the final MATS using the Integrated Planning Model (IPM).
EPA has used this model in the past to analyze the impacts of regulations on the power sector,
including the proposed and final CSAPR and proposed MATS. EPA estimates the annual
incremental compliance costs of the rule to the power sector to be $9.6 billion in 2016
(2007$).4 A description of the methodology used to model the costs and economic impacts to
the power sector is discussed in Chapter 3 of this report. A description of how the employment
impacts associated with this final rule are estimated is provided in Chapter 6 of this report.
1.8 Organization of the Regulatory Impact Analysis
This report presents EPA's analysis of the benefits, costs, and other economic effects of
the final MATS to fulfill the requirements of a Regulatory Impact Analysis (RIA). This RIA
includes the following chapters:
• Chapter 2, Electric Power Sector Profile, describes the industry affected by the rule.
• Chapter 3, Cost, Economic, and Energy Impacts, describes the modeling conducted
to estimate the cost, economic, and energy impacts to the power sector.
• Chapter 4, Mercury and Other HAP Benefits Analysis, describes the methodology
and results of the benefits analysis for mercury and other HAP.
• Chapter 5, Co-Benefits Analysis, describes the methodology and results of the
benefits analysis for PM2.s, Ozone, and other benefit categories.
• Chapter 6, Employment and Economic Impacts, describes the analysis to estimate
the employment impacts and economic impacts associated with the final rule.
• Chapter 7, Statutory and Executive Order Impact Analyses, describes the small
business, unfunded mandates, paperwork reduction act, environmental justice, and
other analyses conducted for the rule to meet statutory and Executive Order
requirements.
4 This total includes compliance costs of $9.4 billion modeled in IPM for coal fired EGUs, monitoring,
recordkeeping, and reporting costs of $158 million, and compliance costs modeled in a separate analysis for oil-
fired EGUs of $56 million.
1-13
-------
• Chapter 8, Comparison of Benefits and Costs, shows a comparison of the total
benefits to total costs of the rule.
1.9 References
Fann, N., C.M. Fulcher, B.J. Hubbell. 2009. The influence of location, source, and emission type
in estimates of the human health benefits of reducing a ton of air pollution. Air dual
Atmos Health 2:169-176.
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). 2010. 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 .
U.S. Environmental Protection Agency (U.S. EPA). 2011a. Proposed Regulatory Impact Analysis
(RIA) for the Toxics Rule. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. March. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011b. Regulatory Impact Analysis for the
Federal Implementation Plans to Reduce Interstate Transport of Fine Particulate Matter
and Ozone in 27 States; Correction of SIP Approvals for 22 States. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. June. Available on the Internet at
.
1-14
-------
CHAPTER 2
ELECTRIC POWER SECTOR PROFILE
2.1 Introduction
This chapter discusses important aspects of the power sector that relate to the final
MATS Rule, including the types of power-sector sources affected by the rule, and provides
background on the power sector and electric generating units (EGUs). In addition, this chapter
provides some historical background on EPA regulation of and future projections for the power
sector. The specific impacts of MATS are discussed in Chapter 3.
2.2 Power Sector Overview
The production and delivery of electricity to customers consists of three distinct
segments: generation, transmission, and distribution.
2.2.1 Generation
Electricity generation is the first process in the delivery of electricity to consumers. Most
of the existing capacity for generating electricity involves creating heat to rotate turbines
which, in turn, create electricity. The power sector consists of over 17,000 generating units,
comprising fossil-fuel-fired units, nuclear units, and hydroelectric and other renewable sources
dispersed throughout the country (see Table 2-1).
Table 2-1. Existing Electricity Generating Capacity by Energy Source, 2009
Energy Source
Coal
Petroleum
Natural Gas
Other Gases
Nuclear
Hydroelectric Conventional
Wind
Solar Thermal and Photovoltaic
Wood and Wood Derived Fuels
Geothermal
Other Biomass
Pumped Storage
Other
Total
Number of Generators
1,436
3,757
5,470
98
104
4,005
620
110
353
222
1,502
151
48
17,876
Generator Nameplate
Capacity (MW)
338,723
63,254
459,803
2,218
106,618
77,910
34,683
640
7,829
3,421
5,007
20,538
1,042
1,121,686
Generator Net
Summer Capacity
(MW)
314,294
56,781
401,272
1,932
101,004
78,518
34,296
619
6,939
2,382
4,317
22,160
888
1,025,402
Source: EIA(2009).
2-1
-------
These electric generating sources provide electricity for commercial, industrial, and
residential uses, each of which consumes roughly a quarter to a third of the total electricity
produced (see Table 2-2). Some of these uses are highly variable, such as heating and air
conditioning in residential and commercial buildings, while others are relatively constant, such
as industrial processes that operate 24 hours a day.
Table 2-2. Total U.S. Electric Power Industry Retail Sales in 2009 (Billion kWh)
Sales/Direct Use (Billion kWh)
Residential
Commercial
Retail Sales
Industrial
Transportation
Direct Use
Total End Use
1,364
1,307
917
8
127
3,723
Share of Total End Use
37%
35%
25%
0.2%
3%
100%
Source: EIA(2009).
In 2009, electric generating sources produced 3,950 billion kWh to meet electricity
demand. Roughly 70 percent of this electricity was produced through the combustion of fossil
fuels, primarily coal and natural gas, with coal accounting for the largest single share (see
Table 2-3).
Table 2-3. Electricity Net Generation in 2009 (Billion kWh)
Coal
Petroleum
Natural Gas
Other Gases
Nuclear
Hydroelectric
Other
Total
Net Generation (Billion kWh)
1,756
39
921
11
799
273
151
3,950
Fuel Source Share
44.5%
1.0%
23.3%
0.3%
20.2%
6.9%
3.8%
100%
Source: EIA(2009).
Note: Retail sales are not equal to net generation because net generation includes net exported electricity and
loss of electricity that occurs through transmission and distribution.
2-2
-------
Coal-fired generating units typically supply "base-load" electricity, the portion of
electricity loads which are continually present, and typically operate throughout the day. Along
with nuclear generation, these coal units meet the part of demand that is relatively constant.
Although much of the coal fleet operates as base load, there can be notable differences across
various facilities (see Table 2-4). For example, coal-fired units less than 100 MW in size compose
37 percent of the total number of coal-fired units, but only 6 percent of total coal-fired
capacity. Gas-fired generation is better able to vary output and is the primary option used to
meet the variable portion of the electricity load and typically supplies "peak" power, when
there is increased demand for electricity (for example, when businesses operate throughout
the day or when people return home from work and run appliances and heating/air-
conditioning), versus late at night or very early in the morning, when demand for electricity is
reduced. However, the evolving economics of the power sector, in particular the increased
natural gas supply and relatively low natural gas prices, have resulted in more gas being utilized
as base load energy. Figure 2-1 shows the distribution and relative size of the fossil-fuel fired
generating capacity across the United States.
Table 2-4. Coal Steam Electricity Generating Units, by Size, Age, Capacity, and Efficiency
(Heat Rate)
Unit Size Grouping
(MW)
0 to 25
>25 to 49
50 to 99
100 to 149
150 to 249
250 and up
Total
No. Units
193
108
162
269
81
453
1,266
% of All
Units
15%
9%
13%
21%
6%
36%
Avg. Age
45
42
47
49
43
34
Avg. Net
Summer
Capacity
(MW)
15
38
75
141
224
532
Total Net
Summer
Capacity
(MW)
2,849
4,081
12,132
38,051
18,184
241,184
316,480
% Total
Capacity
1%
1%
4%
12%
6%
76%
Avg. Heat
Rate
(Btu/kWh)
11,154
11,722
11,328
10,641
10,303
10,193
Source: National Electric Energy Data System (NEEDS) v.4.10
Note: The average heat rate reported is the mean of the heat rate of the units in each size category (as opposed
to a generation-weighted or capacity-weighted average heat rate.) A lower heat rate indicates a higher
level of fuel efficiency. Table is limited to coal-steam units online in 2010 or earlier, and excludes those
units with planned retirements.
2-3
-------
Facility Capacity (MWI
• 25 to 100
« 100 to 500
• 500 to 1,000
• 1,000 to 2,000
• 2,000103,700
Figure 2-1. Fossil Fuel-Fired Electricity Generating Facilities, by Size
Source: National Electric Energy Data System (NEEDS) 4.10
Note: This map displays facilities in the NEEDS 4.10 IPM frame. NEEDS reflects available capacity on-line by the
end of 2011. This includes planned new builds and planned retirements. In areas with a dense
concentration of facilities, some facilities may be obscured.
2.2.2 Transmission
Transmission is the term used to describe the movement of electricity over a network of
high voltage lines, from electric generators to substations where power is stepped down for
local distribution. In the US and Canada, there are three separate interconnected networks of
high voltage transmission lines,1 each operating at a common frequency. Within each of these
transmission networks, there are multiple areas where the operation of power plants is
monitored and controlled to ensure that electricity generation and load are kept in balance. In
some areas, the operation of the transmission system is under the control of a single regional
operator; in others, individual utilities coordinate the operations of their generation,
transmission, and distribution systems to balance their common generation and load needs.
1These three network interconnections are the western US and Canada, corresponding approximately to the area
west of the Rocky Mountains; eastern US and Canada, not including most of Texas; and a third network
operating in most of Texas. These are commonly referred to as the Western Interconnect Region, Eastern
Interconnect Region, and ERCOT, respectively.
2-4
-------
2.2.3 Distribution
Distribution of electricity involves networks of lower voltage lines and substations that
take the higher voltage power from the transmission system and step it down to lower voltage
levels to match the needs of customers. The transmission and distribution system is the classic
example of a natural monopoly, in part because it is not practical to have more than one set of
lines running from the electricity generating sources to substations or from substations to
residences and business.
Transmission has generally been developed by the larger vertically integrated utilities
that typically operate generation and distribution networks. Distribution is handled by a large
number of utilities that often purchase and sell electricity, but do not generate it. Over the last
couple of decades, several jurisdictions in the United States began restructuring the power
industry to separate transmission and distribution from generation, ownership, and operation.
As discussed below, electricity restructuring has focused primarily on efforts to reorganize the
industry to encourage competition in the generation segment of the industry, including
ensuring open access of generation to the transmission and distribution services needed to
deliver power to consumers. In many states, such efforts have also included separating
generation assets from transmission and distribution assets to form distinct economic entities.
Transmission and distribution remain price-regulated throughout the country based on the cost
of service.
2.3 Deregulation and Restructuring
The process of restructuring and deregulation of wholesale and retail electric markets
has changed the structure of the electric power industry. In addition to reorganizing asset
management between companies, restructuring sought a functional unbundling of the
generation, transmission, distribution, and ancillary services the power sector has historically
provided, with the aim of enhancing competition in the generation segment of the industry.
Beginning in the 1970s, government policy shifted against traditional regulatory
approaches and in favor of deregulation for many important industries, including
transportation (notably commercial airlines), communications, and energy, which were all
thought to be natural monopolies (prior to 1970) that warranted governmental control of
pricing. However, deregulation efforts in the power sector were most active during the 1990s.
Some of the primary drivers for deregulation of electric power included the desire for more
efficient investment choices, the economic incentive to provide least-cost electric rates through
market competition, reduced costs of combustion turbine technology that opened the door for
2-5
-------
more companies to sell power with smaller investments, and complexity of monitoring utilities'
cost of service and establishing cost-based rates for various customer classes.
The pace of restructuring in the electric power industry slowed significantly in response
to market volatility in California and financial turmoil associated with bankruptcy filings of key
energy companies. By the end of 2001, restructuring had either been delayed or suspended in
eight states that previously enacted legislation or issued regulatory orders for its
implementation (shown as "Suspended" in Figure 2-2 below). Another 18 other states that had
seriously explored the possibility of deregulation in 2000 reported no legislative or regulatory
activity in 2001 (EIA, 2003) ("Not Active" in Figure 2-2 below). Currently, there are 15 states
where price deregulation of generation (restructuring) has occurred ("Active" in Figure 2-2
below). Power sector restructuring is more or less at a standstill; there have been no recent
proposals to the Federal Energy Regulatory Commission (FERC) for actions aimed at wider
restructuring, and no additional states have recently begun retail deregulation activity.
Electricity Restructuring by State
I—' Not Active
I—I Active
'—' Suspended
Figure 2-2. Status of State Electricity Industry Restructuring Activities
Source: EIA(2010b).
2.4 Emissions of Mercury and Other Hazardous Air Pollutants from Electric Utilities
The burning of fossil fuels, which generates about 70 percent of our electricity
nationwide, results in air emissions of Hazardous Air Pollutants (HAPs): mercury, acid gasses,
and non-mercury metallic particulates. Additionally, S02 and NOX emissions from the power
2-6
-------
sector are important precursors in the formation of fine particles and ozone (NOX only). The
power sector is a major contributor of all of these pollutants.
The Emissions Overview Memorandum Technical Support Document (TSD) to the
proposed air toxics standards (Docket number EPA-HQ-OAR-2009-0234) details the emissions
of mercury and other HAPs emitted by EGUs. In 2005, ECU emissions of mercury accounted for
approximately half of all anthropogenic mercury emissions in the U.S. Table 2-5 shows the
trend in ECU and total anthropogenic mercury emissions from 1990-2005 and ECU mercury
emissions reported in the Utility MACT Information Collection Request (ICR) in 2010.
Table 2-5. U.S. Anthropogenic Mercury Emissions, 1990-2010
EGU Hg Emissions
Non-EGU Hg Emission
Total U.S. Hg Emissions
1990
(tons)
59
205
264
1999
(tons)
49
66
115
2005
(tons)
53
52
105
2010a
(tons)
29
Not Available"
Not Available"
a The estimate of the current level of Hg emissions based on the 2010 ICR database may underestimate total EGU
Hg emissions due to targeting of the 2010 ICR on the best performing EGUs.
b Information on recent U.S. EGU emissions was obtained using an ICR for EGUs only. This same information is not
available for other sources, which were not covered by the ICR.
In 2005, EGUs contributed 82 percent of U.S. hydrogen chloride emissions. Table 2-6
shows the total HCI emissions from EGU and non-EGU sources in 2005 and the EGU HCI
emissions reported in the Utility MACT ICR in 2010.
Table 2-6. U.S. Hydrogen Chloride Emissions, 2005 and 2010
EGU HCI Emissions
Non-EGU HCI Emissions
Total U.S
. HCI Emissions
2005a
(tons)
350,000
78,000
428,000
2010b
(tons)
106,000
Not Available0
Not Available"
a 2005 emissions from the National Air Toxics Assessment Inventory. Available online at
http://www.epa.gov/ttn/atw/nata2005/. EGU emissions were extracted from the total using the MACT code
field (1808).
bThe estimate of the current level of Hg emissions based on 2010 may underestimate the total EGU emissions due
to targeting of the 2010 ICR on the best performing EGUs.
c Information on recent U.S. EGU emissions was obtained using an ICR for EGUs only. This same information is not
available for other sources, which were not covered by the ICR.
2-7
-------
Individual fossil fuel-fired units vary widely in their air emissions levels for HAPs,
particularly when uncontrolled. In 2010, as reported in the Utility MACT ICR, mercury emissions
range from less than 0.3 Ib/trillion Btu (TBtu) to more than 20 Ibs/TBtu. HCI emissions from
coal-fired units range from less than 0.00002 Ib/million Btu (mmBtu) (for a unit with a scrubber)
to over 0.1 Ib/mmBtu. Additionally, emissions of fine particulates less than or equal to 2.5
microns (PM2.5) range from 0.002 Ib/mmBtu to over 0.06 Ib/mmBtu. For an uncontrolled plant,
mercury, acid gas, and particulate emissions are directly related to the elemental profile and
ash content of the coal burned.
Oil-fired units also have a wide range of HAP emissions. Based on the Utility MACT ICR,
Mercury emissions range from less than 0.01 Ib/TBtu to more than 60 Ibs/TBtu. HCI emissions
from oil-fired units range from less than 0.00001 Ib/mmBtu (for a unit with a scrubber) to over
0.003 Ib/mmBtu. Emissions of PM2.5 range from less than 0.004 Ib/mmBtu to over 0.07
Ib/mmBtu.
2.5 Pollution Control Technologies
Acid gas HAPs (e.g., hydrogen chloride (HCI), hydrogen fluoride (HF), sulfur dioxide
(S02)) from coal-fired power plants can be controlled by fuel selection, fuel blending, or post
combustion controls. Fossil fuels, particularly coal, vary widely in the content of pollutants like
chlorine (Cl), fluorine (F), sulfur (S) and other HAPs, making fuel blending and/or switching an
effective method for reducing emissions of HAPs. In general, it is easier to switch fuels within a
coal rank (rather than across a coal rank) due to similar heat contents and other characteristics.
Switching fuels across ranks tends to trigger more costly modifications. As a compromise,
blending is employed when a complete fuel switch adversely affects the unit. EGUs may also
choose to retrofit post combustion controls to achieve superior pollutant removal. Post-
combustion controls typically remove larger proportions of HCI and HF than S02 due to
differences in molecular weight.
Acid gas emissions (including S02) can be reduced with flue gas desulfurization (FGD,
also known as "scrubbers") or with dry sorbent injection (DSI). EGUs may choose either "wet"
or "dry" configurations of scrubbers. Wet scrubbers can use a variety of reagents including
crushed limestone, quick lime, and magnesium-enhanced lime. The choice of reagent affects
performance, size, capital and operating costs. Current wet scrubber technology is capable of
removing at least 99 percent of HF and HCI emissions while simultaneously achieving
96 percent S02 removal. Modern dry FGD technology combines lime-based slurry with a
downstream fabric filter to remove at least 93 percent S02 while also capturing over 99 percent
2-8
-------
HCL and HF. An alternative to scrubber technology is dry sorbent injection (DSI), which injects
an alkaline powdered material (post combustion) to react with the acid gases. The product of
this reaction is removed by particulate matter (PM) control device. DSI technology is most
efficient with a baghouse present downstream but can function with an electrostatic
precipitator (ESP) downstream as well. Under these circumstances, the ESP requires more
reagent per molecule of acid gas removed as compared to a similar operation with a baghouse.
Finally, DSI may employ a multitude of sorbents (trona,2 sodium carbonate, calcium
carbonate—and their bicarbonate counterparts) for a more tailored approach to reduce
emissions based on the source, cost, and unit and fuel characteristics.
Mercury capture and removal requires multiple controls. Upon combustion, mercury
exits the furnace in three forms: elemental, oxidized, and as a particulate. Elemental mercury is
emitted out of the stack. The particulate form is bound to the ash and removed by PM control
equipment such as ESP or fabric filter. A portion of mercury that has converted to oxidized
compounds may be removed by either a wet scrubber or by activated carbon injection (ACI).
Each of these control devices uses a different method to remove the mercury compounds. The
wet FGD system captures oxidized mercury because it is water soluble, while activated carbon
injection provides a unique physical surface to which oxidized mercury can adhere. Mercury
oxidation can occur at multiple locations within a unit as long as an oxidizing agent, generally a
halogen, is present for reaction. This allows the unit operator some latitude in selecting a
control method and injection point based on existing equipment at the particular source. A
halogen can be introduced to the fuel prior to combustion, injected directly into the furnace,
introduced upstream of a selective catalytic reduction (SCR) system,3 or infused with the
activated carbon injections. The unit operator may also increase halogens by blending in higher
chlorine fuels (e.g., Powder River Basin fuel blended with bituminous coal). Operating a wet
FGD for S02 control alongside selective catalytic reduction (SCR) for NOX control with sufficient
halogen present will remove more than 90 percent of the mercury within the flue gas stream.
Alternatively, in the absence of a wet FGD, activated carbon injection (ACI) can be employed for
mercury capture with at least 90 percent removal using a downstream fabric filter. An ESP
results in less efficient mercury removal with ACI.
Non-mercury heavy metals and organics are removed by PM control equipment such as
fabric filters and ESP. Unlike mercury, the heavy metals (e.g., selenium and arsenic) are non-
volatile and affix to the ash. Likewise, any organics surviving the high temperature combustion
2 Trona refers to the chemical compound sodium sesquicarbonate.
3 SCR is primarily used for NOX control, but can also be used to promote mercury oxidation.
2-9
-------
process are non-volatile and bind to the ash. Both control technologies are capable of removing
more than 99 percent of PM2.5 mass from the emissions stream. ESPs sap relatively little energy
from the flue gas but are less flexible for fuel switching, since they are designed for use with a
specific intended fuel. Fuel switching or blending that increases gas flow rate, ash resistivity, or
particle loading may render an existing ESP insufficient for removing particulate matter. ESPs
also suffer from ash re-entrainment, which is the release of particulate matter from the last
compartment due to the self cleaning action. On the other hand, an ESP with sufficient design
margin may succeed with these fuel alterations. Conversely, a fabric filter does not suffer from
these limitations. Moreover, the fabric filter readily lends itself to mercury and acid gas removal
since DSI and ACI operate more efficiently with a baghouse. When considering retrofit PM
control options, a unit with an existing ESP will examine upgrading the precipitator as an
alternative to installing a new fabric filter to achieve emission reductions.
For more detail on the cost and performance assumptions of pollution controls, see the
documentation for the Integrated Planning Model (IPM),4 a dynamic linear programming model
that EPA uses to examine air pollution control policies for various air emissions throughout the
United States for the entire power system.
2.6 HAP Regulation in the Power Sector
2.6.1 Programs Targeting HAP
In 2000, EPA made a finding that it was appropriate and necessary to regulate coal- and
oil-fired EGUs under CAA section 112 and listed EGUs pursuant to CAA section 112(c). This
finding triggered a requirement for EPA to propose regulations to control air toxics emissions,
including mercury, from these facilities.
On January 30, 2004, EPA proposed a rule with two basic approaches for controlling
mercury from power plants. One approach would require power plants to meet emissions
standards reflecting the application of the "maximum achievable control technology" (MACT)
determined according to the procedure set forth in section 112(d) of the Clean Air Act. A
second approach proposed by EPA would create a market-based "cap and trade" program that,
if implemented, would reduce nationwide utility emissions of mercury in two phases under
Section 111 or Section 112 of the Clean Air Act. EPA also proposed to revise its December 2000
finding that it is "appropriate and necessary" to regulate utility hazardous air emissions using
the MACT standards provisions (section 112) of the Clean Air Act.
Documentation for IPM can be found at www.epa.gov/airmarkets/epa-ipm.
2-10
-------
On March 15, 2005, EPA issued the final Clean Air Mercury Rule (CAMR). CAMR
established "standards of performance" limiting mercury emissions from new and existing
utilities and created a market-based cap-and-trade program to reduce nationwide utility
emissions of mercury in two phases. In conjunction with CAMR, EPA published a final rule
(Section 112(n) Revision Rule) that removed EGUs from the list of sources for which regulation
under CAA section 112 was required.
The Section 112(n) Revision Rule was vacated on February 8, 2008, by the U.S. Court of
Appeals for the District of Columbia Circuit. As a result of that vacatur, CAMR was also vacated
and EGUs remained on the list of sources that must be regulated under CAA section 112. This
action finalizes the rule EPA proposed on March 16, 2011 to replace CAMR in response to the
court's decisions.
2.6.2 Programs Targeting SO2 and NOx
Programs to reduce S02 and NOx also impact emissions of mercury and other HAP. At
the federal level, efforts to reduce emissions of S02 have been occurring since 1970. Policy
makers have recognized the need to address these harmful emissions, and incremental steps
have been taken to ensure that the country meets air quality standards. The recently finalized
Cross State Air Pollution Rule (CSAPR) is the next step toward attainment of the national
standards for PM2.5 and ozone.
Even before widespread regulation of S02 and NOX for the power sector, total
suspended particulate matter (TSP) was a related target of state and federal action. Because
larger particulates are visible as dark smoke from smokestacks, most states had regulations by
1970 limiting the opacity of emissions. Requirements for taller smokestacks also mitigated local
impacts of TSP. Notably, such regulations effectively addressed large-diameter, filterable
particulate matter rather than condensable particulate matter (such as PM2.5) associated with
S02 and NOX emissions, which are not visible at the smokestack and have impacts far from their
sources.
Federal regulation of S02 and NOX emissions at power plants began with the 1970 Clean
Air Act. The Act required the Agency to develop New Source Performance Standards (NSPS) for
a number of source categories including coal-fired power plants. The first NSPS for power
plants (subpart D) required new units to limit S02 emissions either by using scrubbers or by
using low sulfur coal. NOX was required to be limited through the use of low NOX burners. A
new NSPS (subpart Da), promulgated in 1978, tightened the standards for S02, requiring
scrubbers on all new units.
2-11
-------
The 1990 Clean Air Act Amendments (CAAA) placed a number of new requirements on
power plants. The Acid Rain Program, established under Title IV of the 1990 CAAA, requires
major reductions of S02 and NOX emissions. The S02 program sets a permanent cap on the
total amount of S02 that can be emitted by electric power plants in the contiguous United
States at about one-half of the amount of S02 these sources emitted in 1980. Using a
market-based cap and trade mechanism allows flexibility for individual combustion units to
select their own methods of compliance with the S02 reduction requirements. The program
uses a more traditional approach to NOX emissions limitations for certain coal-fired electric
utility boilers, with the objective of achieving a 2 million ton reduction from projected NOX
emission levels that would have been emitted in 2000 without implementation of Title IV.
The Acid Rain Program comprises two phases for S02 and NOX. Phase I applied primarily
to the largest coal-fired electric generating sources from 1995 through 1999 for S02 and from
1996 through 1999 for NOX. Phase II for both pollutants began in 2000. For S02, it applies to
thousands of combustion units generating electricity nationwide; for NOX it generally applies to
affected units that burned coal during 1990 through 1995. The Acid Rain Program has led to
the installation of a number of scrubbers on existing coal-fired units as well as significant fuel
switching to lower sulfur coals. Under the NOX provisions of Title IV, most existing coal-fired
units installed low NOX burners.
The CAAA also placed much greater emphasis on control of NOX to reduce ozone
nonattainment. This led to the formation of several regional NOX trading programs as well as
intrastate NOX trading programs in states such as Texas. The northeastern states of the Ozone
Transport Commission (OTC) required existing sources to meet Reasonably Available Control
Technology (RACT) limits on NOX in 1995 and in 1999 began an ozone-season cap and trade
program to achieve deeper reductions. In 1998, EPA promulgated regulations (the NOX SIP Call)
that required 21 states in the eastern United States and the District of Columbia to reduce NOX
emissions that contributed to nonattainment in downwind states using the cap and trade
approach. This program began in May of 2003 and has resulted in the installation of significant
amounts of selective catalytic reduction.
The Clean Air Interstate Rule (CAIR) built on EPA's efforts in the NOX SIP call to address
specifically interstate pollution transport for ozone, and was EPA's first attempt to address
interstate pollution transport for PM2.5. It required significant reductions in emissions of S02
and NOX in 28 states and the District of Columbia (see Figure 6-4 below). EGUs were found to
be a major source of the S02 and NOX emissions which contributed to fine particle
concentrations and ozone problems downwind. Although the D.C. Circuit remanded the rule to
2-12
-------
EPA in 2008, it did so without vacatur, allowing the rule to remain in effect while EPA addressed
the remand. Thus, CAIR continued to help states address ozone and PM2.5 nonattainment and
improve visibility by reducing transported precursors of S02 and NOX through the
implementation of three separate cap and trade compliance programs for annual NOX, ozone
season NOX, and annual S02 emissions from power plants.
Perhaps in anticipation of complying with CAIR, especially the more stringent second
phase that was set to begin in 2015, several sources began installing or planning to install
advanced controls for S02 and NOX to begin operating in the 2010 to 2015 timeframe. Many
EPA New Source Review (NSR) settlements also required controls in those years, as do state
rules in Georgia, Illinois, and Maryland. States like North Carolina, New York, Connecticut,
Massachusetts, and Delaware have also moved to control these emissions to address
nonattainment.
On July 6, 2011, the EPA finalized the Cross-State Air Pollution Rule (CSAPR) to replace
CAIR. The rule requires states to eliminate the portion of their emissions defined as their
"significant contribution" by setting a pollution limit (or budget) for each covered state. The
rule allows air-quality-assured allowance trading among covered sources, utilizing an allowance
market infrastructure based on existing, successful allowance trading programs. The final
CSAPR allows sources to trade emissions allowances with other sources within the same
program (e.g., ozone season NOx) in the same or different states, while firmly constraining any
emissions shifting that may occur by requiring a strict emission ceiling in each state (the budget
plus variability limit). It also includes assurance provisions that ensure each state will make the
emission reductions necessary to fulfill the "good neighbor" provision of the Clean Air Act.
2.7 Revenues, Expenses, and Prices
Due to lower retail electricity sales, total utility operating revenues declined in 2009 to
$276 billion from a peak of almost $300 billion in 2008. However, operating expenses were
appreciably lower and as a result, net income actually rose modestly compared to 2008 (see
Table 2-7). Recent economic events have put downward pressure on electricity demand, thus
dampening electricity prices and consumption (utility revenues), but have also reduced the
price and cost of fossil fuels and other expenses. Electricity sales and revenues associated with
the generation, transmission, and distribution of electricity are expected to rebound and
increase modestly by 2015, where they are projected to be roughly $360 billion (see Table 2-8).
Based on ElA's Annual Energy Outlook 2011, Table 2-8 shows that in the base case, the
power sector is expected to derive revenues of $360 billion in 2015. Table 2-7 shows that
2-13
-------
investor-owned utilities (lOUs) earned income of about 11.5% compared to total revenues in
2009. Assuming the same income ratio from lOUs (with no income kept by public power), and
using the same proportion of power sales from public power as observed in 2009, EPA projects
that the power sector will expend over $320 billion in 2015 alone to generate, transmit, and
distribute electricity to end-use consumers.
Over the past 50 years, real retail electricity prices have ranged from around 7 cents per
kWh in the early 1970s, to around 11 cents, reached in the early 1980s. Generally, retail
electricity prices do not change rapidly and do not display the variability of other energy or
commodity prices, although the frequency at which these prices change varies across different
types of customers. Retail rate regulation has largely insulated consumers from the rising and
falling wholesale electricity price signals whose variation in the marketplace on an hourly, daily,
and seasonal basis is critical for driving lowest-cost matching of supply and demand. In fact, the
real price of electricity today is lower than it was in the early 1960s and 1980s (see Figure 2-3).
2-14
-------
Table 2-7. Revenue and Expense Statistics for Major U.S. Investor-Owned Electric Utilities
for 2009 ($millions)
Utility Operating Revenues
Electric Utility
Other Utility
Utility Operating Expenses
Electric Utility
Operation
Production
Cost of Fuel
Purchased Power
Other
Transmission
Distribution
Customer Accounts
Customer Service
Sales
Administrative and General
Maintenance
Depreciation
Taxes and Other
Other Utility
Net Utility Operating Income
2008
298,962
266,124
32,838
267,263
236,572
175,887
140,974
47,337
84,724
8,937
6,950
3,997
5,286
3,567
225
14,718
14,192
19,049
26,202
30,692
31,699
2009
276,124
249,303
26,822
244,243
219,544
154,925
118,816
40,242
67,630
10,970
6,742
3,947
5,203
3,857
178
15,991
14,092
20,095
29,081
24,698
31,881
Source: EIA(2009).
Note: These data do not include information for public utilities.
Table 2-8. Projected Revenues by Service Category in 2015 for Public Power and Investor-
Owned Utilities (billions)
Generation
Transmission
Distribution
$195
36
129
Total
$360
Source: EIA(2011).
Note: Data are derived by taking either total electricity use (for generation) or sales (transmission and
distribution) and multiplying by forecasted prices by service category from Table 8 (Electricity Supply,
Disposition, Prices, and Emissions).
2-15
-------
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Figure 2-3. National Average Retail Electricity Price (1960-2009)
Source: EIA(2009).
On a state-by-state basis, retail electricity prices vary considerably. The Northeast and
California have average retail prices that can be as much as double those of other states (see
Figure 2-4).
Average Retail Price of Electricity by State, 2009
NV /
10.36 rjr J
CA \ , 6<< CO
1324 \ / Ml
Average Price (cents per kitowatthour)
^1 6.0910 7 35
~ -.37 MS 3;
8 « tn << 38
19.40101308
_j 1309 to !1 21
Note: Data are displayed as S groups of 10 States and the OistcKt of Columbia.
U.S. tola) average price pel fcibwaltltoui is 3.83 GenlE.
Source: U S Energy Information Administration. Form eiA-551. "Annual Eredac
Power IrKisutry Repon.'
Figure 2-4. Average Retail Electricity Price by State (cents/kWh), 2009
Source: EIA(2009).
2-16
-------
2.7.1 Natural Gas Market
The natural gas market in the United States has historically experienced significant price
volatility from year to year, between seasons within a year, and can undergo major price swings
during short-lived weather events (such as cold snaps leading to short-run spikes in heating
demand). Over the last decade, gas prices (both Henry Hub5 prices and delivered prices to the
power sector) have ranged from $3 per mmBtu to as high as $9 on an annual average basis (see
Figure 2-5). During that time, the daily price of natural gas reached as high as $15/mmBtu.
Recent forecasts of natural gas have also experienced considerable revision as new sources of
gas have been discovered and have come to market, although there continues to be some
uncertainty surrounding the precise quantity of the resource base.6
10
EIA Historical Natural Gas Spot Price (Henry Hub)
— EIA Projected (AEO 2011) Natural Gas Spot Price (Henry Hub)
EIA Projected (AEO 2010) Natural Gas Spot Price (Henry Hub)
1990
1995
2000
2005
2010
2015
2020
2025
2030
2035
Figure 2-5. Natural Gas Spot Price, Annual Average (Henry Hub)
Source: EIA (2010a), EIA (2011).
' The Henry Hub is the pricing point for natural gas futures contracts traded on the New York Mercantile Exchange.
It is a point on the natural gas pipeline system that interconnects nine interstate and four intrastate pipelines.
' In August, EIA announced it would lower its previous estimates of recoverable shale gas by nearly 80 percent.
EPA's modeling of the natural gas market is discussed in more detail in Chapter 7 of this RIA.
2-17
-------
2.8 Electricity Demand and Demand Response
Electricity performs a vital and high-value function in the economy. Historically, growth
in electricity consumption has been closely aligned with economic growth. Overall, the U.S.
economy has become more efficient over time, producing more output (GDP) per unit of
energy input, with per capita energy use fairly constant over the past 30 years. The growth rate
of electricity demanded has also been in overall decline for the past sixty years (see Figure 2-8),
with several key drivers that are worth noting. First, there has been a significant structural shift
in the U.S. economy towards less energy-intensive sectors, like services. Second, companies
have strong financial incentives to reduce energy expenditures. Third, companies are
responding to the marketplace and continually develop and bring to market new technologies
that reduce energy consumption. Fourth, other policies, such as energy efficiency standards at
the state and Federal level, have helped address certain market failures. These broader changes
have altered the outlook for future electricity growth (see Figure 2-6).
Figure 2-6. Electricity Growth Rate (3 Year Rolling Average) and Projections from the
Annual Energy Outlook 2011
Source: EIA (2009), EIA (2011).
2-18
-------
Energy efficiency initiatives have become more common, and investments in energy
efficiency are projected to continue to increase for the next 5 to 10 years, driven in part by the
growing number of states that have adopted energy efficiency resource standards.7 These
investments, and other energy efficiency policies at both the state and federal level, create
incentives to reduce energy consumption and peak load. According to EIA, demand-side
management provided actual peak load reductions of 31.7 GW in 2009. For context, the current
coal fleet is roughly 320 GW of capacity.
Demand for electricity, especially in the short run, is not very sensitive to changes in
prices and is considered relatively price inelastic, although some demand reduction does occur
in response to price. With that in mind, EPA modeling does not typically incorporate a "demand
response" in its electric generation modeling (Chapter 3) to the increases in electricity prices
typically projected for EPA rulemakings. Electricity demand is considered to be constant in EPA
modeling applications and the reduction in production costs that would result from lower
demand is not considered in the primary analytical scenario that is modeled. This leads to some
overstatement in the private compliance costs that EPA estimates. Notably, the "compliance
costs" are the changes in the electric power generation costs in the base case and pollution
control options that are evaluated in Chapter 3. In simple terms, it is the resource costs of what
the power industry will directly expend to comply with EPA's requirements.
2.9 References
U.S. Energy Information Administration (U.S. EIA). Electric Power Annual 2003. 2003. Available
online at: http://www.eia.gov/oiaf/archive/aeo03/index.html.
U.S. Energy Information Administration (U.S. EIA). Electric Power Annual 2009. 2009. Available
online at: http://www.eia.doe.gov/cneaf/electricity/epa/epa_sum.html.
U.S. Energy Information Administration (U.S. EIA). Annual Energy Outlook 2010. 2010a.
Available online at: http://www.eia.gov/oiaf/archive/aeolO/index.html.
U.S. Energy Information Administration (U.S. EIA). "Status of Electricity Restructuring by State."
2010b. Available online at:
http://www.eia.gov/cneaf/electricity/page/restructuring/restructu re_elect.html.
U.S. Energy Information Administration (U.S. EIA). Annual Energy Outlook 2011. 2011. Available
online at: http://www.eia.gov/forecasts/aeo/.
7 To the extent that EIA includes these measures in its baseline forecast from the Annual Energy Outlook, EPA has
also incorporated them into the baseline for purposes of assessing the economic impacts of this rule. See AEO
2011 and Chapter 3 and the IPM documentation for more detail.
2-19
-------
CHAPTER 3
COST, ECONOMIC, AND ENERGY IMPACTS
This chapter reports the compliance cost, economic, and energy impact analysis
performed for the Mercury and Air Toxics Standards (MATS). EPA used the Integrated Planning
Model (IPM), developed by ICF Consulting, to conduct its analysis. IPM is a dynamic linear
programming model that can be used to examine air pollution control policies for S02, NOX, Hg,
HCI, and other air pollutants throughout the United States for the entire power system.
Documentation for IPM can be found at http://www.epa.gov/airmarkets/progsregs/epa-ipm,
and updates specific to the MATS modeling are in the "Documentation Supplement for EPA
Base Case v.4.10_MATS - Updates for Final Mercury and Air Toxics Standards (MATS) Rule"
(hereafter IPM 4.10 Supplemental Documentation for MATS).
3.1 Background
Over the last decade, EPA has on several occasions used IPM to consider pollution
control options for reducing power-sector emissions.1 Most recently EPA used IPM extensively
in the development and analysis of the impacts of the Cross-State Air Pollution Rule (CSAPR).2
As discussed in Chapter 2, MATS coincides with a period when many new pollution controls are
being installed. Many are needed for compliance with NSR settlements and state rules, while
others may have been planned in expectation of CAIR and its replacement, the CSAPR.
The emissions scenarios for the RIA reflects the Cross-State Air Pollution Rule (CSAPR) as
finalized in July 2011 and the emissions reductions of SOX, NOX, directly emitted PM, and C02
are consistent with application of federal rules, state rules and statutes, and other binding,
enforceable commitments in place by December 2010 for the analysis timeframe.3
1 Many EPA analyses with IPM have focused on legislative proposals with national scope, such as EPA's IPM
analyses of the Clean Air Planning Act (S.843 in 108th Congress), the Clean Power Act (S. 150 in 109th Congress),
the Clear Skies Act of 2005 (S.131 in 109th Congress), the Clear Skies Act of 2003 (S.485 in 108th Congress), and
the Clear Skies Manager's Mark (of S.131). These analyses are available at EPA's website:
(http://www.epa.gov/airmarkt/progsregs/epa-ipm/index.html). EPA also analyzed several multi-pollutant
reduction scenarios in July 2009 at the request of Senator Tom Carper to illustrate the costs and benefits of
multiple levels of SO2 and NOX control in the power sector.
Additionally, IPM has been used to develop the NOX Budget Trading Program, the Clean Air Interstate Rule
programs, the Clean Air Visibility Programs, and other EPA regulatory programs for the last 15 years.
3 Consistent with the mercury risk deposition modeling for MATS, EPA did not model non-federally enforceable
mercury-specific emissions reduction rules in the base case or MATS policy case (see preamble section III.A).
Note that this approach does not significantly affect SO2 and NOX projections underlying the cost and benefit
results presented in this RIA
3-1
-------
EPA has made these base case assumptions recognizing that the power sector will install
a significant amount of pollution controls in response to several requirements. The inclusion of
CSAPR and other regulatory actions (including federal, state, and local actions) in the base case
is necessary in order to reflect the level of controls that are likely to be in place in response to
other requirements apart from MATS. This base case will provide meaningful projections of
how the power sector will respond to the cumulative regulatory requirements for air emissions
in totality, while isolating the incremental impacts of MATS relative to a base case with other air
emission reduction requirements separate from today's action.
The model's base case features an updated Title IV S02 allowance bank assumption and
incorporates updates related to the Energy Independence and Security Act of 2007. Some
modeling assumptions, most notably the projected demand for electricity, are based on the
2010 Annual Energy Outlook from the Energy Information Administration (EIA). In addition, the
model includes existing policies affecting emissions from the power sector: the Title IV of the
Clean Air Act (the Acid Rain Program); the NOX SIP Call; various New Source Review (NSR)
settlements4; and several state rules5 affecting emissions of S02, NOX, and C02 that were
finalized through June of 2011. IPM includes state rules that have been finalized and/or
approved by a state's legislature or environmental agency, with the exception of non-federal
mercury-specific rules. The IPM 4.10 Supplemental Documentation for MATS contains details
on all of these other legally binding and enforceable commitments for installation and
operation of pollution controls. This chapter focuses on results of EPA's analysis with IPM for
the model's 2015 run-year in connection with the compliance date for MATS.
MATS establishes National Emissions Standards for Hazardous Air Pollutants (NESHAPS)
for the "electric utility steam generating unit" source category, which includes those units that
combust coal or oil for the purpose of generating electricity for sale and distribution through
the national electric grid to the public.
4The NSR settlements include agreements between EPA and Southern Indiana Gas and Electric Company (Vectren),
Public Service Enterprise Group, Tampa Electric Company, We Energies (WEPCO), Virginia Electric & Power
Company (Dominion), Santee Cooper, Minnkota Power Coop, American Electric Power (AEP), East Kentucky
Power Cooperative (EKPC), Nevada Power Company, Illinois Power, Mirant, Ohio Edison, Kentucky Utilities,
Hoosier Energy, Salt River Project, Westar, Puerto Rico Power Authority, Duke Energy, American Municipal Power,
and Dayton Power and Light. These agreements lay out specific NOX, SO2, and other emissions controls for the
fleets of these major Eastern companies by specified dates. Many of the pollution controls are required between
2010 and 2015.
5These include current and future state programs in Alabama, Arizona, California, Colorado, Connecticut,
Delaware, Georgia, Illinois, Kansas, Louisiana, Maine, Maryland, Massachusetts, Michigan, Minnesota, Missouri,
Montana, New Hampshire, New Jersey, New York, North Carolina, Oregon, Pennsylvania, Tennessee, Texas, Utah,
Washington, West Virginia, and Wisconsin the cover certain emissions from the power sector.
3-2
-------
Coal-fired electric utility steam generating units include electric utility steam generating
units that burn coal, coal refuse, or a synthetic gas derived from coal either exclusively, in any
combination together, or in any combination with other supplemental fuels. Examples of
supplemental fuels include petroleum coke and tire-derived fuels. The NESHAP establishes
standards for HAP emissions from both coal- and oil-fired EGUs and will apply to any existing,
new, or reconstructed units located at major or area sources of HAP. Although all HAP are
pollutants of interest, those of particular concern are hydrogen fluoride (HF), hydrogen chloride
(HCI), dioxins/furans, and HAP metals, including antimony, arsenic, beryllium, cadmium,
chromium, cobalt, mercury, manganese, nickel, lead, and selenium.
This rule affects any fossil fuel fired combustion unit of more than 25 megawatts electric
(MWe) that serves a generator that produces electricity for sale. A unit that cogenerates steam
and electricity and supplies more than one-third of its potential electric output capacity and
more than 25 MWe output to any utility power distribution system for sale is also considered
an electric utility steam generating unit. The rule affects roughly 1,400 EGUs: approximately
1,100 existing coal-fired generating units and 300 oil-fired steam units, should those units
combust oil. Of the 600 power plants potentially covered by this rule, about 430 have coal-fired
units only, 30 have both coal- and oil- or gas-fired steam units, and 130 have oil- or gas-fired
steam units only. Note that only steam electric units combusting coal or oil are covered by this
rule.
EPA analyzed for the RIAthe input-based (Ibs/MMBtu) MATS control requirements
shown in Table 3-1. In this analysis, EPA does not model an alternative S02 standard. Coal
steam units with access to lignite in the modeling are subjected to the "Existing coal-fired unit
low Btu virgin coal" standard. For further discussion about the scope and requirements of
MATS, see the preamble or Chapter 1 of this RIA.
3-3
-------
Table 3-1. Emissions Limitations for Coal-Fired and Solid Oil-Derived Fuel-Fired Electric
Utility Steam Generating Units
Subcategory
Filterable Hydrogen Chloride
Participate Matter
Note: Ib/MMBtu = pounds pollutant per million British thermal units fuel input
Ib/TBtu = pounds pollutant per trillion British thermal units fuel input
Ib/MWh = pounds pollutant per megawatt-hour electric output (gross)
Ib/GWh = pounds pollutant per gigawatt-hour electric output (gross)
a Beyond-the-floor limit as discussed elsewhere
b Duct burners on syngas; based on permit levels in comments received
c Duct burners on natural gas; based on permit levels in comments received
d Based on best-performing similar source
e Based on permit levels in comments received
Mercury
Existing coal-fired unit not low Btu
virgin coal
Existing coal-fired unit low Btu
virgin coal
Existing- IGCC
Existing -Solid oil-derived
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.040 Ib/MMBtu
(0.40 Ib/MWh)
0.0080 Ib/MMBtu
(0.090 Ib/MWh)
0.0020 Ib/MMBtu
(0.020 Ib/MWh)
0.0020 Ib/MMBtu
(0.020 Ib/MWh)
0.00050 Ib/MMBtu
(0.0050 Ib/MWh)
0.0050 Ib/MMBtu
(0.080 Ib/MWh)
1.2 Ib/TBtu
(0.020 Ib/GWh)
11.0 Ib/TBtu
(0.20 Ib/GWh)
4.0 lb/TBtua
(0.040 lb/GWha)
2.5 Ib/TBtu
(0.030 Ib/GWh)
0.20 Ib/TBtu
(0.0020 Ib/GWh)
virgin coal
New coal-fired unit low Btu virgin
coal
New -IGCC
New -Solid oil-derived
0.0070 Ib/MWh
0.0070 Ib/MWh
0.070 lb/MWhb
0.090 lb/MWhc
0.020 Ib/MWh
0.40 Ib/GWh
0.40 Ib/GWh
0.0020 lb/MWhd
0.00040 Ib/MWh
0.00020 Ib/GWh
0.040 Ib/GWh
0.0030 lb/GWhe
0.0020 Ib/GWh
3-4
-------
Table 3-2. Emissions Limitations for Liquid Oil-Fired Electric Utility Steam Generating Units
Subcategory
Existing- Liquid oil-
continental
Existing- Liquid oil-
non-continental
New- Liquid oil -
continental
New- Liquid oil -
non-continental
Filterable PM
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.030 Ib/MMBtu
(0.30 Ib/MWh)
0.070 Ib/MWh
0.20 Ib/MWh
Hydrogen Chloride
0.0020 Ib/MMBtu
(0.010 Ib/MWh)
0.00020 Ib/MMBtu
(0.0020 Ib/MWh)
0.00040 Ib/MWh
0.0020 Ib/MWh
Hydrogen Fluoride
0.00040 Ib/MMBtu
(0.0040 Ib/MWh)
0.000060 Ib/MMBtu
(0.00050 Ib/MWh)
0.00040 Ib/MWh
0.00050 Ib/MWh
EPA used the Integrated Planning Model (IPM) v.4.10 to assess the impacts of the MATS
emission limitations for coal-fired electricity generating units (ECU) in the contiguous United
States. IPM modeling did not subject oil-fired units to policy criteria.6 Furthermore, IPM
modeling did not include generation outside the contiguous U.S., where EPA is aware of only 2
facilities that would be subject to the coal-fired requirements of the final rule. Given the limited
number of potentially impacted facilities, limited availability of input data to inform the
modeling, and limited connection to the continental grid, EPA did not model the impacts of the
rule beyond the contiguous U.S.
Mercury emissions are modeled as a function of mercury content of the fuel type(s)
consumed at each plant in concert with that plant's pollutant control configuration. HCI
emissions are projected in a similar fashion using the chlorine content of the fuel(s). For both
mercury and HCI, EGUs in the model must emit at or below the final mercury and HCI emission
rate standards in order to operate from 2015 onwards. EGUs may change fuels and/or install
additional control technology to meet the standard, or they may choose to retire if it is more
economic for the power sector to meet electricity demand with other sources of generation.
See IPM 4.10 documentation and IPM 4.10 Supplemental Documentation for MATS for more
details.
Total PM emissions are calculated exogenously to IPM, using EPA's Source Classification
Code (SCC) and control-based emissions factors. SCC is a classification system that describes a
generating unit's characteristics.
6 EPA did not model the impacts of MATS on oil-fired units using IPM. Rather, EPA performed an analysis of
impacts on oil-fired units for the final rule. The results are summarized in Appendix 3A.
3-5
-------
Instead of emission limitations for the organic HAP, EPA is proposing that if requested,
owners or operators of EGUs submit to the delegated authority or EPA, as appropriate,
documentation showing that an annual performance test meeting the requirements of the rule
was conducted. IPM modeling of the MATS policy assumes compliance with these work practice
standards.
Electricity demand is anticipated to grow by roughly 1 percent per year, and total
electricity demand is projected to be 4,103 billion kWh by 2015. Table 3-3 shows current
electricity generation alongside EPA's base case projection for 2015 generation using IPM. EPA's
IPM modeling for this rule relies on EI ft^'s Annual Energy Outlook for 2010's electric demand
forecast for the US and employs a set of EPA assumptions regarding fuel supplies and the
performance and cost of electric generation technologies as well as pollution controls.7 The
base case includes CSAPR as well as other existing state and federal programs for air emissions
control from electric generating units, with the exception of state mercury rules.
7 Note that projected electricity demand in AEO 2010 is about 2% higher than the AEO 2011 projection in 2015.
Since this RIA assumes higher electricity demand in 2015 than is shown in the latest AEO projection, it is possible
that the model may be taking compliance actions to meet incremental electricity demand that may not actually
occur, and projected compliance costs may therefore be somewhat overstated in this analysis.
3-6
-------
Table 3-3. 2009 U.S. Electricity Net Generation and EPA Base Case Projections for 2015-
2030 (Billion kWh)
Coal
Oil
Natural Gas
Nuclear
Hydroelectric
Non-hydro Renewables
Other
Total
Historical
2009
1,741
36
841
799
267
116
10
3,810
2015
1,982
0.11
710
828
286
252
45
4,103
Base Case
2020
2,002
0.13
847
837
286
289
45
4,307
2030
2,027
0.21
1,185
817
286
333
55
4,702
Source: 2009 data from AEO Annual Energy Review, Table 8.2c Electricity Net Generation: Electric Power Sector
by Plant Type, 1989-2010; Projections from Integrated Planning Model run by EPA, 2011.
3-7
-------
•f
• .
'?'*
•.v
Facility Capacity
(megawatts)
• 2510100
« "OOtoSOO
» 500101,000
• 1.000 to 2.00C
• 2,1X0 to 3,400
Fadlity hasoilunit*
.
v -,
Figure 3-1. Geographic Distribution of Affected Units, by Facility, Size and Fuel Source in
2012
Source/Notes: National Electric Energy Data System (NEEDS 4.10 MATS) (EPA, December 2011) and EPA's
Information Collection Request (ICR) for New and Existing Coal- And Oil-Fired Electric Utility Stream Generation
Units (2010). This map displays facilities that are included in the NEEDS 4.10 MATS data base and that contain at
least one oil-fired steam generating unit or one coal-fired steam generating unit that generates more than 25
megawatts of power. This includes coal-fired units that burn petroleum coke and that turn coal into gas before
burning (using integrated gasification combined cycle or IGCC). NEEDS reflects available capacity on-line by the
end of 2011; this includes committed new builds and committed retirements of old units. Only coal and oil-fired
units are covered by this rule. Some of the oil units displayed on the map are capable of burning oil and/or gas. If
a unit burns only gas, it will not be covered in the rule. In areas with a dense concentration of facilities, the
facilities on the map may overlap and some may be impossible to see. IPM modeling did not include generation
outside the contiguous U.S., where EPA is aware of only two facilities that would be subject to the coal-fired
requirements of the final rule. Given the limited number of potentially impacted facilities, limited availability of
input data to inform the modeling, and limited connection to the continental grid, EPA did not model the
impacts of the rule beyond the contiguous U.S. Facilities outside the contiguous U.S. are displayed based on data
from EPA's 2010 ICR for the rule.
As noted above, IPM has been used for evaluating the economic and emission impacts
of environmental policies for over two decades. The economic modeling presented in this
chapter has been developed for specific analyses of the power sector. Thus, the model has
been designed to reflect the industry as accurately as possible. To that end, EPA uses a series of
capital charge factors in IPM that embody financial terms for the various types of investments
that the power sector considers for meeting future generation and environmental constraints.
3-8
-------
The model applies a discount rate of 6.15% for optimizing the sector's decision-making over
time. IPM's discount rate, designed to represent a broad range of private-sector decisions for
power generation, rates differs from discount rates used in other analyses in this RIA, such as
the benefits analysis which each assume alternative social discount rates of 3% and 7%. These
discount rates represent social rates of time preference, whereas the discount rate in IPM
represents an empirically-informed price of raising capital for the power sector. Like all other
assumed price inputs in IPM, EPA uses the best available information from utilities, financial
institutions, debt rating agencies, and government statistics as the basis for the capital charge
rates and the discount rate used for power sector modeling in IPM.
More detail on IPM can be found in the model documentation, which provides
additional information on the assumptions discussed here as well as all other assumptions and
inputs to the model (http://www.epa.gov/airmarkets/progsregs/epa-ipm). Updates specific to
MATS modeling are also in the IPM 4.10 Supplemental Documentation for MATS.
3.2 Projected Emissions
MATS is anticipated to achieve substantial emissions reductions from the power sector.
Since the technologies available to meet the emission reduction requirements of the rule
reduce multiple air pollutants, EPA expects the rule to yield a broad array of pollutant
reductions from the power sector. The primary pollutants of concern under MATS from the
power sector are mercury, acid gases such as hydrogen chloride (HCI), and HAP metals,
including antimony, arsenic, beryllium, cadmium, chromium, cobalt, mercury, manganese,
nickel, lead, and selenium. EPA has extensively analyzed mercury emissions from the power
sector, and IPM modeling assesses the mercury contents in all coals and the removal
efficiencies of relevant emission control technologies (e.g., ACI). EPA also models emissions and
the pollution control technologies associated with HCI (as a surrogate for acid gas emissions).
Like S02, HCI is removed by both scrubbers and DSI (dry sorbent injection). Projected emissions
are based on both control technology and detailed coal supply curves used in the model that
reflect the chlorine content of coals, which corresponds with the supply region, coal grade, and
sulfur, mercury, and ash content of each coal type. This information is critical for accurately
projecting future HCI emissions, and for understanding how the power sector will respond to a
policy requiring reductions of multiple HAPs.
Generally, existing pollution control technologies reduce emissions across a range of
pollutants. For example, both FGD and SCR can achieve notable reductions in mercury in
addition to their primary targets of S02 and NOX reductions. DSI will reduce HCI emissions while
3-9
-------
also yielding substantial S02 emission reductions, but is not assumed in EPA modeling to result
in mercury reductions. Since there are many avenues to reduce emissions, and because the
power sector is a highly complex and dynamic industry, EPA employs IPM in order to reflect the
relevant components of the power sector accurately, while also providing a sophisticated view
of how the industry could respond to particular policies to reduce emissions. For more detail on
how EPA models emissions from the power sector, including recent updates to include acid
gases, see IPM 4.10 Supplemental Documentation for MATS.
Under MATS, EPA projects annual HCI emissions reductions of 88 percent in 2015, Hg
emissions reductions of 75 percent in 2015, and PM2.5 emissions reductions of 19 percent in
2015 from coal-fired EGUs greater than 25 MW. In addition, EPA projects S02 emission
reductions of 41 percent, and annual C02 reductions of 1 percent from coal-fired EGUs greater
than 25 MW by 2015, relative to the base case (see Table 3-4).8 Mercury emission projections in
EPA's base case are affected by the incidental capture in other pollution control technologies
(such as FGD and SCR) as described above.
Table 3-4. Projected Emissions of SO2, NOX, Mercury, Hydrogen Chloride, PM, and CO2 with
the Base Case and with MATS, 2015
Million Tons
Base
MATS
All EGUs
Covered EGUs
All EGUs
Covered EGUs
S02
3.4
3.3
2.1
1.9
NOX
1.9
1.7
1.9
1.7
Mercury
(Tons)
28.7
26.6
8.8
6.6
Thousand Tons
HCI
48.7
45.3
9.0
5.5
PM2.5
277
270
227
218
C02
(Million Metric
Tonnes)
2,230
1,906
2,215
1,882
Source: Integrated Planning Model run by EPA, 2011
The CO2 emissions reported from IPM account for the direct CO2 emissions from fuel combustion and CO2 created
from chemical reactions in pollution controls to reduced sulfur.
3-10
-------
SO: Emissions (tons)
| | 2015 Base Case
^^| 2015 MATS
Scale: Largest batequals274 tbi
tons of SOa in Texasin 201:5 Bas
Figure 3-2. SO2 Emissions from the Power Sector in 2015 with and without MATS
Source: 2015 emissions include coal steam (including IGCC and petroleum coke) units >25 MW from IPM v4.10
base case and control case projections (EPA, February 2011)
NOx Emissions (tons)
[ | 2015 Base Case
| | 2015 MATS
Scale: Largestbarequalsl 16 thousand
NOX in Texasiii2015 Control Case
Figure 3-3. NOX Emissions from the Power Sector in 2015 with and without MATS
Source: 2015 emissions include coal steam (including IGCC and petroleum coke) units >25 MW from IPM
v4.10_MATS base case and control case projections (EPA, 2011)
3-11
-------
Hg Emissions (tons)
[ | 2015 Base Case
| | 201 5 MATS
Scale: Laigestbarequals33S tons of
Figure 3-4. Mercury Emissions from the Power Sector in 2015 with and without MATS
Source: 2015 emissions include coal steam (including IGCC and petroleum coke) units >25 MW from IPM
v4.10_MATS base case and control case projections (EPA, 2011)
3-12
-------
HC1 Emissions (tons)
[ | 2015 Base Case
| | 2015 MATS
Scale: Largest-barequalsWCO tonsof
HClin Michigan in 201.5 Base Case
Figure 3-5. Hydrogen Chloride Emissions from the Power Sector in 2015 with and without
MATS
Source: 2015 emissions include coal steam (including IGCC and petroleum coke) units >25 MW from IPM
v4.10_MATS base case and control case projections (EPA, 2011)
3.3 Projected Compliance Costs
The power industry's "compliance costs" are represented in this analysis as the change
in electric power generation costs between the base case and policy case in which the sector
pursues pollution control approaches to meet the final HAP emission standards. In simple
terms, these costs are the resource costs of what the power industry will directly expend to
comply with EPA's requirements.
EPA projects that the annual incremental compliance cost of MATS is $9.4 billion in 2015
($2007). The annual incremental cost is the projected additional cost of complying with the
final rule in the year analyzed, and includes the amortized cost of capital investment (at 6.15%)
and the ongoing costs of operating additional pollution controls, investments in new generating
sources, shifts between or amongst various fuels, and other actions associated with
compliance. This projected cost does not include the compliance calculated outside of IPM
modeling, namely the compliance costs for oil-fired EGUs, and monitoring, reporting, and
record-keeping costs. See section 3.14 for further details on these costs. EPA believes that the
3-13
-------
cost assumptions used for the final rule reflect, as closely as possible, the best information
available to the Agency today.
Table 3-5. Annualized Compliance Cost for MATS Requirements on Coal-fired Generation
2015 2020 2030
Annualized Compliance Cost (billions of 2007$) $9.4 $8.6 $7.4
Source: Integrated Planning Model run by EPA, 2011.
EPA's projection of $9.4 billion in additional costs in 2015 should be put into context for
power sector operations. As shown in section 2.7, the power sector is expected in the base case
to expend over $320 billion in 2015 to generate, transmit, and distribute electricity to end-use
consumers. Therefore, the projected costs of compliance with MATS amount to less than a 3%
increase in the cost to meet electricity demand, while securing public health benefits that are
several times more valuable (as described in Chapters 4 and 5).
3.4 Projected Compliance Actions for Emissions Reductions
Fossil fuel-fired electric generating units are projected to achieve HAP emission
reductions through a combination of compliance options. These actions include improved
operation of existing controls, additional pollution control installations, coal switching
(including blending of coals), and generation shifts towards more efficient units and lower-
emitting generation technologies (e.g., some reduction of coal-fired generation with an
increase of generation from natural gas). In addition, there will be some affected sources that
find it uneconomic to invest in new pollution control equipment and will be removed from
service. These facilities are generally amongst the oldest and least efficient power plants, and
typically run infrequently. In order to ensure that any retirements resulting from MATS do not
adversely impact the ability of affected sources and electric utilities from meeting the demand
for electricity, EPA has conducted an analysis of the impacts of projected retirements on
electric reliability. This analysis is discussed in TSD titled: "Resource Adequacy and Reliability in
the IPM Projections for the MATS Rule" which is available in the docket.
The requirements under MATS are largely met through the installation of pollution
controls (see Figure 3-6). To a lesser extent, there is a small degree of shifting within and across
various ranks and types of coals, and a relatively small shift from coal-fired generation to
greater use of natural gas and non-emitting sources of electricity (e.g., hydro and nuclear) (see
Table 3-6). The largest share of emissions reductions occur from coal-fired units installing new
pollution control devices, such as FGD, ACI, and fabric filters; a smaller share of emission
3-14
-------
reductions come from fuel shifts and unit retirements. Mercury emission reductions are largely
driven by SCR/FGD combinations and ACI installations. HCI emission reductions are largely
driven by FGD and DSI installations, which also incidentally provide substantial S02 reductions
in the policy case. Mercury, PM2.5, and HCI emission reductions are also facilitated by the
installation of fabric filters, which boost mercury and HCI removal efficiencies of ACI and DSI,
respectively.
250
200
191
u
OJ
a
m
u
"ro
4-*
o
Wet FGD Dry FGD FGD DSI
Upgrade
SCR
ACI
FF
ESP
Upgrade
Figure 3-6. Operating Pollution Control Capacity on Coal-fired Capacity (by Technology) with
the Base Case and with MATS, 2015 (GW)
Note: The difference between controlled capacity in the base case and under the MATS may not necessarily equal
new retrofit construction, since controlled capacity above reflects incremental operation of dispatchable
controls in 2015. Additionally, existing ACI installed on those units online before 2008 are not included in the
base case to reflect removal of state mercury rules from IPM modeling. For these reasons, and due to rounding,
numbers in the text below may not reflect the increments displayed in this figure. See IPM Documentation for
more information on dispatchable controls.
Source: Integrated Planning Model run by EPA, 2011.
As shown in Figure 3-6, this analysis projects that by 2015, the final rule will drive the
installation of an additional 20 GW of dry FGD (dry scrubbers), 44 GW of DSI, 99 GW of
additional ACI, 102 GW of additional fabric filters, 63 GW of scrubber upgrades, and 34 GW of
ESP upgrades. Furthermore, the final rule results in a 3 GW decrease in retrofit wet FGD
capacity relative to the base, where the S02 allowance price under CSAPR provides an incentive
for the additional S02 reductions achieved by a wet scrubber relative to a dry scrubber.
The difference between operating controlled capacity in the base case and under MATS
in Figure 3-6 may not necessarily equal new retrofit construction, since total controlled capacity
in the figure reflects incremental operation of existing controls that are projected to operate
3-15
-------
under MATS but not under the base case. With respect to the increase in operating ACI, some
of this increase represents existing ACI capacity on units built before 2008. EPA's modeling does
not reflect the presence of state mercury rules, and EPA assumes that ACI controls on units
built before 2008 do not operate in the absence of these rules. In the policy case, these controls
are projected to operate and the projected compliance cost thus reflects the operating cost of
these controls. Since these controls are in existence, EPA does not count their capacity toward
new retrofit construction, nor does EPA's compliance costs projection reflect the capital cost of
these controls (new retrofit capacity is reported in the previous paragraph).
3.5 Projected Generation Mix
Table 3-6 and Figure 3-7 show the generation mix in the base case and in MATS. In 2015,
coal-fired generation is projected to decline slightly and natural-gas-fired generation is
projected to increase slightly relative to the base case. Coal-fired generation is projected to
increase above 2009 actual levels. 2015 natural gas-fired generation is projected to be lower
than 2009, due in large part to the smaller relative difference in delivered natural gas and coal
prices in different areas of the country projected in 2015 than occurred in 2009. The vast
majority (over 98%) of base case coal capacity is projected to remain in service under MATS. In
addition, the operating costs of complying coal-fired units are not so affected as to result in
major changes in the electricity generation mix.
Table 3-6. Generation Mix with the Base Case and the MATS, 2015 (Thousand GWh)
Coal
Oil
Natural Gas
Nuclear
Hydroelectric
Non-hydro Renewables
Other
Total
2009
Historical
1,741
36
841
799
267
116
10
3,810
2015
Base Case
1,982
0.11
710
828
286
252
45
4,103
Policy Case
1,957
0.11
731
831
288
250
45
4,104
Change from
Base
-25
0.00
22
3
2
-1
0.0
1
Percent Change
-1.3%
3.6%
3.1%
0.4%
0.8%
-0.6%
0.0%
0.0%
Note: Numbers may not add due to rounding.
Source: 2009 data from AEO Annual Energy Review, Table 8.2c Electricity Net Generation: Electric Power Sector
by Plant Type, 1989-2010; 2015 projections are from the Integrated Planning Model run by EPA, 2011.
3-16
-------
Generation Mix
I Other
lOil
I Non-Hydro
Renewables
Hydro
Natural Gas
Nuclear
I Coal
Base MATS
2015
Base MATS Base MATS
2020
2030
Figure 3-7. Generation Mix with the Base Case and with MATS, 2015-2030
Source: Integrated Planning Model run by EPA, 2011.
3.6 Projected Withdrawals from Service
Relative to the base case, about 4.7 GW (less than 2 percent) of coal-fired capacity is
projected to be uneconomic to maintain by 2015. This projection considers various regional
factors (e.g., other available capacity and fuel prices) and unit attributes (e.g., efficiency and
age). These projected "uneconomic" units, for the most part, are older, smaller, and less
frequently used generating units that are dispersed throughout the country (see Table 3-7).
Table 3-7. Characteristics of Covered Operational Coal Units and Additional Coal Units
Projected to Withdraw as Uneconomic under MATS, 2015
Average Age
(Years)
Average Capacity
MW
Factor in Base
Withdrawn as Uneconomic
Operational
52
43
129
322
54%
71%
Source: Integrated Planning Model run by EPA, 2011.
These results should be considered "potential" closures. There are a variety of local
factors that could make plant owners decide to keep one or more units projected to be
uneconomic in service. These factors include different costs or demand estimates than what
3-17
-------
was included in the IPM modeling, and local operating conditions or requirements that are on a
smaller scale than that represented in EPA's IPM modeling. To the extent EPA's modeling does
not account for plants that continue to operate due to one or more of these local factors, these
results could be overestimating the capacity removed from service as a result of this rule.
For the final rule, EPA has examined whether the IPM-projected closures may adversely
impact reserve margins and reliability planning. The IPM model is specifically designed to
ensure that generation resource availability is maintained in the projected results subject to
reserve margins in 32 modeling regions for the contiguous US, which must be preserved either
by using existing resources or through the construction of new resources. IPM also addresses
reliable delivery of generation resources by limiting the ability to transfer power between
regions using the bulk power transmission system. Within each model region, IPM assumes that
adequate transmission capacity is available to deliver any resources located in, or transferred
to, the region. The IPM model projects available capacity given certain constraints such as
reserve margins and transmission capability but does not constitute a detailed reliability
analysis. For example, the IPM model does not examine frequency response. For more detail on
IPM's electric load modeling and power system operation, please see IPM documentation
(http://www.epa.gov/airmarkt/progsregs/epa-ipm/index.html) and the TSD on Resource
Adequacy and Reliability in the IPM Projections for the MATS Rule.
Total operational capacity is lower in the policy scenario, primarily as a result of
additional coal projected to be uneconomic to maintain. Since most regions are projected to
have excess capacity above their target reserve margins, most of these withdrawals from
service are absorbed by a reduction in excess reserves. Operational capacity changes from the
base case in 2015 are shown in Table 3-8.
3-18
-------
Table 3-8. Total Generation Capacity by 2015 (GW)
Pulverized Coal
Natural Gas Combined Cycle
Other Oil/Gas
Non-Hydro Renewables
Hydro
Nuclear
Other
Total
2010
317
201
253
31
99
102
5
1,009
Base Case
310
206
233
70
99
104
4
1,026
MATS
305
206
233
70
99
105
4
1,021
Source: 2010 data from EPA's NEEDS v.4.10_PTox. Projections from Integrated Planning Model run by EPA.
Note: "Non-Hydro Renewables" include biomass, geothermal, solar, and wind electric generation capacity. 2015
capacity reflects plant closures planned to occur prior to 2015.
The policy case analyzed maintains resource adequacy in each region projected to
decrease in coal capacity by using excess reserve capacity within the region, reversing base case
withdrawals of non-coal capacity, building new capacity, or by importing excess reserve
capacity from other regions. Although any closure of a large generation facility will need to be
studied to determine potential local reliability concerns, EPA analysis suggests that projected
economic withdrawals from service under the final rule could have little to no overall impact on
electric reliability. Not only are projected withdrawals under MATS limited in scope, but the
existing state of the power sector is also characterized by substantial excess capacity. The
weighted average reserve margin at the national level is projected to be approximately 25% in
the base case, while the North American Electric Reliability Corporation (NERC) recommends a
margin of 15%. EPA projects that MATS would only reduce total operational capacity by less
than one percent in 2015.
Moreover, coal units projected to withdraw as uneconomic are distributed throughout
the power grid with limited effect at the regional level, such that any potential impacts should
not adversely affect reserve margins and should be manageable through the normal industry
processes. For example, in the RFC NERC reliability Region, containing coal-fired generating
area in Pennsylvania, West Virginia and the Midwest, there is a decrease of less than 2% in the
reserve margin in the policy case and a remaining overall reserve margin of over 20%.
Furthermore, subregions may share each other's excess reserves to ensure adequate reserve
margins within a larger reliability region. EPA's IPM modeling accommodates such transfers of
reserves within the assumed limits of reliability of the inter-regional bulk power system. For
3-19
-------
these reasons, the projected closures of coal plants are not expected to raise broad reliability
concerns.
3.7 Projected Capacity Additions
Due in part to a low growth rate anticipated for future electricity demand levels in the
latest EIA forecast, EPA analysis indicates that there is sufficient excess capacity through 2015
to compensate for capacity that is retired from service under MATS. In the short-term, most
new capacity is projected as a mix of wind and natural gas in response to low fuel prices and
other energy policies (such as tax credits and state renewable portfolio standards). In addition,
future electricity demand expectations have trended downwards in recent forecasts, reducing
the need for new capacity in the 2015 timeframe (see Chapter 2 for more discussion on future
electricity demand).
Table 3-9. Total Generation Capacity by 2030 (GW)
Pulverized Coal
Natural Gas Combined Cycle
Other Oil/Gas
Non-Hydro Renewables
Hydro
Nuclear
Other
Total
2010
317
201
253
31
99
102
5
1,009
Base Case
308
275
235
79
99
103
4
1,103
MATS
304
278
235
79
99
103
4
1,102
Change
-3.9
2.9
0.6
0.1
0.0
0.3
0.0
-0.1
Note: "Non-Hydro Renewables" include biomass, geothermal, solar, and wind electric generation capacity.
Source: 2010 data from EPA's NEEDS v.4.10_PTox. Projections from Integrated Planning Model run by EPA.
3.8 Projected Coal Production for the Electric Power Sector
Coal production for electricity generation under MATS is expected to increase from
2009 levels and decline modestly relative to the base case without the rule. The reductions in
emissions from the power sector will be met through the installation and operation of pollution
controls for HAP removal. Many available pollution controls achieve emissions removal rates of
up to 99 percent (e.g., HCI removal by new scrubbers), which allows industry to rely more
heavily on local bituminous coal in the eastern and central parts of the country that has higher
contents of HCI and sulfur, and it is less expensive to transport than western subbituminous
coal. Overall demand for coal is projected to be reduced as a result of MATS, with a slight
3-20
-------
reduction in bituminous coal, and more of a reduction in subbituminous coal (see Tables 3-10
and 3-11). The trend reflects the projected reduced demand for lower-sulfur coal under MATS,
where nearly all units are operating with a post-combustion emissions control. In this case,
because of the additional pollution controls, many of these units no longer find it economic to
pay a transportation premium to purchase lower-sulfur subbituminous coals. Instead, EGUs are
generally projected to shift consumption towards nearby bituminous coal, which can achieve
low emissions when combined with post-combustion emissions controls. This explains the
increase from the base case in coal supplied from the Interior region, which is located in
relatively close proximity to many coal-fired generators subject to MATS. This continues a trend
of increased Interior supply (due to abundant Illinois Basin reserves that are relatively
inexpensive to mine) and decreased Central Appalachian supply which is forecasted to occur in
the base case from historic levels. The decline in Appalachia is a result of an increase in the
relative cost of Central Appalachian extraction due both to rising mining cost (e.g., in 2010
major producers reported mining cost increases up to 15% with this trend continuing into 2011)
and shrinking economically recoverable capacity. Growing international demand for
Appalachian thermal coal is also contributing to its rising price. The increase in lignite use
occurs at units blending subbituminous and lignite coals, and reflects a small shift in blended
balance towards a greater use of lignite.
Table 3-10. 2015 Coal Production for the Electric Power Sector with the Base Case and MATS
(Million Tons)
Supply Area
Appalachia
Interior
West
Waste Coal
Imports
Total
2009
246
129
553
14
942
2015 Base
184
216
554
14
30
998
2015 MATS
172
236
537
13
30
989
Change in 2015
-6%
9%
-3%
-5%
0%
-1%
Source: Production: U.S. Energy Information Administration (EIA), Coal Distribution — Annual (Final), web site
http://www.eia.doe.gov/cneaf/coal/page/coaldistrib/a_distributions.html (posted February 18, 2011); Waste
Coal: U.S. EIA, Monthly Energy Review, January 2011 Edition, Table 6.1 Coal Overview, web site
http://www.eia.doe.gov/emeu/mer/coal.html (posted January 31, 2011). All projections from Integrated
Planning Model run by EPA, 2011.
3-21
-------
Table 3-11. 2015 Power Sector Coal Use with the Base Case and the MATS, by Coal Rank
(TBtu)
Coal Rank
Bituminous
Subbituminous
Lignite
Total
Base
11,314
7,736
849
19,900
MATS
11,248
7,554
895
19,698
Change
-0.6%
-2%
5%
-1%
Source: Integrated Planning Model run by EPA, 2011.
Figure 3-8. Total Coal Production by Coal-Producing Region, 2007 (Million Short Tons)
Note: Regional totals do not include refuse recovery
Source: EIA Annual Coal Report, 2007
3.9 Projected Retail Electricity Prices
EPA's analysis projects a near-term increase in the average retail electricity price of 3.1%
in 2015 falling to 2% by 2020 under the final rule in the contiguous U.S. The projected price
impacts vary by region and are provided in Table 3-12 (see Figure 3-9 for regional
classifications).
Regional retail electricity prices are projected to range from 1 to 6 percent higher with MATS in
2015. The extent of regional retail electricity increases correlates with states that have
considerable coal-fired generation in total generation capacity and that coal-fired generation is
less well-controlled (such as in the ECAR and SPP regions). Retail electricity prices embody
generation, transmission, and distribution costs. IPM modeling projects changes in regional
3-22
-------
wholesale power prices, capacity payments, and actual costs of compliance in areas that are
"cost of service" regions that are combined with EIA regional transmission and distribution
costs to complete the retail price picture.
3-23
-------
Table 3-12. Projected Contiguous U.S. and Regional Retail Electricity Prices with the Base
Case and with the MATS (2007 cents/kWh)
Base Case
ECAR
ERCOT
MAAC
MAIN
MAPP
NY
NE
FRCC
STV
SPP
PNW
RM
CALI
Contiguous
U.S.
Average
2015
8.2
8.9
9.5
8.1
8.0
13.8
12.3
10.2
7.9
7.7
7.1
9.2
13.0
9.0
2020
8.2
8.8
10.4
8.4
7.9
13.4
11.8
9.7
7.8
7.4
6.8
9.5
12.5
9.0
2030
9.8
11.3
12.7
9.7
8.5
16.6
13.8
11.0
8.4
8.1
7.6
11.0
12.7
10.2
2015
8.5
9.2
9.8
8.3
8.5
14.1
12.6
10.4
8.2
8.1
7.3
9.4
13.2
9.3
MATS
2020
8.5
8.8
10.4
8.6
8.3
13.5
11.9
9.8
8.0
7.8
7.0
9.7
12.6
9.2
Percent Change
2030
9.9
11.3
12.7
9.7
8.8
16.6
13.8
11.0
8.6
8.4
7.6
11.1
12.7
10.3
2015
4.5%
3.3%
2.8%
2.8%
5.3%
2.2%
2.0%
2.2%
3.1%
6.3%
2.7%
2.3%
1.3%
3.1%
2020
2.8%
0.6%
0.4%
2.2%
5.6%
0.7%
0.8%
0.9%
2.4%
6.1%
2.6%
1.9%
0.7%
2.0%
2030
1.0%
-0.2%
-0.2%
0.2%
3.4%
-0.1%
0.0%
0.4%
1.6%
4.6%
1.1%
1.1%
0.0%
0.9%
Source: EPA's Retail Electricity Price Model, 2011.
3-24
-------
PNW
CALI RM
Figure 3-9. Retail Price Model Regions
3.10 Projected Fuel Price Impacts
The impacts of the final Rule on coal and natural gas prices before shipment are shown
below in Tables 3-13 and 3-14. Overall, the national average coal price changes are related to
changes in demand for a wide variety of coals based upon a number of parameters (e.g.,
chlorine or mercury content, heat content, proximity to the power plant, etc.), and this national
average captures increases and decreases in coal demand and price at the regional level.
Generally, total demand for coal decreases slightly under MATS, most notably subbituminous
coal, which is by far the least expensive type of coal supplied to the power sector on an MMBtu
basis. This is reflected in the projected average minemouth price of coal, which goes up by
about 3 percent even though total demand for coal is reduced slightly (1 percent reduction).
Notwithstanding the projected "mine-mouth" coal price changes, many units may in fact be
realizing overall fuel cost savings by switching to more local coal supplies (which reduces
transportation costs) after installing additional pollution control equipment. Gas price changes
are directly related the projected increase in natural gas consumption under MATS. This
increase in demand is met by producing additional natural gas at some increase in regional
costs, resulting over time in a small price increase.
3-25
-------
Table 3-13. Average Minemouth and Delivered Coal Prices with the Base Case and with
MATS (2007$/MMBtu)
Minemouth
Delivered
2007
1.27
1.76
Base Case
1.35
2.11
2015
MATS
1.39
2.15
Percent
Change
from Base
2.8%
1.9%
Base Case
1.51
2.29
2030
MATS
1.56
2.33
Percent
Change
from Base
3.3%
1.7%
Source: Historical data from EIA AEO 2010 Reference Case Table 15 (Coal Supply, Distribution, and Prices);
projections from the Integrated Planning Model run by EPA, 2011.
Table 3-14. 2015-2030 Weighted Average Henry Hub (spot) and Delivered Natural Gas Prices
with the Base Case and with MATS (2007$/MMBtu)
Henry Hub
Delivered - Electric Power
Delivered - Residential
Base Case
5.29
5.56
10.94
MATS
5.32
5.60
10.97
Percent Change from Base
0.6%
0.6%
0.3%
Source: Projections from the Integrated Planning Model run by EPA (2011) adjusted to Henry Hub prices using
historical data from EIA AEO 2011 reference case to derive residential prices.
IPM modeling of natural gas prices uses both short- and long-term price signals to
balance supply of and demand in competitive markets for the fuel across the modeled time
horizon. As such, it should be understood that the pattern of IPM natural gas price projections
over time is not a forecast of natural gas prices incurred by end-use consumers at any particular
point in time. The natural gas market in the United States has historically experienced
significant price volatility from year to year, between seasons within a year, and even sees
major price swings during short-lived weather events (such as cold snaps leading to short-run
spikes in heating demand). These short-term price signals are fundamental for allowing the
market to successfully align immediate supply and demand needs; however, end-use
consumers are typically shielded from experiencing these rapid fluctuations in natural gas
prices by retail rate regulation and by hedging through longer-term fuel supply contracts. IPM
assumes these longer-term price arrangements take place "outside of the model" and on top of
the "real-time" shorter-term price variation necessary to align supply and demand. Therefore,
the model's natural gas price projections should not be mistaken for traditionally experienced
consumer price impacts related to natural gas, but a reflection of expected average price
changes over the time period 2015 to 2030.
3-26
-------
For this analysis, in order to represent a natural gas price evolution that end-use
consumers can anticipate under retail rate regulation and/or typical hedging behavior, EPA is
displaying the weighted average of IPM's natural gas price projections for the 2015-2030 time
horizon (see Table 3-14). In that framework, consumer natural gas price impacts are anticipated
to range from 0.3% to 0.6% based on consumer class in response to MATS.
3.11 Key Differences in EPA Model Runs for MATS Modeling
In this analysis, we use the Integrated Planning Model (IPM), which is a multiregional,
dynamic, deterministic linear programming model of the U.S. electric power sector.9 The length
of time required to conduct emissions and photochemical modeling precluded the use of IPM
version 4.10_MATS. Thus the air quality modeling for MATS relied on ECU emission projections
from an interim IPM platform that was subsequently updated during the rulemaking process for
the base case and policy scenario summarized in this chapter. The 2015 base case ECU
emissions projections of mercury, hydrogen chloride, S02, and PM used in air quality modeling
were obtained from an earlier version of IPM, 4.10_FTransport. IPM version 4.10_FTransport
reflects all state rules and consent decrees adopted through December 2010. Units with S02 or
NOX advanced controls (e.g., scrubber, SCR) that were not required to run for compliance with
Title IV, New Source Review (NSR), state settlements, or state-specific rules were allowed in
IPM to decide on the basis of economic efficiency whether to operate those controls. Note that
this base case includes CSAPR, which was finalized in July 2011. Further details on the ECU
emissions inventory used for this proposal can be found in the IPM Documentation.
The results presented in this chapter, from IPM version 4.10_MATS, reflect updates
made to the 4.10_FTransport base case. These revisions are fully documented in the IPM 4.10
Supplemental Documentation for MATS and include: updated assumptions regarding the
removal of HCI by alkaline fly ash in subbituminous and lignite coals; an update to the fuel-
based mercury emission factor for petroleum coke, which was corrected based on re-
examination of the 1999 ICR data; updated capital cost for new nuclear capacity and nuclear life
extension costs; corrected variable operating and maintenance cost (VOM) for ACI retrofits;
adjusted coal rank availability for some units, consistent with EIA From 923 (2008); updated
state rules in Washington and Colorado; and numerous unit-level revisions based on comments
received through the notice and comment process. Additionally, IPM v.4.10_MATS does not
reflect mercury-specific state regulations (see section 1 above).
9 http://www.epa.gov/airmarkt/progsregs/epa-ipm/index.html
3-27
-------
3.12 Projected Primary PM Emissions from Power Plants
IPM does not endogenously model primary PM emissions from power plants. These
emissions are calculated as a function of IPM outputs, emission factors and control
configuration. IPM-projected fuel use (heat input) is multiplied by PM emission factors (based
in part on the presence of PM-relevant pollution control devices) to determine PM emissions.
Primary PM emissions are calculated by adding the filterable PM and condensable PM
emissions.
Filterable PM emissions for each unit are based on historical information regarding
existing emissions controls and types of fuel burned and ash content of the fuel burned, as well
as the projected emission controls (e.g., scrubbers and fabric filters).
Condensable PM emissions are based on plant type, sulfur content of the fuel, and
S02/HCI and PM control configurations. Although EPA's analysis is based on the best available
emission factors, these emission factors do not account for the potential changes in
condensable PM emissions due to the installation and operation of SCRs. The formation of
additional condensable PM (in the form of S03 and H2S04) in units with SCRs depends on a
number of factors, including coal sulfur content, combustion conditions and characteristics of
the catalyst used in the SCR, and is likely to vary widely from unit to unit. SCRs are generally
designed and operated to minimize increases in condensable PM. This limitation means that
IPM post-processing is potentially underestimating condensable PM emissions for units with
SCRs. In contrast, it is possible that IPM post-processing overestimates condensable PM
emissions in a case where the unit is combusting a low-sulfur coal in the presence of a scrubber.
EPA plans to continue improving and updating the PM emission factors and calculation
methodologies. For a more complete description of the methodologies used to post-process
PM emissions from IPM, see "IPM ORL File Generation Methodology" (March, 2011).
3.13 Illustrative Dry Sorbent Injection Sensitivity
Several commenters believe that EPA's IPM modeling assumptions regarding the
efficacy and cost of DSI are based on too little data and are too optimistic. Some commenters
believe that in practice there will be a need for many more FGD scrubbers for MATS compliance
than projected by EPA for effective acid gas control, and at a corresponding higher cost. EPA
disagrees with these opinions for several reasons (see the response to comments document in
the docket) and believes that EPA's modeling assumptions regarding DSI cost and performance
are reasonable.
3-28
-------
However, to examine the potential impacts of limited DSI availability, EPA analyzed a
scenario that limited total DSI capacity to 35 GW in 2015. In this scenario, which reduces the
capacity of DSI by 18 GW compared to the primary MATS scenario, an additional 14 GW of coal
capacity chooses to install scrubbers, and an additional 1.3 GW of capacity is projected to
withdraw from service.
Limiting total DSI capacity to 35 GW results in a $1.2 billion (2007$) increase in
annualized compliance costs in 2015. Additionally, S02 is further reduced in 2015 by an
additional 62,000 tons (a 4.7% increase in S02 reductions and 4.5% increase in health benefits).
3.14 Additional Compliance Costs Analyzed for Covered Units
3.14.1 Compliance Cost for Oil-Fired Units.
As discussed in section 3.1, EPA used IPM to assess impacts of the MATS emission
limitations for coal-fired EGUs but did not use IPM to assess the impacts for oil-fired units. IPM,
with its power system and fuel cost assumptions, predicts many dual fuel units switch to
natural gas and oil-fired units will not operate because IPM focuses on least cost operation of
the power system. However, despite their apparent economic disadvantages, many of these
units have run during many of the past five years (2006-2010). Therefore, EPA conducted a
separate analysis to assess the impacts of the MATS emission limitations for oil-fired units.10
EPA limited this analysis to oil-fired units in the contiguous U.S. Although there are several oil-
fired units in states and territories outside the contiguous U.S., the final MATS emission
limitations (shown in Table 3-2) for non-continental units will likely allow these units to
continue firing residual fuel oil without additional air pollution controls.
For the base case, EPA categorized units by modeled fuels as listed in NEEDS 4.10 (EPA,
December 2010) and assigned each unit the least-cost fuel among its available fuels. For units
with natural gas curtailment provisions that might require the firing of residual fuel oil, EPA
assigned a mixed fuel ratio based on each unit's 2008-2010 weighted average natural gas-to-
fuel oil ratio. For the policy case, EPA assessed three compliance options: (1) switching to
natural gas where available, (2) switching to distillate fuel oil, and (3) installing an electrostatic
precipitator (ESP) capable of 90% particulate removal efficiency. These compliance options
address particulate emissions only. However, there might be additional emission reductions
that result from changes to oil-fired units' generation due to changes in relative generating
costs.
10 Additional details and methodology for the analysis are presented in appendix 3 A.
3-29
-------
Between the base case and policy case, 12 units convert from residual fuel oil to
distillate fuel oil at a cost of approximately $12 million annually (2007$) to meet the MATS
emission limitations for oil-fired units. An additional 11 units, eight of which are subject to
natural gas curtailment, that do not have existing ESP particulate pollution controls install an
ESP at a cost of approximately $44 million annually (2007$) to achieve the MATS emission
limitations for oil-fired units (see Table 3-15). EPA believes the emission impacts from these
potential actions will be relatively small when compared to the full impacts of the MATS
emission limitations because particulate emissions from oil-fired units are a small fraction of
the total particulate emissions from EGUs.
Table 3-15. Cost Impacts of Compliance Actions for Oil-Fired Units
Compliance option
Switch to distillate fuel oil
Install ESP for residual fuel oil
Total
Number of units affected
12
11
23
Capacity of units affected
2,675 MW
4,015 MW
6,690 MW
Annual cost (2007$)
$12 million
$44 million
$56 million
3.14.2 Monitoring, Reporting and Record-keeping Costs
The annual monitoring, reporting, and record-keeping burden for this collection (averaged over
the first 3 years after the effective date of the standards) is estimated to be $158 million. This
includes 698,907 labor hours per year at a total labor cost of $49 million per year, and total
non-labor capital costs of $108 million per year. This estimate includes initial and annual
performance tests, semiannual excess emission reports, developing a monitoring plan,
notifications, and record-keeping. Initial capital expenses to purchase monitoring equipment
for affected units are estimated at a cost of $231 million. This includes 504,629 labor hours at a
total labor cost of $35 million for planning, selection, purchase, installation, configuration, and
certification of the new systems and total non-labor capital costs of $196 million. All burden
estimates are in 2007 dollars and represent the most cost effective monitoring approach for
affected facilities. See Section 7.3, Paperwork Reduction Act.
3.14.3 Total Costs Projected for Covered Units under MATS
EPA used IPM to analyze the compliance cost, and economic and energy impacts of the MATS
rule. IPM estimated the costs for coal-fired electric utility steam generating units that burn coal,
coal refuse, or solid-oil derived fuel. EPA did not use IPM, however, estimate compliance costs
for most oil/gas steam boilers because IPM projection shows least-cost dispatch in an
3-30
-------
environment where oil/gas-fired units are primarily selecting natural gas on an economic basis.
In the separate analysis summarized above, EPA estimates compliance costs for oil-fired EGUs
in a scenario in which these units continue to burn oil as historically observed and thus take
compliance measures to remain on oil. This is a reasonable estimate of compliance costs for
these units, but does not represent a re-balancing of electricity dispatch where these units
combust oil rather than natural gas. Therefore, the summation of IPM-projected compliance
costs for least-cost dispatch with the oil-fired compliance costs and the monitoring, reporting,
and record-keeping costs is a reasonable approximation of total compliance costs, but does not
represent projected compliance costs under an economically efficient dispatch (see Table 3-16).
Table 3-16. Total Costs Projected for Covered Units under MATS, 2015 (billions of 2007$)
2015
IPM Projection $9.4
Monitoring/Reporting/Record-keeping $0.158
Oil-Fired Fleet $0.056
Total $9.6
3.15 Limitations of Analysis
EPA's modeling is based on expert judgment of various input assumptions forvariables
whose outcomes are in fact uncertain. Assumptions for future fuel supplies and electricity
demand growth deserve particular attention because of the importance of these two key model
inputs to the power sector. As a general matter, the Agency reviews the best available
information from engineering studies of air pollution controls to support a reasonable modeling
framework for analyzing the cost, emission changes, and other impacts of regulatory actions.
The IPM-projected annualized cost estimates of private compliance costs provided in
this analysis are meant to show the increase in production (generating) costs to the power
sector in response to the final rule. To estimate these annualized costs, EPA uses a conventional
and widely-accepted approach that applies a capital recovery factor (CRF) multiplier to capital
investments and adds that to the annual incremental operating expenses. The CRF is derived
from estimates of the cost of capital (private discount rate), the amount of insurance coverage
required, local property taxes, and the life of capital. The private compliance costs presented
earlier are EPA's best estimate of the direct private compliance costs of MATS.
The annualized cost of the final rule, as quantified here, is EPA's best assessment of the
cost of implementing the rule. These costs are generated from rigorous economic modeling of
3-31
-------
changes in the power sector due to implementation of MATS. This type of analysis using IPM
has undergone peer review, and federal courts have upheld regulations covering the power
sector that have relied on IPM's cost analysis.
Cost estimates for MATS are based on results from ICF's Integrated Planning Model. The
model minimizes the costs of producing electricity (including abatement costs) while meeting
load demand and other constraints (full documentation for IPM can be found at
http://www.epa.gov/airmarkets/progsregs/epa-ipm and in the IPM 4.10 Supplemental
Documentation for MATS. IPM assumes "perfect foresight" of market conditions over the time
horizon modeled; to the extent that utilities and/or energy regulators misjudge future
conditions affecting the economics of pollution control, costs may be understated as well.
In the policy case modeling, EPA exogenously determines that a subset of covered units
might require a retrofit fabric filter (also known as a baghouse) retrofit, or might need to
upgrade existing ESP control in order to meet the PM standard. EPA's methodology for
assigning these controls to EGUs in policy case modeling is based on historic PM emission rates
and reported control efficiencies, and is explained in the IPM 4.10 Supplemental
Documentation for MATS.
Additionally, this modeling analysis does not take into account the potential for
advancements in the capabilities of pollution control technologies as well as reductions in their
costs over time. In addition, EPA modeling cannot anticipate in advance the full spectrum of
compliance strategies that the power sector may innovate to achieve the required emission
reductions under MATS, which would potentially reduce overall compliance costs. Where
possible, EPA designs regulations to assure environmental performance while preserving
flexibility for affected sources to design their own solutions for compliance. Industry will
employ an array of responses, some of which regulators may not fully anticipate and will
generally lead to lower costs associated with the rule than modeled in this analysis. For
example, unit operators may find opportunities to improve or upgrade existing pollution
control equipment without requiring as many new retrofit devices (i.e., meeting the PM
standard with an existing ESP without requiring installation of a new fabric filter).
With that in mind, MATS establishes emission rates on key HAPs, and although this
analysis projects a specific set of technologies and behaviors as EPA's judgment of least-cost
compliance, the power sector is free to adopt alternative technologies and behaviors to achieve
the same environmental outcome EPA has deemed in the public interest as laid out in the Clean
Air Act. Such regulation serves to promote innovation and the development of new and
3-32
-------
cheaper technologies. As an example, cost estimates of the Acid Rain S02 trading program by
Resources for the Future (RFF) and MIT's Center for Energy and Environmental Policy Research
(CEEPR) have been as much as 83 percent lower than originally projected by the EPA (see
Carlson et al., 2000; Ellerman, 2003). It is important to note that the original analysis for the
Acid Rain Program done by EPA also relied on an optimization model like IPM. Ex ante, EPA cost
estimates of roughly $2.7 to $6.2 billion11 in 1989 were an overestimate of the costs of the
program in part because of the limitation of economic modeling to perfectly anticipate
technological improvement of pollution controls and economic improvement of other
compliance options such as fuel switching. Ex post estimates of the annual cost of the Acid Rain
S02 trading program range from $1.0 to $1.4 billion.
In recognition of this historic pattern of overestimated regulatory cost, EPA's mobile
source program uses adjusted engineering cost estimates of pollution control equipment and
installation costs.12 To date, and including this analysis, EPA has not incorporated a similar
approach into IPM modeling of ECU compliance with environmental constraints. As a result,
this analysis may overstate costs where such cost savings from as-yet untapped improvements
to pollution control technologies may occur in the future. Considering the broad and complex
suite of generating technologies, fuels, and pollution control strategies available to the power
sector, as well as the fundamental role of operating cost in electricity dispatch, it is not possible
to apply a single technology-improving "discount" transformation to the cost projections in this
analysis. The Agency will consider additional methodologies in the future which may inform the
amount by which projected compliance costs could be overstated regarding further
technological development in analyses of power sector regulations.
As configured in this application, IPM does not take into account demand response (i.e.,
consumer reaction to electricity prices). The increased retail electricity prices shown in
Table 3-13 would prompt end users to increase investment in energy efficiency and/or curtail
(to some extent) their use of electricity and encourage them to use substitutes.13 Those
responses would lessen the demand for electricity, resulting in electricity price increases slightly
lower than IPM predicts, which would also reduce generation and emissions. Demand response
would yield certain unquantified cost savings from requiring less electricity to meet the
quantity demanded. To some degree, these saved resource costs will offset the additional costs
11 2010 Phase II cost estimate in $1995.
See regulatory impact analysis for the Tier 2 Regulations for passenger vehicles (1999) and Heavy-Duty Diesel
Vehicle Rules (2000).
13 The degree of substitution/curtailment depends on the costs and performance of the goods that substitute for
more energy consuming goods, which is reflected in the demand elasticity.
3-33
-------
of pollution controls and fuel switching that EPA anticipates from the final rule, although there
could be some increase in social cost resulting from any decrease in electricity consumption.
Although the reduction in electricity use is likely to be small, the cost savings from such a large
industry14 are not insignificant. EIA analysis examining multi-pollutant legislation in 2003
indicated that the annualized costs of MATS may be overstated substantially by not considering
demand response, depending on the magnitude and coverage of the price increases.15
EPA's IPM modeling of MATS reflects the Agency's authority to allow facility-level
compliance with the HAP emission standards rather than require each affected unit at a given
facility to meet the standards separately. This flexibility would offer important cost savings to
facility owners in situations where a subset of affected units at a given facility could be
controlled more cost-effectively such that their "overperformance" would compensate for any
"underperformance" of the rest of the affected units. EPA's modeling in this analysis required
the average emission rate across all affected units at a given facility to meet the standard. This
averaging flexibility has the potential to offer further cost savings beyond this analysis if
particular units find ways to achieve superior pollution control beyond EPA's assumptions of
retrofit technology performance at the modeled costs (which could then reduce the need to
control other units at the same facility).
Additionally, EPA has chosen to express most of the control requirements here as
engineering performance standards (e.g., Ibs/MMBtu of heat input), which provide power plant
operators goals to meet as they see fit in choosing coals with various pollutant concentrations
and pollutant control technologies that they adopt to meet the requirements. Historically, such
an approach encourages industry to engineer cheaper solutions over time to achieve the
pollution controls requirements.
EPA's IPM modeling is based on retrofit technology cost assumptions which reflect the
best available information on current and foreseeable market conditions for pollution control
deployment. In the current economic environment, EPA does not anticipate (and thus this
analysis does not reflect) significant near-term price increases in retrofit pollution control
supply chains in response to MATS. To the extent that such conditions may develop during the
14 Investor-owned utilities alone accounted for nearly $300 billion in revenue in 2008 (EIA).
15 See "Analysis of S. 485, the Clear Skies Act of 2003, and S. 843, the Clean Air Planning Act of 2003." Energy
Information Administration. September, 2003. EIA modeling indicated that the Clear Skies Act of 2003 (a
nationwide cap and trade program for SO2, NOX, and mercury), demand response could lower present value costs
by as much as 47% below what it would have been without an emission constraint similar to the Transport Rule.
3-34
-------
sector's installation of pollution control technologies under the final rule, this analysis may
understate the cost of compliance.
3.16 Significant Energy Impact
MATS would have a significant impact according to E.O. 13211: Actions that Significantly
Affect Energy Supply, Distribution, or Use. Under the provisions of this rule, EPA projects that
approximately 4.7 GW of coal-fired generation (less than 2 percent of all coal-fired capacity and
0.5% of total generation capacity in 2015) may be removed from operation by 2015. These
units are predominantly smaller and less frequently-used generating units dispersed
throughout the area affected by the rule. If current forecasts of either natural gas prices or
electricity demand were revised in the future to be higher, that would create a greater
incentive to keep these units operational.
EPA also projects fuel price increases resulting from MATS. Average retail electricity
price are shown to increase in the contiguous U.S. by 3.1 percent in 2015. This is generally less
of an increase than often occurs with fluctuating fuel prices and other market factors. Related
to this, the average delivered coal price increases by less than 2 percent in 2015 as a result of
shifts within and across coal types. As discussed above in section 8.10, EPA also projects that
electric power sector-delivered natural gas prices will increase by about 0.6% percent over the
2015-2030 timeframe and that natural gas use for electricity generation will increase by less
than 200 billion cubic feet (BCF) in 2015. These impacts are well within the range of price
variability that is regularly experienced in natural gas markets. Finally, the EPA projects coal
production for use by the power sector, a large component of total coal production, will
decrease by 10 million tons in 2015 from base case levels, which is about 1 percent of total coal
produced for the electric power sector in that year. The EPA does not believe that this rule will
have any other impacts (e.g., on oil markets) that exceed the significance criteria.
3.17 References
EIA Annual Coal Report 2008. DOE/EIA-0584 (2008). Available at:
http://www.eia. doe.gov/cneaf/coa l/page/acr/acr_sum. html
EIA Annual Energy Outlook 2003. DOE/EIA-0383 (2003). Available at:
http://www.eia.doe.gov/oiaf/archive/aeo03/index.html
EIA Electric Power Annual 2008. DOE/EIA-0348 (2008). Available at:
http://www.eia.doe.gov/cneaf/electricity/epa/epa_sum.htm
EIA Electric Power Monthly March 2010 with Data for December 2009. DOE/EIA-0226
(2010/03). Available at: http://www.eia.doe.gov/cneaf/electricity/epm/epm_sum.html
3-35
-------
Freme, Fred. 2009. U.S. Coal Supply and Demand: 2008 Review. EIA. Available at:
http://www.eia.doe.gov/cneaf/coal/page/special/tbll.html
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.
Manson, Nelson, and Neumann. 2002. "Assessing the Impact of Progress and Learning Curves
on Clean Air Act Compliance Costs." Industrial Economics Incorporated.
3-36
-------
APPENDIX 3A
COMPLIANCE COSTS FOR OIL-FIRED ELECTRIC GENERATING UNITS
This appendix highlights the supplemental oil-fired electric generating unit (ECU)
compliance cost analysis performed for the Mercury and Air Toxics Standards (MATS). EPA used
the Integrated Planning Model (IPM) to assess the cost, economic, and energy impacts of the
MATS emission limitations on coal-fired EGUs in the contiguous U.S., but did not use IPM to
assess the compliance costs for oil-fired EGUs because IPM focuses on the least cost operation
of the power system and, therefore, predicts the oil-fired units will not operate. These oil-fired
units, however, do not operate on a purely economic basis. Some oil-fired units may operate as
"must run", "black start", or "spinning reserve". In addition, some dual fuel fired units which
IPM predicts will fire natural gas may be required to fire fuel oil when subject to mandatory
curtailment of natural gas supplies.
When practicable, this supplemental analysis for oil-fired EGUs was based on the data
and assumptions used in IPM. Documentation for IPM can be found at
http://www.epa.gov/airmarkets/progsregs/epa-ipm.
3A.1 Methodology and Assumptions
3A. 1.1 Base Case
EPA developed the base case for oil-fired units listed in the National Electric Energy Data
System (hereafter, NEEDS) (EPA, 2010a). NEEDS lists 302 "oil/gas steam" units greater than 25
MW for which distillate fuel oil and/or residual fuel oil are among the modeled fuels (see Table
3A-1).16 For each of these units, EPA projected 2015 heat input and apportioned the heat input
among the NEEDS modeled fuels. EPA used each unit's average annual heat input from 2006-
201017 as a proxy for 2015 heat input. For units not subject to mandatory natural gas
curtailment, EPA assumed the unit fired the least cost fuel available based on regional IPM fuel
cost projections for 2015. For units that may be required to fire fuel oil due to mandatory
natural gas curtailment, EPA apportioned the heat input based on the unit's weighted average
natural gas and fuel oil apportionment from 2008-2010.1S EPA used the three most recent years
because, as a percentage of total heat input, fuel oil heat input has fallen steadily since 2007
(see Figure 3A-1). With increased availability of natural gas in the New York region from new
16 One unit, Charles Poletti unit 001 (ORIS 2491), was removed because the unit retired in 2010 (EPA, 2011).
17 Designated representatives for each of the oil-fired units included in this analysis certify and report hourly heat
input and emission data to EPA under 40CFR Part 75.
18 The units subject to mandatory natural gas curtailment report fuel-apportioned heat input to EPA under 40CFR
Part 75 (Appendix D). EPA categorized "diesel" as distillate fuel oil and "oil" and "other oil" as residual fuel oil.
3A-1
-------
gas supplies and new gas pipelines (FERC, 2011), it is likely this trend will continue even in the
absence of the MATS. Therefore, using a longer historical period might significantly
overestimate the proportion of heat input derived from fuel oil for these units.
Table 3A-1. Oil-fired EGUs by Fuel Type
NEEDS modeled fuel
Distillate fuel oil
Distillate fuel oil, natural gas
Residual fuel oil
Residual fuel oil, distillate fuel oil
Residual fuel oil, natural gas
Residual fuel oil, distillate fuel oil, natural gas
Number of units
10
99
17
15
149
12
Capacity (MW)
814
19,822
5,867
1,187
39,913
3,706
Source: EPA. 2010. National Electricity Energy Data System (NEEDS 4.10). Available at:
http://www.epa.gov/airmarkets/progsregs/epa-ipm/toxics.html.
250
200
Oil-fired heat
input (% of total
heat input)
£
c
o
150 -
= 100
50 -
35%
30%
25%
20%
15%
10%
- 5%
0%
Diesel
I Oil
I Gas
2006
2007
2008
2009
2010
Figure 3A-1. 2006-2010 Heat Input Apportioned by Fuel for Oil-Fired Units Subject to
Mandatory Natural Gas Curtailment
Source/Notes: EPA. 2011. Data and Maps. Available at: http://epa.gov/camddataandmaps/
Power companies are responding to fuel prices, natural gas supplies, and other market
factors by replacing some oil-gas steam units with new combined cycle plants (Neville, J. 2011).
EPA did not, however, factor in the effect of expanded availability of natural gas on the
3A-2
-------
utilization of these oil-fired units. As a result, this analysis likely overestimates the impact of the
MATS emission limitations on oil-fired units.
In the base case, natural gas is the least cost fuel for the majority of units (see Table 3A-
2). However, 41 units are expected to continue burning some amount of residual fuel oil
because the units are subject to mandatory natural gas curtailment or may not have access to
natural gas supplies.19 Of these 41 units, 14 have existing electrostatic precipitator (ESP)
particulate pollution controls.
Table 3A-2. Least Cost NEEDS Modeled Fuels for Oil-fired EGUs
NEEDS modeled fuel
Distillate fuel oil
Residual fuel oil
Natural gas
Natural gas with mandatory curtailment
Number of units
19
23
242
18
Capacity (MW)
1,228
6,640
57,232
6,208
3A. 1.2 Pol icy Case
For the policy case, EPA considered three actions to comply with the MATS emission
limitations: (1) switching to natural gas where available, (2) switching to distillate fuel oil, and
(3) ESP particulate pollution control capable of 90% particulate removal efficiency. EPA
modeled the cost of actions 2 and 3 for each unit in the base case. EPA did not model the cost
of converting to natural gas because, for units with natural gas as a NEEDS modeled fuel, it was
the least cost fuel and therefore the base case fuel for the unit. The cost of switching a unit's
heat input to distillate fuel oil was based on the cost of converting operations, including tank,
line, and pump cleaning and burner atomizer assembly replacement, and the unit's 2015
projected heat input from residual fuel oil multiplied by the cost difference between residual
fuel oil and distillate fuel oil in the region where the unit is located. Conversion costs were
annualized using the methodology described in the IPM documentation (EPA, 2010b).
The cost of installing a flat plate-type ESP on oil-fired model units of various sizes was
calculated using the methodology outlined in EPA's Cost Manual (EPA, 2002) and adjusted to
2010 values using the Chemical Engineering Plant Cost Index (CEPCI). EPA developed non-linear
To ensure the analysis was not likely to underestimate compliance costs, EPA assumed units that do not include
natural gas as a NEEDS modeled fuel do not have access to a natural gas pipeline. The cost of obtaining pipeline
access for these units was assumed to be uneconomical and was not modeled in the analysis.
3A-3
-------
regression power functions similar to those used for costing air pollution controls in IPM. The
cost functions are shown in equations (3A.1)-(3A.3).
Capital costs = 243,494.4 X (MW capacity)0'7800 (3A.1)
Annual fixed costs = 13,883.4 X (MW capacity)0'7294 (3A.2)
Annual variable costs = 8,108.6 X (MWh generation)0'8632 (3A.3)
Capital costs were annualized using the capital cost recovery factor used in the IPM
documentation (EPA, 2010b). Annual variable costs were calculated using the predicted 2015
generation from residual fuel oil based on the unit's base case 2015 residual fuel oil heat input
and the unit's heat rate listed in NEEDS (EPA, 2010a).
3A. 1.3 Cost Sensitivities Related to Mandatory Natural Gas Curtailment
There are 18 dual fuel fired units (i.e., units capable of firing both gas and oil) that are
subject to mandatory natural gas curtailment. Of these units, six have existing ESP particulate
pollution controls installed. For the remaining 12 units, nine fired natural gas for more than 90
percent of their total heat input (see Table 3A-3). Because the MATS emission limits do not
apply to units that fire coal or oil for less than 10 percent of total heat input averaged over
three years or 15 percent in a single year, EPA analyzed historical oil-fired heat input between
2006 and 2010 at these units and found that four dual fuel fired units subject to mandatory
natural gas curtailment did not exceed 15 percent in any single year and averaged less than 10
percent across all three year periods between 2006 and 2010. EPA did not include the cost of
control on these units in the summary results. If these four units were to install ESPs, however,
the annual compliance cost of the MATS emission limits would increase $13 million (2007$).
As noted in 3A.1.1, natural gas supplies to the region are increasing and operating data
for dual fuel fired units subject to mandatory natural gas curtailment indicate that their
proportion of heat input from residual oil is declining. There are four units in addition to those
described in the paragraph above that exceeded 15 percent oil-fired heat input in 2006 and/or
2007, but between 2008 and 2010 did not exceed 15 percent oil-fired heat input in a single year
and averaged below 10 percent across all three years. These units were assigned ESP
particulate pollution controls in this analysis. However, if these four dual fuel fired units do not
install ESPs, the annual compliance cost of the MATS emission limits would decline $16 million
(2007$).
3A-4
-------
Table 3A-3. Percentage of Total Heat Input Derived from Oil for Oil-Fired Units Subject to
Mandatory Natural Gas Curtailment (2008-2010)
Percentage Number of units
< 1.0% 4
1.0% to 4.9% 1
5.0% to 9.9% 4
10.0% to 15.0% 3
3A.2 Results
For the purpose of estimating the impacts of the MATS emission limitations for oil-fired
units, EPA had to make assumptions about the compliance actions oil-fired units will take. Table
3A-4 lists those assumptions based on differences between the base and policy cases. EPA
assumed that the least cost compliance option for 12 residual fuel oil-fired units would be
converting to distillate fuel oil at an annual cost of approximately $12 million (2007$). An
additional 11 units would likely continue to burn residual fuel oil following the installation of an
ESP at a cost of approximately $44 million annually (2007$).
3A-5
-------
Table 3A-4. Costs to Achieve the MATS Emission Limitations for Oil-Fired Units
Unit
Cleary Flood, Units
Jefferies, Unit 1
Jefferies, Unit 2
McManus, Unit 1
McManus, Unit 2
Montville Station, Unit 6
Possum Point, Unit 5
Schuylkill Generating Station, Unit 1
Vienna Operations, Unit 8
William FWyman, Unit 1
William FWyman, Unit 2
Yorktown, Unit 3
Astoria Generating Station, Unit 30
Astoria Generating Station, Unit 40
Astoria Generating Station, Unit 50
BL England, Units
East River, Unit 60
East River, Unit 70
Herbert A Wagner, Unit 4
Middletown, Unit 4
Raven swood, Unit 10
Ravenswood, Unit 20
Ravenswood, Unit 30
Compliance action
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
Distillate fuel oil
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
Annual cost (2007$)
$ 308,000
$ 642,000
$ 673,000
$ 391,000
$ 512,000
$ 3,968,000
$ 119,000
$ 2,113,000
$ 1,741,000
$ 783,000
$ 646,000
$ 119,000
$ 4,214,000
$ 4,132,000
$ 4,202,000
$ 2,155,000
$ 1,844,000
$ 2,336,000
$ 4,352,000
$ 4,391,000
$ 3,904,000
$ 3,898,000
$ 8,322,000
3A-6
-------
3A.3 References
EPA. 2002. EPA Air Pollution Control Cost Manual. Sixth Edition. EPA/452/B-02-001. Available
at: http://www.epa.gOV/ttn/catc/dirl/c_allchs.pdf
EPA. 2010a. National Electric Energy Data System (NEEDS 4.10). Available at:
http://www.epa.gov/airmarkets/progsregs/epa-ipm/toxics.html
EPA. 2010b. Documentation for EPA Base Case v. 4.10. Chapter 8: Financial Assumptions.
Available at: http://www.epa.gov/airmarkets/progsregs/epa-
ipm/docs/v410/Chapter8.pdf
EPA. 2011. Data and Maps. Available at: http://epa.gov/camddataandmaps/
FERC. 2011. Major Pipeline Projects Pending (Onshore). Available at:
http://www.ferc.gov/industries/gas/indus-act/pipelines/pending-projects.asp
Neville, J. 2011. "Top Plant: Astoria II Combined Cycle Plant, Queens, New York," Power
Magazine. September.
3A-7
-------
CHAPTER 4
MERCURY AND OTHER HAP BENEFITS ANALYSIS
4.1 Introduction
This chapter provides an analysis of the benefits of the proposed Toxics Rule from
mercury and reductions of other HAP. Our efforts at quantifying the toxics benefits of this rule
focus on quantifying and estimating the welfare benefits of reducing mercury emissions
because mercury is the only HAP controlled by this rule for which there are sufficient available
analytic tools to conduct a national-scale benefits assessment.
This analysis of the benefits of reduced mercury exposure from EGUs as a result of the
rule is not changed from that provided for the proposed rule. It uses the same baseline and
control cases for mercury deposition as was used to estimate mercury benefits in the Mercury
and Air Toxics Rule proposal. EPA determined that it was reasonable to not update the mercury
benefits assessment for the final rule because of the small magnitude of the quantified mercury
benefits in the proposal, and the small difference (approximately 2 tons) in mercury emissions
reductions between the proposed and final rules. It is not expected that mercury benefits
would be substantially changed, and given the small magnitude of the benefits, any changes
would not meaningfully affect the overall benefits of the rule, nor impact the benefit-cost
comparison. An assessment of how forecast ECU mercury emissions changed between the
baseline used at proposal and the baseline used for the costs and co-benefits analysis, and
between the regulation as proposed and the regulation as finalized, is described in Appendix
5A.
This analysis builds on the methodologies developed previously for the 2005 Clean Air
Mercury Rule (CAMR). This is a national scale assessment which focuses on the exposures to
methylmercury in populations who consume self-caught freshwater fish (recreational fishers
and their families). While there are other routes of exposure, including self-caught saltwater
fish and commercially purchased fresh and saltwater fish, these exposures are not evaluated
because (1) for self-caught saltwater fish, we are unable to estimate the reduction in fish tissue
methylmercury that would be associated with reductions in mercury deposition from U.S.
EGUs, and (2) for commercially purchased ocean fish, it is nearly impossible to determine the
source of the methylmercury in those fish, and thus we could not attribute mercury levels to
U.S. EGUs.
This benefits analysis focuses on reductions in lost IQ points in the population, because
of the discrete nature of the effect, and because we are able to assign an economic value to IQ
4-1
-------
points. There are other neurological effects associated with exposures to methylmercury,
including impacts on motor skills and attention/behavior and therefore, risk estimates based on
IQ will not cover these additional endpoints and therefore could lead to an underestimate of
overall neurodevelopmental impacts. In addition, the NRC (2001) noted that "there remains
some uncertainty about the possibility of other health effects at low levels of exposure. In
particular, there are indications of immune and cardiovascular effects, as well as neurological
effects emerging later in life, that have not been adequately studied." These limitations suggest
that the benefits of mercury reductions are understated by our analysis; however, the
magnitude of the additional benefits is highly uncertain.
In Section 4.2, we discuss the potential health effects of mercury. Section 4.3 provides a
discussion of mercury in the environment, including potential impacts on wildlife. Section 4.4
describes the resulting change in mercury deposition from air quality modeling of the proposed
Toxics rule. Section 4.5 presents information on key data and assumptions used in conducting
the benefits analysis. Section 4.6 presents information on a dose-response function that relates
mercury consumption in women of childbearing with changes in IQ seen in children that were
exposed prenatally. IQ is used as a surrogate for the neurobehavioral endpoints that EPA relied
upon for setting the methylmercury reference dose (RfD). Section 4.7 presents exposure
modeling and benefit methodologies applied to a no-threshold model (i.e., a model that
assumes no threshold in effects at low doses of mercury exposure). Section 4.8 presents the
final benefits and risk estimates for recreational freshwater anglers and selected high-risk
subpopulations. Section 4.9 presents a qualitative description of the benefits from reductions in
HAPs other than mercury that will take place as a result of the Toxics Rule.
For this benefits assessment, EPA chose to focus on quantification of intelligence
quotient (IQ) decrements associated with prenatal mercury exposure as the endpoint for
quantification and valuation of mercury health benefits. Reasons for this focus on IQ included
the availability of thoroughly-reviewed, high-quality epidemiological studies assessing IQor
related cognitive outcomes suitable for IQ estimation, and the availability of well-established
methods and data for economic valuation of avoided IQ deficits, as applied in EPA's previous
benefits analyses for childhood lead exposure.
The quantitative estimates of human health benefits and risk levels provided in Section
4.2 is a national-scale assessment of economic benefits associated with avoided IQ loss due to
reduced methylmercury (MeHg) exposure among recreational freshwater anglers. Modeled risk
levels, in terms of IQ loss, for six high-risk subpopulations as a means of estimating potential
4-2
-------
disproportionate impacts on demographic groups with traditionally subsistence or near-
subsistence rates offish consumption are presented in Chapter 7 Section 7.11.
The first analysis (Section 4.2.1) estimates benefits from avoided IQ loss under various
regulatory scenarios for all recreational freshwater anglers in the 48 contiguous U.S. states. The
average effect on individual avoided IQ loss in 2016 is 0.00209 IQ points, with total nationwide
benefits estimated between $0.5 and $6.1 million.1 In contrast, the subpopulations analyses
(Section 7.12.2) focus on specific demographic groups with relatively high levels offish
consumption. For example, an African-American child in the Southeast born in 2016 to a
mother consuming fish at the 90th percentile of published subsistence-like levels is estimated
to experience a loss of 7.711 IQ points as a result of in-utero MeHg exposure from all sources in
the absence of a Toxics Rule.2The implementation of the Toxics Rule would reduce the
expected IQ loss for this child by an estimated 0.176 IQ points.
4.2 Impact of Mercury on Human Health
4.2.1 Introduction
Mercury is a persistent, bioaccumulative toxic metal that is emitted from power plants
in three forms: gaseous elemental Hg (Hg°), oxidized Hg compounds (Hg+2), and particle-bound
Hg (HgP). Elemental Hg does not quickly deposit or chemically react in the atmosphere,
resulting in residence times that are long enough to contribute to global scale deposition.
Oxidized Hg and HgP deposit quickly from the atmosphere impacting local and regional areas in
proximity to sources. 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.
1Monetized benefits estimates are for an immediate change in MeHg levels in fish. If a lag in the response of MeHg
levels in fish was accounted for, the monetized benefits could be significantly lower, depending on the length of
the lag and the discount rate used. As noted in the discussion of the Mercury Maps modeling, the relationship
between deposition and fish tissue MeHg is proportional in equilibrium, but the MMaps approach does not
provide any information on the time lag of response. Depending on the watershed studied, the lag time
between changes in mercury deposition and changes in the MeHg levels in fish has been shown to range from
XX
2We do note that overall confidence in IQ loss estimates above approximately 7 points decreases because we
begin to apply the underlying IQ loss function at exposure levels (ppm hair levels) above those reflected in
epidemiological studies used to derive those functions. The 39.1 ppm was the highest measured ppm level in
the Faroes Island study, while ~86 was the highest value in the New Zealand study (USEPA, 2005) (a 7 IQ points
loss is approximately associated with a 40 ppm hair level given the concentration-response function we are
using).
4-3
-------
Although Hg is toxic to humans when it is inhaled or ingested, we focus in this rulemaking on
exposure to MeHg through ingestion of fish, as it is the primary route for human exposures in
the U.S., and potential health risks do not likely result from Hg inhalation exposures associated
with Hg emissions from utilities.
In 2000, the National Research Council (NRC) of the NAS issued the NAS Study, which
provides a thorough review of the effects of MeHg on human health. There are numerous
studies that have been published more recently that report effects on neurologic and other
endpoints.
4.2.2 Neurologic Effects
In its review of the literature, the NAS found neurodevelopmental effects to be the most
sensitive and best documented endpoints and appropriate for establishing an RfD (NRC, 2000);
in particular NAS supported the use of results from neurobehavioral or neuropsychological
tests. The NAS report (NRC, 2000) noted that studies in animals reported sensory effects as well
as effects on brain development and memory functions and support the conclusions based on
epidemiology studies. The NAS noted that their recommended endpoints for an RfD are
associated with the ability of children to learn and to succeed in school. They concluded the
following: "The population at highest risk is the children of women who consumed large
amounts of fish and seafood during pregnancy. The committee concludes that the risk to that
population is likely to be sufficient to result in an increase in the number of children who have
to struggle to keep up in school."
4.2.3 Cardiovascular Impacts
The NAS summarized data on cardiovascular effects available up to 2000. Based on
these and other studies, the NRC (2000) concluded that "Although the data base is not as
extensive for cardiovascular effects as it is for other end points (i.e. neurologic effects) the
cardiovascular system appears to be a target for MeHg toxicity in humans and animals." The
NRC also stated that "additional studies are needed to better characterize the effect of
methylmercury exposure on blood pressure and cardiovascular function at various stages of
life."
Additional cardiovascular studies have been published since 2000. EPA did not to
develop a quantitative dose-response assessment for cardiovascular effects associated with
MeHg exposures, as EPA finds there is no consensus among scientists on the dose-response
4-4
-------
functions for these effects. In addition, there is inconsistency among available studies as to the
association between MeHg exposure and various cardiovascular system effects. The
pharmacokinetics of some of the exposure measures (such as toenail Hg levels) are not well
understood. The studies have not yet received the review and scrutiny of neurotoxicity studies.
4.2.4 Genotoxic Effects
The Mercury Study noted that MeHg is not a potent mutagen but is capable of causing
chromosomal damage in a number of experimental systems. The NAS concluded that evidence
that human exposure to MeHg caused genetic damage is inconclusive; they note that some
earlier studies showing chromosomal damage in lymphocytes may not have controlled
sufficiently for potential confounders. One study of adults living in the Tapajos River region in
Brazil (Amorim et al., 2000) reported a direct relationship between MeHg concentration in hair
and DNA damage in lymphocytes; as well as effects on chromosomes. Long-term MeHg
exposures in this population were believed to occur through consumption of fish, suggesting
that genotoxic effects (largely chromosomal aberrations) may result from dietary, chronic
MeHg exposures similar to and above those seen in the Faroes and Seychelles populations.
4.2.5 Immunotoxic Effects
Although exposure to some forms of Hg can result in a decrease in immune activity or
an autoimmune response (ATSDR, 1999), evidence for immunotoxic effects of MeHg is limited
(NRC, 2000).
4.2.6 Other Human Toxicity Data
Based on limited human and animal data, MeHg is classified as a "possible" human
carcinogen by the International Agency for Research on Cancer (IARC, 1994) and in IRIS (USEPA,
2002). The existing evidence supporting the possibility of carcinogenic effects in humans from
low-dose chronic exposures is tenuous. Multiple human epidemiological studies have found no
significant association between Hg exposure and overall cancer incidence, although a few
studies have shown an association between Hg exposure and specific types of cancer incidence
(e.g., acute leukemia and liver cancer) (NAS, 2000).
4-5
-------
4.3 Impact of Mercury on Ecosystems and Wildlife
4.3.1 Introduction
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
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
4-6
-------
population effects are difficult to make at this time. In the sections that follow numerous
effects have been identified at differing exposure levels.
4.3.2 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.
The Integrated Science Assessment for Oxides of Nitrogen and Sulfur—Ecological
Criteria (EPA, 2008) presents information regarding the possible complementary effects of
sulfur and mercury deposition. The ISA has concluded that there is a causal relationship
between sulfur deposition and increased mercury methylation in wetlands and aquatic
environments. This suggests that lowering the rate of sulfur deposition would also reduce
mercury methylation thus alleviating the effects of aquatic acidification as well as the effects of
mercury on fish.
4.3.3 Effects on Birds
In addition to effects on fish, mercury also affects avian species. In previous reports
(EPA, 1997 and EPA, 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, Burgess,
2008) and reduced survival (Mitro et al., 2008). Additionally Evers et al. (2008) report that they
believe that results from their study integrating the effects on the endpoints listed above and
evidence from other studies the weight of evidence indicates that population-level effects
negatively impacting population viability occur in parts of Maine and New Hampshire, and
potentially in broad areas of the loon's range.
4-7
-------
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
diverse 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 at chronic exposure to
levels of dietary MeHg commonly encountered by wild birds (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 resulted in a shortage of partners for females and the effect of homosexual
breeding were 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 experience
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.
Insectivorous birds have also been shown to suffer adverse effects due to current levels
of 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). While these effects were small and would
likely have little effect on population viability in such a short-lived species.
4.3.4 Effects on Mammals
In mammals adverse effects of methylmercury exposure have been observed in mink
and river otter, both fish eating species, collected in the wild in the northeast where
atmospheric deposition from municipal waste incinerators and electric utilities are the largest
sources (USEPA, 1999). For otter from Maine and Vermont maximum concentrations of Hg in
fur nearly equal or exceed a concentration associated with mortality. Concentrations of Hg in
liver for mink in Massachusetts/ Connecticut and the levels in fur from mink in Maine exceed
4-8
-------
concentrations associated with acute mortality (Yates, 2005). Adverse sub-lethal 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).
The studies cited here provide a glimpse of the scope of mercury effects on wildlife
particularly reproductive and survival effects at current exposure levels. 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) such as recreational fishing, bird watching and wildlife
viewing. EPA is not, however, currently able to quantify or monetize the benefits of reducing
mercury exposures affecting provision of ecosystem services adversely affected by mercury
depostion.
4.4 Mercury Risk and Exposure Analyses—Data Inputs and Assumptions
4.4.1 Introduction
This section provides information regarding key data inputs and assumptions used in
this assessment. The section begins with a description of the populations modeled in this
assessment, follows with information about the data used to estimate MeHg concentrations in
fish, and closes with a summary of the science and related assumptions used in this assessment
to link changes in modeled mercury deposition to changes in fish tissue concentrations.
4.4.2 Data Inputs
4.4.2.1 Populations Assessed For the National Aggregate Estimates of Exposed Populations in
Freshwater Fishing Households
The main source of data for identifying the size and location of the potentially exposed
populations is the Census 2000 data, summarized at the tract-level. There are roughly 64,500
tracts in the continental United States, with populations generally ranging between 1,500 and
8,000 inhabitants. For the national aggregate analysis of exposure levels, the specific
population of interest drawn from these data is the number of women aged 15 to 44 (i.e.,
childbearing age) in each tract. To predict populations in later years (2005 and 2016), we
applied county-level population growth projections for the corresponding population category
(Woods and Poole, 2008) to the 2000 tract-level data. To specifically estimate the portion of
4-9
-------
these populations that are pregnant in any given year, we applied state-level 2006 fertility rate
(live births per 1,000 women aged 15 to 44 years) data from U.S. Vital Statistics (DHHS, 2009).
Two main sources of national-level recreation activity data are available and suitable for
estimating the size and spatial distribution of freshwater recreational angler populations and
activities in the United States:
• the National Survey of Fishing, Hunting, and Wildlife-Associated Recreation
(FHWAR), maintained by the Department of the Interior (DOI) (DOI and DOC, 1992,
1997, 2002, 2007) and
• the National Survey of Recreation and the Environment (USDA, 1994).
FHWAR Angler Data. The FHWAR, conducted by the U.S. Census Bureau about every
5 years since 1955, includes data on the number and characteristics of participants as well as
time and money spent on hunting, fishing, and wildlife watching. The most recent survey and
report are for recreational activities conducted in 2006 (DOI and DOC, 2007). Data from this
report were used to provide the most recent estimate of the percentage of the resident
population in each state (16 years old or older) that engaged in freshwater fishing during the
year. As shown in Table 4-1, these percentages vary from 3% (New Jersey) to 27% (Minnesota).
The methodology for assessing mercury exposures also requires a further breakdown of
freshwater fishing activities into two categories: rivers (including rivers and streams) and lakes
(including lakes, ponds, reservoirs, and other flat water). Data at this level of detail are not
reported in the summary national reports for the FHWAR; however, they are available from the
FHWAR survey household-level data. For this analysis, data from a previous analysis and
summary of the 2001 FHWAR household-level survey data (EPA, 2005) were used to provide
estimates of the percentage of freshwater fishing days by residents in each state that were to
either the lake or river category3. As shown in Table 4-1, the highest percentage going to lakes
is in Minnesota (89%) and the highest to rivers is in Oregon (61%).
3Although the total number of fishing trips varies from year to year, there is little reason to expect that the ratio of
river trips to lake trips would have changed significantly since 2001. For this reason, despite information on the
type of waterbody visited being collected on the 2006 FWHAR survey, given resource and timetable limitations
we did not update this input to the analysis.
4-10
-------
Table 4-1. Summary of FWHAR State-Level Recreational Fishing Characteristics
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
Freshwater Anglers as
Percentage of State Population3
15.7%
7.0%
19.5%
4.1%
13.2%
6.4%
5.0%
7.9%
12.6%
18.4%
7.3%
12.3%
16.8%
14.8%
17.5%
14.2%
19.4%
5.5%
5.1%
14.2%
26.9%
19.6%
18.9%
22.8%
12.3%
5.9%
8.9%
3.1%
Percentage of Freshwater
Lakes
59.9%
79.2%
81.1%
53.5%
63.7%
58.7%
52.8%
67.4%
70.4%
44.4%
76.4%
77.8%
55.1%
84.7%
80.0%
71.2%
73.7%
40.7%
75.5%
85.6%
89.0%
79.0%
80.2%
46.8%
80.6%
80.5%
67.9%
68.9%
Fishing Tripsb
Rivers
40.1%
20.8%
18.9%
46.5%
36.3%
41.3%
47.2%
32.6%
29.6%
55.6%
23.6%
22.2%
44.9%
15.3%
20.0%
28.8%
26.3%
59.3%
24.5%
14.4%
11.0%
21.0%
19.8%
53.2%
19.4%
19.5%
32.1%
31.1%
(continued)
4-11
-------
Table 4-1. Summary of FWHAR State-Level Recreational Fishing Characteristics (continued)
State
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Freshwater Anglers as
Percentage of State Population3
10.9%
4.7%
10.7%
17.3%
11.8%
18.8%
13.6%
8.1%
4.4%
14.2%
14.6%
13.8%
9.7%
15.6%
12.6%
7.5%
9.5%
19.7%
22.8%
23.5%
Percentage of Freshwater
Lakes
56.1%
67.2%
68.7%
87.2%
78.8%
83.1%
39.0%
44.0%
73.5%
75.6%
69.7%
68.6%
79.3%
68.0%
71.1%
70.4%
50.0%
50.1%
79.5%
64.0%
Fishing Tripsb
Rivers
43.9%
32.8%
31.3%
12.8%
21.2%
16.9%
61.0%
56.0%
26.5%
24.4%
30.3%
31.4%
20.7%
32.0%
28.9%
29.6%
50.0%
49.9%
20.5%
36.0%
a Based on FHWAR 2006 data for residents 16 years and older.
b Based on FHWAR 2001 data for residents 16 years and older.
NSRE Angler Data. The NSRE, formerly known as the National Recreation Survey (NRS),
is a nationally administered survey, which has been conducted periodically since 1962. It is
designed to assess outdoor recreation participation in the United States and elicit information
regarding people's opinions about their natural environment. The NSRE sample of freshwater
anglers is smaller than the FHWAR sample, but it is nonetheless a useful resource because it
provides a wide variety of information about fishing activities. Importantly, it includes relatively
detailed information about the nature and location of recent freshwater trips. Because the
4-12
-------
sampling procedure is designed to be representative, inferences may be drawn about the
relative popularity of particular types of freshwater bodies (e.g., lakes, rivers) among the
general public and the average distance traveled to reach these sites. Although more recent
NSRE surveys have been conducted in 2000 and 2009, data from 1994 survey (NSRE, 1994) is
used for this analysis because it contains the most detailed information regarding fishing trip
destinations.
The NSRE 1994 elicited information from respondents about the most recent fishing trip.
One of the main advantages of NSRE 1994 is that it includes geocoded data for reported fishing
destinations. To specify the location of the last fishing trip, respondents were asked to provide
the name of the waterbody, the nearest town to the waterbody, and an estimate of the
distance and direction from their home to the waterbody. Appendix B describes how these data
were used in this analysis to estimate the percentage of freshwater fishing trips that were in
different distance intervals from respondents' homes. Using the demographic data from the
NSRE, these estimates were further differentiated according to the income level and urban
versus nonurban location of the respondents.
4.4.3 Mercury Concentrations in Freshwater Fish
4.4.3.1 Data Sources for Fish Tissue Concentrations
To characterize the spatial distribution of mercury concentration estimates in
freshwater fish across the country, we compiled data from three main sources, which are
described below.
National Listing of Fish Advisory (NLFA) database. The NLFA, managed by EPA
(http://water.epa.gov/scitech/swguidance/fishshellfish/fishadvisories/), collects and compiles
fish tissue sample data from all 50 states and from tribes across the United States. In particular,
it contains data for over 43,000 mercury fish tissue samples collected from 1995 to 2007.
U.S. Geologic Survey (USGS) compilation of mercury datasets. As part of its
Environmental Mercury Mapping and Analysis (EMMA) program, USGS compiled mercury fish
tissue sample data from a wide variety of sources (including the NLFA) and has posted these
data at http://emmma.usgs.gov/datasets.aspx. The compilation includes (1) state-agency
collected and reported data (including Delaware, Iowa, Indiana, Louisiana, Minnesota, Ohio,
South Carolina, Virginia, Wisconsin, and West Virginia) from over 40,000 fish tissue samples,
covering the period 1995 to 2007 and (2) over 10,000 fish tissue samples from several other
sources, including the National Fish Tissue Survey, the National Pesticide Monitoring Program
4-13
-------
(NPMP), the National Contaminant Biomonitoring Program (NCBP), the Biomonitoring of
Environmental Status and Trends (BEST) datasets of the USFWS and USGS
(http://www.cerc.cr.usgs.gov/data/data.htm), and the Environmental Monitoring and Analysis
Program (EMAP) (http://www.epa.gov/emap/).
EPA's National River and Stream Assessment (NRSA) study data. These data include
nearly 600 fish tissue mercury samples collected at randomly selected freshwater sites across
the United States during the period 2008 to 2009.
4.4.3.2 Approach for Compiling Fish Tissue Datasetfor Exposure Analysis
Data from these three datasets were combined into a single master fish tissue dataset
covering the period 1995 to 2009. One problem encountered in combining these datasets is the
potential duplication of samples in the NLFA and USGS state-collected data. Unfortunately,
these two datasets do not contain directly comparable and unique identifiers that allow
duplicate samples to be easily identified and removed. Therefore, as an alternative, the
samples from these two datasets were subdivided into data groups according to the year and
state in which they were collected. If both datasets contained a data group for the same year
and the same state, then the data group with the fewer number of observations was excluded
from the master data.
The following criteria were also applied to exclude data from the master fish tissue
dataset to be used in the analysis. Samples were excluded if they:
• did not include useable latitude-longitude coordinates for spatial identification;
• were located at sites outside the tidal boundaries of the continental United States
(i.e., if they were not sampled from freshwater sites);
• did not come from fish species found in freshwater; or
• did not come from sampled fish that were at least 7 inches in length (i.e., unlikely to
be consumed).
Each remaining sample was then categorized as either a river or lake sample based on
information about the sampling site location. First, specific character strings in the site names
(e.g., "river," "creek," "lake," "pond," and "reservoir") were used to classify sites. Second,
remaining sites were categorized based on a CIS analysis that linked the sites' latitude-
longitude coordinates to the nearest waterbody and its category.
4-14
-------
• The resulting master fish tissue mercury concentration dataset contains 26,940
sample concentration estimates from 3,876 river sites and 23,206 estimates from
2,167 lake sites.
• A new dataset was then created by spatially grouping and averaging the river and
lake concentration estimates at the HUC-12 sub-watershed level. First, all of the
mercury sampling sites included in the master data were mapped and matched to
the HUC-12 sub-watersheds in which they are located. A total of 3,884 HUC-12s in
the continental United States (4.6%) contain at least one river or lake mercury
sample.4 Second, site-specific average mercury concentration values were generated
by computing the mean concentration estimate at each site. Third, HUC-level
average lake concentration estimates were computed as the mean of the site-
specific average lake concentration estimates for each HUC containing at least one
lake sampling site (1,396 HUCs). Fourth, HUC-level average river concentration
estimates were computed as the mean of the site-specific average river
concentration estimates for each HUC containing at least one river sampling site
(2,655 HUCs).
4.4.3.3 Summary of Fish Tissue Mercury Concentration Estimates Used in the Exposure Analysis
The resulting HUC-level mercury concentration dataset is summarized in Table 4-2. The
average HUC-level mercury concentration estimate for lakes is 0.29 ppm and for rivers is 0.26
ppm. The large standard deviations and ranges reported in the table also reflect the
considerable spatial variation in lake and river concentration estimates across samples. As
described below, the analysis uses this inter-watershed spatial variation (rather than just the
average point estimate across watersheds) to estimate mercury exposures However, in this
analysis, exposure estimates were only generated for populations linked to these HUCs
containing at least one river or lake mercury fish tissue sample.
4This number excludes 15 HUC-12s containing mercury samples. These HUC-12s were excluded from the analysis
due to their proximity to potentially significant non-air sources of mercury, including gold mines or non-EGU
mercury sources included in the 2008 Toxic Release Inventory.
4-15
-------
Table 4-2. Summary of HUC-level Average Mercury Fish Tissue Concentration Estimates
Lake Fish Tissue Concentrations
HUC-level average mercury concentration (ppm)
Number of lake samples per HUC
Number of lake sampling sites per HUC
River Fish Tissue Concentrations
HUC-level average mercury concentration (ppm)
Number of river samples per HUC
Number of river sampling sites per HUC
Na
1,396
1,396
1,396
2,655
2,655
2,655
Mean
0.286
16.62
1.55
0.261
10.15
1.46
Std. Dev.
0.231
31.61
1.97
0.259
22.45
1.10
Min
0.000
1
1
0.006
1
1
Max
3.56
458
33
4.97
288
16
Number of HUC-12s with at least one river or lake sampling site.
4.5 Linking Changes in Modeled Mercury Deposition to Changes in Fish Tissue
Concentrations
4.5.1 Introduction
In the United States, humans are exposed to MeHg mainly by consuming fish that
contain MeHg. Accordingly, to estimate changes in human exposure EPA must analyze how
changes in Hg deposition from U.S. coal-fired power plants translate into changes in MeHg
concentrations in fish. Quantifying the linkage between different levels of Hg deposition and
fish tissue MeHg concentration is an important step in the risk assessment process and the
focus of the material described in this section.
To effectively estimate fish MeHg concentrations in a given ecosystem, it is important to
understand that the behavior of Hg in aquatic ecosystems is a complex function of the
chemistry, biology, and physical dynamics of different ecosystems. The majority (95 to 97
percent) of the Hg that enters lakes, rivers, and estuaries from direct atmospheric deposition is
in the inorganic form (Lin and Pehkonen, 1999). Microbes convert a small fraction of the pool of
inorganic Hg in the water and sediments of these ecosystems into the organic form of Hg
(MeHg). MeHg is the only form of Hg that biomagnifies in organisms (Bloom, 1992). Ecosystem-
specific factors that affect both the bioavailability of inorganic Hg to methylating microbes (e.g.,
sulfide, dissolved organic carbon) and the activity of the microbes themselves (e.g.,
temperature, organic carbon, redox status) determine the rate of MeHg production and
subsequent accumulation in fish (Benoit et al., 2003). The extent of MeHg bioaccumulation is
also affected by the number of trophic levels in the food web (e.g., piscivorous fish populations)
4-16
-------
because MeHg biomagnifies as large piscivorous fish eat smaller organisms (Watras and Bloom,
1992; Wren and MacCrimmon, 1986). These and other factors can result in considerable
variability in fish MeHg levels among ecosystems at the regional and local scale.
4.5.2 Use of Mercury Maps to Project Changes in Fish Tissue Concentrations
To analyze the relationship between Hg deposition and MeHg concentrations in fish in
freshwater aquatic ecosystems across the U.S. for the national scale benefits assessment, EPA
applied EPA's Office of Water's Mercury Maps (MMaps) approach (US EPA, 2001a). MMaps
implements a simplified form of the IEM-2M model applied in EPA's Mercury Study Report to
Congress (USEPA, 1997). By simplifying the assumptions inherent in the freshwater ecosystem
models that were described in the Report to Congress, the MMaps model showed that these
models converge at a steady-state solution for MeHg concentrations in fish that are
proportional to changes in Hg inputs from atmospheric deposition (i.e., over the long term, fish
concentrations are expected to decline proportionally to declines in atmospheric loading to a
waterbody).
MMaps has several limitations:
1. The MMaps approach is based on the assumption of a linear, steady-state
relationship between concentrations of MeHg in fish and present day air deposition
mercury inputs. We expect that this condition will likely not be met in many
waterbodies because of recent changes in mercury inputs and other environmental
variables that affect mercury bioaccumulation. For example, the US has recently
reduced human-caused emissions while international emissions have increased.
2. The requirement that environmental conditions remain constant over the time
required to reach steady state inherent in the MMaps methodology may not be met,
particularly in systems that respond slowly to changes in mercury inputs.
3. Many water bodies, particularly in areas of historic gold and mercury mining, contain
significant non-air sources of mercury. The MMaps methodology will yield biased
results when applied to such waterbodies. As a simple illustrative example, if we
have mercury deposition of 100 at a given location and a MeHg fish concentration of
6 in a local fish tissue sample, and a new emissions rule reduces deposition by half to
50, then, in the absence of other non-air deposition sources, we would assume that
the MeHg fish concentration is reduced by the same proportion, to 3 ((50 / 100) x 6).
However, if total pre-control mercury loading to the system is actually 100 plus
another unaccounted for source (for example, an additional 100 due to area gold
mining), then the MeHg fish concentration of 6 is actually due to 200 in total
mercury loading. In this case, reducing mercury air deposition from 100 to 50 would
only reduce the total loading by 25%, to 150, which, based on the MMaps
4-17
-------
methodology, would result in a MeHg fish concentration of 4.5 ((150 / 200) x 6)
rather than 3. In areas where non-air sources of mercury load are unaccounted for,
MMaps-based estimates of changes in MeHg fish tissue concentrations due to
reduced mercury air emissions would therefore be biased high.
4. Finally, MMaps does not account for a calculation of the time lag between a
reduction in mercury deposition and a reduction in the MeHg concentrations in fish
and, as noted earlier, depending on the nature of the watersheds and waterbodies
involved, the temporal response time for fish tissue MeHg levels following a change
in mercury deposition can range from years to decades depending on the attributes
of the watershed and waterbody involved5 Research has suggested that fish tissue
MeHg levels in some locations may display a multi-phase response following a
discrete change in mercury deposition, with the first phase lasting a few years to a
decade or more and primarily involving changes in aerial loading directly to the
waterbody and the second phase lasting decade (to a century or more) and
reflecting longer-term changes in watershed erosion and runoff to the waterbody
(Knights et al., 2009, Harris et al., 2007).
This methodology therefore applies only to situations where air deposition is the sole
significant source of Hg to a water body, and where the physical, chemical, and biological
characteristics of the ecosystem remain constant over time. EPA recognizes that concentrations
of MeHg in fish across all ecosystems may not reach steady state and that ecosystem conditions
affecting mercury dynamics are unlikely to remain constant over time. EPA further recognizes
that many water bodies, particularly in areas of historic gold and Hg mining in western states,
contain significant non-air sources of Hg. Finally, EPA recognizes that MMaps does not account
for the time lag between a reduction in Hg deposition and a reduction in the MeHg
concentrations in fish. While acknowledging these limitations, EPA is unaware of any other tool
for performing a national-scale assessment of the change in fish MeHg concentrations resulting
from reductions in atmospheric deposition of Hg. The following paragraphs provide additional
details on the above limitations, as well as a brief assessment of the degree to which conditions
match those assumptions.
The MMaps model represents a reduced form of the IEM-2M and MCM models used in
the Mercury Study Report to Congress (USEPA, 1997), as well as the subsequent Dynamic MCM
(D-MCM) model (Harris et al., 1996). That is, the equations of these mercury fate and transport
5As noted in footnote 1 of this chapter,monetized benefits estimates are for an immediate change in MeHg levels
in fish (i.e., the potential lag period associated with fully realizing fish tissue MeHg levels was not reflected in
benefits modeling). If a lag in the response of MeHg levels in fish were assumed, the monetized benefits could
be significantly lower, depending on the length of the lag and the discount rate used. MMaps approach does
not account for the time lag of response.
4-18
-------
models are reduced to steady state and consolidated into a single equilibrium equation
equating the ratio of future/current air deposition rates to future/current fish tissue
concentrations.
Though plainly stated, the steady-state assumption is a compilation of a number of
individual conditions. For example, fish tissue data may not represent average, steady-state
concentrations for two major reasons:
• Fish tissue and deposition rate data for the base period are not at steady state.
Where deposition rates have recently changed, the watershed or waterbody may
not have had sufficient time to fully respond. The pool of mercury in different media
could be sufficiently large relative to release rates, and thus needs more time to
achieve a new equilibrium. This is more likely to occur in deeper lakes and lakes with
large catchments where turnover rates are longer and where the watershed
provides significant inputs of mercury.
• Fish tissue data do not represent average conditions (or conditions of interest for
forecast fish levels). Methylation and bioaccumulation are variable and dynamic
processes. If fish are sampled during a period of high or low methylation or
bioaccumulation, they would not be representative of the average, steady-state or
dynamic equilibrium conditions of the waterbody. This effect is significantly more
pronounced in small and juvenile fish. Examples include tissue data collected during
a drought or during conditions offish starvation. Other examples include areas in
which seasonal fluctuations in fish mercury levels are significant due, for example,
from seasonal runoff of contaminated soils from abandoned gold and mercury
mines or areas geologically rich in mercury. In such a case, MMaps predictions would
be valid for similar conditions (e.g. wet year/dry year, or season) in the future, rather
than typical or average conditions. Alternatively, sufficient fish tissue would need to
be collected to get an average concentration that represents a baseline dynamic
equilibrium.
Other ecosystem conditions might cause projections from the MMaps approach to be
inaccurate for a particular ecosystem. Watershed and waterbody conditions can undergo
significant changes in capacity to transport, methylate, and bioaccumulate mercury. Examples
of this include regions where sulfate and/or acid deposition rates are changing (in turn affecting
MeHg production independently of total mercury loading), and where the trophic status of a
waterbody is changing. A number of other water quality parameters have been correlated with
increased fish tissue concentrations (e.g. low pH, high DOC, lower algal concentrations), but
these relationships are highly variable among different waterbodies. MMaps will be biased
when waterbody characteristics change between when fish were initially sampled, and the new
conditions of the waterbody.
4-19
-------
As stated above, the relationship between the change in mercury deposition from air to
the change in fish tissue concentration holds only when air deposition is the predominant
source of the mercury load to a waterbody. Due to this requirement in the model, the national
application of the MMaps approach screened out those watersheds that either contained
active gold mines or had other substantial non-US ECU anthropogenic releases of mercury.
Identification of watersheds with gold mines was based on a 2005 USGS data set characterizing
mineral and metal operations in the United States. The data represent commodities monitored
by the National Minerals Information Center of the USGS, and the operations included are
those considered active in 2003 (online link: http://tin.er.usgs.gov/mineplant/). EPA considered
the 25th percentile US-EGU emission level to be a reasonable screen for additional substantial
non-US ECU releases to a given watershed. The identification of watersheds with substantial
non-EGU anthropogenic loadings was based on a TRI-net query for 2008 of non-EGU mercury
sources with total annual on-site Hg loading (all media) of 39.7 pounds or more. This threshold
value corresponds to the 25th percentile annual US-EGU mercury emission value as
characterized in the 2005 NATA. It should be noted that MMaps was designed to address an
important, but very specific issue—that of eventual response of fish tissue to air deposition
reductions. As such it responds to a need to understand how mercury reductions, independent
of other changes in the environment, will impact fish contamination and human health. More
complex models are required in cases where more complete descriptions are needed. A
dynamic model is essential for modeling waterbody recovery during the period in which
waterbody response lags reductions in mercury loads. A dynamic model is also essential for
understanding seasonal fluctuations, as well as year-to-year fluctuations due to meteorological
variability. Finally, a more complex model would be essential for assessing the impact of other
watershed and water quality changes (e.g., erosion, wetlands coverage, and acid deposition)
that might affect mercury bioaccumulation in fish. These complex models are used to derive
the MMaps approach, and are themselves based on a number of assumptions. While these
assumptions are considered reasonable given the state of the science of environmental
modeling and mercury in the environment, the validity of assumptions inherent in both the
MMaps approach and dynamic ecosystem scale models will need to be reevaluated as the
science of mercury fate and transport evolves.
The MMaps methodology was peer reviewed by a set of national experts in the fate and
transport of mercury in watersheds (US EPA, 2001a). While two reviewers felt it could be used
to predict future fish tissue concentrations, a third cautioned it should not be considered a
robust predictor until scientific data can be generated to validate the approach. Reviewers
systematically identified a set of implicit assumptions that compose the steady state
4-20
-------
assumption in the MMaps approach. They pointed out that due to evolving and complex nature
of the science of mercury, some features of the complex models are assumptions themselves,
and thus cannot be wholly relied upon as ultimate predictors of mercury fate and transport.
The reviewers pointed out that there is limited scientific information to directly verify this
approach, and that some scientific data appears to refute individual components of the overall
steady state assumption. One reviewer did perform a D-MCM and MMaps comparison, and
found that, under these assumptions, MMaps model did produce comparable steady-state
results as the D-MCM model. There was considerable discussion about how best to aggregate
the data, to scale up to a deposition reduction requirement, from fish-specific and waterbody
specific information. The description of the approach and the methodologies as applied in this
analysis are largely consistent with the peer review recommendations.
The MMaps report (US EPA, 2001a) presented a national-scale application of Mercury
Maps to determine the percent reductions in air deposition that would be needed in
watersheds across the country for average fish tissue concentrations to achieve the national
MeHg criterion. In this national-scale assessment, fish tissue concentrations were aggregated at
the scale of large watersheds, thus presenting average results for each watershed. The use of
other scales of aggregation, e.g., waterbody specific, is consistent with the MMaps approach to
the degree to which different mercury loads can be discerned.
4.5.3 The Science of Mercury Processes and Variability in Aquatic Ecosystems
The set of physical, chemical, and biological processes controlling mercury fate in
watersheds and water bodies can be grouped into specific categories: mercury cycle chemistry;
mercury processes in the atmosphere, soils and water; bioavailability of mercury in water; and
mercury accumulation in the food web. The following is a review of these categories, discussing
the related scientific developments that have added to our understanding of mercury
processes. This review builds upon the work previously summarized in EPA's Mercury Report to
Congress (USEPA, 1997).
4.5.3.1 Mercury Cycle Chemistry
Mercury occurs naturally in the environment as several different chemical species. The
majority of mercury in the atmosphere (95-97%) is present in a neutral, elemental state (Hg°)
(Lin and Pehkonen, 1999), while in water, sediments and soils the majority of mercury is found
in the oxidized, divalent state (Hg(ll)) (Morel et al., 1998). A small fraction (percent) of this pool
of divalent mercury is transformed by microbes into MeHg (CH3Hg(ll)/ MeHg) (Jackson, 1998).
MeHg is retained in fish tissue and is the only form of mercury that biomagnifies in aquatic food
4-21
-------
webs (Kidd et al., 1995). As a result, MeHg concentrations in higher trophic level organisms
such as piscivorous fish, birds and wildlife are often 104-106 times higher than aqueous MeHg
concentrations (Jackson, 1998). Transformations among mercury species within and between
environmental media result in a complicated chemical cycle. Mercury emissions from both
natural and anthropogenic sources are predominantly as Hg(ll) species and Hg° (Landis and
Keeler, 2002; Seigneur et al., 2004). Anthropogenic point sources of mercury consist of
combustion (e.g., utility boilers, municipal waste combustors, commercial/industrial boilers,
medical waste incinerators) and manufacturing sources (e.g., chlor-alkali, cement, pulp and
paper manufacturing) (USEPA, 1997). Natural sources of mercury arise from geothermic
emissions such as crustal degassing in the deep ocean and volcanoes as well as dissolution of
mercury from geologic sources (Rasmussen, 1994).
4.5.3.2 Mercury Processes in the Atmosphere
The relative contributions of local, regional and long range sources of mercury to fish
mercury levels in a given water body are strongly affected by the speciation of natural and
anthropogenic emissions sources. Elemental mercury is oxidized in the atmosphere to form the
more soluble mercuric ion (Hg(ll)) (Schroeder et al., 1989). Particulate and reactive gaseous
phases of Hg(ll) are the principle forms of mercury deposited onto terrestrial and aquatic
systems because they are more efficiently scavenged from the atmosphere through wet and
dry deposition than HgO (Lindberg and Stratton, 1998). Because Hg(ll) species or reactive
gaseous mercury (RGM) and particulate mercury (Hg(p)) in the atmosphere tend to be
deposited more locally than Hg°, differences in the species of mercury emitted affect whether it
is deposited locally or travels longer distances in the atmosphere (Landis et al., 2004).
4.5.3.3 Mercury Processes in Soils
A portion of the mercury deposited in terrestrial systems is re-emitted to the
atmosphere. On soil surfaces, sunlight may reduce deposited Hg(ll) to Hg°, which may then
evade back to the atmosphere (Carpi and Lindberg, 1997; Frescholtz and Gustin, 2004; Scholtz
et al., 2003). Significant amounts of mercury can be co-deposited to soil surfaces in throughfall
and litterfall of forested ecosystems (St. Louis et al., 2001), and exchange of gaseous Hg° by
vegetation has been observed (e.g., (Gustin et al., 2004).
Hg(ll) has a strong affinity for organic compounds such that inorganic Hg in soils and
wetlands is predominantly bound to dissolved organic matter (Mierle and Ingram, 1991). MeHg
likewise forms stable complexes with solid and dissolved organic matter (Hintelmann and
Evans, 1997). These complexes can dominate MeHg speciation under aerobic conditions
4-22
-------
(Karlsson and Skyllberg, 2003). Truly dissolved and dissolved organic carbon (DOC)-complexed
Hg(ll) and MeHg are transported by percolation to shallow groundwater, and by runoff to
adjacent surface waters (Ravichandran, 2004). Sorbed Hg(ll) and MeHg are transported by
erosion fluxes to depositional areas on the watershed and to adjacent surface waters (e.g.,
(Hurley etal., 1998).
Concentrations of MeHg in soils are generally very low. In contrast, wetlands are areas
of enhanced MeHg production and account for a significant fraction of the external MeHg
inputs to surface waters that have watersheds with a large portion of wetland coverage (e.g.,
St. Louis et al., 2001). Accordingly, there is a positive relationship between MeHg yield and
percent wetland coverage (Hurley et al., 1995). Hydrology exerts an important control on the
magnitude and flux of MeHg in wetland ecosystems (Branfireun and Roulet, 2002), as well as
the transport of inorganic mercury deposited in a given watershed to surface waters (Babiarz
etal., 2001).
4.5.3.4 Mercury Processes in Water
In a water body, deposited Hg(ll) is reduced to Hg° by ultraviolet and visible wavelengths
of sunlight as well as microbially mediated reduction pathways (Amyot et al., 2000; Mason
et al., 1995). In turn, Hg° is oxidized back to Hg(ll), driven by sunlight as well as by "dark"
chemical or biochemical processes (Lalonde et al., 2001; Zhang and Lindberg, 2001). Driven by
wind and water currents, dissolved Hg° in the water column is volatilized, which can be a
significant removal mechanism for mercury in surface waters and a net source of mercury to
the atmosphere (Siciliano et al., 2002).
In the water column and sediments, Hg(ll) partitions strongly to silts and biotic solids,
sorbs weakly to sands, and complexes strongly with dissolved and particulate organic material.
The abundance of various inorganic ligands (e.g., OH", CI", S2-, DOC) in freshwater and saltwater
ecosystems plays an important role in both oxidation and reduction of inorganic mercury as
well as its bioavailability to methylating microbes. For example, reduction of Hg(ll) is
hypothesized to be a function of the predominance of Hg(OH)2, which is inversely correlated
with pH (Mason et al., 1995). Reduction of Hg(ll) to Hg° and subsequent volatilization from the
water column is important because it effectively reduces the pool of inorganic mercury that
could potentially undergo conversion to MeHg.
Hg(ll) and MeHg sorbed to solids settle out of the water column and accumulate on the
surface of the benthic sediment layer. Surficial sediments interact with the water column via
resuspension and bioturbation. The burial of sediments below the surficial zone can be a
4-23
-------
significant removal mechanism for contaminants in surface sediments (e.g., Gobas et al., 1998;
Gobas et al., 1995). The depth of the active sediment layer is a highly sensitive parameter for
predicting the temporal response of different ecosystems to changes in mercury loading in
environmental fate models. This is because the reservoir of Hg(ll) potentially available for
conversion to MeHg in the sediments is a function of the depth and volume of the active
sediment layer. The compartment conducive for methylation is similarly affected (Harris and
Hutchison, 2003; Sunderland et al., 2004). Physical characteristics of different ecosystem types
affect estuarine mixing and sediment resuspension, which also affect the production of MeHg
in the water and sediments (Rolfhus et al., 2003; Sunderland et al., 2004; Tseng et al., 2001).
4.5.3.5 Bioavailability of Inorganic Mercury to Methylating Microbes
The amount of bioavailable MeHg in water and sediments of aquatic systems is a
function of the relative rates of mercury methylation and demethylation. In the water, MeHg is
degraded by two microbial processes and sunlight (Barkay et al., 2003; Sellers et al., 1996).
Recent research has shown that demethylating Hg-resistant bacteria may adapt to systems that
are highly contaminated with total mercury, helping to explain the paradox of low MeHg and
fish Hg levels in these systems (Schaefer et al., 2004).
Mass balances for a variety of lakes and coastal ecosystems show that in situ production
of MeHg is often one of the main sources of MeHg in the water and sediments (Benoit et al.,
1998; Bigham and Vandal, 1994; Gbundgo-Tugbawa and Driscoll, 1998; Gilmour et al., 1998;
Mason et al., 1999). Sulfate-reducing bacteria (SRB) are thought to be the principle agents
responsible for the majority of MeHg production in aquatic systems (Beyers et al., 1999;
Compeau and Bartha, 1987; Gilmour and Henry, 1991). SRB thrive in the redoxocline, where the
maximum gradient between oxic and anoxic conditions exists (Hintelmann et al., 2000). Thus, in
addition to the presence of bioavailable Hg(ll), MeHg production and accumulation in aquatic
systems is a function of the geochemical parameters that enhance or inhibit the activity of
methylating microbes, especially sulfur concentrations, redox potential (Eh) and the
composition and availability of organic carbon.
A number of factors affect the bioavailabilty of Hg(ll). A strong inverse relationship
between complexation of Hg(ll) by sulfides and MeHg production has been demonstrated in a
number of studies (Benoit et al., 1999a; Benoit et al., 1999b; Craig and Bartlett, 1978; Craig and
Moreton, 1986). Passive diffusion of dissolved, neutral inorganic mercury species is
hypothesized as one of the main modes of entry across the cell membranes of methylating
microbes (Benoit et al., 1999a; Benoit et al., 2003; Benoit et al., 1999b). Thus, the formation of
4-24
-------
neutral, dissolved mercury species such as HgCI2, Hg(OH)2, HgCIOH, and HgS°(aq.), which
depend on the availability of constituent ligands in the surface and interstitial waters, may
strongly influence the availability of inorganic mercury to SRB, although our understanding of
the forms of mercury that are bioavailable to methylating microbes is currently incomplete
(Benoit et al., 2001; Benoit et al., 1999a; King et al., 2001). Additional detail is provided below
on the relationship between sulfur deposition and mercury methylation.
Changes in the bioavailability of inorganic mercury and the activity of methylating
microbes as a function of sulfur, carbon and ecosystem specific characteristics mean that
ecosystem changes and anthropogenic "stresses" that do not result in a direct increase in
mercury loading to the ecosystem but alter the rate of MeHg formation may also affect
mercury levels in organisms (Grieb et al., 1990). Because mercury concentrations in fish can
increase even when there has been no change in the total amount of mercury deposited in the
ecosystem, environmental changes such as eutrophication, which may alter microbial activity
and the chemical dynamics of mercury within an ecosystem, must be considered together with
emission control strategies to effectively manage mercury accumulation in the food web.
Recent research indicates that the bioavailability or reactivity of newly deposited Hg(ll)
may be greater than older "legacy" mercury in the system (Hintelmann et al., 2002). These
results suggest that lakes receiving the bulk of their mercury directly from deposition to the
lake surface (e.g., some seepage lakes) would see fish mercury concentrations respond more
rapidly to changes in atmospheric deposition than lakes receiving most of their mercury from
watershed runoff. The implications of these data are also that systems with a greater surface
area to watershed area ratio that receive most of their inputs directly from the atmosphere
(e.g., seepage lakes) may respond more rapidly to changes in emissions and deposition of
mercury than those receiving significant inputs of mercury from the catchment area.
Sulfur and Mercury Methylation. EPA's 2008 Integrated Science Assessment (ISA) for
Oxides of Nitrogen and Sulfur-Ecological Criteria (Final Report) concluded that evidence is
sufficient to infer a casual relationship between sulfur deposition and increased mercury
methylation in wetlands and aquatic environments. 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 4-1). Therefore, the correlation between S042" deposition
4-25
-------
Spatial Factors
Lake/Reservoir
Sediment Disturbance
Upstream Wetlands
Upstream Forested Land
Upstream Erosion
Upstream Urban Land
Mercury
+
SRB
+
Sulfate
s
z
Biogeochemical Factors
Organic Matter
Sulfide
Salinity
Anoxia
Temperature
Methylmercury
Does not promote methylalion
Promotes methylalion
Figure 4-1. Spatial and Biogeochemical Factors Influencing MeHg Production
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 (EPA, 2008, Sections 4.4.1 and 4.5).
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 4-2 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
4-3 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).
4-26
-------
Figure 4-2. Preliminary USGS Map of Mercury Methylation-Sensitive Watersheds Derived
from More Than 55,000 Water Quality Sites aqnd 2,500 Watersheds (Myers et al., 2007)
Interdependent biogeochemical factors preclude the existence of simple sulfate-related
mercury methylation models (see Figure 4-2). 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 mercurysulfide complexes and S042",
4-27
-------
increased mercury methylation downstream may still occur in sulfate-enriched ecosystems with
increased organic matter and/or downstream transport capabilities.
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.
4.5.3.6 Mercury Accumulation in the Food Web
Dissolved Hg(ll) and MeHg accumulate in aquatic vegetation, phytoplankton, and
benthic invertebrates. Unlike Hg(ll), MeHg biomagnifies though each successive trophic level in
both benthic and pelagic food chains such that mercury in predatory, freshwater fish is found
almost exclusively as MeHg (Bloom, 1992; Watras et al., 1998). Thus, trophic position and food-
chain complexity plays an important role in MeHg bioaccumulation (Kidd et al., 1995). The
chemical and physical characteristics of different ecosystems affect MeHg uptake at the base of
the food chain, driving bioaccumulation at higher trophic levels. At the base of pelagic
freshwater food-webs, MeHg uptake by plankton is thought to be a combination of passive
diffusion and facilitated transport (Laporte et al., 2002; Watras et al., 1998). Uptake of MeHg by
plankton can be enhanced or inhibited by the presence of different ligands bound to MeHg
(Lawson and Mason, 1998). Similarly, the assimilation efficiency of MeHg at the base of the
food chain is also affected by the type of dissolved MeHg-complexes in the water and
sediments. This may be a function of differences in the ability of organisms to solubilize MeHg
through digestive processes with different MeHg complexes (Lawrence and Mason, 2001;
Leaner and Mason, 2002). The presence of organic ligands and high concentrations of DOC in
aquatic ecosystems are generally thought to limit MeHg uptake by biota (Driscoll et al., 1995;
Sunda and Huntsman, 1998; Watras et al., 1998).
In fish, MeHg bioaccumulation is a function of several uptake (diet, gills) and elimination
pathways (excretion, growth dilution) (Gilmour et al., 1998; Greenfield et al., 2001). As a result,
the highest mercury concentrations for a given fish species correspond to smaller, long-lived
4-28
-------
fish that accumulate MeHg over their life span with minimal growth dilution (e.g., (Doyon et al.,
1998). In general, higher mercury concentrations are expected in top predators, which are
often large fish relative to other species in a waterbody.
4.5.4 Summary
In the United States, humans are exposed to MeHg mainly by consuming fish that
contain MeHg. Aquatic ecosystems respond to changes in mercury deposition in a highly
variable manner as a function of differences in their chemical, biological and physical
properties. Depending on the characteristics of a given ecosystem, methylating microbes
convert a small but variable fraction of the inorganic mercury in the sediments and water
derived from human activities and natural sources into MeHg. 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.
For this analysis, EPA used the Mercury Maps (MMaps) model to estimate changes in
freshwater fish mercury concentrations resulting from changes in mercury deposition after
regulation of mercury emissions from U.S. coal-fired power plants. MMaps, a simplified form of
the IEM-2M model applied in EPA's 1997 Mercury Study Report to Congress, is a static model
that assumes a proportional relationship between declines in atmospheric mercury deposition
and concentrations in fish at steady state. This means, for example, that a 50% decrease in
mercury deposition rates is projected to lead to a 50% decrease in mercury concentrations in
fish. MMaps does not consider the dynamics of relevant ecosystem specific factors that can
affect the methylation and bioaccumulation in fish in different water bodies over time, nor does
it consider the inputs of non-air sources to the watershed. In all cases, the MMaps model does
not address the lag time of different ecosystems to reach steady state (i.e., when fish mercury
concentrations reflect changes in atmospheric deposition). In addition, applying the MMaps
model assumes that atmospheric deposition is the principle source of mercury to the
waterbodies being investigated and environmental factors that affect MeHg production and
accumulation in organisms will remain constant, allowing each ecosystem to reach steady state.
While MMaps has several limitations, EPA knows of no alternative tool for performing a
national-scale assessment of such changes.
4-29
-------
4.6 Analysis of the Dose-Response Relationship Between Maternal Mercury Body Burden
and Childhood IQ
4.6.1 Introduction
In considering possible health end points for quantification and monetization, EPA
reviewed the scientific literature on the health effects of mercury, including the "lexicological
Effects of Methylmercury," published by the National Research Council (NRC) in 2000 (NRC,
2000).
EPA chose to focus on quantification of intelligence quotient (IQ) decrements associated
with prenatal mercury exposure as the initial endpoint for quantification and valuation of
mercury health benefits. Reasons for this initial focus on IQ included the availability of
thoroughly-reviewed, high-quality epidemiological studies assessing IQ or related cognitive
outcomes suitable for IQ estimation, and the availability of well-established methods and data
for economic valuation of avoided IQ deficits, as applied in EPA's previous benefits analyses for
childhood lead exposure. In the "Peer Review of EPA's Draft National-Scale Mercury Risk
Assessment" (SAB, 2011 available at:
http://yosemite.epa.gov/sab/sabproduct.nsf/ea5d9a9b55cc319285256cbdOQ5a472e/aaf67ae4
ddl99409852578cb006bcb04!OpenDocument) the Science Advisory Board noted that a
number of measures of potential neurodevelopmental effects of methylmercury exist, some of
which have greater sensitivity than IQ loss. However, none were viewed by the Panel as
suitable for quantitative risk estimation with a reasonable degree of scientific certainty at the
present time, and none were recommended for incorporation into the analysis. IQ score has
not been the most sensitive indicator of methylmercury's neurotoxicity in the populations
studied. The Faroe Island study the most sensitive indicators were in the domains of language
(Boston Naming), attention (continuous performance) and memory (California Verbal Learning
Test), neuropsychological tests that are not subtests of IQ tests and are not highly correlated
with global IQ. In the Seychelles study, the Psychomotor Development Index has been most
sensitive measure and, while this is a component of the Bailey Scales of Infant Development, it
is not highly correlated with cognitive measures. While the Panel agreed that the
concentration-response function for IQ loss used in the risk assessment is appropriate, IQ loss is
not a sensitive response to methylmercury and its use likely underestimates the impact of
reducing methylmercury in water bodies.
Epidemiological studies of prenatal mercury exposure conducted in the Faroe Islands
(Grandjean et al., 1997), New Zealand (Kjellstrom et al., 1989; Crump et al., 1998), and the
Seychelles Islands (Davidson et al., 1998; Myers et al., 2003) have examined
4-30
-------
neurodevelopmental outcomes through the administration of tests of cognitive functioning.
Each of these studies included some but not all of the following tests: full-scale IQ, performance
IQ, problem solving, social and adaptive behavior, language functions, motor skills, attention,
memory and other functions. The NRC reviewed the studies and determined that "Each of the
studies was well designed and carefully conducted, and each examined prenatal MeHg
exposures within the range of the general U.S. population exposures" (NRC, 2000).
As part of previous analyses, EPA attempted to identify the appropriate dose-response
coefficients from the Faroe Islands, New Zealand, and Seychelles Islands studies, and devised a
statistical approach for combining those coefficients to provide an integrated estimate of the IQ
dose-response coefficient.
For this assessment, EPA used a more recently revised estimate of the IQ dose-response
function, based on a peer-reviewed study by Axelrad et al. (2007) ("the Axelrad study"). The
Axelrad study estimated a dose-response relationship between maternal mercury body burden
and subsequent childhood decrements in IQ using a Bayesian hierarchical model to integrate
data from the Faroe Islands, New Zealand, and Seychelles Islands studies.
The Axelrad study used a linear model that goes through the origin to fit population-
level dose-response relationships to the pooled data from the three studies. The application of
a linear model should not be interpreted to suggest that any of the three studies used have
data showing health effects from MeHg exposure at or below the RfD. The RfD is an estimate of
a daily exposure to the human population (including sensitive subgroups) that is likely to be
without an appreciable risk of deleterious effects during a lifetime (EPA, 2002). EPA believes
that exposures at or below the RfD are unlikely to be associated with appreciable risk of
deleterious effects. It is important to note, however, that the RfD does not define an exposure
level corresponding to zero risk; mercury exposure near or below the RfD could pose a very low
level of risk which EPA deems to be non-appreciable. It is also important to note that the RfD
does not define a bright line, above which individuals are necessarily at risk of adverse effect.
Use of a linear model that goes through the origin, rather than one that reflects a threshold
effect is technically more simple and practical. It associates an increment of IQ benefit with a
given reduction in exposure. A linear model allows us to estimate the benefits of reductions in
exposure due to power plants without a complete assessment of other sources of exposure.
Other models would require information on the joint distribution of exposure from power
plants and other sources to estimate the benefits of reducing the exposure due to power
plants, which would require much more precise information about consumption patterns.
4-31
-------
4.6.2 Epidemiological Studies of Mercury and Neurodevelopmental Effects
The IQ dose-response estimate is based on data from three major prospective studies
investigating potential neurotoxicity of low-level, chronic mercury exposure: the Faroe Islands
study, the New Zealand study, and the Seychelles Child Development Study.
In assembling the New Zealand sample, Kjellstrom et al. (1989) ascertained the fish
consumption of 10,930 of 16,293 pregnant women in the study area. They identified 935
women who reportedly consumed fish at least 3 times per week. Hair samples were obtained
from these women, and 73 were found to have a hair mercury level of 6 parts per million (ppm)
or greater. In this group, the mean was 8.3 ppm, with a range of 6 to 86 ppm, although only one
woman had a level greater than 20 ppm. Each woman with 6 ppm hair mercury or greater was
matched to 3 controls—one with hair mercury between 3-6 ppm, one with hair mercury less
than 3 ppm and high fish consumption, and one with hair mercury less than 3 ppm and low fish
consumption. Ethnic group, age, smoking, residence time in New Zealand, and child sex were
also used to select controls. The final study group included 237 children, including 57 fully
matched sets of 4 children. Although children were assessed at 4 and 6 years of age, only the
data collected at the older age is considered in this analysis, as the reliability and validity of
neurodevelopmental testing generally increases with child age.
The Faroe Islands investigators assembled a birth cohort of 1,353 newborns recruited
from three hospitals over a 21-month period in 1986-1987. In 1,022 women, two biomarkers of
prenatal mercury exposure were collected: cord-blood mercury, and maternal hair mercury at
delivery. Neurodevelopmental assessments of 917 children were conducted at age 7
(Grandjean et al., 1997). For these 917 children, the geometric mean concentration of mercury
in cord-blood was 22.6 parts per billion (ppb) (inter-quartile range 13.1-40.5 ppb, full range
0.9-351 ppb). The geometric mean concentration of mercury in maternal hair was 4.2 ppm
(inter-quartile range: 2.5-7.7 ppm, full range 0.2-39.1 ppm) (Budtz-Jorgensen et al., 2004a).
Neurodevelopmental assessments of the children were conducted at age 7 years (Grandjean
etal., 1997).
In assembling the Seychelles Child Development Study sample, investigators obtained
hair samples from 779 pregnant women and ultimately enrolled a study sample consisting of
740 newborns. The mean maternal hair mercury level was 6.8 ppm (range 0.9-25.8 ppm)
(Davidson et al., 1998). Neurodevelopmental assessments were conducted when the children
were 6.5, 19, 29, and 66 months, and at 9 years. The mean maternal hair mercury level for the
4-32
-------
643 children who participated in the assessment at age 9 years was 6.9 ppm (standard
deviation 4.5 ppm) (Myers et al., 2003).
4.6.3 Statistical Analysis
Previous statistical analysis conducted by Ryan (2005) produced a dose-response
relationship, integrating data from all three studies, with a central estimate of an IQ change of
-0.13 IQ points (95% confidence interval -0.28, -0.03) for every ppm of mercury in maternal
hair. Axelrad et al. (2007) conducted a more recent statistical analysis integrating data from the
Faroe Islands, New Zealand, and Seychelles Islands studies to produce a single estimate of the
IQ dose-response relationship, which is used in this RIA. Additional details of the analysis are
reported in the Axelrad study and in its Supplemental Material (available at
http://www.ehponline.org/docs/2007/9303/suppl.pdf). The information is summarized below.
The Axelrad study used a Bayesian hierarchical statistical model to estimate the
integrated dose-response coefficient. This is similar to the approach used by the NRC panel to
calculate a benchmark dose value integrating data from all three studies (NRC, 2000). The
model makes use of dose-response coefficients for IQ, and also considered all other cognitive
endpoints reported in the three studies in an effort to obtain more robust estimates of the IQ
relationship that account for within-study (endpoint-to-endpoint) variability as well as
variability across studies.
The Axelrad study assumed a linear relationship between mercury body burdens and
neurodevelopmental outcomes, in keeping with the recommendation of the NRC committee
(NRC, 2000). In the New Zealand and Seychelles Islands studies, all information necessary for
the model was obtained from the published papers, including linear regression coefficients
(Crump et al., 1998; Myers et al., 2003). The Faroe Islands publications, however, reported
results with cord blood and maternal hair mercury transformed to the log scale and provided
no results of linear models (Grandjean et al., 1997, 1999). A report by the Faroe Islands
investigators (Budtz-Jorgensen et al., 2005) provided the additional details needed for the
analysis.
The Wechsler Intelligence Scales for Children (WISC) is a standard test of childhood IQ
that was used in each of the three studies. The version of the test administered in the
Seychelles Islands (3rd ed.; WISC-III) was different from the earlier version used in New Zealand
and the Faroe Islands (revised ed.; WISC-R). In a sample of approximately 200 children, the
correlation between the Full-Scale IQ scores for the two versions was 0.89; thus the WISC-R and
4-33
-------
WISC-III appear to measure the same constructs and generate scores with similar dispersion
(Wechsler, 1991).
The WISC-R includes 10 core subtests and three supplementary subtests. For the Faroe
Islands study, the investigators administered only three subtests of the WISC-R: Digit Span and
Similarities (core subtests) and Block Design (a supplementary subtest). The Axelrad study used
data for these three subtests to estimate an IQ-mercury coefficient for the Faroe Islands
cohort. The Faroe Islands investigators fit data for these three subtests in a structural equation
model (SEM) to estimate a standardized coefficient for a hypothetical Full-Scale IQ (Budtz-
Jorgensen et al., 2005). In the SEM analysis of IQ, the three WISC-R subtests are viewed as
representative of an underlying latent IQ variable.
To estimate the association between mercury and IQ using information from the three
studies, the Axelrad study used a hierarchical random-effects model that includes study-to-
study as well as endpoint-to-endpoint variability. Axelrad et al. (2007) implemented the model
with a Bayesian approach, using WinBUGS version 1.4 (http://www.mrc-bsu.cam.ac.uk/bugs/).
Although the Axelrad study's Bayesian analysis yields highest posterior density (HPD) intervals,
the authors refer to these as confidence intervals to aid in the interpretation of results (Axelrad
etal., 2007).
The integrated analysis produced a central estimate of-0.18 (95% Cl, -0.378 to -0.009)
IQ points for each part per million maternal hair mercury, similar to the results found for both
the Faroe Islands and Seychelles studies, and lower than the estimate found in the New Zealand
study. This central estimate was used as the basis for estimating IQ loss associated with
prenatal MeHg exposure in this assessment.
4.6.4 Strengths and Limitations of the IQ Dose-Response Analysis
The Axelrad study produced an estimate of the relationship between maternal mercury
body burdens during pregnancy and childhood IQs that incorporates data from all three
epidemiologic studies judged by the NRC to be of high quality and suitable for risk assessment.
The statistical approach makes use of all the available data (including information on results for
related tests of cognitive function), and can be used to produce population-based estimates of
a health outcome that can be readily monetized for use in benefit-cost analysis.6
6There is limited evidence directly linking IQ and methylmercury exposure in the three large epidemiological
studies that were evaluated by the NAS and EPA. Based on its evaluation of the three studies, EPA believes that
children who are prenatally exposed to low concentrations of methylmercury may be at increased risk of poor
4-34
-------
There are several aspects of IQ as a metric for neurodevelopmental effects in this
benefit-cost analysis that are important to recognize. Full-Scale IQ is a composite index that
averages a child's performance across many functional domains, providing a good overall
picture of cognitive health. An extensive body of data documents the predictive validity of full-
scale IQ, as measured at school age, and late outcomes such as academic and occupational
success (Neisser et al., 1996). In addition, methods are readily available for valuing shifts in IQ
and thus conducting a benefits analysis of interventions that shift the IQ distribution in a
population. Methods for monetization of the other tests administered in the three studies have
not been developed.
It is important to recognize, however, that full-scale IQ might not be the cognitive
endpoint that is most sensitive to prenatal mercury exposure. Significant inverse associations
were found, in both the New Zealand and Faroe Islands studies, between prenatal mercury
levels and neurobehavioral endpoints other than IQ. If the effects of mercury are highly focal,
affecting only specific cognitive functions, taking full-scale IQ as the primary endpoint for a
benefits analysis might underestimate the impacts. In averaging performance over diverse
functions in order to compute full-scale IQ, the specific effects of mercury on only certain of
these functions would be "diluted," and the estimated magnitude of the change in performance
per unit change in the mercury biomarker would be underestimated.
Moreover, it is well known that there may be substantial deficits in cognitive wellbeing
even in individuals with normal or above average IQ. The criterion most frequently used to
identify children with learning disabilities for the purposes of assignment to special education
services is a discrepancy between IQ and achievement. Specifically, the child's achievement in
reading, math, or other academic areas is significantly lower than what would be expected,
given his or her full-scale IQ. Thus, there are deficits in cognitive functioning that are not
captured by IQ scores. For example, two of the most sensitive endpoints in the Faroe Islands
study were the Boston Naming Test, which assesses word retrieval, and the California Verbal
Learning Test-Children, which assesses the acquisition and retention of information presented
verbally. Depending on the severity of the deficits, a child who has deficits in either of these
skills could be at a considerable disadvantage in the classroom setting and at substantial
educational risk. Neither of these abilities is directly assessed by the WISC-R or WISC-III,
however, and so do not explicitly contribute to a child's IQ score. Therefore, benefits
performance on neurobehavioral tests, such as those measuring attention, fine motor function, language skills,
visual-spatial abilities (like drawing), and verbal memory. For this analysis, EPA is adopting IQ as a surrogate for
the neurobehavioral endpoints that NAS and EPA relied upon for the RfD.
4-35
-------
calculations relying solely on IQ decrements are likely to underestimate the benefits to
cognitive functioning of reduced mercury exposures. In additions, impacts on other
neurological domains (such as motor skills and attention/behavior) are not represented by IQ
scores and thus are also excluded from the benefits analysis.
As discussed above, the Faroe Islands study did not include testing for full-scale IQ. For
the Axelrad study, an estimate of a dose-response coefficient for full-scale IQ was estimated
using the three subtests. While this extrapolation introduces some uncertainty, information has
been presented that demonstrates a high correlation between the subtests and full-scale IQ
scores.
While the Seychelles and New Zealand studies use maternal hair mercury as the
exposure biomarker, the Faroe Islands study uses cord blood mercury. For purposes of the
integrated analysis, it was necessary to express results from all three studies in the same terms.
Several studies have examined the relationship between hair mercury and blood mercury, and
have reported hairblood ratios typically in the range of 200 to 300 (see ATSDR, 1999, pages
249-252 for a review). However, these studies generally do not use cord blood mercury, which
is the exposure metric in the Faroe Islands study. One analysis found that mercury
concentrations in cord blood are, on average, 70 percent higher than those in maternal blood
(Stern and Smith, 2003). For conversion of Faroe Islands data from cord blood mercury to
maternal hair mercury, the Axelrad study used data specific to this population, indicating a
median maternal haircord blood mercury ratio of 200 (Budtz-Jorgensen et al., 2004a).
One uncertainty concerning the New Zealand study is the strong influence of one child
in the study population with a particularly high maternal hair mercury level. Published analyses
of the New Zealand study presented results with data for this child both included and excluded
(Crump et al., 1998). In keeping with the conclusions of the NRC (2000), the integrated dose-
response analysis in the Axelrad study made use of the dose-response coefficients calculated
with this child omitted. A sensitivity analysis using the New Zealand coefficient with this child
included results in an integrated dose-response coefficient that is reduced in magnitude by 25
percent (-0.125 versus a primary central estimate of -0.18).
Some uncertainty is also associated with the Seychelles study due to the exclusion of
some members of the cohort from the data reported by Myers et al. (2003) and used as input
to this integrated dose-response analysis. The Seychelles researchers did not include a small
number of outliers (defined as observations with model residuals exceeding 3 standard
deviation units), and no results are available for the full cohort. However, the authors report
4-36
-------
that "In all cases, the association between prenatal MeHg exposure and the endpoint was the
same, irrespective of whether outliers were included" (Myers et al., 2003).
Finally, the integrated dose-response analysis assumes the exposures assigned to each
study subject are accurate representations of true exposure. In reality, there is likely to be
some discrepancy between measured and actual exposures, for example, due to variation in
hair length. Alternatively, the true exposure of interest may have been during the first trimester
of pregnancy, whereas exposures in maternal hair and cord blood measured at birth reflect
exposures later in pregnancy. Presence of exposure measurement error could introduce a bias
in the results, most likely towards the null (Budtz-Jorgensen et al., 2004b).
4.6.5 Possible Confounding from Long-Chained Polyunsaturated Fatty Acids
Maternal consumption of fish during pregnancy exposes the fetus to long-chain
polyunsaturated fatty acids (LCPUFAs), believed to be beneficial for fetal brain development,
and to the neurotoxicant MeHg (Helland et al., 2003; Daniels et al., 2004; Dunstan et al., 2006;
Judge et al., 2007). Reports from the Seychelles Islands study cohort have suggested a negative
impact of MeHg exposure, accompanied by a simultaneous beneficial effect of omega-3
LCPUFAs on children's development (Davidson et al., 2008; Strain et al., 2008). It is unclear
whether this result was evidence for independent influences of MeHg and LCPUFAs or effect
modification. A recent study by Lynch et al. (2010) used varying coefficient models to
characterize the interaction of mercury and nutritional covariates (Hastie and Tibshirani, 1993),
including omega-3 LCPUFAs, using data from the Seychelles Islands study.
The Seychelles Islands study cohort of mother-child pairs had fish consumption
averaging 9 meals per week. Lynch et al., (2010) assessed maternal nutritional status for five
different nutritional covariates known to be present in fish (n-3 LCPUFA, n-6 LCPUFA, iron
status, iodine status, and choline) and associated with children's neurological development. The
study also included prenatal MeHg exposure (measured in maternal hair).
Lynch et al., (2010) examined two child neurodevelopmental outcomes (Bayley Scales
Infant Development-ll (BSID-II) Mental Developmental Index (MDI) and Psychomotor
Developmental Index (PDI)), each administered at 9 and at 30 months. The varying coefficient
models allowed the possible interactions between each nutritional component and MeHg to be
modeled as a smoothly varying function of MeHg as an effect modifier. Iron, iodine, choline,
and omega-6 LCPUFAs had little or no observable modulation at different MeHg exposures. In
contrast the omega-3 LCPUFA docosahexaenoic acid had beneficial effects on the BSID-II PDI
that were reduced or absent at higher MeHg exposures. The results from Lynch et al. (2010)
4-37
-------
suggest a potentially useful modeling method that could shed further light on the issue of
interactions between nutritional covariates.
A recent study by Rice et al. (2010) considered possible confounding in a probabilistic
assessment of the health benefits of reducing MeHg exposure in the United States. In deciding
on a dose-response relationship between MeHg exposure and effects on IQ loss, the authors
chose to use the central estimate from the Axelrad study, noting however that Axelrad et al.
(2007) did not explicitly consider possible confounding of the MeHg-IQ relationship by the
concurrent consumption of LCPUFAs that might enhance cognitive development and bias
downward the observed regression coefficient estimates from the Faroe Islands, New Zealand,
and Seychelles Islands studies. Rice et al. (2010) therefore multiplied the central estimate from
Axelrad et al. (2010) by an adjustment factor to offset the possible downward bias from
inadequate confounder control. A factor of 1.5 was selected "to acknowledge the recent
argument of Budtz-Jorgensen et al. (2007) that the parameter estimates from the three
epidemiological studies may be biased downward by a factor of approximately 2 because of
failure to adequately control for confounding" (Rice et al., 2010).
There remains uncertainty with respect to the nature and magnitude of potential
confounding between LCPUFAs and MeHg, and the associated effects on childhood
neurodevelopment due to maternal ingestion during pregnancy. Additional research is needed
to provide further clarity on this issue, but recent studies such as those referenced above
reinforce the view that fish consumption during pregnancy should be approached as a case of
multiple exposures to nutrients and to MeHg, with a complex and potentially interactive set of
risks and benefits related to infant development. Due to the remaining uncertainty regarding
the potential confounding between LCPUFAs and MeHg exposure, we have not incorporated
any factors or other quantitative adjustments into this assessment.
4.7 Mercury Benefits Analysis Modeling Methodology
4.7.1 Introduction
This section describes the methodology used to model fishing behavior and associated
MeHg exposure levels. The methodology incorporates data, assumptions, and analytical
techniques already described in previous sections. Sections 4.7.2 and 4.7.3 below describe
elements of the methodology applied to develop a national-scale estimate of benefits
associated with avoided IQ loss among freshwater recreational anglers. Chapter 7 section 7.11
describes a variation of the methodology used to estimate risk levels (as measured by IQ loss)
among modeled high-risk subpopulations.
4-38
-------
4.7.2 Estimation of Exposed Populations and Fishing Behaviors
This section describes the methodology used to estimate the average daily ingestion of
mercury (g/day) through noncommercial freshwater fish consumption (Hgl) for selected
populations of interest. Because the primary measurable health effect of concern-
developmental neurological abnormalities in children—occurs as a result of in-utero exposures
to mercury, the specific population of interest in this case is prenatally exposed children. To
identify and estimate the size of this exposed population, the benefits analysis focuses on
pregnant women in freshwater recreational angler households.
Generally speaking, estimating mercury exposures for this exposure pathway and
population of interest requires three main components:
Nj = size of the exposed population of interest i (annual number of pregnant
women in freshwater angler households during the year),
CHgj = average concentration (ppm) of methyl mercury in noncommercial freshwater
fish filets consumed by population i, and
Q = average daily consumption rate (gm/day) of noncommercial freshwater fish
by population i.
The flow diagram in Figure 4-3 illustrates the approach used to estimate the first two
components of this equation—A//and CHg/. It shows the spatial scale of the data used to
estimate these components and describes how these components are interrelated. For the
third component—C,~recommendations from EPA's Environmental Exposure Factors
Handbook (EPA, 1997) were used to estimate an average consumption rate estimate for
recreationally caught freshwater fish.
4-39
-------
Description
Spatial Scale Data Source!Computation
Number of females of
childbearing age
Fertility rate
Number of pregnant women
Number of angler residents (age > 15)
Population (age > 15)
Portion of population that is anglers
Number of prenatally exposed children
tany exposed cnimren i • _
in angler households I "
NPA in demographic group i,
distance], and waterbody type k
Portion of freshwater fishing days by residents
to waterbody type k (k = lake (I) or river (r))
Portion of freshwater fishing trips by residents
in demographic group i (i = 1—4) to sites in
distance interval] (j = 0-10, >10-20, >25-50,
or>50-100)
Portion of resident anglers in demographic
group i (i = 1—4)
Average Hg fish tissue concentration by
waterbody type (k) and distance (j)
NPAijk= NPA * pi * eg * ck
Census Tract
State
Census Tract
State
National
Census Tract
Census Tract
Census
Vital Stats
Eq.(4.1)
State
State
State
Census Tract
Census Tract
FHWAR
Census
Eq. (4.2)
Eq. (4.2)
Eq. (4.3)
FHWAR
NSRE
Census
NLFWA/
USGS/GIS
Figure 4-3. Methodology for Estimating and Linking Exposed Populations and Levels of Mercury Exposure
-------
First, 2000 Census data (U.S. Census Bureau, Census 2000 Summary File 3, Detailed
Tables, United States) were used to define the size, age, gender distribution, and income of the
populations within each census tract in the 48 contiguous U.S. states.
1. Estimating the number of pregnant women (NP) living in the census tract as
NP = NF * fs, (4.1)
where
NF = number of females aged 15 to 44 in the tract (Census 2000) and
fs = state-level general fertility rate (average number of live births in a year per
1,000 women aged 15 to 44) (2006 Vital Statistics).
2. Estimating the annual number of prenatally exposed children in angler households
(NPA)as
NPA = NP*(NAS/NS), (4.2)
where
NAS = state-level number of angler residents (FHWAR) and
Ns = adult population of state s (Census).
Using Eq. (4.2) to estimate NPA implies that (1) the fraction of pregnant women in a
state who are in freshwater angler households is equal to the fraction of households in the
state that include freshwater anglers (i.e., pregnant women are no more or less likely than the
rest of the state population to live in households with freshwater anglers) and (2) the fraction
of households in the state that includes freshwater anglers is equal to the fraction of adult
residents in the state who are freshwater anglers.
To estimate NPA for years after 2000, it was assumed that state-level fertility rates (fs)
and angler participation rates (NAS/NS) would remain constant; however, the number of women
of childbearing age in each block (NF) was increased based on county-level population growth
projections (Woods and Poole, 2008). In other words, for the period 2000 to 2016, the
estimated NPA for each census tract was assumed to increase at the same rate as the projected
annual population growth rates for females 15 to 44 in their corresponding counties.
41
-------
100 Miles
Census Data:
Number of Females Aged 15-44
Percentage of Population with
Income >$50k and <$50k
Urban v. Rural Classification
NJ
Ha Fish
Tissue
Sample Data:
Average
River
Hg Cone.
50-100 miles
(CHgr4)
Average
River
Hg Cone.
20-50 miles
(CHgr3)
Average
River
Hg Cone.
10-20 miles
(CHgr2)
Average
River
Hg Cone.
0-10 miles
(CHgr1)
Average
Lake
Hg Cone.
0-10 miles
(CHgM)
Average
Lake
Hg Cone.
10-20 miles
(CHg,2)
Average
Lake
Hg Cone.
20-50 miles
(CHg,3)
Average
Lake
Hg Cone.
50-100 miles
(CHg,4)
Figure 4-4. Linking Census Tracts to Demographic Data and Mercury Fish Tissue Samples
-------
Fourth, to match exposed populations in each tract with mercury concentrations, we
first divided the exposed population into four distinct demographic groups (i = 1 - 4):
urban/low income, urban/high income, nonurban/low income, and nonurban/high income. To
estimate the portion of households in each demographic group (PJ for i = 1 - 4), tract-level
Census data were used to specify (1) the percentage of the population in each tract that resides
in an urban area and (2) the percentage with household income less than $50,000 (i.e., the
portion in the low-income group).
In addition, it was assumed that
1. each exposed individual in a census tract is associated with freshwater fishing in a
single distance interval and a single waterbody type (i.e., all the fish they consume
comes from the same distance and type of waterbody),7 and
2. the exposed populations in each census tract (rather than just the fishing trips) are
distributed across the distance intervals and waterbody types according to the
estimated proportions (i.e., parameters c, e, and p shown in Figure 4-4).
More specifically, a maximum of 32 separate exposed subpopulations were defined for each
census tract:
NPAijk=NPA*pi*eij*ck(foralli,j, and k) (4.3)
for
i = 1-4 demographic subgroup in the census tract,
j = 1-4 distance interval, and
k = lake or river.
(See Figure 4-3 for definitions of p\, BJJ, and Ck).
Using this approach, we were able to separately match each subpopulation NPAjjk with
the census tract's average mercury concentration for the corresponding distance and
waterbody category (CHgjk).
7An alternative would be to assume that all anglers in the census tract have the same distribution of trips across
distance intervals and water types. This assumption would imply no variation in per-capita mercury exposures
within a census tract, but it would not affect the estimates of total exposure and total IQ losses in the tract.
43
-------
To approximate the percentage freshwater fishing trips (and exposed individuals) from
each census tract matched to each waterbody type (q or cr), we used state-level averages.
These averages were calculated for each state, based on the portion of residents' freshwater
fishing trips that are to each waterbody type, based on 2001 FHWAR data.
Data from NSRE 1994 were used to approximate the percentage of freshwater fishing
trips (and exposed individuals) matched to different distances from anglers' residential location.
Four distance intervals were defined as 0-10 miles, >10-20 miles, >20-50 miles, and >50-100
miles. Based on self-reported trip distance information from nearly 2,000 respondents (see
Appendix B for details), each of these distance categories was associated with roughly 20% of
the reported trips in the NSRE sample. Four distinct demographic groups were also found to
have significantly different average travel distances for freshwater fishing in the NSRE sample:
high-income urban, high-income rural, low-income urban, and low-income rural. An annual
household income threshold of $50,000 (in 2000 dollars) was used to define high and low
income, because it is close to the median value for both the NSRE sample and the U.S.
population. The portion of trips for each demographic group (i = 1 - 4) to each distance interval
(j = 1 -4) is defined as e\-s. The estimated values for e^ are reported in Appendix B.
To estimate average daily mercury ingestion rates for each exposed subpopulation n=ijk,
we applied the following equation:
Hgln = CHgFCn* Cn= (CHgn* CCF) * Cn (4.4)
where
Hgl = average daily mercury ingestion rate (jag/day);
CHg = average mercury concentration in uncooked freshwater fish (ppm);
CCF = cooking conversion factor: ratio of mercury concentration in cooked fish to
mercury concentration in uncooked fish (= 1.5);
CHgFC = average mercury concentration in cooked freshwater fish (ppm); and
C = average daily self-caught freshwater cooked fish consumption rate
(gm/day) = 8 gm/day.
To determine an appropriate daily fish consumption rate (C) for the analysis, EPA
conducted an extensive review of existing literature characterizing self-caught freshwater fish
consumption. Based on this review, it was decided that the ingestion rates for recreational
44
-------
freshwater fishers, specified as "recommended" in EPA's Environmental Exposure Factors
Handbook (EPA, 1997) (mean of 8 gm/day and 95th percentile of 25 gm/day), represented the
most appropriate values to use in this analysis. These recommended values were derived based
on ingestion rates from four studies conducted in Maine, Michigan, and Lake Ontario (Ebert
et al., 1992; Connelly et al., 1996; West et al., 1989; West et al., 1993), which measured annual
average daily intake rates for self-caught freshwater fish by all recreational fishers including
consumers and non-consumers of fish. The mean values presented in these four studies ranged
from 5 to 17 gm/day, while the 95th percent values ranged from 13 to 39 gm/day (Note: the
39 gm/day value actually represents a 96th percent value). The EPA "recommended values"
were developed by considering the range and spread of means and 95th percentile values
presented in the four studies. EPA recognizes that using mean and 95th percentile consumption
rates based on these four studies may not be representative of fishing behavior across the
entire 48-state study area and that regional trends in consumption may differ from the values
used in this analysis. Moreover, rates of consumption by pregnant women in freshwater angler
households may be different from those of the recreational fishers themselves. However, EPA
believes that these four studies do represent the best available data for developing recreational
fisher ingestion rates in the United States.
Because the consumption rate estimate C is for cooked fish and the mercury
concentrations are estimated for uncooked filet, a conversion factor (CCF) was applied to
estimate mercury concentrations in cooked fish. Cooking fish tends to reduce the overall weight
of fish by approximately one-third (Great Lakes Sport Fish Advisory Task Force, 1993). Because
volatilization of mercury is unlikely to occur during cooking, the overall amount of mercury will
stay unchanged during cooking, and the concentration of mercury will increase by a factor of
roughly 1.5 (Morgan, Berry, and Graves, 1997).
4.7.3 Estimation of Lost Future Earnings
Estimating the IQ decrements in children that result from mothers' ingestion of mercury
required two steps. First, based on the estimated average daily maternal ingestion rate, the
expected mercury concentration in the hair of exposed pregnant women was estimated as
follows:
CHgHn = (0.08)-1 * (Hgln/W), (4.5)
where
CHgH = average mercury concentration in maternal hair (ppm) and
45
-------
W = average body weight for female adults below age 45 (= 64 kg).
This conversion rate between average daily ingestion rate and maternal hair
concentration is based on the one compartment model developed by Swartout and Rice (2000).
The 2002 EPA Workshop on Methyl mercury Neurotoxicity recommended that this one
compartment model might be better suited than the PBPK model in modeling dose-response
(EPA, 2002). The average body weight estimate (W) was based on EPA's Exposure Factor
Handbook (EPA, 1997).
Second, to estimate the expected IQ decrement in offspring resulting from in-utero
exposure to mercury through mothers' fish consumption, the following dose-response
relationship was applied:
dlQn = 0.18*CHgHn, (4.6)
where
dlQ = IQ decrement in exposed mother/child (IQ pts).
The 0.18 dose-response coefficient in this equation is based on the summary findings reported
in Axelrad et al. (2007).
The valuation approach used to assess monetary losses due to IQ decrements is based
on an approach applied in previous EPA analyses (EPA, 2008). The approach expresses the loss
to an affected individual resulting from IQ decrements in terms of foregone future earnings (net
of changes in education costs) for that individual. These losses were estimated using the
following equation:
Vn = VIQ*dlQi, (4.7)
where
V = present value of net loss per exposed mother/child (2006 dollars) and
VIQ = net loss per change in IQ point.
The net loss per IQ point decrement is estimated based on the following relationship:
VIQ=(z* PVY)-(s* PVS), (4.8)
46
-------
where
PVY = median present value of lifetime earnings,
PVS = present value of education costs per additional year of schooling,
z = percentage change in PVY per 1-point change in IQ, and
s = years of additional schooling per 1-point increase in IQ.
The estimate for PVY is derived using earnings and labor force participation rate data
from the 2006 Current Population Survey (CPS) and assuming (1) an individual born today
would begin working at age 16 and retire at age 67; (2) the growth rate of wages is 1% per year,
adjusted for survival probabilities and labor force participation by age; and (3) lifetime earnings
are discounted back to the year of birth. Using a 3% discount rate, the resulting present value of
median lifetime earnings is $555,427 in 2006 dollars.
Estimates of the average effect of a 1-point increase in IQ on lifetime earnings (z) range
from a 1.76% increase (Schwartz, 1994) to a 2.379% increase (Salkever, 1995). The percentage
increases in the two studies reflect both the direct impact of IQ on hourly wages and indirect
effects on annual earnings as the result of additional schooling and increased labor force
participation. The estimate for s is based on Schwartz (1994) who reports an increase of 0.131
years of schooling per IQ point.
In addition to this positive net effect on earnings, an increase in IQ is also assumed to
have a positive effect on the amount of time spent in school (s) and on associated costs (PVS).
The range of estimate for s is based on Schwartz (1994) who reports an increase of 0.131 years
of schooling per IQ point and Salkever (1995) who reports an increase of 0.1007 years.
The estimate for PVS is derived using an estimate of $16,425 per additional year of
schooling in 1992 dollars (EPA, 2005), which is based on U.S. Department of Education data
reflecting both direct annual expenditures per student and annual average opportunity cost
(i.e., lost income from being in school). We assume these costs are incurred when an individual
born today turns 19, based on an average 12.9 years of education among people aged 25 and
over in the United States. Discounting at a 3% rate to the year of birth results in an estimate of
$13,453 per additional year of schooling in 2006 dollars.
To incorporate (1) uncertainty regarding the size of z and (2) different assumptions
regarding the discount rate, the resulting value estimates for the average net loss per IQ point
47
-------
decrement (VIQ) are expressed as a range. Assuming a 3% discount rate, VIQ ranges from
$8,013 (using the Schwartz estimate for z and s) to $11,859 (using the Salkever estimates). With
a 7% discount rate assumption, the VIQ estimates range from $893 to $1,958.
4.8 Mercury Benefits and Risk Analysis Results
4.8.1 Baseline Incidence
Applying the methodology described in Section 4.7, we first used CIS to link census tract
centroids in the continental United States with HUC-12 watersheds containing mercury fish
tissue sample data for 1995 to 2007. We found that, out of the 64,500 tracts in the 48-state
area, almost all of them are located within 100 miles of at least one HUC-12 with freshwater
mercury fish tissue sampling data. Therefore, very few tracts were entirely excluded from the
analysis due to a lack of sampling data within 100 miles. Table 4-4 reports the number of tracts
linked to HUC-level river or lake mercury concentration estimates within each distance interval.
As expected, this number decreases as the size of the distance interval decreases. For example,
33% are within 10 miles of a HUC-12 containing a lake sample, and 52% are within 10 miles of a
HUC-12 containing a river sample.
Table 4-4 also reports the average river and lake HUC-level fish tissue mercury
concentrations found within each distance interval. Assuming that the 1995 to 2007 samples
are representative of baseline conditions in 2005, the distance-specific mean lake
concentrations range from 0.26 to 0.3 ppm, and the mean river concentrations vary from 0.25
to 0.27 ppm.
Table 4-4 also reports corresponding river and lake mercury concentration estimates for
a 2016 base case scenario. This scenario represents total mercury deposition from all global
natural and anthropogenic sources based on projected 2016 conditions, including future
anticipated regulations (e.g., Transport Rule). As described in Section 4.4, CMAQ air quality
modeling runs were used to estimate average mercury deposition levels by HUC-12 sub-
watershed under both the 2005 base case and the 2016 base case scenarios. For this analysis, it
is assumed that HUC-level fish tissue mercury concentrations would change (between the two
scenarios) by the same percentage as the change in modeled deposition levels. Overall, the
mean concentrations decline by 6% to 9% in the 2005 base case compared with the 2016 base
case scenarios.
With these tract-level mercury concentration estimates, we then estimated the size of
the exposed populations (NPA) in 2005 and 2016. These estimates are reported in Table 4-5. As
48
-------
described in Section 4.7.2, a separate exposed population (NPAjk) was estimated for each
distance interval (j = 1-4) and waterbody (k = lake or river) combination at each tract. If
mercury concentration data were not available for a specific distance-waterbody combination,
then the corresponding exposed population for the tract (NPAjk) was not included in the
analysis. Consequently, the exposed population estimates reported in Table 4-5 are best
interpreted as lower-bound estimates of the total exposed population. Excluding potentially
exposed populations from the analysis because of missing/unavailable mercury concentration
data reduced the total exposed population estimate by roughly 44%. These excluded
populations include the portions of the tract-level exposed populations that were matched with
fishing trip travel distances that either (1) did not overlap with at least one HUC-12 with
sampling data or (2) were greater than 100 miles (see Appendix C). For 2005, there were
estimated to be 239,174 prenatally exposed children, and for 2016 the estimate is 244,286
prenatally exposed children.
49
-------
Table 4-4. Summary of Baseline Mercury Fish Tissue Concentrations
2005 Base Case
Distance from Tract Centroid
Lake Sampling Sites
0-10 miles
>10-20 miles
>20-50 miles
>50-100 miles
River Sampling Sites
0-10 miles
>10-20 miles
>20-50 miles
>50-100 miles
Na
20,998
35,149
55,885
61,820
33,342
44,493
54,970
62,868
Min
(ppm)
0.000
0.000
0.000
0.000
0.006
0.006
0.019
0.023
Mean
(ppm)
0.297
0.285
0.289
0.264
0.246
0.269
0.270
0.267
Max
(ppm)
3.561
3.561
3.561
2.333
4.967
4.967
4.480
4.967
Median
(ppm)
0.198
0.209
0.223
0.241
0.185
0.195
0.203
0.214
Min
(ppm)
0.000
0.000
0.000
0.000
0.005
0.005
0.019
0.022
2016 Base Case
Mean
(ppm)
0.276
0.264
0.270
0.247
0.224
0.247
0.251
0.251
Max
(ppm)
3.420
3.420
3.420
2.251
4.924
4.924
4.441
4.924
Median
(ppm)
0.178
0.187
0.202
0.227
0.168
0.174
0.183
0.192
Number of tracts (out of 64,419) with at least one HUC-12 with sample data in the distance interval.
-------
Table 4-5. Baseline Levels of Mercury Exposure and IQ Impacts Due to Freshwater Self-Caught Fish Consumption
2005 Base Case
State
Total
AL
AR
AZ
CA
CO
CT
DC
DE
FL
GA
IA
ID
IL
IN
KS
KY
Number of
Census
Tracts with
Hg Samples
w/in 100
Miles
63,978
1,081
623
1,097
6,801
1,045
812
181
196
3,144
1,614
791
280
2,950
1,409
716
993
Number of
Prenatally Exposed
Children (NPA)
Mean
per
Tract
3.74
5.51
6.45
3.17
1.19
3.53
2.47
2.23
1.77
3.28
8.38
6.39
6.30
2.33
5.47
2.08
4.99
Total in
State
239,174
5,956
4,017
3,476
8,089
3,693
2,003
404
348
10,299
13,525
5,052
1,765
6,884
7,711
1,490
4,954
Average
Maternal
Daily
Mercury
Ingestion
(Hgl)
(ug/day)
3.04
3.28
3.80
2.21
6.04
1.20
4.58
1.67
1.98
5.24
3.14
1.21
2.43
1.83
2.20
2.38
2.19
Average IQ
Loss per
Exposed
Child (dlQ)
0.11
0.12
0.13
0.08
0.21
0.04
0.16
0.06
0.07
0.18
0.11
0.04
0.09
0.06
0.08
0.08
0.08
Number of
Prenatally Exposed
Children (NPA)
Total IQ
Point
Losses
25,544.9
685.9
537.1
269.8
1,716.4
155.3
322.2
23.7
24.2
1,897.5
1,494.8
215.3
150.9
442.3
596.7
124.8
381.9
Mean
per
Tract
3.82
5.53
6.55
3.75
1.26
3.92
2.38
2.03
1.79
3.71
8.74
6.18
7.13
2.32
5.51
2.06
4.92
Total in
State
244,286
5,981
4,084
4,117
8,599
4,101
1,929
367
352
11,651
14,111
4,888
1,996
6,831
7,759
1,478
4,889
2016 Base Case
Average
Maternal
Daily
Mercury
Ingestion
(Hgl)
(ug/day)
2.84
3.04
3.66
2.18
5.74
1.18
4.29
1.35
1.71
5.17
2.88
1.15
2.31
1.49
1.90
2.34
1.90
Average
IQ Loss
per
Exposed
Child (dlQ)
0.10
0.11
0.13
0.08
0.20
0.04
0.15
0.05
0.06
0.18
0.10
0.04
0.08
0.05
0.07
0.08
0.07
Total IQ
Point
Losses
24,419.4
638.3
525.9
316.3
1,734.0
169.8
291.3
17.4
21.2
2,118.9
1,431.0
197.5
162.3
356.9
519.2
121.8
326.1
(continued)
-------
Table 4-5. Baseline Levels of Mercury Exposure and IQ Impacts Due to Freshwater Self-Caught Fish Consumption (continued)
NJ
2005 Base Case
State
LA
MA
MD
ME
Ml
MN
MO
MS
MT
NC
ND
NE
NH
NJ
NM
NV
NY
Number of
Census
Tracts with
Hg Samples
w/in 100
Miles
1,103
1,357
1,210
344
2,701
1,294
1,311
604
267
1,554
224
500
272
1,930
244
471
4,791
Number of
Prenatally Exposed
Children (NPA)
Mean
per
Tract
6.91
1.81
2.23
4.66
3.89
11.53
3.66
9.18
3.62
5.13
2.89
3.97
3.68
1.02
1.75
1.70
1.41
Total in
State
7,623
2,456
2,703
1,602
10,520
14,915
4,796
5,546
965
7,976
647
1,984
1,001
1,965
426
803
6,770
Average
Maternal
Daily
Mercury
Ingestion
(Hgl)
(ug/day)
3.82
5.40
2.16
5.12
2.72
2.86
1.80
5.11
2.40
3.29
3.43
1.60
5.53
3.28
1.74
3.78
3.86
Average IQ
Loss per
Exposed
Child (dlQ)
0.13
0.19
0.08
0.18
0.10
0.10
0.06
0.18
0.08
0.12
0.12
0.06
0.19
0.12
0.06
0.13
0.14
Number of
Prenatally Exposed
Children (NPA)
Total IQ
Point
Losses
1,022.9
466.0
204.8
288.3
1,005.0
1,501.2
302.7
996.2
81.5
921.5
78.1
111.9
194.5
226.5
26.0
106.8
918.4
Mean
per
Tract
6.59
1.74
2.35
4.31
3.79
11.71
3.75
9.32
3.68
5.33
2.79
4.03
3.71
1.00
1.89
2.09
1.35
Total in
State
7,269
2,359
2,840
1,484
10,234
15,157
4,911
5,632
984
8,280
626
2,014
1,010
1,936
461
985
6,486
2016 Base Case
Average
Maternal
Daily
Mercury
Ingestion
(Hgl)
(ug/day)
3.77
5.04
1.76
5.05
2.37
2.77
1.70
4.98
2.38
2.95
3.41
1.56
5.39
2.98
1.77
3.60
3.54
Average
IQ Loss
per
Exposed
Child (dlQ)
0.13
0.18
0.06
0.18
0.08
0.10
0.06
0.18
0.08
0.10
0.12
0.05
0.19
0.10
0.06
0.13
0.12
Total IQ
Point
Losses
962.6
417.7
176.2
263.4
854.0
1,474.7
294.2
986.9
82.3
859.1
74.9
110.5
191.2
202.7
28.6
124.8
807.0
(continued)
-------
Table 4-5. Baseline Levels of Mercury Exposure and IQ Impacts Due to Freshwater Self-Caught Fish Consumption (continued)
u>
2005 Base Case
State
OH
OK
OR
PA
Rl
SC
SD
TN
TX
UT
VA
VT
WA
Wl
WV
WY
Number of
Census
Tracts with
Hg Samples
w/in 100
Miles
2,923
987
754
3,116
233
864
225
1,253
4,310
482
1,524
179
1,315
1,313
466
124
Number of
Prenatally Exposed
Children (NPA)
Mean
per
Tract
4.11
5.65
5.14
2.40
1.55
7.39
3.29
4.95
3.97
3.95
3.66
3.50
3.67
8.03
6.53
4.13
Total in
State
12,015
5,580
3,877
7,485
361
6,388
740
6,204
17,127
1,905
5,580
627
4,823
10,543
3,042
512
Average
Maternal
Daily
Mercury
Ingestion
(Hgl)
(ug/day)
1.61
3.07
2.80
2.30
6.01
4.43
1.77
3.01
2.83
2.05
2.61
3.85
1.69
2.77
2.10
1.97
Average IQ
Loss per
Exposed
Child (dlQ)
0.06
0.11
0.10
0.08
0.21
0.16
0.06
0.11
0.10
0.07
0.09
0.14
0.06
0.10
0.07
0.07
Number of
Prenatally Exposed
Children (NPA)
Total IQ
Point
Losses
678.8
602.9
382.1
605.9
76.2
995.4
45.9
656.7
1,701.2
137.3
512.7
84.8
287.2
1,026.2
224.3
35.5
Mean
per
Tract
3.93
5.73
5.43
2.31
1.53
7.59
3.20
5.06
4.32
4.68
3.82
3.37
3.90
7.85
6.10
3.99
Total in
State
11,489
5,653
4,095
7,194
356
6,559
719
6,335
18,633
2,254
5,820
604
5,133
10,309
2,840
495
2016 Base Case
Average
Maternal
Daily
Mercury
Ingestion
(Hgl)
(ug/day)
1.30
3.03
2.81
1.91
5.15
4.08
1.72
2.76
2.67
2.06
2.19
3.70
1.68
2.59
1.66
1.97
Average
IQ Loss
per
Exposed
Child (dlQ)
0.05
0.11
0.10
0.07
0.18
0.14
0.06
0.10
0.09
0.07
0.08
0.13
0.06
0.09
0.06
0.07
Total IQ
Point
Losses
527.0
601.4
404.3
482.2
64.5
941.0
43.6
615.5
1,748.9
163.5
448.7
78.6
302.8
938.1
166.1
34.3
-------
For each exposed population, we then estimated their average mercury ingestion rate
(Hgl) using Equation (4.4) and the IQ loss associated with this exposure level. As reported in
Table 4-5, in 2005, the average estimated mercury ingestion rate for the population of exposed
pregnant women was 3.04 ug/day. For 2016, the ingestion rate was estimated to be 2.84
ug/day (6.6% lower). The corresponding average IQ loss per prenatally exposed child was 0.11
in 2005 and 0.10 in 2016. Multiplying these average IQ losses by the size of the exposed
population, the total loss in IQ points due to mercury exposures through consumption of self-
caught freshwater fish was estimated to be 25,545 in 2005. For the 2016 base case, the total
decrease in IQ points was estimated to be 24,419 (4.4% lower).
4.8.2 IQ Loss and Economic Valuation Estimates
In addition to the base case scenarios described above, CMAQ air quality modeling runs
were used to estimate average mercury deposition levels for three emissions control scenarios:
• 2005 EGU Zero-Out. This scenario represents total mercury deposition from all
global natural and anthropogenic sources except for U.S. EGUs based on current-day
conditions.
• 2016 EGU Zero-Out. This scenario represents total mercury deposition from all
global natural and anthropogenic sources except for U.S. EGUs based on projected
2016 conditions, including future anticipated regulations (e.g., Transport Rule).
• 2016 Toxics Rule. This scenario represents total mercury deposition from all global
natural and anthropogenic sources based on projected 2016 conditions, including
future anticipated regulations (e.g., Transport Rule) and the Toxics Rule.
For these three scenarios, it was again assumed that the HUC-level fish tissue mercury
concentrations would change (relative to the 2005 base case) by the same percentage as the
change in modeled deposition levels.
Mercury exposure and IQ loss estimates were then derived for these three scenarios,
using the exposed population estimates for the relevant year (2005 or 2016) and the
corresponding mercury concentration estimates for the relevant emission scenario (zero-out or
Toxics Rule). In addition, the valuation methodology summarized in Section 4.7.2 (in particular,
Equation [4.7]) was applied to estimate the present value of IQ loss estimates for the two base
case and three emissions control scenarios.
To assess the aggregate benefits of reductions in EGU emissions, we evaluated five
emission reduction scenarios.
54
-------
• 2005 ECU zero-out (relative to 2005 base case)
• 2016 base case (relative to 2005 base case)
• 2016 ECU zero-out (relative to 2016 base case)
• 2016 Toxics Rule (relative to 2005 base case)
• 2016 Toxics Rule (relative to 2016 base case)
The benefits of each emission reduction scenario are calculated as the difference (i.e.,
decrease) in total present value of IQ losses between the selected emission control scenario
and the selected base case scenario.
4.8.3 Primary Results for National Analysis of Exposures from Recreational Freshwater Fish
Consumption
Table 4-6 summarizes the aggregate national IQ and present-value loss estimates for the
two base case and three emission control scenarios. The highest losses are estimated for the
2005 base case. For the population of prenatally exposed children included in the analysis
(almost 240,000, as reported in Table 4-5), mercury exposures under baseline conditions during
the year 2005 are estimated to have resulted in more than 25,500 IQ points lost. Assuming a 3%
discount rate, the present value of these losses ranges from $210 million to $290 million.8'This
range of total loss estimates is based on the range of per-IQ-point value (VIQ) estimates
summarized in Section 4.7.3. These losses represent expected present value of declines in
future net earnings over the entire lifetimes of the children who are prenatally exposed during
the year 2005. With a 7% discount rate, the present value range is considerably lower: $23
million to $51 million.
The lowest losses are estimated to result from the 2016 zero-out scenario, with total IQ
losses of less than 24,000 among roughly 244,000 prenatally exposed children and present
values of these losses ranging from $200 to $290 million (3% discount rate).
For the five emission reduction scenarios described above, Table 4-7 reports estimates
of aggregate nationwide benefits associated with reductions in mercury exposures and
resulting reductions in IQ losses. Most importantly, the benefits of the 2016 Toxics Rule
Monetized benefits estimates are for an immediate change in MeHg levels in fish. If a lag in the response of MeHg
levels in fish were assumed, the monetized benefits could be significantly lower, depending on the length of the
lag and the discount rate used. As noted in the discussion of the Mercury Maps modeling, the relationship
between deposition and fish tissue MeHg is proportional in equilibrium, but the MMaps approach does not
provide any information on the time lag of response.
55
-------
Table 4-6. Summary Estimates of the Aggregate Size and Present Value of IQ Losses Under
Alternative Base Case and Emissions Control Scenarios
Average IQ Loss per
Prenatally Exposed Child
Scenario (dlQ)
2005 base case 0.1068
2005 ECU zero-out 0.0985
2016 base case 0.1000
2016 ECU zero-out 0.0971
2016 Toxics Rule 0.0979
Value of Total IQ Losses in 2016
(millions of 2007$)
Total IQ Losses from
One Year of Exposure 3% Discount Rate
25,545
23,561
24,419
23,722
23,909
$210
$190
$200
$200
$200
to $310
to $290
to $300
to $290
to $290
7% Discount
Rate
$23
$22
$22
$22
$22
to
to
to
to
to
$51
$47
$49
$48
$48
Table 4-7. Aggregate Benefit Estimates for Reductions IQ Losses Associated with
Alternative Emissions Reduction Scenarios
Decrease in
Decrease in
Average IQ _ . ._
Total IQ
Emission Reduction Scenario ,, Losses from
Prenatally _
. y One Year of
Exposed
Child (dlQ)a EXP°SUre
(relative to 2005 base case) ' '
2016 base case
(relative to 2005 base case)
2016 ECU zero-out
(relative to 2016 base case)
2016 Toxics Rule n nnnn->
/ i *• * ™r,c u i 0.00893 1,636
(relative to 2005 base case)
2016 Toxics Rule
(relative to 2016 base case)
Value of Total IQ Losses in 2016
(millions of 2007$)
3% Discount Rate
$16 to
$9.3 to
$5.7 to
$13 to
$4.2 to
$24
$14
$8.5
$20
$6.2
7% Discount Rate
$1.8 to
$1.0 to
$0.6 to
$1.5 to
$0.47 to
$4.0
$2.3
$1.4
$3.3
$1.0
As reported in Table 4-5, the estimated number of prenatally exposed children is 239,174 in 2005 and 244,286 in
2016.
scenario (relative to the 2016 base case) are estimated to range between $4.2 million and $6.2
million (assuming a 3% discount rate), because of an estimated 511 point reduction in IQ losses.
These benefits are 73% as large as the benefits of the 2016 zero-out scenario (relative to the
56
-------
same 2016 base case). Relative to the 2005 base case, the benefits of the 2016 Toxics Rule
scenario range from $13 million to $20 million (3% discount). Despite growth in the exposed
population from 2005 to 2016, the changes from the 2005 base case to the 2016 base case
account for 69% of these benefits, while the changes from the 2016 base case to the 2016
Toxics Rule account for 31%.
4.8.5 Discussion of Assumptions, Limitations, and Uncertainties
Uncertainty regarding the model results and estimates reported in Section 4.8 can arise
from several sources. Some of the uncertainty can be attributed to model uncertainty. For
example, to estimate exposures a number of different modeling approaches have been
selected and combined. The separate model components are summarized in Figure 4-4 and
equations (4.) to (4.8), each of which simplifies potentially complex processes. The results,
therefore, depend importantly on how these models are selected, specified, and combined.
Another important source of uncertainty can be characterized as input or parameter
uncertainties. Each of the modeling components discussed in this report requires summary data
and estimates of key model parameters. For example, estimating IQ losses associated with
consumption of freshwater fish requires estimates of the size of the exposed population of
interest, the average mercury concentrations in consumed fish, the freshwater fish
consumption rate for the exposed population, and the concentration-response relationship
between mercury ingestion and IQ loss. All of these inputs are measured with some degree of
uncertainty and can affect, to differing degrees, the confidence range of our summary results.
The discussion below identifies and highlights some of the key model parameters, characterizes
the source and extent of uncertainties associated with them, and characterizes the potential
effects of these uncertainties on the model results.
To organize this discussion, we discuss different components of the modeling
framework separately. This section first discusses issues related to estimating the mercury
concentrations and then those related to estimating the exposed population. After that, it
discusses issues related to matching these two components and then concludes by discussing
the estimation of mercury ingestion through fish consumption.
4.8.5.1 Mercury Concentration Estimates
As described in Section 4.2.2, the mercury concentration estimates for the analysis
come from several different sources, including fish tissue sample data from the National Listing
of Fish Advisories (NLFA) and several other state- and national-level sources. These estimates
57
-------
were then used to approximate mercury concentrations across the study area. Some of the key
assumptions, limitations, and uncertainties associated with these estimates are the following:
• The fish tissue sampling data from various sources are subject to measurement and
reporting error and variability. The NLFA is the largest and most detailed source of
data on mercury in fish; however, even this system was not centrally designed (e.g.,
by EPA) using a common set of sampling and analytical methods. Rather, states
collected the data primarily to support the development of advisories, and the data
are submitted voluntarily to EPA. Each state uses different methods and criteria for
sampling and allocates different levels of resources to their monitoring programs. In
addition, there are uncertainties regarding the precise locations (lat/long
coordinates) of some of the samples. The heterogeneity and potential errors across
state sampling programs can bias the results in any direction and contribute to
uncertainty.
• The fish tissue sampling data were assigned as either lake or river samples, based on
the site name and/or the location coordinates mapped to the nearest type of
waterbody. This process also involves measurement error and may have resulted in
misclassifications for some of the samples. These errors are not expected to bias
results, but they contribute to uncertainty.
• The mercury concentration estimates used in the model were based on simple
temporal and spatial averages of reported fish tissue samples. This approach
assumes that the mercury samples are representative of "local" conditions (i.e.,
within the same HUC-12) in similar waterbodies (i.e., rivers or lakes). However, even
though states use a variety of approaches to monitor and sample fish tissue
contaminants, in some cases, the sampling sites are selected to target areas with
high levels of angler activity and/or a high level of pollution potential. To the extent
that sample selection procedures favor areas with relatively high mercury, the
spatial extrapolation methods used in this report will tend to overstate exposures.
These approaches also implicitly assume that mercury concentration estimates are
strongly spatially correlated, such that closer sampling sites (i.e., from the same HUC
or distance interval) provide more information about mercury concentrations than
more distant sites. To the extent that spatial correlation is weaker than assumed,
this will increase the degree of uncertainty in the modeling results.
• To generate average mercury fish tissue concentration estimates, all available
samples from the three main data sources (1995-2009) and from freshwater fish
larger the 7 inches were included in the analysis. Smaller fish were excluded to
better approximate concentrations in the types of fish that are more likely to be
consumed, and samples from years before 1995 were excluded to better represent
more recent conditions. Even with these sample selection procedures, average
concentration estimates from the retained samples may still under or overestimate
actual concentrations in currently consumed fish.
58
-------
4.8.5.2 Exposed Population Estimates
The methods described in Section 4.7 to estimate the total exposed population of
interest in 2005 and 2016 involve the following key assumptions, limitations, and uncertainties:
• The approach relies on data from the FHWAR to estimate state-level freshwater
angler activity levels, including freshwater fishing participation rates and lake-to-
river trip ratios. Each of these data elements is measured with some error in the
FHWAR, but they are based on a relatively large sample. More importantly the state-
level averages are applied to each modeled census tract in the state; therefore, the
model fails to capture within-state variation in these factors, which contributes to
uncertainty in the model estimates.
• The analysis also uses state-level fertility rate data to approximate the rate of
pregnancy among women of childbearing age in angler households for a smaller
geographic area. The state-level fertility rates from the National Vital Statistics are
estimated with relatively little error; however, applying these rates to specific
census tracts (and specifically to women in angler households) does involve
considerably more uncertainty.
• The approach assumes that, in each census tract, the percentage of women who live
in freshwater angler households (i.e., households with at least one freshwater
angler) is equal to the percentage of the state adult population that fishes. Applying
the state-level participation rate to approximate the conditions at a block level
creates uncertainty. More importantly, however, using individual-based fishing
participation rates to approximate household rates is likely to underestimate the
percentage of women living in freshwater angler households.9 Unfortunately, data
on household participation levels in freshwater fishing are not readily available.
• Census tract populations are only included in the model if they are matched to
distance intervals and waterbody types that have spatial overlap with at least one
HUC-12 sub-watershed containing a mercury concentrations estimate for that
waterbody type. By design, this approach undercounts the exposed population (by
roughly 40 to 45%) and, therefore, leads to underestimates of national aggregate
baseline exposures and risks and underestimates of the risk reductions and benefits
resulting from mercury emission reductions.
• All of the tract-level population estimates are based on Census 2000 data, which are
projected forward to 2005 and 2016 using county-level growth projections for the
subpopulations of interest from Woods and Poole (2008). Therefore, the 2005 and
2016 population estimates incorporate uncertainty from both the growth
9For example, hypothetically if one out of every three members in each household fished, the population rate
would be 33%, but the household rate would be 100%.
59
-------
projections themselves and from transferring the county-level growth estimates to
the tract level.
The purpose of the analysis of potentially high risk subpopulations is not to estimate the
size of the exposed population but rather to characterize the distribution of individual-level
risks in the subpopulations of interest. Nevertheless, the size and spatial distribution of the
total population in each group was used as a proxy for characterizing the spatial distribution of
pregnant women in freshwater fishing households in each group.
The main assumption underlying this approach is that the expected proportion of the
subgroup's population in each Census tract that consists of pregnant women in fishing
households is the same across the selected census tracts. The main limitation of this
assumption is that it does not allow or account for spatial variation in (1) the percentage of the
subpopulation that are women of childbearing age, (2) the percentage of these women that are
pregnant (i.e., fertility rate) and (3) the freshwater angler participation rates for the subgroups
of interest. Unfortunately, spatially varying data for the last component (fishing participation
rates among the subpopulations of interest) are not readily available. This assumption is not
expected to bias the results but it does contribute to uncertainty in the estimated distributions
of individual-level risks.
4.8.5.3 Matching of Exposed Populations to Mercury Concentrations
The methods described in Section 4.7 to match the exposed population estimates with
the corresponding mercury concentration estimates involve the following key assumptions,
limitations, and uncertainties:
• For the aggregate benefits analysis, tract-level exposed populations are assigned to
waterbody types based on state-level ratios of lake-to-river fishing days (from the
FHWAR). They are further assigned to distance intervals based on observed travel
distance patterns in national fishing data (NSRE, 1994). Both of these assignment
methods involve uncertainty, but particularly the second method because it is based
on much more aggregate data and on a much smaller and more dated sample of
anglers. This approach does not take into account the physical characteristics of the
area in which the population is located. In particular, the allocation of exposures to
lakes or rivers at different distances from each census tract does not take into
account the presence or number of these waterbodies in each distance interval.
Using these state and national level estimates to represent conditions at a local (i.e.,
census tract) level increases uncertainty in the model results, but it is not expected
to bias the results in either direction.
60
-------
• For the analysis of potentially high-risk populations, these methods and assumptions
were slightly modified. In particular, because these analyses focus on low-income
and/or subsistence fishing populations, all trips were assumed to occur within 20
miles of the census tract. Unfortunately, it is difficult to evaluate the accuracy of this
restriction due to limited data on travel distances for the subgroups of interest.
One potentially important factor that is not included for matching populations and
mercury concentrations is the effect offish consumption advisories on fishing behavior.
Evidence summarized in Jakus, McGuinness, and Krupnick (2002) suggests that awareness of
advisories by anglers is relatively low (less than 50%), and even those who are aware do not
always alter their fishing behavior. Nonetheless, anglers are less likely to fish in areas with
advisories. Unfortunately, we were not able to reliably quantify the reduction and
redistribution of fishing trips in either model to account for fish advisories. By excluding these
effects, the model estimates are likely to overstate mercury exposures.
4.8.5.4 Fish Consumption Estimates
One of the most influential variables in both modeling approaches is the rate of self-
caught freshwater fish consumption. The following key assumptions, limitation, and
uncertainties are associated with the methods used:
• For the aggregate analysis we have assumed 8 g/day for the general population in
freshwater angler households (based on recommendations in EPA's EFH).
Unfortunately, data are not available to reliably vary this rate with respect to
characteristics of the population across the entire study area. Uncertainty regarding
the true average fish consumption rate has a direct effect on uncertainty for the
aggregate exposure and benefit estimates. Because a single consumption rate is
applied uniformly across the entire exposed population and because it is a
multiplicative factor in the model, the two uncertainties are directly proportional to
one another. The recommended 8 g/day rate is based on four studies with mean
estimates ranging from 5 g/day (37% less than 8) to 17 g/day (113% more than 8). If
it is assumed that this range of estimates represents the uncertainty in the mean
freshwater fish consumption rate for the study population, then the resulting
uncertainty range for the estimated mean mercury ingestion level (and resulting IQ
loss) will also be between -37% and +113% of the mean mercury ingestion level.
• To analyze the distributions of individual-level risks in potentially high risk
subpopulations, we applied empirical distributions offish consumption rates for
specific subpopulations. One of the main limitations of this approach is that these
empirical distributions are based on relatively small and localized samples. In
particular, the estimated distribution of consumption rates for low-income African
American subsistence/recreational fishers in the Southeastern U.S. (see Table 4-3) is
based on a very small sample (N=39) drawn from one location (Columbia, SC). The
61
-------
sample sizes for the other groups, particularly the Hispanic (N= 45) and Laotian
(N=54) populations are also small; therefore, there is considerable uncertainty
regarding how well these empirical consumption rate distributions reflect actual
rates of consumption in the subpopulations of interest.
Another related and potentially influential variable in the modeling approach is the
assumed conversion factor for mercury concentrations between uncooked and cooked fish.
Studies have found that cooking fish tends to reduce the overall weight of fish by approximately
one-third (Great Lakes Sport Fish Advisory Task Force, 1993) without affecting the overall
amount of mercury. But these conversion rates depend on cooking practices and types offish.
Uncertainty regarding this conversion factor also has a proportionate effect on the modeling
results.
4.8.5.5 Measurement and Valuation of IQ Related Effects
The models for estimating and valuing IQ effects involve three main steps. The first step
is translating maternal mercury ingestion rates to mercury levels in hair. The second step is
translating differences in hair mercury concentrations during pregnancy to IQ changes in
offspring. The third step is translating IQ losses into expected reductions in lifetime earnings. As
discussed below, each of these steps also involves the following assumptions, limiations, and
uncertainties:.
• The conversion of mercury ingestion rate to mercury concentration in hair is based
on uncertainty analysis of a toxicokinetic model for estimating reference dose
(Swartout and Rice, 2000). The conversion factor was estimated by considering the
variability and uncertainty in various inputs used in deriving the dose including body
weight, hair-to-blood mercury ratio, half-life of MeHg in blood, and others.
Therefore, there is uncertainty regarding the conversion factor between hair
mercury concentration and mercury ingestion rate. Although, the median
conversion factor (0.08 u.g/kg-day/hair-ppm) is used, the 90% confidence interval is
from 0.037 to 0.16 u.g/kg day/hair-ppm. Any change in the conversion factor will
proportionately affect the benefits results because of the linearity of the model.
• The dose-response model used to estimate neurological effects on children because
of maternal mercury body burden is susceptible to various uncertainties. In
particular, there are three main concerns. First, there are other cognitive end-points
that have stronger association with MeHg than IQ point losses. Therefore, using IQ
points as a primary end point in the benefits assessment may underestimate the
impacts. Second, blood-to-hair ratio for mercury is uncertain, which can cause the
results from analyses based on mercury concentration in blood to be uncertain.
Third, uncertainty is associated with the epidemiological studies used in deriving the
dose-response models.
62
-------
With regard to the relationship between prenatal methylmercury exposure and
childhood IQ loss, we expect greater uncertainty in associated estimates of IQ loss as
exposure levels increase beyond those observed in the primary studies (i.e., Faroe
Islands, New Zealand, Seychelles Islands studies) used to derive the dose-response
function. In particular, high-end total exposure estimates for some of the
subsistence-level fishing subpopulations included in this assessment likely exceed
levels observed in the three primary studies.
To parameterize the dose-response relationship between maternal hair
concentrations and IQ loss for this analysis, we applied the results of an integrative
study by Axelrad et al. (2007). The implications of applying this study include the
following:
- This approach may confound potentially positive cognitive effects offish
consumption and, more specifically, omega-3 fatty acids. Results from Rice
(2010) offer a reasonable, but highly uncertain, estimate for offsetting the
possible downward bias resulting from the positive confounding effects of fatty
acids. Rice's high coefficient reflects the central estimate of Axelrad but adjusted
upwards by a factor of 1.5 to "acknowledge the recent argument of Budtz-
Jorgensen (2007) that the parameter estimates from these three epidemiological
studies (Faroe Islands, Seychelles Islands, New Zealand) may be biased
downward by a factor of approximately 2 because of failure to adequately
control for confounding." A third study, Oken (2008), analyzes a cohort in
Massachusetts and also seems to support a higher "Axelrad-plus" coefficient
range due to evidence of fatty acid confounding (i.e., positive cognitive effects of
fatty acids in fish may have previously led to underestimates of mercury-
attributable IQ loss). This study offers further qualitative support for a higher-
end estimate but is limited by the fact that it did not control for the children's
home environment, which is generally a significant factor in early cognitive
development.
- The dose-response coefficient from the Axelrad et al. study is sensitive to the
exclusion of one outlier data point from the Seychelles study. Including the
outlier would reduce the effect size by about 25 percent. If this outlier actually
reflects the true response for a subset of the populations, then risks (as
modeled) could be biased high specifically for this subpopulation
- Because the dose-response coefficient is applied uniformly across the entire
exposed population and is a multiplicative factor in the model, the uncertainty in
this parameter has a directly proportional effect on the reported risk and benefit
estimates. In other words, adjusting the absolute value of the dose-response
coefficient upward by a factor of 1.5 (i.e., based on Rice, 2010) would yield
reductions in IQ losses and benefits from mercury emission reductions that are
also greater by a factor of 1.5.
63
-------
• The valuation of IQ losses is based on a unit-value approach developed by EPA,
which estimates that the average effect of a 1-point reduction in IQ is to reduce the
present value of net future earnings. Three key assumptions of this unit-value
approach are that (1) there is a linear relationship between IQ changes and net
earnings losses, (2) the unit value applies to even very small changes in IQ, and
(3) the unit value will remain constant (in real present value terms) for several years
into the future. Each of these assumptions contributes to uncertainty in the result. In
particular the unit value estimate is itself subject to two main sources of uncertainty.
- The first source is directly related to uncertainties regarding the average
reductions in future earnings and years in school as a result of IQ changes. The
average percentage change estimates are subject to statistical error, modeling
uncertainties, and variability across the population. To address these
uncertainties we have included in the analysis and reported results a range of
values for this parameter, based on statistical analyses by Salkever (1995) and
Schwartz (1994).
- The second main source of uncertainty is the estimates of average lifetime
earnings and costs of schooling. Both of these estimates are derived from
national statistics from the early 1990s, but they are also subject to statistical
error, modeling uncertainties, and variability across the population. It is also
worth noting that the lost future earnings estimates do not include present value
estimates for nonwage/nonsalary earnings (i.e., fringe benefits) and household
(nonmarket) production. Based on the results of Grosse et al. (2009), including
these factors would increase the present value of median earnings (both explicit
and implicit) by a factor of roughly 1.9. However, it is not known whether IQ
changes have a similar effect on these other (implicit) earnings.
4.8.5.6 Unqualified Benefits
In addition to the uncertainties discussed above associated with the benefit analysis of
reducing exposures to MeHg from recreational freshwater angling, we are unable to quantify
several additional benefits, which adds to the uncertainties in the final estimate of benefits.
Table 4-20 displays the health and ecosystem effects associated with MeHg exposure
that are discussed in Section 4.2.2 for which we are currently unable to quantify. We note that
specifically with regard to health effects, the NRC (2000) provided the following observation:
"Neurodevelopmental effects are the most extensively studied sensitive end point for MeHg
exposure, but there remains some uncertainty about the possibility of other health effects at
low levels of exposure. In particular, there are indications of immune and cardiovascular
effects, as well as neurological effects emerging later in life, that have not been adequately
studied."
64
-------
Table 4-8. Unquantified Health and Ecosystem Effects Associated with Exposure to Mercury
Category of Health or Ecosystem Effect Potential Health or Ecosystem Outcomes
Neurologic Effects Impaired cognitive development
Problems with language
Abnormal social development
Other Health Effects3 Associations with genetic, autoimmune and cardiovascular effects
Ecological Effects3 Survival, reproductive, behavioral, and neurological effects in wildlife
(birds, fish, and mammals)
3 These are potential effects and are not quantified because the literature is either contradictory or incomplete.
In addition to the health and ecosystem effects that we are not able to quantify, we are
currently unable to quantify exposures to other segments of the U.S. population including
consumption of commercial seafood and freshwater fish (produced domestically as well as
imported from foreign sources) and consumption of recreationally caught seafood from
estuaries, coastal waters, and the deep ocean. These consumption pathways impact additional
recreational anglers who are not modeled in our benefits analysis as well as the general U.S.
population. Reductions in domestic fish tissue concentrations can also impact the health of
foreign consumers (consuming U.S. exports). Because of technical/theoretical limitations in the
science, EPA is unable to quantify the benefits associated with several of these fish
consumption pathways. For example, reductions in U.S. power plant emissions will result in a
lowering of the global burden of elemental mercury, which will likely produce some degree of
reduction in mercury concentrations for fish sourced from the open ocean and freshwater and
estuarine waterbodies in foreign countries. In the case of mercury reductions for fish in the
open ocean, complexities associated with modeling the linkage between changes in air
deposition of mercury and reductions in biomagnification and bioaccumulation up the food
chain (including open ocean dilution and the extensive migration patterns of certain high-
consumption fish such as tuna) prevent the modeling of fish obtained from the open ocean. In
the case of commercial fish obtained from foreign freshwater and estuarine waterbodies,
although technical challenges are associated with modeling long-range transport of elemental
mercury and the subsequent impacts to fish in these distant locations, additional complexities
such as accurately modeling patterns of harvesting and their linkages to commercial
consumption in the United States prevent inclusion of foreign-sourced freshwater and
estuarine fish in the primary benefits analysis.
Finally, with regard to commercially-produced freshwater fish sourced in the United
States (i.e., fish from catfish, bass, and trout farms), we are unable to accurately quantify
65
-------
effects from this consumption pathway because many of the fish farms operating in the United
States use feed that is not part of the aquatic food web of the waterbody containing the fish
farm (e.g., use of agricultural-based supplemental feed). In addition, many of the farms involve
artificial "constructed" waterbody environments that are atypical of aquatic environments
found in the regions where those farms are located, thereby limiting the applicability of
Mercury Maps' assumption in linking changes to mercury deposition to changes in mercury fish
tissue concentrations (e.g., waterbodies may have restricted or absent watersheds and
modified aquatic chemistry, which can effect methylation rates and impact time scales for
reaching steady-state mercury fish tissue concentrations following reductions in mercury
deposition). Some research indicates that the recycling of water at fish farms can magnify the
mercury concentration because the system does not remove mercury as it is recycled, while
newly deposited mercury is added to the system. Thus, additional research on aquaculture
farms is necessary before a benefits analysis can be conducted.
Exclusion of these commercial pathways means that this benefits analysis, although
covering an important source of exposure to domestic mercury emissions (recreational
freshwater anglers), excludes a large and potentially important group of individuals.
Recreational freshwater consumption accounts for approximately 10 to 17% of total U.S. fish
consumption, and 90% is derived from commercial sources (domestic seafood, aquaculture,
and imports) (EPA, 2005).
In conclusion, several unquantified benefits associated with this analysis add to the
overall uncertainty in estimating total benefits. To the extent that the proposed rule will reduce
mercury deposition from power plants over estuarine areas, coastal, and open ocean waters,
there would be a subsequent reduction in mercury fish tissue concentrations in these different
waterbodies and an associated benefit from avoided decrements in IQ and other known health
and ecosystem effects.
4.8.6 Overall Conclusions
4.8.6.1 Total Baseline Incidence of IQ Loss: Self-Caught Fish Consumption among Recreational
Freshwater Anglers
• Out of 64,500 census tracts in the continental U.S., 63,978 are located within 100
miles of at least one HUC-12 watershed with freshwater mercury fish tissue
sampling data, and therefore were included in the modeling of IQ loss among
recreational freshwater anglers.
66
-------
• Approximately 240,000 prenatally exposed children were modeled, with an average
IQ loss of 0.11 and 0.10 IQ points, respectively, from self-caught freshwater fish
consumption for the 2005 and 2016 base case scenarios.
• The highest estimated state-specific average IQ loss among children of freshwater
recreational anglers is 0.21 IQ points under the 2005 base case scenario, in both
California and Rhode Island.
• Total estimated IQ loss from self-caught freshwater fish consumption among
children of recreational anglers is estimated at 25,555 and 24,419 IQ points,
respectively, for the 2005 and 2016 base case scenarios.
• The present economic value of baseline IQ loss for 2005 ranges from $210 million to
$310 million, assuming a 3% discount rate, and from $23 million to $51 million,
assuming a 7% discount rate.
• The present economic value of baseline IQ loss for 2016 ranges from $200 million to
$300 million, assuming a 3% discount rate, and from $22 million to $49 million,
assuming a 7% discount rate.
4.8.6.2 Avoided IQ Loss and Economic Benefits due to Regulatory Action: Self-Caught Fish
Consumption among Recreational Freshwater Anglers
* Eliminating all mercury air emissions from U.S. EGUs in 2016 would result in an
estimated 0.00893 fewer IQ points lost per prenatally exposed child from self-caught
freshwater fish consumption, as compared with the 2005 base case scenario.
• The present economic value of avoided IQ loss from eliminating all mercury air
emissions from U.S. EGUs in 2016 is estimated at a range of $5.7 million to $8.5
million, assuming a 3% discount rate, and $0.6 million to $1.4 million, assuming a 7%
discount rate.
• Reduced mercury air emissions due to implementation of the Toxics Rule in 2016
would result in an estimated 0.00209 fewer IQ points lost per prenatally exposed
child from self-caught freshwater fish consumption, as compared with the 2016 base
case scenario.
• The present economic value of avoided IQ loss from reduced mercury air emissions
due to implementation of the Toxics Rule in 2016 is estimated at a range of $4.2
million to $6.2 million, assuming a 3% discount rate, and $0.47 million to $1 million,
assuming a 7% discount rate.
67
-------
4.9 Benefits Associated with Reductions in Other HAP than Mercury
Even though emissions of air toxics from all sources in the U.S. declined by
approximately 42 percent since 1990, the 2005 National-Scale Air Toxics Assessment (NATA)
predicts that most Americans are exposed to ambient concentrations of air toxics at levels that
have the potential to cause adverse health effects (U.S. EPA, 2011d).10 The levels of air toxics to
which people are exposed vary depending on where people live and work and the kinds of
activities in which they engage. In order to identify and prioritize air toxics, emission source
types and locations that are of greatest potential concern, U.S. EPA conducts the NATA.n The
most recent NATA was conducted for calendar year 2005 and was released in March 2011.
NATA includes four steps:
1) Compiling a national emissions inventory of air toxics emissions from outdoor
sources
2) Estimating ambient and exposure concentrations of air toxics across the United
States
3) Estimating population exposures across the United States
4) Characterizing potential public health risk due to inhalation of air toxics including
both cancer and noncancer effects
Based on the 2005 NATA, EPA estimates that about 5 percent of census tracts
nationwide have increased cancer risks greater than 100 in a million. The average national
cancer risk is about 50 in a million. Nationwide, the key pollutants that contribute most to the
overall cancer risks are formaldehyde and benzene.12'13 Secondary formation (e.g.,
formaldehyde forming from other emitted pollutants) was the largest contributor to cancer
risks, while stationary, mobile and background sources contribute almost equal portions of the
remaining cancer risk.
iaThe 2005 NATA is available on the Internet at http://www.epa.gov/ttn/atw/nata2005/.
nThe NATA modeling framework has a number of limitations that prevent its use as the sole basis for setting
regulatory standards. These limitations and uncertainties are discussed on the 2005 NATA website. Even so,
this modeling framework is very useful in identifying air toxic pollutants and sources of greatest concern,
setting regulatory priorities, and informing the decision making process. U.S. EPA. (2011) 2005 National-Scale
Air Toxics Assessment, http://www.epa.gov/ttn/atw/nata2005/
"Details on EPA's approach to characterization of cancer risks and uncertainties associated with the 2005 NATA
risk estimates can be found at http://www.epa.gov/ttn/atw/natal999/riskbg.htmlftZ2.
13Details about the overall confidence of certainty ranking of the individual pieces of NATA assessments including
both quantitative (e.g., model-to-monitor ratios) and qualitative (e.g., quality of data, review of emission
inventories) judgments can be found at http://www.epa.gov/ttn/atw/nata/roy/pagel6.html.
68
-------
Noncancer health effects can result from chronic,14 subchronic,15 or acute16 inhalation
exposures to air toxics, and include neurological, cardiovascular, liver, kidney, and respiratory
effects as well as effects on the immune and reproductive systems. According to the 2005
NAT A, about three-fourths of the U.S. population was exposed to an average chronic
concentration of air toxics that has the potential for adverse noncancer respiratory health
effects. Results from the 2005 NATA indicate that acrolein is the primary driver for noncancer
respiratory risk.
Figure 4-5and Figure 46 depict the estimated census tract-level carcinogenic risk and
noncancer respiratory hazard from the assessment. It is important to note that large reductions
in HAP emissions may not necessarily translate into significant reductions in health risk because
toxicity varies by pollutant, and exposures may or may not exceed levels of concern. For
example, acetaldehyde mass emissions are more than double acrolein emissions on a national
basis, according to EPA's 2005 National Emissions Inventory (NEI). However, the Integrated Risk
Information System (IRIS) reference concentration (RfC) for acrolein is considerably lower than
that for acetaldehyde, suggesting that acrolein could be potentially more toxic than
acetaldehyde. 17 Thus, it is important to account for the toxicity and exposure, as well as the
mass of the targeted emissions.
Due to methodology and data limitations, we were unable to estimate the benefits
associated with the hazardous air pollutants that would be reduced as a result of these rules..
In a few previous analyses of the benefits of reductions in HAPs, EPA has quantified the benefits
of potential reductions in the incidences of cancer and non-cancer risk (e.g., U.S. EPA, 1995). In
those analyses, EPA relied on unit risk factors (URF) developed through risk assessment
procedures.18 These URFs are designed to be conservative, and as such, are more likely to
represent the high end of the distribution of risk rather than a best or most likely estimate of
risk. As the purpose of a benefit analysis is to describe the benefits most likely to occur from a
14Chronic exposure is defined in the glossary of the Integrated Risk Information (IRIS) database
(http://www.epa.gov/iris) as repeated exposure by the oral, dermal, or inhalation route for more than
approximately 10% of the life span in humans (more than approximately 90 days to 2 years in typically used
laboratory animal species).
15Defined in the IRIS database as repeated exposure by the oral, dermal, or inhalation route for more than 30 days,
up to approximately 10% of the life span in humans (more than 30 days up to approximately 90 days in typically
used laboratory animal species).
16Defined in the IRIS database as exposure by the oral, dermal, or inhalation route for 24 hours or less.
"Details on the derivation of IRIS values and available supporting documentation for individual chemicals (as well
as chemical values comparisons) can be found at http://cfpub.epa.gov/ncea/iris/compare.cfm.
18The unit risk factor is a quantitative estimate of the carcinogenic potency of a pollutant, often expressed as the
probability of contracting cancer from a 70-year lifetime continuous exposure to a concentration of one u.g/m3
of a pollutant.
69
-------
reduction in pollution, use of high-end, conservative risk estimates would overestimate the
benefits of the regulation. While we used high-end risk estimates in past analyses, advice from
the EPA's Science Advisory Board (SAB) recommended that we avoid using high-end estimates
Figure 4-5. Estimated Chronic Census Tract Carcinogenic Risk from HAP Exposure from
Outdoor Sources (2005 NATA)
70
-------
Total Respiratory
Hazard Index
0-1
1-5
^B 5-10
•• 10-15
j^H 15-20
^B >20
Zero Population Tracts
Figure 4-6. Estimated Chronic Census Tract Noncancer (Respiratory) Risk from HAP
Exposure from Outdoor Sources (2005 NATA)
in benefit analyses (U.S. EPA-SAB, 2002). Since this time, EPA has continued to develop better
methods for analyzing the benefits of reductions in HAPs.
As part of the second prospective analysis of the benefits and costs of the Clean Air Act
(U.S. EPA, 2011a), EPA conducted a case study analysis of the health effects associated with
reducing exposure to benzene in Houston from implementation of the Clean Air Act (lEc, 2009).
While reviewing the draft report, EPA's Advisory Council on Clean Air Compliance Analysis
concluded that "the challenges for assessing progress in health improvement as a result of
reductions in emissions of hazardous air pollutants (HAPs) are daunting...due to a lack of
exposure-response functions, uncertainties in emissions inventories and background levels, the
difficulty of extrapolating risk estimates to low doses and the challenges of tracking health
progress for diseases, such as cancer, that have long latency periods" (U.S. EPA-SAB, 2008).
71
-------
In 2009, EPA convened a workshop to address the inherent complexities, limitations,
and uncertainties in current methods to quantify the benefits of reducing HAPs.
Recommendations from this workshop included identifying research priorities, focusing on
susceptible and vulnerable populations, and improving dose-response relationships (Gwinn
etal.,2011).
In summary, monetization of the benefits of reductions in cancer incidences requires
several important inputs, including central estimates of cancer risks, estimates of exposure to
carcinogenic HAPs, and estimates of the value of an avoided case of cancer (fatal and non-
fatal). Due to methodology and data limitations, we did not attempt to monetize the health
benefits of reductions in HAPs in this analysis. Instead, we provide a qualitative analysis of the
health effects associated with the HAPs anticipated to be reduced by these rules and we
summarize the results of the residual risk assessment for the Risk and Technology Review (RTR).
EPA remains committed to improving methods for estimating HAP benefits by continuing to
explore additional concepts of benefits, including changes in the distribution of risk.
Available emissions data show that several different HAPs are emitted from oil and
natural gas operations, either from equipment leaks, processing, compressing, transmission and
distribution, or storage tanks. Emissions of eight HAPs make up a large percentage the total
HAP emissions by mass from the oil and gas sector: toluene, hexane, benzene, xylenes (mixed),
ethylene glycol, methanol, ethyl benzene, and 2,2,4-trimethylpentane (U.S. EPA, 2011a). In the
subsequent sections, we describe the health effects associated with the main HAPs of concern
from the oil and natural gas sector: benzene, toluene, carbonyl sulfide, ethyl benzene, mixed
xylenes, and n-hexane. These rules combined are anticipated to avoid or reduce 58,000 tons of
HAPs per year. With the data available, it was not possible to estimate the tons of each
individual HAP that would be reduced.
EPA conducted a residual risk assessment for the NESHAP rule (U.S. EPA, 2011c). The
results for oil and gas production indicate that maximum lifetime individual cancer risks could
be 30 in-a-million for existing sources before and after controls with a cancer incidence of 0.02
before and after controls. For existing natural gas transmission and storage, the maximum
individual cancer risk decreases from 90-in-a-million before controls to 20-in-a-million after
controls with a cancer incidence that decreases from 0.001 before controls to 0.0002 after
controls. Benzene is the primary cancer risk driver. The results also indicate that significant
noncancer impacts from existing sources are unlikely, especially after controls. EPA did not
conduct a risk assessment for new sources affected by the NSPS. However, it is important to
note that the magnitude of the HAP emissions avoided by new sources with the NSPS are more
72
-------
than an order of magnitude higher than the HAP emissions reduced from existing sources with
the NESHAP.
4.9.1 Hazards
Emissions data collected during development of this proposed rule show that HCI
emissions represent the predominant HAP emitted by industrial boilers. Coal- and oil-fired
EGUs emit lesser amounts of HF, chlorine, metals (As, Cd, Cr, Hg, Mn, Ni, and Pb), and organic
HAP emissions. Although numerous organic HAP may be emitted from coal- and oil-fired EGUs,
only a few account for essentially all the mass of organic HAP emissions. These organic HAP are
formaldehyde, benzene, and acetaldehyde.
Exposure to high levels of these HAP is associated with a variety of adverse health
effects. These adverse health effects include chronic health disorders (e.g., irritation of the
lung, skin, and mucus membranes, effects on the central nervous system, and damage to the
kidneys), and acute health disorders (e.g., lung irritation and congestion, alimentary effects
such as nausea and vomiting, and effects on the kidney and central nervous system). We have
classified three of the HAP as human carcinogens and five as probable human carcinogens. The
following sections briefly discuss the main health effects information we have regarding the key
HAPs emitted by EGUs.
4.9.1.1 Acetaldehyde
Acetaldehyde is classified in EPA's IRIS database as a probable human carcinogen, based
on nasal tumors in rats, and is considered toxic by the inhalation, oral, and intravenous
routes.19 Acetaldehyde is reasonably anticipated to be a human carcinogen by the U.S.
Department of Health and Human Services (DHHS) in the 11th Report on Carcinogens and is
classified as possibly carcinogenic to humans (Group 2B) by the IARC.20'21 The primary
19U.S. Environmental Protection Agency (U.S. EPA). 1991. Integrated Risk Information System File of Acetaldehyde.
Research and Development, National Center for Environmental Assessment, Washington, DC. This material is
available electronically at http://www.epa.gov/iris/subst/0290.htm.
20U.S. Department of Health and Human Services National Toxicology Program llth Report on Carcinogens
available at: http://ntp.niehs.nih.gov/go/16183.
^International Agency for Research on Cancer (IARC). 1999. Re-evaluation of some organic chemicals, hydrazine,
and hydrogen peroxide. IARC Monographs on the Evaluation of Carcinogenic Risk of Chemical to Humans, Vol
71. Lyon, France.
73
-------
noncancer effects of exposure to acetaldehyde vapors include irritation of the eyes, skin, and
respiratory tract.22
4.9.1.2 Arsenic
Arsenic, a naturally occurring element, is found throughout the environment and is
considered toxic through the oral, inhalation and dermal routes. Acute (short-term) high-level
inhalation exposure to As dust or fumes has resulted in gastrointestinal effects (nausea,
diarrhea, abdominal pain, and gastrointestinal hemorrhage); central and peripheral nervous
system disorders have occurred in workers acutely exposed to inorganic As. Chronic (long-term)
inhalation exposure to inorganic As in humans is associated with irritation of the skin and
mucous membranes. Chronic inhalation can also lead to conjunctivitis, irritation of the throat
and respiratory tract and perforation of the nasal septum.23 Chronic oral exposure has resulted
in gastrointestinal effects, anemia, peripheral neuropathy, skin lesions, hyperpigmentation, and
liver or kidney damage in humans. Inorganic As exposure in humans, by the inhalation route,
has been shown to be strongly associated with lung cancer, while ingestion of inorganic As in
humans has been linked to a form of skin cancer and also to bladder, liver, and lung cancer. EPA
has classified inorganic As as a Group A, human carcinogen.24
4.9.1.3 Benzene
The EPA's IRIS database lists benzene as a known human carcinogen (causing leukemia)
by all routes of exposure, and concludes that exposure is associated with additional health
effects, including genetic changes in both humans and animals and increased proliferation of
bone marrow cells in mice.25'26'27 EPA states in its IRIS database that data indicate a causal
22U.S. Environmental Protection Agency (U.S. EPA). 1991. Integrated Risk Information System File of Acetaldehyde.
Research and Development, National Center for Environmental Assessment, Washington, DC. This material is
available electronically at http://www.epa.gov/iris/subst/0290.htm.
23Agency for Toxic Substances and Disease Registry (ATSDR). Medical Management Guidelines for Arsenic. Atlanta,
GA: U.S. Department of Health and Human Services. Available on the Internet at
24U.S. Environmental Protection Agency (U.S. EPA). 1998. Integrated Risk Information System File for Arsenic.
Research and Development, National Center for Environmental Assessment, Washington, DC. This material is
available electronically at: http://www.epa.gov/iris/subst/0278.htm.
25U.S. Environmental Protection Agency (U.S. EPA). 2000. Integrated Risk Information System File for Benzene.
Research and Development, National Center for Environmental Assessment, Washington, DC. This material is
available electronically at: http://www.epa.gov/iris/subst/0276.htm.
^International Agency for Research on Cancer, IARC monographs on the evaluation of carcinogenic risk of
chemicals to humans, Volume 29, Some industrial chemicals and dyestuffs, International Agency for Research
on Cancer, World Health Organization, Lyon, France, p. 345-389,1982.
27lrons, R.D.; Stillman, W.S.; Colagiovanni, D.B.; Henry, V.A. (1992) Synergistic action of the benzene metabolite
hydroquinone on myelopoietic stimulating activity of granulocyte/macrophage colony-stimulating factor in
vitro, Proc. Natl. Acad. Sci. 89:3691-3695.
74
-------
relationship between benzene exposure and acute lymphocytic leukemia and suggest a
relationship between benzene exposure and chronic non-lymphocytic leukemia and chronic
lymphocytic leukemia. The IARC has determined that benzene is a human carcinogen and the
DHHS has characterized benzene as a known human carcinogen.28'29
A number of adverse noncancer health effects including blood disorders, such as
preleukemia and aplastic anemia, have also been associated with long-term exposure to
benzene.30'31
4.9.1.4 Cadmium
Breathing air with lower levels of Cd over long periods of time (for years) results in a
build-up of Cd in the kidney, and if sufficiently high, may result in kidney disease. Lung cancer
has been found in some studies of workers exposed to Cd in the air and studies of rats that
inhaled Cd. The U.S. DHHS has determined that Cd and Cd compounds are known human
carcinogens. The IARC has determined that Cd is carcinogenic to humans. EPA has determined
that Cd is a probable human carcinogen.32
4.9.1.5 Chlorine
The acute (short term) toxic effects of CI2 are primarily due to its corrosive properties.
Chlorine is a strong oxidant that upon contact with water moist tissue (e.g., eyes, skin, and
upper respiratory tract) can produce major tissue damage.33 Chronic inhalation exposure to low
concentrations of CI2 (1 to 10 parts per million, ppm) may cause eye and nasal irritation, sore
throat, and coughing. Chronic exposure to CI2, usually in the workplace, has been reported to
cause corrosion of the teeth. Inhalation of higher concentrations of CI2 gas (greater than
15 ppm) can rapidly lead to respiratory distress with airway constriction and accumulation of
fluid in the lungs (pulmonary edema). Exposed individuals may have immediate onset of rapid
breathing, blue discoloration of the skin, wheezing, rales or hemoptysis (coughing up blood or
^International Agency for Research on Cancer (IARC). 1987. Monographs on the evaluation of carcinogenic risk of
chemicals to humans, Volume 29, Supplement 7, Some industrial chemicals and dyestuffs, World Health
Organization, Lyon, France.
29U.S. Department of Health and Human Services National Toxicology Program llth Report on Carcinogens
available at: http://ntp.niehs.nih.gov/go/16183.
30Aksoy, M. (1989). Hematotoxicity and carcinogenicity of benzene. Environ. Health Perspect. 82: 193-197.
31Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art Reviews. 3: 541-554.
32Agency for Toxic Substances and Disease Registry (ATSDR). 2008. Public Health Statement for Cadmium. CAS#
1306-19-0. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service. Available on the
Internet at .
33Agency for Toxic Substances and Disease Registry (ATSDR). Medical Management Guidelines for Chlorine.
Atlanta, GA: U.S. Department of Health and Human Services.
http://www.atsdr.cdc.gov/mmg/mmg.asp?id=198&tid=36.
75
-------
blood-stain sputum). Intoxication with high concentrations of CI2 may induce lung collapse.
Exposure to CI2 can lead to reactive airways dysfunction syndrome (RADS), a chemical irritant-
induced type of asthma. Dermal exposure to CI2 may cause irritation, burns, inflammation and
blisters. EPA has not classified CI2 with respect to carcinogenicity.
4.9.1.6 Chromium
Chromium may be emitted in two forms, trivalent Cr (Cr+3) or hexavalent Cr (Cr+6). The
respiratory tract is the major target organ for Cr+6 toxicity, for acute and chronic inhalation
exposures. Shortness of breath, coughing, and wheezing have been reported from acute
exposure to Cr+6, while perforations and ulcerations of the septum, bronchitis, decreased
pulmonary function, pneumonia, and other respiratory effects have been noted from chronic
exposures. Limited human studies suggest that Cr+6 inhalation exposure may be associated with
complications during pregnancy and childbirth, but there are no supporting data from animal
studies reporting reproductive effects from inhalation exposure to Cr+6. Human and animal
studies have clearly established the carcinogenic potential of Cr+6 by the inhalation route,
resulting in an increased risk of lung cancer. EPA has classified Cr+6 as a Group A, human
carcinogen. Trivalent Cr is less toxic than Cr+6. The respiratory tract is also the major target
organ for Cr+3 toxicity, similar to Cr+6. EPA has not classified Cr+3 with respect to carcinogenicity.
4.9.1.7 Formaldehyde
Since 1987, EPA has classified formaldehyde as a probable human carcinogen based on
evidence in humans and in rats, mice, hamsters, and monkeys.34 EPA is currently reviewing
recently published epidemiological data. After reviewing the currently available epidemiological
evidence, the IARC (2006) characterized the human evidence for formaldehyde carcinogenicity
as "sufficient," based upon the data on nasopharyngeal cancers; the epidemiologic evidence on
leukemia was characterized as "strong."35 EPA is reviewing the recent work cited above from
the NCI and NIOSH, as well as the analysis by the CUT Centers for Health Research and other
studies, as part of a reassessment of the human hazard and dose-response associated with
formaldehyde.
Formaldehyde exposure also causes a range of noncancer health effects, including
irritation of the eyes (burning and watering of the eyes), nose and throat. Effects from repeated
exposure in humans include respiratory tract irritation, chronic bronchitis and nasal epithelial
34U.S. EPA. 1987. Assessment of Health Risks to Garment Workers and Certain Home Residents from Exposure to
Formaldehyde, Office of Pesticides and Toxic Substances, April 1987.
iternational Agency for Research on Cancer (2006) Formaldehyde, 2-B
ol. Monographs Volume 88. World Health Organization, Lyon, France.
^International Agency for Research on Cancer (2006) Formaldehyde, 2-Butoxyethanol and l-tert-Butoxypropan-2-
76
-------
lesions such as metaplasia and loss of cilia. Animal studies suggest that formaldehyde may also
cause airway inflammation—including eosinophil infiltration into the airways. There are several
studies that suggest that formaldehyde may increase the risk of asthma—particularly in the
young.36'37
4.9.1.8 Hydrogen Chloride
Hydrogen chloride is a corrosive gas that can cause irritation of the mucous membranes
of the nose, throat, and respiratory tract. Brief exposure to 35 ppm causes throat irritation, and
levels of 50 to 100 ppm are barely tolerable for 1 hour.38 The greatest impact is on the upper
respiratory tract; exposure to high concentrations can rapidly lead to swelling and spasm of the
throat and suffocation. Most seriously exposed persons have immediate onset of rapid
breathing, blue coloring of the skin, and narrowing of the bronchioles. Exposure to HCI can lead
to RADS, a chemically- or irritant-induced type of asthma. Children may be more vulnerable to
corrosive agents than adults because of the relatively smaller diameter of their airways.
Children may also be more vulnerable to gas exposure because of increased minute ventilation
per kg and failure to evacuate an area promptly when exposed. Hydrogen chloride has not been
classified for carcinogenic effects.39
4.9.1.9 Hydrogen Fluoride
Acute (short-term) inhalation exposure to gaseous HF can cause severe respiratory
damage in humans, including severe irritation and pulmonary edema. Chronic (long-term) oral
exposure to fluoride at low levels has a beneficial effect of dental cavity prevention and may
also be useful for the treatment of osteoporosis. Exposure to higher levels of fluoride may
36Agency for Toxic Substances and Disease Registry (ATSDR). 1999. Toxicological profile for Formaldehyde. Atlanta,
GA: U.S. Department of Health and Human Services, Public Health Service.
http://www.atsdr.cdc.gov/toxprofiles/tplll.html
37WHO (2002) Concise International Chemical Assessment Document 40: Formaldehyde. Published under the joint
sponsorship of the United Nations Environment Programme, the International Labour Organization, and the
World Health Organization, and produced within the framework of the Inter-Organization Programme for the
Sound Management of Chemicals. Geneva.
38Agency for Toxic Substances and Disease Registry (ATSDR). Medical Management Guidelines for Hydrogen
Chloride. Atlanta, GA: U.S. Department of Health and Human Services. Available online at
http://www.atsdr. cdc.gov/mmg/mmg. asp?id=758&tid=147#bookmark02.
39U.S. Environmental Protection Agency (U.S. EPA). 1995. Integrated Risk Information System File of Hydrogen
Chloride. Research and Development, National Center for Environmental Assessment, Washington, DC. This
material is available electronically at .http://www.epa.gov/iris/subst/0396.htm.
77
-------
cause dental fluorosis. One study reported menstrual irregularities in women occupationally
exposed to fluoride via inhalation. The EPA has not classified HF for carcinogenicity40.
4.9.1.10 Lead
The main target for Pb toxicity is the nervous system, both in adults and children. Long-
term exposure of adults to Pb at work has resulted in decreased performance in some tests that
measure functions of the nervous system. Lead exposure may also cause weakness in fingers,
wrists, or ankles. Lead exposure also causes small increases in blood pressure, particularly in
middle-aged and older people. Lead exposure may also cause anemia.
Children are more sensitive to the health effects of Pb than adults. No safe blood Pb
level in children has been determined. At lower levels of exposure, Pb can affect a child's
mental and physical growth. Fetuses exposed to Pb in the womb may be born prematurely and
have lower weights at birth. Exposure in the womb, in infancy, or in early childhood also may
slow mental development and cause lower intelligence later in childhood. There is evidence
that these effects may persist beyond childhood.41
There are insufficient data from epidemiologic studies alone to conclude that Pb causes
cancer (is carcinogenic) in humans. The DHHS has determined that Pb and Pb compounds are
reasonably anticipated to be human carcinogens based on limited evidence from studies in
humans and sufficient evidence from animal studies, and the EPA has determined that Pb is a
probable human carcinogen.
4.9.1.11 Manganese
Health effects in humans have been associated with both deficiencies and excess
intakes of Mn. Chronic exposure to high levels of Mn by inhalation in humans results primarily
in central nervous system effects. Visual reaction time, hand steadiness, and eye-hand
coordination were affected in chronically-exposed workers. Manganism, characterized by
feelings of weakness and lethargy, tremors, a masklike face, and psychological disturbances,
may result from chronic exposure to higher levels. Impotence and loss of libido have been
40U.S. Environmental Protection Agency. Health Issue Assessment: Summary Review of Health Effects Associated
with Hydrogen Fluoride and Related Compounds. EPA/600/8-89/002F. Environmental Criteria and Assessment
Office, Office of Health and Environmental Assessment, Office of Research and Development, Cincinnati, OH.
1989.
41Agency for Toxic Substances and Disease Registry (ATSDR). 2007. Public Health Statement for Lead. CAS#: 7439-
92-1. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service. Available on the
Internet at < http://www.atsdr.cdc.gov/ToxProfiles/phsl3.html>.
78
-------
noted in male workers afflicted with manganism attributed to inhalation exposures. The EPA
has classified Mn in Group D, not classifiable as to carcinogenicity in humans.42
4.9.1.12 Nickel
Respiratory effects have been reported in humans from inhalation exposure to Ni. No
information is available regarding the reproductive or developmental effects of Ni in humans,
but animal studies have reported such effects. Human and animal studies have reported an
increased risk of lung and nasal cancers from exposure to Ni refinery dusts and nickel
subsulfide. The EPA has classified nickel subsulfide as a human carcinogen and nickel carbonyl
as a probable human carcinogen.43'44 The IARC has classified Ni compounds as carcinogenic to
humans.45
4.9.1.13 Selenium
Acute exposure to elemental Se, hydrogen selenide, and selenium dioxide (Se02) by
inhalation results primarily in respiratory effects, such as irritation of the mucous membranes,
pulmonary edema, severe bronchitis, and bronchial pneumonia. One Se compound, selenium
sulfide, is carcinogenic in animals exposed orally. EPA has classified elemental Se as a Group D,
not classifiable as to human carcinogenicity, and selenium sulfide as a Group B2, probable
human carcinogen.
4.10 References
Adams, Evan M., and Frederick, Peter C. 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, 2008.
Agency for Toxic Substances and Disease Registry (ATSDR). 1999. Toxicological Profile for
Mercury. U.S. Department of Health and Human Services, Public Health Service, Atlanta,
GA.
Amorim, M.I.M., D. Mergler, M.O. Bahia, H. Dubeau, D. Miranda, J. Lebel, R.R. Burbano, and M.
Lucotte. 2000. Cytogenetic damage related to low levels of methyl mercury
contamination in the Brazilian Amazon. An. Acad. Bras. Cienc. 72(4): 497-507.
42U.S. Environmental Protection Agency. Integrated Risk Information System (IRIS) on Manganese. National Center
for Environmental Assessment, Office of Research and Development, Washington, DC. 1999.
43U.S. Environmental Protection Agency. Integrated Risk Information System (IRIS) on Nickel Subsulfide. National
44
Center for Environmental Assessment, Office of Research and Development, Washington, DC. 1999.
U.S. Environmental Protection Agency. Integrated Risk Information System (IRIS) on Nickel Carbonyl. National
Center for Environmental Assessment, Office of Research and Development, Washington, DC. 1999.
ickel (IARC Summary & Evaluation , Volume 49,1990),
http://www.inchem.org/documents/iarc/vol49/nickel.html
79
-------
Amyot, M., Lean, D.R.S., Poissant, L. and Doyon, M.-R., 2000. Distribution and transformation of
elemental mercury in the St. Lawrence River and Lake Ontario. Canadian Journal of
Fisheries and Aquatic Sciences, 57 (Suppl. 1): 155-163.
ATSDR (1999). Toxicological Profile for Mercury. Agency for Toxic Substances and Disease
Registry.
Axelrad, D. A.; Bellinger, D. C.; Ryan, L. M.; Woodruff, T. J. (2007). Dose-response relationship of
prenatal mercury exposure and IQ: an integrative analysis of epidemiologic data.
Environmental Health Perspectives. 2007, 115, 609-615.
Babiarz, C.L. et al., 2001. Partitioning of total mercury and methylmercury to the colloidal phase
in freshwaters. Environmental Science and Technology, 35(24): 4773-4782.
Barkay, T., Miller, S.M. and Summers, A.O., 2003. Bacterial mercury resistance from atoms to
ecosystems. FEMS Microbiology Reviews, 27: 355-384.
Benoit, J.M., Gilmour, C.C. and Mason, R.P., 2001. The influence of sulfide on solid-phase
mercury bioavailability for methylation by pure cultures of Desulfobulbus propionicus
(Ipr3). Environmental Science and Technology, 35(1): 127-132.
Benoit, J.M., Gilmour, C.C., Heyes, A., Mason, R.P. and Miller, C, 2003. Geochemical and
Biological Controls over Methylmercury Production and Degradation in Aquatic Systems,
Biogeochemistry of Environmentally Important Trace Metals. ACS Symposium Series
835.
Benoit, J.M., Gilmour, C.C., Mason, R.P. and Heyes, A., 1999a. Sulfide controls on mercury
speciation and bioavailability to methylating bacteria in sediment pore waters.
Environmental Science and Technology, 33(6): 951-957.
Benoit, J.M., Gilmour, C.C., Mason, R.P., Riedel, G.S. and Reidel, G.F., 1998. Behavior of mercury
in the Patuxent River estuary. Biogeochemistry, 40: 249-265.
Benoit, J.M., Mason, R.P. and Gilmour, C.C., 1999b. Estimation of mercury-sulfide speciation in
sediment pore waters using octanol-water partitioning and implications for availability
to methylating bacteria. Environmental Toxicology and Chemistry, 18(10): 2138-2141.
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.
80
-------
Beyers, D.W., Rice, J.A. and Clements, W.H., 1999. Evaluating biological significance of chemical
exposure to fish using a bioenergetics-based stressor-response model. Canadian Journal
of Fisheries and Aquatic Sciences, 56: 823-829.
Bigham, G.N. and Vandal, G.M., 1994. A drainage basin perspective of mercury transport and
bioaccumulation: Onondaga Lake, New York, Twelfth International Neurotoxicology
Conference, Hot Springs, Arkansas USA.
Bloom, N.S., 1992. On the chemical form of mercury in edible fish and marine invertebrate
tissue. Canadian Journal of Fisheries and Aquatic Sciences, 49: 1010-1017.
Branfireun, B. and Roulet, N., 2002. Controls on the fate and transport of methylmercury in a
boreal headwater catchment, northwestern Ontario, Canada. Hydrology and Earth
System Sciences, 6(4): 785-794.
Brasso, Rebecka L, and Cristol, Daniel A. Effects of mercury exposure in the reproductive
success of tree swallows (Tachycineta bicolor). Ecotoxicology. 17:133-141, 2008.
Budtz-Jorgensen E, Grandjean P, Jorgensen P, Weihe P, Keiding N (2004a). Association between
mercury concentrations in blood and hair in methylmercury-exposed subjects at
different ages. Environmental Research, 95(3):385-93.
Budtz-Jorgensen E, Keiding N, Grandjean P (2004b). Effects of exposure imprecision on
estimation of the benchmark dose. Risk Analysis, 24(6):1689-96.
Budtz-Jorgensen, E.; Grandjean, P.; Weihe, P. (2007). Separation of risks and benefits of 16
seafood intake. Environmental Health Perspectives. Vol. 115, 323-327.
Burger, J. (2002). Daily consumption of wild fish and game: Exposures of high end
recreationalists, International Journal of Environmental Health Research, 12:4, p. 343-
354.
Burgess, Neil M., and Meyer, Michael W. Methylmercury exposure associated with reduced
productivity in common loons. Ecotoxicology. 17:83-91, 2008.
Carpi, A. and Lindberg, S.E., 1997. Sunlight-mediated emission of elemental mercury from soil
amended with municipal sewage sludge. Environmental Science and Technology, 31(7):
2085-2091.
Compeau, G.C. and Bartha, R., 1987. Effect of salinity on mercury-methylating activity of sulfate
reducing bacteria in estuarine sediments. Applied and Environmental Microbiology, 53:
261-265.
Connelly, N.A., B.A. Knuth, and T.L. Brown. 1996. "Sportfish Consumption Patterns of Lake
Ontario Anglers and the Relationship to Health Advisories." North American Journal of
Fisheries Management 16:90-101.
81
-------
Craig, P.J. and Bartlett, P.D., 1978. The role of hydrogen sulphide in environmental transport of
mercury. Nature, 275: 635-637.
Craig, P.J. and Moreton, P.A., 1986. Total mercury, methyl mercury and sulphide levels in British
estuarine sediments-Ill. Water Research, 20(9): 1111-1118.
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.
Crump KS, Kjellstrom T, Shipp AM, Silvers A, Stewart A (1998). Influence of prenatal mercury
exposure upon scholastic and psychological test performance: Benchmark analysis of a
New Zealand cohort. Risk Analysis, 18:701-713.
Crump, Kate L., and Trudeau, Vance L. Mercury-induced reproductive impairment in fish.
Environmental Toxicology and Chemistry. Vol. 28, No. 5, 2009.
Daniels, J.L., Longnecker,M.P., Rowland, A.S., et al., (2004). Fish intake during pregnancy and
early cognitive development of offspring. Epidemiology 15, 394-402.
Davidson PW, Myers GJ, Cox C, Axtell C, Shamlaye C, Sloane-Reeves J, Cernichiari E, Needham L,
Choi A, Wang Y, Berlin M, Clarkson TW (1998). Effects of prenatal and postnatal
methylmercury exposure from fish consumption on neurodevelopment: outcomes at 66
months of age in the Seychelles Child Development Study. Journal of the American
Medical Association, 280(8) :701-7.
Davidson, P.W., G. Myers, C.C. Cox,C.F. Shamlaye, D.O.Marsh, M.A.Tanner, M. Berlin, J. Sloane-
Reeves, E. Chernichiari,, 0. Choisy, A. Choi and T.W. Clarkson. 1995. Longitudinal
neurodevelopment study of Seychellois children following in utero exposure to
methylemrcury form maternal fish ingestion: outcomes at 19 and 29 months.
NeuroToxicology 16:677-688.
Davidson, P.W., Strain, J.J., Myers, G.J., et al., (2008). Neurodevelopmental effects of maternal
nutritional status and exposure to methylmercury from eating fish during pregnancy.
NeuroToxicology 29, 767-775.
Dellinger, JA (2004). Exposure assessment and initial intervention regarding fish consumption of
tribal members in the Upper Great Lakes Region in the United States. Environmental
Research 95 (2004) p. 325-340.
Dominici F, Samet JM, Zeger SL (2000). Combining evidence on air pollution and daily mortality
from the 20 largest US cities: a hierarchical modeling strategy. Journal of the Royal
Statistical Society A, 163:263-284.
82
-------
Doyon, J.-F., Schetagne, R. and Verdon, R., 1998. Different mercury bioaccumulation rates
between sympatric populations of dwarf and normal lake whitefish (Coregonus
clupeaformis) in the La Grande complex watershed, James Bay, Quebec.
Biogeochemistry, 40: 203-216.
Drevnick, P.E., D.E. Canfield, P.R. Gorski, A.L.C. Shinneman, D.R. Engstrom, D.C.G. Muir, G.R.
Smith, P.J. Garrison, L.B. Cleckner, J.P. Hurley, R.B. Noble, R.R. Otter, and J.T. Oris. 2007.
Deposition and cycling of sulfur controls mercury accumulation in Isle Royale fish.
Environmental Science and Technology 41(21):7266-7272.
Driscoll, C.T. et al., 1995. The role of dissolved organic carbon in the chemistry and
bioavailability of mercury in remote Adirondack lakes. Water, Air, Soil Pollution, 80: 499-
508.
Dunstan, J.A., Simmer, K., Dixon, G., et al., (2006). Cognitive assessment of children at age 2.5
years after maternal fish oil supplementation in pregnancy: a randomized controlled
trial. Archives of Diseases in Childhood. Fetal Neonatal Ed; December 21, 2006.
Ebert, E., N. Harrington, K. Boyle, J. Knight, J. and R. Keenan. 1994. "Estimating Consumption of
Freshwater Fish among Maine Anglers." North American Journal of Fisheries
Management 13:737-745.
EPA, 1997. U.S. Environmental Protection Agency, Volume I - General Factors Exposure Factors
Handbook Update to Exposure Factors Handbook, EPA/600/8-89/043—May 1989,
EPA/600/P-95/002Fa, August 1997.
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.
Evers, 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. Adverse effects from environmental mercury loads on
breeding common loons. Ecotoxicology. 17:69-81, 2008.
Frederick, Peter, and Jayasena, Nilmini. 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, 2010.
Frescholtz, T. and Gustin, M.S., 2004. Soil and foliar mercury emission as a function of soil
concentration. Water, Air, and Soil Pollution, 155: 223-237.
83
-------
Gbundgo-Tugbawa and Driscoll, 1998. Application of the regional mercury cycling model
(RMCM) to predict the fate and remediation of mercury in Onondaga Lake, New York.
Water, Air, and Soil Pollution, 105: 417-426.
Gilmour, C.C. and Henry, E.A., 1991. Mercury methylation in aquatic systems affected by acid
deposition. Environmental Pollution, 71:131-169.
Gilmour, C.C. et al., 1998. Methylmercury concentrations and production rates across a trophic
gradient in the northern Everglades. Biogeochemistry, 40: 327-345.
Gobas, F.A.P.C., Pasternak, J.P., Lien, K. and Duncan, R.K., 1998. Development and field
validation of a multimedia exposure model for waste load allocation in aquatic
ecosystems: application to 2,3,7,8-tetrachloro-p-dioxin and 2,3,7,8-
tetrachlorodibenzofuran in the Fraser River watershed. Environmental Science and
Technology, 32: 2442-2449.
Gobas, F.A.P.C., Z'Graggen, M.N. and Zhang, X., 1995. Time response of the Lake Ontario
ecosystem to virtual elimination of PCBs. Environmental Science and Technology, 29(8):
2038-2046.
Gorissen, Leen, Snoeijs, Tinne, Van Duyse, Els, and Eens, Marcel. Heavy metal pollution affects
dawn singing behavior in a small passerine bird. Oecologia. 145: 540-509, 2005.
Grandjean P, Weihe P, White R, Debes F, Araki S, Yokiyama K, Murata K, Sorensen N, Dahl R,
Jorgensen P. Cognitive deficit in 7-year-old children with prenatal exposure to
Methylmercury. Neurotoxicol Teratol. 1997. 19:(6)417-428.
Grandjean, Phillippe, Esben Budtz-J0rgensen, Roberta F. White, Poul J. J0rgensen, Pal Weihe,
Frodi Debes, and Niels Keding (1999). Methylmercury Exposure Biomarkers as Indicators
of Neurotoxicity in Children Aged 7 Years. American Journal of Epidemiology. Vol. 150
(3): 301-305.
Great Lakes Sport Fish Advisory Task Force. September 1993. Protocol for a Uniform Great
Lakes Sport Fish Consumption Advisory.
Greenfield, B.K., Hrabik, T.R., Harvey, C.J. and Carpenter, S.R., 2001. Predicting mercury levels in
yellow perch: use of water chemistry, trophic ecology, and spatial traits. Canadian
Journal Fisheries and Aquatic Sciences, 58: 1419-1429.
Grieb, T.M. et al., 1990. Factors affecting mercury accumulation in fish in the upper Michigan
peninsula. Environmental Toxicology and Chemistry, 9: 919-930.
Grosse, Scott D., Kurt V Krueger, Mercy Mvundura (2009). Economic Productivity by Age and
Sex: 2007 Estimates for the United States. Medical Care: July 2009 - Volume 47 - Issue
7_Supplement_l - pp S94-S103, doi: 10.1097/MLR.Ob013e31819c9571.
84
-------
Gustin, M. et al., 2004. Application of controlled mesocosms for understanding mercury air-soil-
plant exchange. Environmental Science and Technology, 38: 6044-6050.
Hallinger, Kelly K., Cornell, Kerri L, Brasso, Rebecka L, and Cristol, Daniel A. Mercury exposure
and survival in free-living tree swallows (Tachycineta bicolor). Ecotoxicology. Doi:
10.1007/S10646-010-0554-4, 2010.
Harmon, S.M., J.K. King, J.B. Gladden, and L.A. Newman. 2007. Using sulfate-amended sediment
slurry batch reactors to evaluate mercury methylation. Archives of Environmental
Contamination and Toxicology 52:326-333.
Harris, R., 2007. Ecosystem Responses to Mercury: Indicators of Change. CFC Press ISBN:
97808493888927.
Harris, R. and Hutchison, D., 2003. Factors Affecting the Predicted Response of Fish Mercury
Concentrations to Changes in Mercury Loading. 1005521, Electric Power Research
Institute, Palo Alto, CA.
Harris, R., Gherini, S. and Hudson, R., 1996. Regional Mercury Cycling Model: A Model for
Mercury Cycling in Lakes, R-MCM Version 1.0 Draft User Guide and Technical Reference,
Electric Power Research Institute, Wisconsin Department of Natural Resources,
Lafayette, California.
Harris., R. C, John W. M. Rudd, Marc Amyot, Christopher L. Babiarz, Ken G. Beaty, Paul J.
Blanchfield, R. A. Bodaly, Brian A. Branfireun, Cynthia C. Gilmour, Jennifer A. Graydon,
Andrew Heyes, Holger Hintelmann, James P. Hurley, Carol A. Kelly, David P.
Krabbenhoft, Steve E. Lindberg, Robert P. Mason, Michael J. Paterson, Cheryl L.
Podemski, Art Robinson, Ken A. Sandilands, George R. Southworth, Vincent L. St. Louis,
and Michael T. TateRudd, J. W. M., Amyot M., et al., Whole-Ecosystem study Shows
Rapid Fish-Mercury Response to Changes in Mercury Deposition. Proceedings of the
National Academy of Sciences Early Edition, PNAS 2007 104 (42) pp. 16586-16591;
(published ahead of print September 27, 2007).
Hastie, T.J., Tibshirani, R., (1993). Varying coefficient models. Journal of the Royal Statistical
Society: Series B 58, 379-396.
Hawley, Dana M., Hallinger, Kelly K., Cristol, Daniel A. Compromised immune competence in
free-living tree swallows exposed to mercury. Ecotoxicology. 18:499-503, 2009.
Helland, I.B., Smith, L., Saarem, K., et al., (2003). Maternal supplementation with very long
chain n-3 fatty acids during pregnancy and lactation augments children's IQ at 4 years of
age. Pediatrics 111, E39-E44.
85
-------
Hintelmann, H. and Evans, R.D., 1997. Application of stable isotopes in environmental tracer
studies - measurement of monomethylmercury by isotope dilution ICP-MS and
detection of species transformation. Fresenius Journal of Analytical Chemistry, 358: 378-
385.
Hintelmann, H. et al., 2002. Reactivity and mobility of new and old mercury deposition in a
boreal forest ecosystem during the first year of the METAALICUS study. Environmental
Science and Technology, 36: 5034-5040.
Hintelmann, H., Keppel-Jones, K. and Evans, R.D., 2000. Constants of mercury methylation and
demethylation rates in sediments and comparison of tracer and ambient mercury
availability. Environmental Toxicology and Chemistry, 19(9): 2204-2211.
Hoffman, David J., Henny, Charles J., Hill, Elwood F., Grover, Robert A., Kaiser, James L,
Stebbins, Katherine R. 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, 2009.
Hrabik, T.R., and C.J. Watras. 2002. Recent declines in mercury concentration in a freshwater
fishery: isolating the effects of de-acidification and decreased atmospheric mercury
deposition in Little Rock Lake. Science of the Total Environment 297:229-237.
Hurley, J. et al., 1995. Influences of watershed characteristics on mercury levels in Wisconsin
rivers. Environmental Science and Technology, 29(7): 1867-1875.
Hurley, J.P., Cowell, S.E., Shafer, M.M. and Hughes, P.E., 1998. Tributary loading of mercury to
Lake Michigan: Importance of seasonal events and phase partitioning. Science of the
Total Environment, 213: 129-137.
Hutchinson, J.W. LaBaugh, R.G. Sayre, and S.E. Schwarzbach. 2007. USGS goals for the coming
decade. Science 318:200-201.
International Agency for Research on Cancer (IARC). 1994. IARC Monographs on the Evaluation
of Carcinogenic Risks to Humans and their Supplements: Beryllium, Cadmium, Mercury,
and Exposures in the Glass Manufacturing Industry. Vol. 58. Jalili, H.A., and A.H. Abbasi.
1961. Poisoning by ethyl mercury toluene sulphonanilide. Br. J. Indust. Med.
18(0ct.):303-308 (as cited in NRC 2000).
Jackson, T.A., 1998. Mercury in aquatic ecosystems. In: W.J. Langston and M.J. Bebianno
(Editors), Metal Metabolism in Aquatic Environments. Chapman & Hall, London, pp. 77-
158.
Jakus, P., M. McGuinness, and A. Krupnick. 2002. "The Benefits and Costs of Fish Consumption
Advisories for Mercury." Discussion Paper 02-55. Washington, DC: Resources for the
Future.
86
-------
Judge, M.P., Harel, 0., Lammi-Keefe, C.J., et al., (2007). Maternal consumption of a
docosahexaenoic acid-containing functional food during pregnancy: benefit for infant
performance on problem-solving but not on recognition memory tasks at age 9 months.
American Journal of Clinical Nutrition 85, 1572-1577.
Karlsson, T. and Skyllberg, U., 2003. Bonding of ppb levels of methyl mercury to reduced sulfur
groups in soil organic matter. Environmental Science and Technology, 37: 4912-4918.
Kidd, K., Hesslein, R., Fudge, R. and Mallard, K., 1995. The influence of trophic level as measured
by delta-N-15 on mercury concentrations in fresh-water organisms. Water, Air, and Soil
Pollution, 80(1-4): 1011-1015.
et al.Kjellstrom, T., P. Kennedy, S. Wallis, A. Stewart, L. Friberg, B. Lind, P. Witherspoon, and C.
Mantell. 1989. Physical and mental development of children with prenatal exposure to
mercury from fish. Stage 2: Interviews and psychological tests at age 6. National
Swedish Environmental Protection Board Report No. 3642.
Knights, D. C, Sunderland, E. M., Barber, M. C., Johnston J. M., and Ambrose, R. B., Application
of Ecosystem-Scale Fate and Bioaccumulation Models to Predict Fish Mercury Response
Times to Changes in Atmospheric Deposition. Environmental Toxicology and Chemistry,
Vol 28, No. 4., pp. 881-893.
Lalonde, J., Amyot, M., Kraepiel, A. and Morel, F., 2001. Photooxidation of Hg(0) in artificial and
natural waters. Environmental Science and Technology, 35: 1367-1372.
Landis, M. and Keeler, G., 2002. Atmospheric mercury deposition to Lake Michigan during the
Lake Michigan mass balance study. Environmental Science and Technology, 36(21):
4518-4524.
Landis, M.S., Lynam, M. and Stevens, R.K., 2004. The Monitoring and Modeling of Mercury
Species in Support of Local Regional and Global Modeling. In: N. Pirrone and K.R.
Mahaffey (Editors), Dynamics of Mercury Pollution on Regional and Global Scales.
Kluwer Academic Publishers, New York, NY.
Laporte, J.-M., Andres, S. and Mason, R.P., 2002. Effect of ligands and other metals on the
uptake of mercury and methylmercury across the gills and the intestine of the blue crab
(Callinectes sapidus). Comparative Biochemistry and Physiology Part C, 131: 185-196.
Lawrence, A.L. and Mason, R.P., 2001. Factors controlling the bioaccumulation of mercury and
methylmercury by the estuarine amphipod Leptocheirus plumulosus. Environmental
Pollution, 111:217-231.
Lawson, N.M. and Mason, R.P., 1998. Accumulation of mercury in estuarine food chains.
Biogeochemistry, 40: 235-247.
87
-------
Leaner, J.J. and Mason, R.P., 2002. Factors controlling the bioavailability of ingested
methylmercury to channel catfish and atlantic sturgeon. Environmental Science and
Technology, 36: 5124-5129.
Lin, C.J. and Pehkonen, S.O., 1999. The chemistry of atmospheric mercury: a review.
Atmospheric Environment, 33: 2067-2079.
Lindberg, S.E. and Stratton, J.E., 1998. Atmospheric mercury speciation: concentrations and
behavior of reactive gaseous mercury in ambient air. Environmental Science and
Technology, 32(1): 49-57.
Lynch, Miranda L., Li-Shan Huang, Christopher Cox b, J.J. Strain, Gary J. Myers, Maxine P.
Bonham, Conrad F. Shamlaye, Abbie Stokes-Riner, Julie M.W. Wallace, Emeir M. Duffy,
Thomas W. Clarkson, Philip W. Davidson, (2010). Varying coefficient function models to
explore interactions between maternal nutritional status and prenatal methylmercury
toxicity in the Seychelles Child Development Nutrition Study. Environmental Research,
doi:10.1016/j.envres.2010.09.005.
Marsh, D.O., T.W. Clarkson, C. Cox, et al. 1987. Fetal methylmercury poisoning: relationship
between concentration in single strands of maternal-hair and child effects. Arch. Neurol.
44:1017-1022. (as cited in EPA 2002 IRIS documentation.)
Mason, R.P. et al., 1999. Mercury in the Chesapeake Bay. Marine Chemistry, 65: 77-96.
Mason, R.P., Morel, F.M.M. and Hemond, H.F., 1995. The role of microorganisms in elemental
mercury formation in natural waters. Water, Air, and Soil Pollution, 80: 775-787.
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.
Mclntyre, J.W., Barr, J.F. 1997Common Loon (Gavia immer) in: Pool A, Gill F (eds) The Birds of
North America. Academy of Natural Sciences, Philadelphia, PA, 313
Mierle, G. and Ingram, R., 1991. The role of humic substance in the mobilization of mercury
from watersheds. Water, Air, and Soil Pollution, 56: 349-357.
Mitro, Matthew G., Evers, David C., Meyer, Michael W., and Piper, Walter H. Common loon
survival rates and mercury in New England and Wisconsin. Journal of Wildlife
Management. 72(3): 665-673, 2008.
Morel, F., Kraepiel, A.M.L. and Amyot, M., 1998. The chemical cycle and bioaccumulation of
mercury. Annual Reviews of Ecological Systems, 29: 543-566.
88
-------
Morgan, J.N., M.R. Berry, and R.L. Graves. 1997. "Effects of Commonly Used Cooking Practices
on Total Mercury Concentration in Fish and Their Impact on Exposure Assessments."
Journal of Exposure Analysis and Environmental Epidemiology 7(1):119-133.
Moya, J. 2004. Overview of fish consumption rates in the United States. Human and Ecological
Risk Assessment: An International Journal 10, no. 6: 1195-1211.
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.
Myers GJ, Davidson PW, Cox, C, Shamlaye CF, Palumbo D, Cernichiari E, Sloane-Reeves J,
Wilding GE, Kost J, Huang LS, Clarkson TW (2003). Prenatal methylmercury exposure
from ocean fish consumption in the Seychelles child development study. Lancet,
361:1686-1692.
Myers, M.D., M.A. Ayers, J.S. Baron, P.R. Beauchemin, K.T. Gallagher, M.B. Goldhaber, D.R.
National Research Council (NRC). 2000. Toxicological Effects of Methylmercury. Committee on
the Toxicological Effects of Methylmercury, Board on Environmental Studies and
Toxicology, Commission on Life Sciences, National Research Council. National Academy
Press, Washington, DC.
Neisser U, Boodoo G, Bouchard TJ, et al. (1996). Intelligence: Knowns and unknowns. American
Psychologist, 51:77-101.
NRC (2000). Toxicological Effects of Methylmercury. National Research Council. Washington,
DC: National Academies Press.
Oken E., K.P. Kleinman, W.E. Berland, S.R. Simon, J.W. Rich-Edwards, and M.W. Gillman. 2003.
"Decline in Fish Consumption Among Pregnant Women After a National Mercury
Advisory." Obstetrics and Gynecology 102(2):346-351.
Rasmussen, P.E., 1994. Current methods of estimating atmospheric mercury fluxes in remote
areas. Environmental Science and Technology, 28(13): 2233-2241.
Ravichandran, M., 2004. Interactions between mercury and dissolved organic matter - a review.
Chemosphere, 55: 319-331.
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.
Rice GE, Hammitt JK, Evans JS. (2010). A probabilistic characterization of the health benefits of
reducing methyl mercury intake in the United States. Environmental Science Technology.
2010Jull;44(13):5216-24
89
-------
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.
Rolfhus, K. et al., 2003. Distribution and fluxes of total and methylmercury in Lake Superior.
Environmental Science and Technology, 37(5): 865-872.
Ryan, LM (2005). Effects of Prenatal Methylmercury on Childhood IQ: A Synthesis of Three
Studies. Report to the U.S. Environmental Protection Agency.
Salkever, D. 1995. "Updated Estimates of Earnings Benefits from Reduced Lead Exposure of
Children to Environmental Lead." Environmental Research 70:1-6.
Schaefer, J. et al., 2004. Role of the bacterial organomercury lyase (MerB) in controlling
methylmercury accumulation in mercury-contaminated natural waters. Environmental
Science and Technology, 38: 4304-4311.
Scheuhammer, Anton M., Meyer Michael W., Sandheinrich, Mark B., and Murray, Michael W.
Effects of environmental methylmercury on the health of wild birds, mammals, and fish.
Ambio.Vol.36, No.l, 2007.
Scholtz, M.T., Heyst, B.J.V. and Schroeder, W.H., 2003. Modeling of mercury emissions from
background soils. Science of the Total Environment, 304: 185-207.
Schroeder, W.H., Munthe, J. and Lindqvist, 0., 1989. Cycling of mercury between water air and
soil compartments of the environment. Water, Air, and Soil Pollution, 48: 337-347.
Schwartz, Joel (1994). Societal Benefits of Reducing Lead Exposure. Environmental Research 66,
105-124.
Seigneur, C., Jayaraghavan, K., Lohman, K., Karamchandani, P. and Scott, C., 2004. Global
Source Attribution for Mercury Deposition in the United States. Environmental Science
and Technology, 38: 555-569.
Sellers, P., Kelly, C.A., Rudd, J.W.M. and MacHutchon, A.R., 1996. Photodegradation of
methylmercury in lakes. Nature, 380: 694.
Sepulveda, Maria S., Frederick, Peter C., Spalding, Marilyn G., and Williams, Gary E. Jr. Mercury
contamination in free-ranging great egret nestlings (Ardea albus) from southern Florida,
USA. Environmental Toxicology and Chemistry. Vol. 18, No.5,1999.
Shilling, Fraser, Aubrey White, Lucas Lippert, Mark Lubell (2010). Contaminated fish
consumption in California's Central Valley Delta. Environmental Research 110, p. 334-
344.
90
-------
Siciliano, S., O'Driscoll, N. and Lean, D., 2002. Microbial reduction and oxidation of mercury in
freshwater lakes. Environmental Science and Technology, 36(14): 3064-3068.
St. Louis, V. et al., 2001. Importance of the forest canopy to fluxes of methyl mercury and total
mercury to boreal ecosystems. Environmental Science and Technology, 35: 3089-3098.
Stern, AH, Smith AE (2003). An assessment of the cord blood:maternal blood methylmercury
ratio: Implications for risk assessment. Environmental Health Perspectives, 111:1465-
1470.
Strain, J.J., Davidson, P.W., Bonham, M.P., et al., (2008). Associations of maternal long-chain
polyunsaturated fatty acids, methylmercury, and infant development in the Seychelles
Child Development Nutrition Study. NeuroToxicology 29, 776-782.
Sunda, W.G. and Huntsman, S.A., 1998. Processes regulating cellular metal accumulation and
physiological effects: Phytoplankton as model systems. Science of the Total
Environment, 219: 165-181.
Sunderland, E.M. et al., 2004. Speciation and bioavailability of mercury in well-mixed estuarine
sediments. Marine Chemistry, 90: 91-105.
Swartout, J., and G. Rice. 2000. "Uncertainty Analysis of the Estimated Ingestion Rates Used to
Derive the Methylmercury Reference Dose." Drug and Chemical Toxicology 23(1):293-
306. 11-41
Tseng, C.M., Amouroux, D., Abril, G. and Donard, O.F.X., 2001. Speciation of mercury in a fluid
mud profile of a highly turbid macrotidal estuary (Gironde, France). Environmental
Science and Technology, 35(13): 2627-2633.
U.S. Department of the Interior (DOI), Fish and Wildlife Service and U.S. Department of
Commerce, Bureau of the Census. 1992. 1991 National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation. Washington, DC: U.S. Government Printing Office.
U.S. Department of the Interior (DOI), Fish and Wildlife Service and U.S. Department of
Commerce, Bureau of the Census. 1997. 1996 National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation. Washington, DC: U.S. Government Printing Office.
U.S. Department of the Interior (DOI), Fish and Wildlife Service and U.S. Department of
Commerce, Bureau of the Census. 2002. 2001 National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation. Washington, DC: U.S. Government Printing Office.
U.S. Department of the Interior (DOI), Fish and Wildlife Service and U.S. Department of
Commerce, Bureau of the Census. 2007. 2006 National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation. Washington, DC: U.S. Government Printing Office.
91
-------
U.S. Environmental Protection Agency (EPA), 1997. Mercury Study Report to Congress. EPA-
452/R-97-005, Office of Air Quality Planning and Standards United States Environmental
Protection Agency, Washington.
U.S. Environmental Protection Agency (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.
U.S. Environmental Protection Agency (EPA) (1997). Volume I - General Factors Exposure
Factors Handbook Update to Exposure Factors Handbook, EPA/600/8-89/043—May
1989, EPA/600/P-95/002Fa, August 1997.
U.S. Environmental Protection Agency (U.S. EPA). 1999. 1999 National Emission Inventory
Documentation and Data—Final Version 3.0; Hazardous Air Pollutants Inventory—
FinalNEI Version 3; HAPS Summary Files. (12 December
2006;www. epa.gov/ttn/chief/net/1999inventory. html)
U.S. Environmental Protection Agency (EPA), 2001a. Mercury Maps. A Quantitative Spatial Link
Between Air Deposition and Fish Tissue. Final Report. EPA/823/R-01/009, USEPA,
Washington, D.C.
U.S. Environmental Protection Agency (EPA), 2001b. Methylmercury fish tissue residue
criterion, United States Environmental Protection Agency, Office of Water, 4304 EPA-
823-F-01-001, January 2001,
www.epa.gov/waterscience/criteria/methylmercurv/factsheet.html.
U.S. Environmental Protection Agency (EPA) (2002). Mercury Neurotoxicity Workshop Notes.
Washington, DC. November 4, 2002.
http://www.epa.gov/ttn/ecas/regdata/Benefits/mercuryworkshop.pdf.
U.S. Environmental Protection Agency (EPA 2002). Mercury Neurotoxicity Workshop Notes;
available at: www.epa.gov/ttn/ecas/benefits.html
U.S. Environmental Protection Agency (EPA). 2002 (date of most recent revision of on-line
materials; website accessed January 2005). Integrated Risk Information System (IRIS).
Methylmercury. U.S. EPA Office of Research and Development, National Center for
Environmental Assessment. Oral RfD and inhalation RfC assessments last revised
7/27/2001. Carcinogenicity assessment last revised 5/1/1995. Available online at
http://www.epa.gov/iris/subst/0073.htm
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 (EPA) (2005). Regulatory Impact Analysis of the Clean Air
Mercury Rule, Research Triangle Park, NC. June 2008. EPA-452/R-05-003.
U.S EPA (Environmental Protection Agency). 2008. Integrated Science Assessment (ISA) for
Oxides of Nitrogen and Sulfur—Ecological Criteria (Final Report). EPA/600/R-08/082F.
U.S. Environmental Protection Agency, National Center for Environmental Assessment-
RTP Division, Office of Research and Development, Research Triangle Park, N.C.
Available at http://cfpub.epa.gov/ncea/cfm/recorddisplay.cfm?deid+201485.
U.S. Environmental Protection Agency (EPA) (2008). Proposed Lead NAAQS Regulatory Impact
Analysis Office of Air Quality Planning and Standards, Research Triangle Park, NC. June
2008.
U.S. Environmental Protection Agency (EPA). 2008. Integrated Science Assessment (ISA) for
Oxides of Nitrogen and Sulfur-Ecological Criteria (Final Report). EPA/600/R- 08/082F.
U.S. Environmental Protection Agency, National Center for Environmental Assessment-
RTP Division, Office of Research and Development, Research Triangle Park, NC. Available
at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=201485.
United States Department of Health and Human Services (DHHS), Centers for Disease Control
and Prevention (CDC), National Center for Health Statistics (NCHS), Division of Vital
Statistics, Natality public-use data 2003-2006, on CDC WONDER Online Database, March
2009. Accessed at http://wonder.cdc.gov/natality-current.html on Nov 3, 2010 4:44:23
PM
USDA Forest Service (1994). "National Survey on Recreation and the Environment: 1994-95."
Washington DC: USDA Forest Service.
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 (Eptesicus
fuscus) near a contaminated river. Ecotoxicology. 19(7), 1277-1284.
Watras, C.J. and Bloom, N.S., 1992. Mercury and methylmercury in individual zooplankton:
implications for bioaccumulation. Limnol. Oceanogr., 37: 1313-1318.
Watras, C.J. et al., 1998. Bioaccumulation of mercury in pelagic freshwater food webs. Science
of the Total Environment, 219(2-3): 183-208.
Wechsler D (1991). WISC-III Manual. San Antonio: The Psychological Corporation.
93
-------
West, P.C., J.M. Fly, R. Marans, F. Larkin, and D. Rosenblatt. May 1993.1991-92 Michigan Sport
Anglers Fish Consumption study. Prepared by the University of Michigan, School of
Natural Resources for the Michigan Department of Natural Resources, Ann Arbor, Ml.
Technical Report No. 6.
West, P.C., M.J. Fly, R. Marans, and F. Larkin. 1989. Michigan Sport Anglers Fish Consumption
Survey. A report to the Michigan Toxic Substance Control Commission. Michigan
Department of Management and Budget Contract No. 87-20141.
Woods & Poole Economics, Inc. 2008. Population by Single Year of Age CD. CD-ROM. Woods &
Poole Economics, Inc.
Wren, C.D. and MacCrimmon, H.R., 1986. Comparative bioaccumulation of mercury in two
adjacent freshwater ecosystems. Water Research, 6: 763-769.
Yates, David E., Mayack, David T., Munney, Kenneth, Evers David C., Major, Andrew, Kaur,
Taranjit, and Taylor, Robert J. Mercury levels in mink (Mustela vison) and river otter
(Lonra canadensis) from northeastern North America. Ecotoxicology. 14, 263-274, 2005.
Zhang, H. and Lindberg, S.E., 2001. Sunlight and lron(lll)-lnduced Photochemical Production of
Dissolved Gaseous Mercury in Freshwater. Environmental Science and Technology, 35:
928-935.
94
-------
APPENDIX 4A
ANALYSIS OF TRIP TRAVEL DISTANCE FOR RECREATIONAL FRESHWATER ANGLERS
As described in Section 3.7.7, the method used to estimate exposures to mercury in
freshwater fish requires information about how far individuals typically travel for freshwater
fishing. This appendix describes the data and methods used to analyze travel distance patterns
by freshwater anglers, and it reports the results that were used to estimate exposures.
4A.1 Data
To conduct an analysis of trip travel distance for freshwater anglers, we used data from
the NSRE 1994. As described previously, this 16,000-person survey elicited information on
water-based recreation activities—specifically boating, fishing, swimming, and wildlife
viewing—during the previous year. Respondents were asked about their most recent trip taken
in each of the four categories. Of particular interest to this analysis is data concerning fishing
trip characteristics for all respondents who fished in freshwater bodies during the previous
year. Of the 3,220 respondents who had reported fishing, 2,482 visited either a lake, pond,
river, or stream on their most recent trip.
The fishing module elicited location information about most recent fishing trip taken
during the preceding 12 months. This trip was recorded as either a single- or multiday trip to a
specific water body ("site") identified by the respondent. Subsequently, a series of questions
were asked to gather location data on the specific site visited, including the site name, the state
in which the site was located, and the name of the city or town nearest the site. To identify
potential determinants of travel distance for a freshwater fishing trip, we analyzed the 2,384
available responses to the following survey question: "What was the one way travel distance, in
miles from your home, to your destination on *site*?" Table C-l presents summary statistics
for travel distance, which are reported separately for single-day, multiday, and aggregated
trips. As would be expected, median travel distance varied according to trip type, from 20 miles
for a single-day trip to almost 140 miles for a multiday trip. Across both trip types, the average
travel distance was slightly less than 100 miles.
4A.2 Analysis of Travel Distance Data
The influence of multiple demographic characteristics on travel distance was tested
using multivariate regression analysis. Table C-2 reports descriptive statistics for the anglers
included in this analysis. As indicated by the table, over 90 percent of the sample is white;
males comprise a higher percentage of the sample (62 percent) than females. More than half
4A-1
-------
Table 4A-1. Reported Trip Travel Distance for Freshwater Anglers (Miles)
All trip types
Single-day trips only
Multiday trips only
N
2,384
1,791
586
Mina
0
0
3
P5
2
2
18
P25
10
10
70
P50
20
20
138
Mean
91.9
41
248.2
P75
45
45
300
P95
125
125
850
Max
3,000
1,100
3,000
a Seven respondents reported traveling 0 miles for their most recent trip; all were described as single-day trips.
Note: Ninety-eight respondents who visited freshwater bodies on their most recent fishing trip did not report the
travel distance.
Table 4A-2. Demographic Characteristics of Freshwater Anglers3
N
Gender 2,267
Race 2,250
Education 2,262
Work status 2,263
Geography 2,237
Region 2,205
Frequency
62%
91%
4%
2%
2%
11%
34%
55%
75%
23%
37%
41%
13%
33%
31%
23%
Male
White
Black
Hispanic
Other
Less than high school degree
High school degree/equivalent
Some college or more
Employed
Urban
Suburban
Rural
Northeast
South
Midwest
West
a In total, 2,384 respondents reported information on trip travel distance to a freshwater destination.
Note: Values may not add to 100 percent due to rounding.
4A-2
-------
the sample had completed at least some college and three-fourths of the sample reported
being employed. The survey asked respondents to classify their place of residence as either
rural, suburban, or urban. Approximately 40 percent described their area as rural, 37 percent as
suburban, and 23 percent as urban. Respondents were assigned to a U.S. Census geographic
region by matching their zip code to a corresponding state. The states were then aggregated to
the appropriate Census region (http://www.census.gov/geo/www/us_regdiv.pdf). The majority
of respondents resided in the South and Midwest, followed by the West and Northeast.
Table C-3 presents additional characteristics on the demographic distribution of the
sample. The average age of respondents was 38 years, while household size averaged
approximately three members, with less than one person under the age of six. Respondents'
average weekly leisure time was 28 hours. However, this varied significantly across the sample,
from zero to 168 hours. In the survey, family income is reported as a categorical variable, with
respondents selecting the income range that reflected family income in the previous year. The
midpoint of this range was taken to produce a continuous income variable. Subsequently, this
value was converted to (2000$) using the consumer price index. Median (mean) income was
estimated to be $57,325 ($66,496) annually.
Table 4A-3. Demographic Characteristics of Freshwater Anglers
Age
Household size
Persons <6 yrs
Persons >16 yrs
Weekly leisure time (hrs)
Family income (2,000$)
N
2,245
2,255
2,270
2,254
2,025
1,851
Mean
38.4
3.1
0.3
2.2
27.7
66,496
SD
14.5
1.5
0.7
0.9
23.9
57,324
Min
16
1
1
0
0
8,938
Max
92
10
5
7
168
208,547
Multivariate regression analysis was used to identify determinants of travel distance to
freshwater fishing sites. The dependent variable in this analysis was the miles traveled to the
most recent freshwater fishing site. The explanatory variables included several demographic
and geographic characteristics of the respondents.
Separate regressions were conducted for the full sample (1), single-day trips only (2),
and multiday trips only (3). The results are reported in Table C-4. Family income was estimated
4A-3
-------
Table 4A-4. OLS Regression Results for Determinants of Reported Trip Travel Distance (Miles)
Variable Description
CONSTANT
AGE
GENDER
EDUC
MINORITY
FAMILY INCOME (log)
URBAN
SUBURBAN
NEAST
MIDWEST
WEST
(1)
Full Sample (both single-
and multiday trips)
Coefficient t-stat
0.6966 1.54
0.0044 1.83*
0.0572 0.83
0.1729 2.48**
-0.0437 -0.36
0.187 4.41**
0.3491 3.95**
0.3422 4.48**
-0.0387 -0.36
0.3856 4.65**
0.6103 6.73**
R2 = 0.077
N = 1,798
(2)
Single-Day
Trips Only
Coefficient t-stat
1.7954 3.89**
0.0011 0.44
0.0173 0.25
0.1552 2.21**
0.0228 0.19
0.0827 1.92*
0.2799 3.12**
0.193 2.50**
-0.2549 -2.42**
0.1 1.21
0.3374 3.59**
R2 = 0.041
N = 1,360
(3)
Multiday
Trips Only
Coefficient t-stat
2.2493 3.26**
0.001 0.28
0.1446 1.39
0.128 1.22
-0.1391 -0.76
0.1759 2.78**
0.2121 1.62*
0.4298 3.67**
0.1525 0.89
0.4923 3.63**
0.3239 2.32**
R2 = 0.112
N = 434
** = significant at 5 percent level.
* = significant at 10 percent level.
to have a positive and highly significant effect in all three models. Dummy variables for urban
and suburban location were also found to have positive and highly significant effects in all
models. These results suggest that wealthier anglers and those living in or near metropolitan
areas tend to travel further to fishing sites, relative to less-wealthy anglers and those living in
rural areas. In models (1) and (2) dummy variables for the Midwest and West regions also had
positive and highly significant effects on trip travel distance, relative to the South region. The
Northeast region did not have a statistically significant effect on distance traveled. Education
was estimated to be positively and significantly related to distance traveled in the first and
second models. (Note that the respondent's level of education, recorded in the survey as a
categorical variable, was receded as a continuous variable for the regression analysis.) Neither
age, race, nor gender had significant effects (at a 5 percent level) on travel distance in any of
the models.
4A-4
-------
4A.3 Summary Results Applied in the Population Centroid Approach
Given the high significance of geographic area and family income across the regressions,
nonparametric results (frequency distributions) were generated for four mutually exclusive
subgroups of respondents and five travel distance categories. The results are reported in
Table C-5. Respondents were categorized into the four following groups:
• Gl: family income >$50,000 (in 2000 dollars) and urban or suburban resident
- (N = 452 for single-day trips)
- (N = 649 for single- and multiday trips)
• G2: family income <$50,000 and urban or suburban resident
- (N = 329 for single-day trips)
- (N = 417 for single- and multiday trips)
• G3: family income >$50,000 and rural resident
- (N = 295 for single-day trips
- (N = 376 for single- and multiday trips)
• G4: family income <$50,000 and rural resident
- (N = 309 for single-day trips)
- (N = 386 for single- and multiday trips)
These categories were selected because they match categories that can be easily
identified in Census data and because they split the sample into roughly similar group sizes.
Travel distance was categorized into ranges reported in the first column of Table C-5. The
results are consistent with those generated from the regression analysis. Among respondents
on single-day trips, the number that traveled longer distances (greater than 100 miles)
increased from the low-income rural cohort (5 percent) to the higher-income urban/suburban
cohort (11 percent). The same pattern holds for those taking either a single- or multiday trip.
The number traveling longer distances more than doubled, from 11 percent among low-income
rural respondents to 27 percent among high-income urban/suburban respondents. These
results indicate higher-income urban/suburban anglers travel greater distances to freshwater
destinations than lower-income urban/suburban anglers and rural anglers.
As described in Section 3.7, the trip frequency estimates reported in Table C-5 for the
full sample were used in the population centroid approach to weight exposures to mercury in
4A-5
-------
fish according to distance from the Census tract centroid, income levels in the tract, and
whether the tract is predominantly rural or urban/suburban.
Table 4A-5. Travel Distance Frequencies by Demographic Group (Percentage in Each Distance
Category)
Travel Distance (mi)
(Gl) (G2)
High-Income and Low-Income and
Urban/Suburban Urban/Suburban
Resident Resident
(G3)
High-Income and
Rural Resident
(G4)
Low-Income and
Rural Resident
Single-day trips only (N = 1,385)
N
Distance <10 mi
>10 mi to 20 mi
>20 mi to 50 mi
>50 mi to 100 mi
Distance >100 mi
Full sample (both single- and
N
Distance <10 mi
>10 mi to 20 mi
>20 mi to 50 mi
>50 mi to 100 mi
Distance >100 mi
(N=452)
23%
18%
31%
17%
11%
multiday trips) (N = 1,828)
(N = 649)
16%
13%
24%
19%
27%
(N = 329)
32%
23%
20%
19%
6%
(N = 417)
26%
18%
18%
19%
18%
(N = 295)
31%
22%
28%
14%
5%
(N = 376)
24%
18%
25%
16%
17%
(N = 309)
34%
24%
26%
11%
5%
(N = 386)
29%
21%
25%
14%
11%
4A-6
-------
CHAPTER 5
HEALTH AND WELFARE CO-BENEFITS
Synopsis
Implementation of HAP emissions controls required by this rule is expected to have
ancillary co-benefits, including lower overall ambient concentrations of S02, N02, PM2.5 and
ozone across the U.S. Pollutants such as S02, NOX, and direct PM2.5 contribute to ambient PM2.5
levels in the atmosphere, and NOx contributes to ambient ozone concentrations. Furthermore,
this rule is expected to reduce C02 emissions affecting climate change. These health and
welfare co-benefits comprise a significant share of the total monetized benefits from this rule.
This chapter provides estimates for this subset of the expected annual health and climate co-
benefits of this rule in 2016.
Due to limits in available air quality modeling, the quantified co-benefits of this rule
consist of only PM2.5-related health co-benefits from reductions in S02 (a precursor to PM2.5
formation) and direct PM2.5 and climate co-benefits from reductions in C02. These co-benefits
are estimated by applying a benefit-per-ton (BPT) approach described below to estimated
reductions in S02 and direct PM2.5 emissions reported in Chapter 3. The monetized co-benefits
assessment omits several important categories of benefits, including health and ecological co-
benefits from reducing exposure to ozone, ecosystem co-benefits for reducing nitrogen and
sulfate deposition, and the direct health co-benefits from reducing exposure to ozone, S02 and
N02. We describe these co-benefits qualitatively in Section 5.5.
We estimate the monetized health and climate co-benefits of MATS to be $37 billion to
$90 billion at a 3% discount rate and $33 billion to $81 billion at a 7% discount rate in 2016,
depending on the epidemiological function used to estimate reductions in premature mortality.
All estimates are in 2007$.
5.1 Overview
The analysis in this chapter aims to characterize the co-benefits of the Mercury and Air
Toxics Standards by answering two key questions:
1. What are the health effects of changes in ambient particulate matter (PM2.5)
resulting from reductions in directly-emitted PM2.5 and S02?
2. What is the economic value of these effects?
5-1
-------
Additionally, this chapter describes health effects that are not quantified for this rule,
unquantified welfare effects, and visibility co-benefits.
In implementing these rules, emission controls may lead to reductions in ambient PM2.5
below the National Ambient Air Quality Standards (NAAQS) for PM in some areas and assist
other areas with attaining the PM NAAQS. Because the PM NAAQS RIAs also calculate PM
benefits, there are important differences worth noting in the design and analytical objectives of
each RIA. The NAAQS RIAs illustrate the potential costs and benefits of attaining a new air
quality standard nationwide based on an array of emission control strategies for different
sources. In short, NAAQS RIAs hypothesize, but do not predict, the control strategies that States
may choose to enact when implementing a NAAQS. The setting of a NAAQS does not directly
result in costs or benefits, and as such, the 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. However, some costs and benefits estimated in this RIA
account for the same air quality improvements as estimated in the illustrative PM2.5 NAAQS RIA.
By contrast, the emission reductions for this rule are from a specific class of well-
characterized sources. In general, EPA is more confident in the magnitude and location of the
emission reductions for these rules. It is important to note that emission reductions anticipated
from these rules do not result in emission increases elsewhere (other than potential energy
disbenefits). Emission reductions achieved under these and other promulgated rules will
ultimately be reflected in the baseline of future NAAQS analyses, which would reduce the
incremental costs and benefits associated with attaining the NAAQS. EPA remains forward
looking towards the next iteration of the 5-year review cycle for the NAAQS, and as a result
does not issue updated RIAs for existing NAAQS that retroactively update the baseline for
NAAQS implementation. For more information on the relationship between the NAAQS and
rules such as analyzed here, please see Section 1.2.4 of the S02 NAAQS RIA (U.S. EPA, 2010a).
To estimate a subset of the co-benefits from reducing PM2.5 exposure, EPA used an
approach that is consistent with the approach utilized to estimate the co-benefits of the
proposed MATS (U.S. EPA 2011a) and the Cross-State Air Pollution Rule (U.S. EPA 2011b). In this
analysis we consider an array of health impacts attributable to changes in PM2.5 air quality. The
2009 PM2.5 Integrated Science Assessment (U.S. EPA, 2009a) identified the human health
effects associated with these ambient pollutants, which include premature mortality and a
variety of morbidity effects associated with acute and chronic exposures.
5-2
-------
Table 5-1 summarizes the total monetized co-benefits of the rule in 2016. This table
reflects the economic value of the change in PM2.5-related human health impacts and the
monetized value of C02 reductions occurring as a result of the Mercury and Air Toxics
Standards.
Table 5-1. Estimated Monetized Co-benefits of the Mercury and Air Toxics Standards in
2016 (billions of 2007$)a
Benefits Estimate Eastern U.S.b Western U.S. Total
Pope et al. (2002) PM25 mortality estimate
Using a 3% discount rate $35+B Sl-l+B $37+B
($2.8-$110) ($0.03-$3.4) ($3.2-$110)
Using a 7% discount rate $32+B Sl-O+B $33+B
($2.5-$98) ($0.03-$3.1) ($2.9-$100)
Laden et al. (2006) PM2.5 mortality estimate
Using a 3% discount rate $87+B $2.7+B $90+B
($7.5-$250) ($0.1-$7.9) ($8.0-$260)
Using a 7% discount rate $78+B $2.4+B $81+B
($6.8-$230) ($0.1-$7.2) ($7.3-$240)
For notational purposes, unquantified benefits are indicated with a "B" to represent the sum of additional
monetary benefits and disbenefits. 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. A detailed
listing of unquantified health and welfare effects is provided in Tables 5-2 and 5-3. Estimates here are subject to
uncertainties discussed further in the body of the document. Estimates are rounded to two significant figures.
Value of total co-benefits includes CO2-related co benefits discounted at 3%.
b Includes Texas and those states to the north and east.
Tables 5-2 and 5-3 summarize the human health and environmental co-benefits
categories contained within the total monetized benefits estimate, and those categories that
were unquantified due to limited data or time. It is important to emphasize that the list of
unquantified benefit categories is not exhaustive, nor is quantification of each effect complete.
In order to identify the most meaningful human health and environmental co-benefits, 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 Integrated Science Assessment. This 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, some effects that were identified
in previous lists of unquantified benefits in other RIAs have been dropped (e.g., UVb exposure).
In addition, some quantified effects represent only a partial accounting of likely impacts due to
limitations in the currently available data (e.g., climate effects from C02, etc).
5-3
-------
Table 5-2. Human Health Effects of Pollutants Affected by the Mercury and Air Toxics
Standards
Benefits Category
Effect Has
Been
Specific Effect Quantified
Effect Has
Been
Monetized More Information
Improved Human Health
Reduced incidence of
premature mortality
from exposure to PM2.5
Reduced incidence of
morbidity from
exposure to PM2.5
Reduced incidence of
mortality from
exposure to ozone
Reduced incidence of
morbidity from
exposure to ozone
Adult premature mortality based on cohort S
study estimates and expert elicitation
estimates (age >25 or age >30)
Infant mortality (age <1) S
Non-fatal heart attacks (age > 18) S
Hospital admissions— respiratory (all ages) S
Hospital admissions— cardiovascular (age S
Emergency room visits for asthma (<18) S
Acute bronchitis (age 8-12) S
Lower respiratory symptoms (age 7-14) S
Upper respiratory symptoms (asthmatics S
age 9-11)
Asthma exacerbation (asthmatics age 6-18) S
Lost work days (age 18-65) S
Minor restricted-activity days (age 18-65) S
Chronic Bronchitis (age >26) S
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
Premature mortality based on short-term —
study estimates (all ages)
Premature mortality based on long-term —
study estimates (age 30-99)
Hospital admissions— respiratory causes —
(age > 65)
Hospital admissions— respiratory causes —
(age <2)
Emergency room visits for asthma (all ages) —
Minor restricted-activity days (age 18-65) —
S Section 5.4
S Section 5.4
S Section 5.4
S Section 5.4
S Section 5.4
•S Section 5.4
S Section 5.4
S Section 5.4
•S Section 5.4
S Section 5.4
S Section 5.4
S Section 5.4
S Section 5.4
- PM ISAb
- PM ISAb
- PM ISAb'c
- PM ISAb'c
— Ozone CD, Draft
Ozone ISA3
— Ozone CD, Draft
Ozone ISA3
— Ozone CD, Draft
Ozone ISA3
— Ozone CD, Draft
Ozone ISA3
— Ozone CD, Draft
Ozone ISA3
— Ozone CD, Draft
Ozone ISA3
(continued)
5-4
-------
Table 5-2. Human Health Effects of Pollutants Affected by the Mercury and Air Toxics
Standards (continued)
Benefits Category
Reduced incidence of
morbidity from
exposure to NO2
Reduced incidence of
morbidity from
exposure to SO2
Reduced incidence of
morbidity from
exposure to
methylmercury
(through reduced
mercury deposition as
well as the role of
sulfate in methylation)
Effect Has
Been
Specific Effect Quantified
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 —
Asthma hospital admissions (all ages) —
Chronic lung disease hospital admissions —
(age > 65)
Respiratory emergency department visits —
(all ages)
Asthma exacerbation (asthmatics age 4-18) —
Acute respiratory symptoms (age 7-14) —
Premature mortality —
Other respiratory effects (e.g., airway —
hyperresponsiveness and inflammation,
lung function, other ages and populations)
Respiratory hospital admissions (age > 65) —
Asthma emergency room visits (all ages) —
Asthma exacerbation (asthmatics age 4-12) —
Acute respiratory symptoms (age 7-14) —
Premature mortality —
Other respiratory effects (e.g., airway —
hyperresponsiveness and inflammation,
lung function, other ages and populations)
Neurologic effects - IQ loss ^
Other neurologic effects (e.g., —
developmental delays, memory, behavior)
Cardiovascular effects —
Genotoxic, immunologic, and other toxic —
effects
Effect Has
Been
Monetized More Information
— Ozone CD, Draft
Ozone ISA3
— Ozone CD, Draft
Ozone ISA3
— Ozone CD, Draft
Ozone ISAb
— Ozone CD, Draft
Ozone ISA0
— Ozone CD, Draft
Ozone ISA0
- NO2 ISA3
- NO2 ISA3
- NO2 ISA3
- NO2 ISA3
- NO2 ISA3
- NO2 ISAb'°
- NO2 ISAb'c
- SO2 ISA3
- SO2 ISA3
- SO2 ISA3
- SO2 ISA3
S02 ISAb'c
- SO2 ISAb'c
S IRIS; NRC, 20003
- IRIS; NRC, 2000b
- IRIS; NRC, 2000b'c
- IRIS; NRC, 2000b'c
We assess these co-benefits qualitatively due to time and resource limitations for this analysis.
We assess these co-benefits qualitatively because we do not have sufficient confidence in available data or methods.
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.
5-5
-------
Table 5-3. Environmental Effects of Pollutants Affected by the Mercury and Air Toxics
Standards
Benefits Category
Effect Has
Been
Specific Effect Quantified
Effect Has
Been
Monetized
More
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
Visibility in Class I areas in SE, SW, and
CA regions
Visibility in Class I areas in other regions
Visibility in 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)
PM ISA
PM ISA3
PM ISA3
Section 5.6
Section 5.6
IPCCb
PM ISAb
PM ISAb
PM ISAb
Ozone CD, Draft
Ozone ISAb
Ozone CD, Draft
Ozone ISA3
Ozone CD, Draft
Ozone ISA3'0
Ozone CD, Draft
Ozone ISA
Ozone CD, Draft
Ozone ISAb
Ozone CD, Draft
Ozone ISAb
Ozone CD, Draft
Ozone ISAb
Ozone CD, Draft
Ozone ISAb
(continued)
5-6
-------
Table 5-3. Environmental Effects of Pollutants Affected by the Mercury and Air Toxics
Standards (continued)
Benefits Category
Reduced effects from
acid deposition
Reduced effects from
nutrient enrichment
Reduced vegetation
effects from ambient
exposure to SO2 and
NOX
Reduced ecosystem
effects from exposure
to methylmercury
(through reduced
mercury deposition as
well as the role of
sulfate in methylation)
Effect Has
Been
Specific Effect Quantified
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)
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)
Injury to vegetation from SO2 exposure —
Injury to vegetation from NOX —
exposure
Effects on fish, birds, and mammals —
(e.g., reproductive effects)
Commercial, subsistence and —
recreational fishing
Effect Has
Been More
Monetized Information
- NOx SOx ISA3
- NOx SOx ISAb
- NOx SOx ISAb
- NOx SOx ISAb
NOx SOx ISAb
- NOx SOx ISAb
- NOx SOx ISAb
- NOx SOx ISAb
- NOx SOx ISAb
NOx SOx ISAb
- NOx SOx ISAb
- NOx SOx ISAb
- NOx SOx ISAb
— Mercury Study
RTCb'c
— Mercury Study
RTCb
a We assess these co-benefits qualitatively due to time and resource limitations for this analysis.
b We assess these co-benefits qualitatively because we do not have sufficient confidence in available data or
methods.
c 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.
The co-benefits analysis in this chapter relies on an array of data inputs—including air
quality modeling, health impact functions and valuation functions among others—which are
themselves subject to uncertainty and may also contribute to the overall uncertainty in this
analysis. As a means of characterizing this uncertainty we employ two primary techniques. First,
5-7
-------
we use Monte Carlo methods for characterizing random sampling error associated with the
concentration response functions from epidemiological studies and economic valuation
functions. Second, because this characterization of random statistical error may omit important
sources of uncertainty we also employ the results of an expert elicitation on the relationship
between premature mortality and ambient PM2.5 concentration (Roman et al., 2008). This
provides additional insight into the likelihood of different outcomes and about the state of
knowledge regarding the co-benefits estimates. Both approaches have different strengths and
weaknesses, which are fully described in Chapter 5 of the PM NAAQS RIA (U.S. EPA, 2006a).
While the contributions from additional data inputs to uncertainty in the results are not
quantified here, this analysis employs best practices in every aspect of its development.
Given that co-benefits of reductions in premature mortality are a dominant share of the
overall monetized co-benefits, more focus on uncertainty in mortality-related co-benefits gives
us greater confidence in our uncertainty characterization surrounding total PM2.5-related co-
benefits. Additional sensitivity analyses have been performed for the 2006 PM NAAQS RIA, and
were not specifically included here as the results would be similar and would not change the
conclusions of the analyses to support this rule. In particular, these analyses characterized the
sensitivity of the monetized co-benefits to the specification of alternate cessation lags and
income growth adjustment factors. As shown in these RIAs, the estimated co-benefits increased
or decreased in proportion to the specification of alternate income growth adjustments and
cessation lags. Therefore, readers can infer the sensitivity of the results in this RIA to these
parameters by referring to the sensitivity analyses in the PM NAAQS RIA (2006d) and Ozone
NAAQS RIA (2008a). For example, based on the results from previous analyses, the use of an
alternate lag structure would change the PM2.5-related mortality co-benefits discounted at 3%
discounted by between 10.4% and -27%; when discounted at 7%, these co-benefits change by
between 31% and -49%. When applying higher and lower income growth adjustments, the
monetary value of PM2.5 -related premature changes between 30% and -10%; the value of
chronic endpoints change between 5% and -2% and the value of acute endpoints change
between 6% and -7%.
Additionally, in this RIA we binned the estimated population exposed to projected
future baseline PM2.5 air quality levels for comparison against the "Lowest Measured Level"
(LML) of PM2.5 air quality in the mortality studies. The purpose of this analysis is to show
whether the estimated premature deaths associated with reduced PM2.5 exposure occur at or
above the range of ambient PM2.5 observations studied in Pope et al. (2002) and Laden et al.
(2006), which are the two epidemiological studies that EPA uses to estimate PM2.5-related
5-8
-------
premature mortality co-benefits. We found that a significant proportion of the avoided PM-
related premature deaths we estimated in this analysis occurred among populations exposed at
or above the LML of each study in the baseline, increasing our confidence in our estimate of the
magnitude of the PM-related premature deaths avoided. Approximately 11% of the avoided
premature deaths occur at or above an annual mean PM2.5 level of 10 u.g/m3 (the LML of the
Laden et al. 2006 study), and about 73% occur at or above an annual mean PM2.s level of 7.5
u.g/m3(the LML of the Pope et al. 2002 study). As we model avoided premature deaths among
populations exposed to levels of PM2.s that are successively lower than the LML of each study
our confidence in the results diminishes.
5.2 Benefits Analysis Methods
We follow a "damage-function" approach in calculating health co-benefits of the
modeled changes in environmental quality. This approach estimates changes in individual
health and welfare endpoints (specific effects that can be associated with changes in air quality)
and estimates values of those changes assuming independence between the values of
individual endpoints. Total benefits are calculated simply as the sum of the values for all non-
overlapping health and welfare 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; Hubbell et al., 2009; 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 health and
welfare impact analysis must first be conducted to convert air quality changes into effects that
can be assigned dollar values.
We note at the outset that 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 recent health impact analyses, our
estimates are based on the best available methods of benefits transfer. Benefits transfer is a
means 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.
5-9
-------
5.2.1 Health Impact Assessment
Health Impact Assessment (HIA) quantifies changes in the incidence of adverse health
impacts resulting from changes in human exposure to specific pollutants, such as PM2.5. 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 £rogram (BenMAP) can systematize health impact analyses by applying a database of
key input parameters, including health impact functions and population projections. Analysts
have applied the HIA approach to estimate human health impacts resulting from hypothetical
changes in pollutant levels (Hubbell et al. 2005; Davidson et al. 2007, Tagaris et al. 2009). EPA
and others have relied upon this method to predict future changes in health impacts expected
to result from the implementation of regulations affecting air quality (e.g. U.S. EPA, 2008a). For
this assessment, the HIAs are 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 impacts for coal miners.
The HIA approach used in this analysis involves three basic steps: (1) utilizing CAMx-
generated projections of PM2.s and ozone air quality 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 = y0 - (e^x - 1) • Pop
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. Tools such as BenMAP can systematize the HIA calculation process,
allowing users to draw upon a library of existing air quality monitoring data, population data
and health impact functions.
Figure 5-1 provides a simplified overview of this approach.
5-10
-------
Baseline Air Quality
Post-Polio,- Scenario Air Quality
—-
Incremental Air Quality
Improvement
Effect _
"Estimate
Mo italic-
Reduction
Figure 5-1. Illustration of BenMAP Approach
5.2.2 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. However, epidemiological studies generally provide estimates of the relative
risks of a particular health effect avoided due to a reduction in air pollution. A convenient way
to use this data in a consistent framework is to convert probabilities to units of avoided
statistical incidences. 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 $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.
5-11
-------
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 as a primary
estimate. 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 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 (Abt Associates, 2010) to estimate the health
impacts and monetized health co-benefits for the Mercury and Air Toxics Standards. Figure 5-2
shows the data inputs and outputs for the BenMAP model.
Census
Population Data
Modeled Baseline
and Post-Control
2016 Ambient
PM25 and O,
Concentrations
PM2 5 & O3 Health
Functions
Economic
Valuation
Functions
2016
Population
Projections
PMZ5&03
Incremental Air
Quality Change
PM2 5 & CyRelated
Health Impacts
Monetized PM
2.5
Woods & Poole
Population
Projections
Background
Incidence and
Prevalence Rates
and O3-related
Benefits
Blue identifies a user-selected input within the BenMAP model
Green identifies a data input generated outside of the BenMAP model
Figure 5-2. Data Inputs and Outputs for the BenMAP Model
5-12
-------
5.2.3 Adjusting the Results of the PM2.s co-benefits Analysis to Account for the Emission
Reductions in the Final Mercury and Air Toxics Standards
As described in Chapter 3 of this RIA, EPA finalized the rule requirements after the
completion of the air quality modeling for this rule. These changes to the rule affected both the
overall level and distribution of PM2.5 precursor emissions across the U.S., which in turn affect
the level of PM2.5 co-benefits. We determined that the geographic distribution of emissions
reductions resulting from the final rule requirements were sufficiently similar to the modeled
interim emissions reductions that we could adjust our co-benefits estimates to reflect these
emission changes by applying benefit per-ton estimates generated using the modeled air
quality changes.
Benefit per-ton (BPT) estimates quantify the health impacts and monetized human
health co-benefits of an incremental change in air pollution precursor emissions. In
circumstances where we are unable to perform air quality modeling because of resource or
time constraints, this approach can provide a reasonable estimate of the co-benefits of
emission reductions. EPA has used the BPT technique in previous RIAs, including the recent
Ozone NAAQS RIA (U.S. EPA, 2008a), the N02 NAAQS RIA (U.S. EPA, 2010b), the proposed
Mercury and Air Toxics Standards RIA (U.S. EPA 2011a), and the Cross-State Air Pollution Rule
(U.S. EPA, 2011b).
For this co-benefits analysis we created per-ton estimates of PM2.5-related incidence-
and monetized co-benefits based on the co-benefits of the air quality modeled scenario. Our
approach here is methodologically consistent with the technique reported in Fann, Fulcher &
Hubbell (2009), but adjusted for this analysis to better match the spatial distribution of air
quality changes expected under the Mercury and Air Toxics Standards. To derive the BPT
estimates for this analysis, we:
1. Quantified the PM2.s-related human and monetized health co-benefits ofS02 and
direct PM2.s changes for Eastern and Western states. We first estimated the health
impacts and monetized co-benefits of reductions in directly emitted PM2.5 and
particulate sulfate.1 MATS is expected to reduce both S02 and NOX emissions. In
general S02 is a precursor to particulate sulfate and NOX is a precursor to particulate
nitrate. However, there are also several interactions between the PM2.5 precursors
which cannot be easily quantified. For example, under conditions in which S02 levels
are reduced by a substantial margin, "nitrate replacement" may occur. This occurs
1 Consistent with advice from the Health Effects Subcommittee of the Science Advisory Board (U.S. EPA-SAB,
2010), we assume that each PM species is equally toxic. We quantify the change in incidence for each PM
component by applying risk coefficients based on undifferentiated PM2.5 mass.
5-13
-------
when particulate ammonium sulfate concentrations are reduced, thereby freeing up
excess gaseous ammonia. The excess ammonia is then available to react with
gaseous nitric acid to form particulate nitrate when meteorological conditions are
conducive (cold temperatures and high humidity). The impact of nitrate replacement
is also affected by concurrent NOX reductions. NOX reductions can lead to decreases
in nitrate, which competes with the process of nitrate replacement. NOX reductions
can also lead to reductions in photochemical by-products which can reduce both
particulate sulfate and secondary organic carbon PM concentrations.
We found that reductions in NOX and SOX resulting from MATS led to significant
decreases in particulate sulfate and small increases in particulate nitrate in some
locations, indicating that nitrate replacement limited the nitrate decreases from NOX
reductions in some locations. Reductions in directly emitted crustal and
carbonaceous PM2.5 (elemental carbon and organic carbon) were fairly modest.
Carbonaceous PM2.5 decreased slightly in the eastern US but did not significantly
change in the western US. We elected not to generate a NOX BPT for three reasons:
(a) reductions in NOX emissions for this rule were relatively small; (b) previous EPA
modeling indicates that PM2.5 formation is less sensitive to NOX emission reductions
on a per-u.g/m3 basis (Fann, Fulcher and Hubbell, 2009); and (c) particulate nitrate
formation is governed by complex non-linear chemistry that is difficult to
characterize using BPT estimates that are derived from a single air quality modeling
run which includes both NOX and S02 reductions. Additional modeling runs with S02
and NOX emissions changes modeled separately can provide information that can be
used to estimate NOx benefits, and these runs have been conducted for other
sectors, but have not been conducted for this rule. For the modeled scenario,
sulfate reductions contributed 95% of the health co-benefits of all PM2.5
components, with an additional 5% from direct PM2.5 reductions (see Appendix 5C).
Health co-benefits of sulfate reductions were two orders of magnitude larger than
the health disbenefits of nitrate increases. Thus, the S02 emission reductions are the
main driver for the health co-benefits of this rule.
2. Divided the health impacts and monetized co-benefits by the emission reduction for
the air quality modeling in the corresponding geographic area. For the reasons
described above, we quantified BPT estimates for S02 and directly emitted PM2.5
(separately for carbonaceous and crustal). For S02, we generated an array of eastern
and western BPT estimates by dividing the particulate sulfate-related co-benefits in
the eastern and western U.S. by the total S02-related emission reductions in these
two areas. As the chemistry of nitrate formation is complex and non-linear, nitrate
impacts were excluded from the BPT analysis. Nitrates can be reduced when NOX
emissions are reduced or increased when S02 emissions are reduced. The increased
nitrate health impacts in the modeled interim scenario were two orders of
magnitude smaller than the sulfate health benefits. Thus, we estimate that including
nitrate health impacts on the calculation for S02 BPT would reduce the S02 BPT by 1-
2%, with a similar magnitude impact on the total health benefits of the rule.
5-14
-------
Carbonaceous and crustal PM2.5 BPT estimates were similarly generated using the
co-benefits and emissions of those species.
The resulting BPT estimates (listed in Table 5C-3) were then multiplied by the projected
S02 (1.33 million tons), carbonaceous PM2.5 (6,100 tons), and crustal PM2.5 (39,000 tons)
emission reductions for the final policy to produce an estimate of the PM2.5-related health
impacts and monetized co-benefits. Due to time limitations, direct PM2.5 benefits are based on
direct PM2.5 emission reductions from an earlier policy scenario. However, since direct PM2.5
benefits contribute only approximately 5% to the total PM2.5 health co-benefits of this rule, and
differences between direct PM2.5 emission reductions between the earlier and final policy
scenarios are expected to be modest, use of earlier PM2.5 emission changes is unlikely to
materially affect the results. Additional details on the BPT methodology and derivation are
given in Appendix 5C.
An implicit assumption in our approach is that the size and distribution of S02 emissions,
and the relative levels of NOX and S02 emissions, are fairly similar in the modeled and revised
policy cases. In general, the modeled and revised policy cases achieve roughly similar levels of
S02 reductions (1.42 versus 1.33 million tons, respectively) with a similar distribution among
states. However, for some states (notably Alabama, Colorado, Louisiana, Michigan, Missouri,
North Dakota, Oklahoma, and Texas), S02 emission reductions were lower for the final case
versus the interim case. By far, the greatest difference in S02 emission reductions was in
Michigan where the final case emission reduction was 70% lower than for the interim case. In a
few states (notably Arkansas, Ohio, and South Carolina), S02 emission reductions were slightly
larger for the final case versus the interim case. Since differences between the interim and final
cases are not concentrated in any particular region of the country and the overall distribution of
emission reductions is similar, we conclude that it is reasonable to apply BPT values derived
from the interim case to the final case. While NOX emissions reductions decreased by 70%
between the interim and final cases (141,000 vs. 46,000 tons), the impact of NOX on PM2.5
concentrations and mortality is very minor relative to the impact of S02 emission reductions.
Therefore, differences in the magnitude and distribution of NOX emission reductions are likely
to have only a minor effect on results.
We did not develop ozone BPT estimates for this rule for two reasons. First, the overall
level of ozone-related co-benefits in the modeled case is relatively small compared to those
associated with PM2.5 reductions, due in part to the modest NOX emission reductions. Second,
the complex non-linear chemistry of ozone formation introduces uncertainty to the
5-15
-------
development and application of BPT estimates. Taken together, these factors argued against
developing ozone BPT estimates for this RIA.
As there is no analogous approach for estimating visibility co-benefits using the BPT
approach, visibility co-benefits are calculated for the modeled interim policy scenario only and
are not included in estimate of co-benefits for the final rule. However, since the magnitude of
S02 emission reductions did not significantly change in the visibility study areas between the
interim and final emissions scenarios, we expect the visibility benefit for the final policy
scenario would be similar to that calculated for the interim policy scenario ($1.1 billion in total
for the U.S., using 2007$; see Appendix 5C).
5.3 Uncertainty Characterization
As for any complex analysis using estimated parameters and inputs from numerous
models, there are likely to be many sources of uncertainty affecting estimated results, including
emission inventories, air quality models (with their associated parameters and inputs),
epidemiological health effect estimates, estimates of values (both from WTP and COI studies),
population estimates, income estimates, and estimates of the future state of the world (i.e.,
regulations, technology, and human behavior). Each of these inputs may be uncertain and,
depending on its role in the co-benefits analysis, may have a disproportionately large impact on
estimates of total monetized co-benefits. For example, emissions estimates are used in the first
stage of the analysis. As such, any uncertainty in emissions estimates will be propagated
through the entire analysis. When compounded with uncertainty in later stages, small
uncertainties in emission levels can lead to large impacts on total monetized co-benefits.
The National Research Council (NRC) (2002, 2008) highlighted the need for EPA to
conduct rigorous quantitative analysis of uncertainty in its benefits estimates and to present
these estimates to decision makers in ways that foster an appropriate appreciation of their
inherent uncertainty. In general, the NRC concluded that EPA's methodology for calculating the
benefits of reducing air pollution is reasonable and informative in spite of inherent
uncertainties. Since the publication of these reports, EPA continues to improve the
characterization of uncertainties for both health incidence and benefits estimates. We use both
Monte Carlo analysis and expert-derived concentration-response functions to assess
uncertainty quantitatively, as well as to provide a qualitative assessment for those aspects that
we are unable to address quantitatively.
First, we used Monte Carlo methods to characterize both sampling error and variability
across the economic valuation functions, including random sampling error associated with the
5-16
-------
concentration response functions from epidemiological studies and random effects modeling.
Monte Carlo simulation uses random sampling from distributions of parameters to characterize
the effects of uncertainty on output variables, such as incidence of premature mortality.
Specifically, we used Monte Carlo methods to generate confidence intervals around the
estimated health impact and dollar benefits. The reported standard errors in the
epidemiological studies determined the distributions for individual effect estimates.
Second, because characterization of random statistical error omits important sources of
uncertainty (e.g., in the functional form of the model—e.g., whether or not a threshold may
exist), we also incorporate the results of an expert elicitation on the relationship between
premature mortality and ambient PM2.s concentration (Roman et al., 2008). Use of the expert
elicitation and incorporation of the standard errors approaches provide insights into the
likelihood of different outcomes and about the state of knowledge regarding the benefits
estimates. However, there remain significant unquantified uncertainties present in upstream
inputs including emission and air quality. Both uncertainty characterization approaches have
different strengths and weaknesses, as detailed in Chapters of the PM NAAQS RIA (U.S. EPA,
2006a).
In benefit analyses of air pollution regulations conducted to date, the estimated impact
of reductions in premature mortality has accounted for 85% to 95% of total monetized benefits.
Therefore, it is particularly important to attempt to characterize the uncertainties associated
with reductions in premature mortality. The health impact functions used to estimate avoided
premature deaths associated with reductions in ozone have associated standard errors that
represent the statistical errors around the effect estimates in the underlying epidemiological
studies. In our results, we report credible intervals based on these standard errors, reflecting
the uncertainty in the estimated change in incidence of avoided premature deaths. We also
provide multiple estimates, to reflect model uncertainty between alternative study designs.
For premature mortality associated with exposure to PM, we follow the same approach
used in the RIA for 2006 PM NAAQS (U.S. EPA, 2006a), presenting two empirical estimates of
premature deaths avoided, and a set of twelve estimates based on results of the expert
elicitation study. Even these multiple characterizations, including confidence intervals, omit the
contribution to overall uncertainty of uncertainty in air quality changes, baseline incidence
rates, populations exposed and transferability of the effect estimate to diverse locations.
Furthermore, the approach presented here does not 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
5-17
-------
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.
EPA estimates PM-related mortality without assuming a health effect threshold at low
concentrations, based on the current body of scientific literature (U.S. EPA-SAB, 2009a, U.S.
EPA-SAB, 2009b). However, as we model mortality impacts among populations exposed to
levels of PM2.s that are successively lower than the lowest measured level (LML) in each
epidemiology study our confidence in the results diminishes. In addition to the uncertainty
analyses described above, we therefore include an assessment of the mortality benefits
accruing to populations exposed to baseline PM2.5 concentrations above the LML in the two
main epidemiology studies used to quantify benefits (see Section 5.7). Based on the modeled
interim baseline which is approximately equivalent to the final baseline (see Appendix 5A), 11%
and 73% of the estimated avoided mortality impacts occur at or above an annual mean PM2.5
level of 10 u.g/m3 (the LML of the Laden et al. 2006 study) and 7.5 u.g/m3(the LML of the Pope
et al. 2002 study), respectively.
Key sources of uncertainty in the PM2.5 health impact assessment include:
• gaps in scientific data and inquiry;
• variability in estimated relationships, such as epidemiological effect estimates,
introduced through differences in study design and statistical modeling;
• errors in measurement and projection for variables such as population growth rates;
• errors due to misspecification of model structures, including the use of surrogate
variables, such as using PMi0 when PM2.5 is not available, excluded variables, and
simplification of complex functions;
• biases due to omissions or other research limitations; and
• additional uncertainties from benefits transfer method using BPT estimates.
In Table 5-4 we summarize some of the key uncertainties in the benefits analysis.
5-18
-------
Table 5-4. Primary Sources of Uncertainty in the Benefits Analysis
1. Uncertainties Associated with Impact Functions
• The value of the ozone or PM effect estimate in each impact function.
• Application of a single impact function to pollutant changes and populations in all locations.
• Similarity of future-year impact functions to current impact functions.
• Correct functional form of each impact function.
• Extrapolation of effect estimates beyond the range of ozone or PM concentrations observed in the source
epidemiological study.
• Application of impact functions only to those subpopulations matching the original study population.
2. Uncertainties Associated with CAMx-Modeled Ozone and PM Concentrations
• Responsiveness of the models to changes in precursor emissions from the control policy.
• Projections of future levels of precursor emissions, especially ammonia and crustal materials.
• Lack of ozone and PM2.5 monitors in all rural areas requires extrapolation of observed ozone data from
urban to rural areas.
3. Uncertainties Associated with PM Mortality Risk
• Limited scientific literature supporting a direct biological mechanism for observed epidemiological
evidence.
• Direct causal agents within the complex mixture of PM have not been identified.
• The extent to which adverse health effects are associated with low-level exposures that occur many times
in the year versus peak exposures.
• The extent to which effects reported in the long-term exposure studies are associated with historically
higher levels of PM rather than the levels occurring during the period of study.
• Reliability of the PM^.s monitoring data in reflecting actual PM^.s exposures.
4. Uncertainties Associated with Possible Lagged Effects
• The portion of the PM-related long-term exposure mortality effects associated with changes in annual PM
levels that would occur in a single year is uncertain as well as the portion that might occur in subsequent
years.
5. Uncertainties Associated with Baseline Incidence Rates
• Some baseline incidence rates are not location specific (e.g., those taken from studies) and therefore may
not accurately represent the actual location-specific rates.
• Current baseline incidence rates may not approximate well baseline incidence rates in 2016.
• Projected population and demographics may not represent well future-year population and demographics.
6. Uncertainties Associated with Economic Valuation
• Unit dollar values associated with health and welfare endpoints are only estimates of mean WTP and
therefore have uncertainty surrounding them.
• Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates because
of differences in income or other factors.
7. Uncertainties Associated with Aggregation of Monetized Benefits
• Health and welfare benefits estimates are limited to the available impact functions. Thus, unquantified or
unmonetized benefits are not included.
PM2.5 mortality benefits represent a substantial proportion of total monetized co-
benefits (over 90%), and these estimates have following key assumptions and uncertainties.
1. The PM2.5-related co-benefits were derived through a benefit per-ton approach,
which does not fully reflect local variability in population density, meteorology,
exposure, baseline health incidence rates, or other local factors that might lead to
5-19
-------
an over-estimate or under-estimate of the actual co-benefits of controlling PM
precursors. In addition, differences in the distribution of emissions reductions
across states between the modeled scenario and the final rule scenario add
uncertainty to the final benefits estimates.
2. 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 via transported precursors emitted from EGUs may differ
significantly from direct PM2.5 released from diesel engines and other industrial
sources, but the scientific evidence is not yet sufficient to allow differential effects
estimates by particle type.
3. We assume that the health impact function for fine particles is linear within the
range of ambient concentrations under consideration. Thus, the estimates include
health co-benefits from reducing fine particles in areas with varied concentrations of
PM2.5, including both regions that are in attainment with fine particle standard and
those that do not meet the standard down to the lowest modeled concentrations.
5.4 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, effect coefficients, incidence rates and economic valuation. Our
approach here is generally consistent with the Regulatory Impact Analysis for the Cross-State
Air Pollution Rule (U.S. EPA, 2011b).
5.4.1 Demographic Data
Quantified and monetized human health impacts depend on the demographic
characteristics of the population, including age, location, and income. We use projections based
on economic forecasting models developed by Woods and Poole, Inc. (Woods and Poole, 2008).
The Woods and Poole (WP) database contains county-level projections of population by age,
sex, and race out to 2030. 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:
1. First, national-level variables such as income, employment, and populations are
forecasted.
5-20
-------
2. Second, employment projections are made for 172 economic areas defined by the
Bureau of Economic Analysis, 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.
3. 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.
4. 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 2016 based on historical rates of mortality, fertility,
and migration.
5.4.2 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 EPA's Integrated
Science Assessments (which replace the Criteria Document), with input and advice from the
EPA Science Advisory Board - Health Effects Subcommittee (SAB-HES), a scientific review panel
specifically established to provide advice on the use of the scientific literature in developing
benefits analyses for air pollution regulations (http://www.epa.gov/sab/). 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 which can be used to evaluate
population-level impacts of reductions in ambient pollution levels in a health impact
assessment.
5-21
-------
For the data-derived estimates, we relied on the published scientific literature to
ascertain the relationship between PM and 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. The selection of C-R
functions for the benefits analysis is guided by the goal of achieving a balance between
comprehensiveness and scientific defensibility. In general, the use of results from more than a
single study can provide a more robust estimate of the relationship between a pollutant and a
given health effect. However, there are often differences between studies examining the same
endpoint, making it difficult to pool the results in a consistent manner. For example, studies
may examine different pollutants or different age groups. For this reason, we consider very
carefully the set of studies available examining each endpoint and select a consistent subset
that provides a good balance of population coverage and match with the pollutant of interest.
In many cases, either because of a lack of multiple studies, consistency problems, or clear
superiority in the quality or comprehensiveness of one study over others, a single published
study is selected as the basis of the effect estimate.
Table 5-5. Criteria Used When Selecting C-R Functions
Consideration
Comments
Peer-Reviewed Peer-reviewed research is preferred to research that has not undergone the peer-review
Research process.
Study Type 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.
Study Period Studies examining a relatively longer period of time (and therefore having more data) are
preferred, because they have greater statistical power to detect effects. More recent
studies 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.
(continued)
5-22
-------
Table 5-5. Criteria Used when Selecting C-R Functions (continued)
Consideration
Comments
Population Attributes
Study Size
Study Location
Pollutants Included in
Model
Measure of PM
Economically Valuable
Health Effects
Non-overlapping
Endpoints
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 C-R 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, either 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.
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.
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 Technical Appendices provides details of the procedures used to
combine multiple impact functions (Abt Associates, 2010). In general, we used fixed or random
effects models to pool estimates from different 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
5-23
-------
determined whether the data suggest that we should reject this null hypothesis, in which case
we would use the random effects model. Pooled impact functions are used to estimate hospital
admissions and asthma exacerbations. For more details on methods used to pool incidence
estimates, see the BenMAP Manual Appendices (Abt Associates, 2010), which are available with
the BenMAP software at http://www.epa.gov/benmap.html.
Effect estimates selected for a given health endpoint were applied consistently across all
locations nationwide. This applies to both impact functions defined by a single effect estimate
and those defined by a pooling of multiple effect estimates. Although the effect estimate may,
in fact, vary from one location to another (e.g., because of differences in population
susceptibilities or differences in the composition of PM), location-specific effect estimates are
generally not available.
The specific studies from which effect estimates for the primary analysis are drawn are
included in Table 5-6. 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. Below we provide the basis for selecting these studies.
5.4.2.1 PM2.5 Premature Mortality Effect Coefficients
Both long- and short-term exposures to ambient levels of PM2.s air pollution have been
associated with increased risk of premature mortality. The size of the mortality risk 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.
Table 5-6. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts3
Endpoint
Pollutant
Study
Study
Population
Premature Mortality
Premature mortality-
cohort study, all-cause
Premature mortality, total
exposures
Premature mortality— all-
cause
PM2.5
(annual avg)
PM2.5
(annual avg)
PM2.5
(annual avg)
Pope et al. (2002)
Laden et al. (2006)
Expert Elicitation (Roman et al., 2008)
Woodruff etal. (2006)
>29 years
>25 years
>24 years
Infant (<1 year)
(continued)
5-24
-------
Table 5-6. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts3
(continued)
Endpoint
Pollutant
Study
Study
Population
Chronic Illness
Chronic bronchitis
Non-fatal heart attacks
Hospital Admissions
Respiratory
Cardiovascular
Asthma-related ER visits
PM2.5
(annual avg)
PM2.5
(24-hour avg)
PM2.5
(24-hour avg)
PM2.5
(24-hour avg)
PM2.5
(24-hour avg)
PM2.5
(24-hour avg)
PM2.5
(24-hour avg)
PM2.5
(24-hour avg)
PM2.5
(24-hour avg)
Abbey et al. (1995)
Peters et al. (2001)
Pooled estimate:
Moolgavkar (2003)-ICD 490-496 (COPD)
Ito (2003)-ICD 490-496 (COPD)
Moolgavkar (2000)-ICD 490-496 (COPD)
Ito (2003)— ICD 480-486 (pneumonia)
Sheppard (2003)-ICD 493 (asthma)
Pooled estimate:
Moolgavkar (2003)-ICD 390-429 (all
cardiovascular)
Ito (2003)-ICD 410-414, 427-428 (ischemic
heart disease, dysrhythmia, heart failure)
Moolgavkar (2000)-ICD 390-429 (all
cardiovascular)
Norris et al. (1999)
>26 years
Adults (>18
years)
>64 years
20-64 years
>64 years
<65 years
>64 years
20-64 years
0-18 years
Other Health Endpoints
Acute bronchitis
PM2.5
(annual avg)
Dockeryetal. (1996)
8-12 years
Upper respiratory symptoms PM10
(24-hour avg)
Lower respiratory symptoms PM25
(24-hour avg)
Pope etal. (1991)
Schwartz and Neas (2000)
Asthma exacerbations
PM2.5
(24-hour avg)
Asthmatics, 9-11
years
7-14 years
Pooled estimate: 6-18 years
Ostro et al. (2001) (cough, wheeze and shortness
of breath)
Vedal etal. (1998) (cough)
(continued)
5-25
-------
Table 5-6. Health Endpoints and Epidemiological Studies Used to Quantify Health Impacts3
(continued)
Endpoint
Work loss days
Pollutant Study
PM2.5 Ostro (1987)
(24-hour avg)
Study
Population
18-65 years
Minor Restricted Activity PM25 Ostro and Rothschild (1989) 18-65 years
Days (MRADs) (24-hour avg)
a Studies or air quality metrics highlighted in blue represent updates incorporated since the 2005 CAIR RIA
The original study populations were 8 to 13 for the Ostro et al. (2001) study and 6 to 13 for the Vedal et al.
(1998) study. Based on advice from the Science Advisory Board Health Effects Subcommittee (SAB-HES), we
extended the applied population to 6 to 18, reflecting the common biological basis for the effect in children in
the broader age group. See: U.S. Science Advisory Board. 2004. Advisory 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-04-004. See also National Research Council (NRC). 2002. Estimating the Public Health
Benefits of Proposed Air Pollution Regulations. Washington, DC: The National Academies Press.
Although a number of uncertainties remain to be addressed by continued research
(NRC, 2002), a substantial body of published scientific literature documents the correlation
between elevated PM2.5 concentrations and increased mortality rates (U.S. EPA, 2009a). 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 PM exposures over multiple years (i.e., long-term exposures) and
community-level annual mortality rates. Researchers have found statistically significant
associations between PM2.5 and premature mortality using both types of studies. In general, the
risk estimates based on the cohort studies are larger than those derived from time-series
studies. Cohort analyses are thought to better capture the full public health impact of exposure
to air pollution over time, because they account for the effects of long-term exposures and
possibly some component of short-term exposures (Kunzli et al., 2001; NRC, 2002). This section
discusses some of the issues surrounding the estimation of PM2.5-related premature mortality.
To demonstrate the sensitivity of the benefits estimates to the specific sources of information
regarding the impact of PM2.5 exposures on the risk of premature death, we are providing
estimates in our results tables based on studies derived from the epidemiological literature and
from the EPA sponsored expert elicitation. The epidemiological studies from which these
estimates are drawn are described below. The expert elicitation project and the derivation of
effect estimates from the expert elicitation results are described in the 2006 PM2.5 NAAQS RIA
and Roman et al. (2008). In the interest of brevity we do not repeat those details here.
5-26
-------
However, Figure 5-13 summarizes the estimated PM2.5-related premature mortalities avoided
using risk estimates drawn from the expert elicitation.
Over a dozen epidemiological studies have found significant associations between
various measures of long-term exposure to PM and elevated rates of annual mortality,
beginning with Lave and Seskin (1977). Most of the published studies found positive (but not
always statistically significant) associations with available PM indices such as total suspended
particles (TSP). However, exploration of alternative model specifications sometimes raised
questions about causal relationships (e.g., Lipfert et al., 1989). These early "ecological cross-
sectional" studies (Lave and Seskin, 1977; Ozkaynak and Thurston, 1987) were criticized for a
number of methodological limitations, particularly for inadequate control at the individual level
for variables that are potentially important in causing mortality, such as wealth, smoking, and
diet.
Over the last 17 years, several studies using "prospective cohort" designs have been
published that appear to be consistent with the earlier body of literature. These new
"prospective cohort" studies reflect a significant improvement over the earlier work because
they include individual level information with respect to health status and residence. The most
extensive analyses have been based on data from two prospective cohort groups, often
referred to as the Harvard "Six-Cities Study" (Dockery et al., 1993; Laden et al., 2006) and the
"American Cancer Society or ACS study" (Pope et al., 1995; Pope et al., 2002; Pope et al., 2004,
Krewski et al. 2009); these studies have found consistent relationships between fine particle
indicators and premature mortality across multiple locations in the United States. A third major
data set comes from the California-based 7th Day Adventist Study (e.g., Abbey et al., 1999),
which reported associations between long-term PM exposure and mortality in men. Results
from this cohort, however, have been inconsistent, and the air quality results are not
geographically representative of most of the United States, and the lifestyle of the population is
not reflective of much of the U.S. population. Analysis is also available for a cohort of adult
male veterans diagnosed with hypertension has been examined (Lipfert et al., 2000; Lipfert
et al., 2003, 2006). The characteristics of this group differ from the cohorts in the Six-Cities,
ACS, and 7th Day Adventist studies with respect to income, race, health status, and smoking
status. Unlike previous long-term analyses, this study found some associations between
mortality and ozone but found inconsistent results for PM indicators. Because of the selective
nature of the population in the veteran's cohort, we have chosen not to include any effect
estimates from the Lipfert et al. (2000) study in our co-benefits assessment.
5-27
-------
Given their consistent results and broad geographic coverage, and importance in
informing the NAAQS development process, the Six-Cities and ACS data have been particularly
important in benefits analyses. The credibility of these two studies is further enhanced by the
fact that the initial published studies (Pope et al., 1995 and 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) (Krewski et al., 2000). The final results of the
reanalysis were then independently peer reviewed by a Special Panel of the HEI Health Review
Committee. The results of these reanalyses confirmed and expanded the conclusions of the
original investigators. While the HEI reexamination lends credibility to the original studies, it
also highlights sensitivities concerning the relative impact of various pollutants, such as S02, the
potential role of education in mediating the association between pollution and mortality, and
the influence of spatial correlation modeling. Further confirmation and extension of the
findings of the 1993 Six City Study and the 1995 ACS study were recently completed using more
recent air quality and a longer follow-up period for the ACS cohort was published over the past
several years (Pope et al., 2002, 2004; Laden et al., 2006, Krewski et al. 2009). The follow up to
the Harvard Six City Study both confirmed the effect size from the first analysis and provided
additional confirmation that reductions in PM2.5 are likely to result in reductions in the risk of
premature death. This additional evidence stems from the observed reductions in PM2.5 in each
city during the extended follow-up period. Laden et al. (2006) found that mortality rates
consistently went down at a rate proportionate to the observed reductions in PM2.5.
A number of additional analyses have been conducted on the ACS cohort data (Jerrett et
al., 2009; Pope et al., 2009). These studies have continued to find a strong significant
relationship between PM2.5 and mortality outcomes and life expectancy. Specifically, much of
the recent research has suggested a stronger relationship between cardiovascular mortality and
lung cancer mortality with PM2.5, and a less significant relationship between respiratory-related
mortality and PM2.5. The extended analyses of the ACS cohort data (Krewski et al. 2009)
provides additional refinements to the analysis of PM-related mortality by (a) extend the
follow-up period by 2 years to the year 2000, for a total of 18 years; (b) incorporate ecological.,
or neighborhood-level co-variates so as to better estimate personal exposure; (c) perform an
extensive spatial analysis using land use regression modeling. These additional refinements may
make this analysis well-suited for the assessment of PM-related mortality for EPA benefits
analyses.
In developing and improving the methods for estimating and valuing the potential
reductions in mortality risk over the years, EPA consulted with the SAB-HES. That panel
5-28
-------
recommended using long-term prospective cohort studies in estimating mortality risk reduction
(U.S. EPA-SAB, 1999). This recommendation has been confirmed by a report from the National
Research Council, which stated that "it is essential to use the cohort studies in benefits analysis
to capture all important effects from air pollution exposure" (NRC, 2002, p. 108). More
specifically, the SAB recommended emphasis on the ACS study because it includes a much
larger sample size and longer exposure interval and covers more locations (e.g., 50 cities
compared to the Six Cities Study) than other studies of its kind. Because of the refinements in
the extended follow-up analysis, the SAB-HES recommended using the Pope et al. (2002) study
as the basis for the primary mortality estimate for adults and suggests that alternate estimates
of mortality generated using other cohort and time-series studies could be included as part of
the sensitivity analysis (U.S. EPA-SAB, 2004a). The PM NAAQS Risk and Exposure Assessment
(U.S. EPA, 2010c) utilized risk coefficients drawn from the Krewski et al. (2009) study. In a
December of 2009 consultation with the SAB-HES, the Agency proposed utilizing the Krewski
et al. (2009) extended analysis of the ACS cohort data. The panel is scheduled to issue an
advisory in early 2010.
As noted above, since 2004 SAB review, an extended follow-up of the Harvard Six cities
study has been published (Laden et al., 2006) and in recent RIAs (see for example the Cross-
State Air Pollution Rule RIA, U.S. EPA 2011b), we have included this estimate of mortality
impacts based on application of the C-R function derived from this study. We use this specific
estimate to represent the Six Cities study because it both reflects among the most up-to-date
science and was cited by many of the experts in their elicitation responses. It is clear from the
expert elicitation that the results published in Laden et al. (2006) are potentially influential, and
in fact the expert elicitation results encompass within their range the estimates from both the
Pope et al. (2002) and Laden et al. (2006) studies (see Figure 5-3). These are logical choices for
anchor points in our presentation because, while both studies are well designed and peer
reviewed, there are strengths and weaknesses inherent in each, which we believe argues for
using both studies to generate benefits estimates.
5.4.2.2 Chronic Bronchitis (CB)
CB 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. CB affects an estimated 5 percent of the U.S.
population (American Lung Association, 1999). A limited number of studies have estimated the
impact of air pollution on new incidences of CB. Schwartz (1993) and Abbey et al. (1995)
provide evidence that long-term PM exposure gives rise to the development of CB in the United
States. Because PM2.s reductions are expected from MATS, this analysis uses only the Abbey et
5-29
-------
al. (1995) study, because it is the only study focusing on the relationship between PM2.5 and
new incidences of CB.
5.4.2.3 Non-fatal Myocardial Infarctions (Heart Attacks)
Non-fatal heart attacks have been linked with short-term exposures to PM2.5 in the
United States (Peters et al., 2001) and other countries (Poloniecki et al., 1997). We used a
recent study by Peters et al. (2001) as the basis for the impact function estimating the
relationship between PM2.s and non-fatal heart attacks. Peters et al. is the only available U.S.
study to provide a specific estimate for heart attacks. Other studies, such as Samet et al. (2000)
and Moolgavkar (2000), show a consistent relationship between all cardiovascular hospital
admissions, including those for non-fatal heart attacks, and PM. Given the lasting impact of a
heart attack on long-term health costs and earnings, we provide a separate estimate for non-
fatal heart attacks. The estimate used in the MATS analysis is based on the single available U.S.
effect estimate. The finding of a specific impact on heart attacks is consistent with hospital
admission and other studies showing relationships between fine particles and cardiovascular
effects both within and outside the United States. Several epidemiologic studies (Liao et al.,
1999; Gold et al., 2000; Magari et al., 2001) have shown that heart rate variability (an indicator
of how much the heart is able to speed up or slow down in response to momentary stresses) is
negatively related to PM levels. Heart rate variability is a risk factor for heart attacks and other
coronary heart diseases (Carthenon et al., 2002; Dekker et al., 2000; Liao et al., 1997; Tsuji et
al., 1996). As such, significant impacts of PM on heart rate variability are consistent with an
increased risk of heart attacks.
5.4.2.4 Hospital and Emergency Room Admissions
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. Because of this, many of the hospital admission endpoints use pooled impact
functions based on the results of a number of studies. In addition, some studies have examined
the relationship between air pollution and emergency room visits. Since most emergency room
visits do not result in an admission to the hospital (the majority of people going to the
emergency room are treated and return home), we treat hospital admissions and emergency
room visits separately, taking account of the fraction of emergency room visits that are
admitted to the hospital.
The two main groups of hospital admissions estimated in this analysis are respiratory
admissions and cardiovascular admissions. There is not much evidence linking ozone or PM
5-30
-------
with other types of hospital admissions. The only type of emergency room visits that have been
consistently linked to ozone and PM in the United States are asthma-related visits.
To estimate avoided incidences of cardiovascular hospital admissions associated with
PM2.5, we used studies by Moolgavkar (2003) and Ito (2003). Additional published studies show
a statistically significant relationship between PMi0 and cardiovascular hospital admissions.
However, given that the control options we are analyzing are expected to reduce primarily
PM2.5, we focus on the two studies that examine PM2.5. Both of these studies provide an effect
estimate for populations over 65, allowing us to pool the impact functions for this age group.
Only Moolgavkar (2000) provided a separate effect estimate for populations 20 to 64.2 Total
cardiovascular hospital admissions are thus the sum of the pooled estimate for populations
over 65 and the single study estimate for populations 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.
To estimate total avoided incidences of respiratory hospital admissions, we used impact
functions for several respiratory causes, including chronic obstructive pulmonary disease
(COPD), pneumonia, and asthma. As with cardiovascular admissions, additional published
studies show a statistically significant relationship between PMi0 and respiratory hospital
admissions. We used only those focusing on PM2.5. Both Moolgavkar (2000) and Ito (2003)
provide effect estimates for COPD in populations over 65, allowing us to pool the impact
functions for this group. Only Moolgavkar (2000) provides a separate effect estimate for
populations 20 to 64. Total COPD hospital admissions are thus the sum of the pooled estimate
for populations over 65 and the single study estimate for populations 20 to 64. Only Ito (2003)
estimated pneumonia and only for the population 65 and older. In addition, Sheppard (2003)
provided an effect estimate for asthma hospital admissions for populations under age 65. Total
avoided incidences of PM-related respiratory-related hospital admissions are the sum of COPD,
pneumonia, and asthma admissions.
2 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 percent) 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.
5-31
-------
To estimate the effects of PM air pollution reductions on asthma-related ER visits, we
use the effect estimate from a study of children 18 and under by Norris et al. (1999). As noted
earlier, there is another study by Schwartz examining a broader age group (less than 65), but
the Schwartz study focused on PMi0 rather than PM2.5. We selected the Norris et al. (1999)
effect estimate because it better matched the pollutant of interest. Because children tend to
have higher rates of hospitalization for asthma relative to adults under 65, we will likely capture
the majority of the impact of PM2.s on asthma emergency room visits in populations under 65,
although there may still be significant impacts in the adult population under 65.
To estimate avoided incidences of respiratory hospital admissions associated with
ozone, we used a number of studies examining hospital admissions for a range of respiratory
illnesses, including pneumonia and COPD. Two age groups, adults over 65 and children under 2,
were examined. For adults over 65, Schwartz (1995) provides effect estimates for two different
cities relating ozone and hospital admissions for all respiratory causes (defined as ICD codes
460-519). Impact functions based on these studies were pooled first before being pooled with
other studies. Two studies (Moolgavkar et al., 1997; Schwartz, 1994a) examine ozone and
pneumonia hospital admissions in Minneapolis. One additional study (Schwartz, 1994b)
examines ozone and pneumonia hospital admissions in Detroit. The impact functions for
Minneapolis were pooled together first, and the resulting impact function was then pooled with
the impact function for Detroit. This avoids assigning too much weight to the information
coming from one city. For COPD hospital admissions, two studies are available: Moolgavkar
et al. (1997), conducted in Minneapolis, and Schwartz (1994b), conducted in Detroit. These two
studies were pooled together. To estimate total respiratory hospital admissions for adults over
65, COPD admissions were added to pneumonia admissions, and the result was pooled with the
Schwartz (1995) estimate of total respiratory admissions. Burnett et al. (2001) is the only study
providing an effect estimate for respiratory hospital admissions in children under 2.
We used two studies as the source of the concentration-response functions we used to
estimate the effects of ozone exposure on asthma-related emergency room (ER) visits: Peel et
al. (2005) and Wilson et al. (2005). We estimated the change in ER visits using the effect
estimate(s) from each study and then pooled the results using the random effects pooling
technique (see Abt, 2005). The Peel et al. (2005) study estimated asthma-related ER visits for all
ages in Atlanta, using air quality data from 1993 to 2000. Using Poisson generalized estimating
equations, the authors found a marginal association between the maximum daily 8-hour
average ozone level and ER visits for asthma over a 3-day moving average (lags of 0,1, and 2
days) in a single pollutant model. Wilson et al. (2005) examined the relationship between ER
5-32
-------
visits for respiratory illnesses and asthma and air pollution for all people residing in Portland,
Maine from 1998-2000 and Manchester, New Hampshire from 1996-2000. For all models used
in the analysis, the authors restricted the ozone data incorporated into the model to the
months ozone levels are usually measured, the spring-summer months (April through
September). Using the generalized additive model, Wilson et al. (2005) found a significant
association between the maximum daily 8-hour average ozone level and ER visits for asthma in
Portland, but found no significant association for Manchester. Similar to the approach used to
generate effect estimates for hospital admissions, we used random effects pooling to combine
the results across the individual study estimates for ER visits for asthma. The Peel et al. (2005)
and Wilson et al. (2005) Manchester estimates were not significant at the 95 percent level, and
thus, the confidence interval for the pooled incidence estimate based on these studies includes
negative values. This is an artifact of the statistical power of the studies, and the negative
values in the tails of the estimated effect distributions do not represent improvements in health
as ozone concentrations are increased. Instead, these should be viewed as a measure of
uncertainty due to limitations in the statistical power of the study. We included both hospital
admissions and ER visits as separate endpoints associated with ozone exposure because our
estimates of hospital admission costs do not include the costs of ER visits and most asthma ER
visits do not result in a hospital admission.
5.4.2.5 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 ambient levels of
ozone and PM. The sources for the effect estimates used to quantify these effects are described
below.
Around 4 percent of U.S. children between the ages of 5 and 17 experience episodes of
acute bronchitis annually (American Lung Association, 2002c). 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,3 with the exception of cough, most acute
bronchitis symptoms abate within 7 to 10 days. Incidence of episodes of acute bronchitis in
children between the ages of 5 and 17 were estimated using an effect estimate developed from
Dockeryetal. (1996).
See http://www.nlm.nih.gov/medlineplus/ency/article/000124.htm, accessed January 2002.
5-33
-------
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).
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. Incidences
of upper respiratory symptoms in asthmatic children aged 9 to 11 are estimated using an effect
estimate developed from Pope et al. (1991).
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. Most studies examining school absence rates have found little or no association with
PM2.5, but several studies have found a significant association between ozone levels and school
absence rates. We used two recent studies, Gilliland et al. (2001) and Chen et al. (2000), to
estimate changes in absences (school loss days) due to changes in ozone levels. The Gilliland et
al. study estimated the incidence of new periods of absence, while the Chen et al. study
examined absence on a given day. We converted the Gilliland estimate to days of absence by
multiplying the absence periods by the average duration of an absence. We estimated an
average duration of school absence of 1.6 days by dividing the average daily school absence
rate from Chen et al. (2000) and Ransom and Pope (1992) by the episodic absence rate from
Gilliland et al. (2001). This provides estimates from Chen et al. (2000) and Gilliland et al. (2001),
which can be pooled to provide an overall estimate.
Minor Restricted Activity Days (MRAD) occur 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.s and ozone on MRAD was estimated using an effect
estimate derived from Ostro and Rothschild (1989).
For this analysis, we have followed the SAB-HES recommendations regarding asthma
exacerbations in developing the primary estimate. To prevent double-counting, we focused the
estimation on asthma exacerbations occurring in children and excluded adults from the
5-34
-------
calculation.4 Asthma exacerbations occurring in adults are assumed to be captured in the
general population endpoints such as work loss days and MRADs. Consequently, if we had
included an adult-specific asthma exacerbation estimate, we would likely double-count
incidence for this endpoint. However, because the general population endpoints do not cover
children (with regard to asthmatic effects), an analysis focused specifically on asthma
exacerbations for children (6 to 18 years of age) could be conducted without concern for
double-counting.
To characterize asthma exacerbations in children, we selected two studies (Ostro et al.,
2001; Vedal et al., 1998) that followed panels of asthmatic children. Ostro et al. (2001) followed
a group of 138 African-American children in Los Angeles for 13 weeks, recording daily
occurrences of respiratory symptoms associated with asthma exacerbations (e.g., shortness of
breath, wheeze, and cough). This study found a statistically significant association between
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). Vedal et al. (1998) followed a group of elementary school children, including 74
asthmatics, located on the west coast of Vancouver Island for 18 months including
measurements of daily peak expiratory flow (PEF) and the tracking of respiratory symptoms
(e.g., cough, phlegm, wheeze, chest tightness) through the use of daily diaries. Association
between PMi0 and respiratory symptoms for the asthmatic population was only reported for
two endpoints: cough and PEF. Because it is difficult to translate PEF measures into clearly
4 Estimating asthma exacerbations associated with air pollution exposures is difficult, due to concerns about
double counting of benefits. Concerns over double counting stem from the fact that studies of the general
population also include asthmatics, so estimates based solely on the asthmatic population cannot be directly
added to the general population numbers without double counting. In one specific case (upper respiratory
symptoms in children), the only study available is limited to asthmatic children, so this endpoint can be readily
included in the calculation of total benefits. However, other endpoints, such as lower respiratory symptoms and
MRADs, are estimated for the total population that includes asthmatics. Therefore, to simply add predictions of
asthma-related symptoms generated for the population of asthmatics to these total population-based estimates
could result in double counting, especially if they evaluate similar endpoints. The SAB-HES, in commenting on
the analytical blueprint for 812, acknowledged these challenges in evaluating asthmatic symptoms and
appropriately adding them into the primary analysis (SAB-HES, 2004). However, despite these challenges, the
SAB-HES recommends the addition of asthma-related symptoms (i.e., asthma exacerbations) to the primary
analysis, provided that the studies use the panel study approach and that they have comparable design and
baseline frequencies in both asthma prevalence and exacerbation rates. Note also, that the SAB-HES, while
supporting the incorporation of asthma exacerbation estimates, does not believe that the association between
ambient air pollution, including ozone and PM, and the new onset of asthma is sufficiently strong to support
inclusion of this asthma-related endpoint in the primary estimate.
5-35
-------
defined health endpoints that can be monetized, we only included the cough-related effect
estimate from this study in quantifying asthma exacerbations. 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 pooled the separate
incidence estimates for shortness of breath, wheeze, and cough generated using effect
estimates from the Ostro et al. study, because each of these endpoints is aimed at capturing
the same overall endpoint (asthma exacerbations) and there could be overlap in their
predictions. The pooled estimate from the Ostro et al. study is then pooled with the cough-
related estimate generated using the Vedal study. The rationale for this second pooling step is
similar to the first; both studies are attempting to quantify the same overall endpoint (asthma
exacerbations).
5.4.3 Baseline Incidence 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 ppb decrease in daily ozone levels might, in turn,
decrease hospital admissions by 3 percent. The baseline incidence of the health effect is
necessary to convert this relative change into a number of cases. A baseline incidence rate is
the estimate of the number of cases of the health effect per year in the assessment location, as
it corresponds to baseline pollutant levels in that location. To derive the total baseline
incidence per year, this rate must be multiplied by the corresponding population number. For
example, if the baseline incidence rate is the number of cases per year per million people, that
number must be multiplied by the millions of people in the total population.
Table 5-7 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. Rates for mortality, hospitalizations, asthma
ER visits, and non-fatal myocardial infarction (heart attacks) have been updated since the MATS
Proposal RIA, consistent with the Cross-State Air Pollution Rule RIA (U.S. EPA 2011b).
5-36
-------
Table 5-7. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population
Endpoint
Rates
Parameter
Value
Source
Mortality
Hospitalizations
Asthma ER Visits
Chronic Bronchitis
Daily or annual mortality rate
projected to 2015
Daily hospitalization rate
Daily asthma ER visit rate
Annual prevalence rate per
person
• Aged 18-44
• Aged 45-64
• Aged 65 and older
Annual incidence rate per
person
Age-, cause-, and county-
specific rate
Age-, region-, state-,
county- and cause-
specific rate
Age-, region-, state-,
county- and cause-
specific rate
0.0367
0.0505
0.0587
0.00378
CDC Wonder (2004-2006)
U.S. Census bureau
2007 HCUP data files3
2007 HCUP data files
1999 NHIS (American Lung
Association, 2002b, Table 4)
Abbey et al. (1995, Table 3)
(continued)
5-37
-------
Table 5-7. Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population (continued)
Rates
Endpoint
Parameter
Value
Source
Non-fatal Myocardial
Infarction (heart
attacks)
Asthma Exacerbations
Acute Bronchitis
Lower Respiratory
Symptoms
Upper Respiratory
Symptoms
Daily non-fatal myocardial
infarction incidence rate per
person, 18+
Incidence among asthmatic
African-American children
• daily wheeze
• daily cough
• daily dyspnea
Annual bronchitis incidence
rate, children
Daily lower respiratory
symptom incidence among
children15
Daily upper respiratory
symptom incidence among
asthmatic children
Age-, region-, state-, and
county- specific rate
2007 HCUP data files3;
adjusted by 0.93 for
probability of surviving after
28 days (Rosamond et al.,
1999)
Ostro et al. (2001)
0.076
0.067
0.037
0.043
0.0012
0.3419
American Lung Association
(2002c, Table 11)
Schwartz et al. (1994,
Table 2)
Popeetal. (1991, Table 2)
Work Loss Days
School Loss Days
Minor Restricted-
Activity Days
Daily WLD incidence rate per
person (18-65)
• Aged 18-24
• Aged 25-44
• Aged 45-64
Rate per person per year,
assuming 180 school days per
year
Daily MRAD incidence rate per
person
0.00540
0.00678
0.00492
9.9
0.02137
1996 HIS (Adams,
Hendershot, and Marano,
1999, Table 41); U.S. Bureau
of the Census (2000)
National Center for
Education Statistics (1996)
and 1996 HIS (Adams etal.,
1999, Table 47);
Ostro and Rothschild (1989,
p. 243)
a 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.
b Lower respiratory symptoms are defined as two or more of the following: cough, chest pain, phlegm, and
wheeze.
The baseline incidence rates for hospital and emergency department visits that we
applied in this analysis are an improvement over the rates we used in the proposal analysis in
two ways. First, these data are newer, and so are a more recent representation of the rates at
which populations of different ages, and in different locations, visit the hospital and emergency
5-38
-------
department for illnesses that may be air pollution related. Second, these newer data are also
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. Newer and more spatially resolved incidence rates
are likely to yield a more reliable estimate of air pollution-related hospitalizations and
emergency department visits. Consistent with the proposal RIA, we continue to use county-
level mortality rates. We have projected mortality rates such that future mortality rates are
consistent with our projections of population growth (Abt Associates, 2010).
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 5-8 lists the prevalence rates used to determine the applicable
population for asthma symptom endpoints. Note that these reflect current asthma prevalence
and assume no change in prevalence rates in future years. These rates have all been updated
since the MATS proposal RIA, consistent with the Cross-State Air Pollution Rule (U.S. EPA
2011b).
Table 5-8. Asthma Prevalence Rates Used for this Analysis3
Asthma Prevalence Rates
Population Group Value Source
All Ages 0.0780 American Lung Association (2010, Table 7)
< 18 0.0941
5-17 0.1070
18^14 0.0719
45-64 0.0745
65+ 0.0716
African American, 5 to 17 0.1776 American Lung Association (2010, Table 9)
African American, <18 0.1553 American Lung Association15
a Seeftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/NHIS/2000/.
b Calculated by ALA for U.S. EPA, based on NHIS data (CDC, 2009)
5.4.4 Economic Valuation Estimates
Reductions in ambient concentrations of air pollution generally lower the risk of future
adverse health effects for a large population. Therefore, the appropriate economic measure is
WTP for changes in risk of a health effect rather than WTP for a health effect that would occur
5-39
-------
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 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.5
WTP estimates generally are not available for some health effects, such as hospital
admissions. In these cases, we used the cost of treating or mitigating the effect as a primary
estimate. These cost-of-illness (COI) estimates generally 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 Tables 5-10 through 5-12. All values are in constant year 2006 dollars, adjusted for
growth in real income out to 2016 using projections provided by Standard and Poor's. Economic
theory argues that WTP for most goods (such as environmental protection) will increase if real
income increases. Many of the valuation studies used in this analysis were conducted in the late
1980s and early 1990s. Because real income has grown since the studies were conducted,
people's willingness to pay for reductions in the risk of premature death and disease likely has
grown as well. We did not adjust cost of illness-based values because they are based on current
costs. Similarly, we did not adjust the value of school absences, because that value is based on
current wage rates. For these two reasons, these cost of illness estimates may underestimate
the economic value of avoided health impacts in 2016. The discussion below provides
additional details on ozone and PM2.5-related related endpoints.
5.4.4.1 Mortality Valuation
Following the advice of the EEAC of the SAB, EPA currently uses the VSL approach in
calculating the primary estimate of mortality co-benefits, because we believe this calculation
5 To comply with Circular A-4, EPA provides monetized benefits using discount rates of 3% and 7% (OMB, 2003).
These benefits are estimated for a specific analysis year (i.e., 2016), and most of the PM benefits occur within
that year with two exceptions: acute myocardial infarctions (AMIs) and premature mortality. For AMIs, we
assume 5 years of follow-up medical costs and lost wages. For premature mortality, we assume that there is a
"cessation" lag between PM exposures and the total realization of changes in health effects. Although the
structure of the lag is uncertain, 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). Changes in the lag assumptions do not change the total
number of estimated deaths but rather the timing of those deaths. Therefore, discounting only affects the AMI
costs after the analysis year and the valuation of premature mortalities that occur after the analysis year. As
such, the monetized benefits using a 7% discount rate are only approximately 10% less than the monetized
benefits using a 3% discount rate.
5-40
-------
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 value of statistical life (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 the
Science Advisory Board (SAB) or other peer-review group.
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 EPA and the SAB on combining estimates from the various
data sources. In addition, the Agency consulted several times with the Science Advisory Board
Environmental Economics Advisory Committee (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)6 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
6 In EPA's recently revised Economic Guidelines (U.S. EPA, 2010d), 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. Therefore, this report does not represent final agency policy.
5-41
-------
1991. The mean VSL across these studies is $6.3 million (2000$).7 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.
The Agency anticipates presenting results from this effort to the SAB-EEAC in Spring 2010 and
that draft guidance will be available shortly thereafter.
As indicated in the previous section on quantification of premature mortality benefits,
we assumed for this analysis that some of the incidences of premature mortality related to PM
exposures occur in a distributed fashion over the 20 years following exposure. To take this into
account in the valuation of reductions in premature mortality, we applied an annual 3%
discount rate to the value of premature mortality occurring in future years.8
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. 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, we use a single VSL for all reductions in mortality
risk.
Although there are several differences between the labor market studies EPA uses to
derive a VSL estimate and the PM air pollution context addressed here, those differences in the
affected populations and the nature of the risks imply both upward and downward
adjustments. Table 5-9 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, EPA believes it is reasonable to continue to use the $6.3 million value
while acknowledging the significant limitations and uncertainties in the available literature.
7 In this analysis, we adjust the VSL to account for a different currency year (2007$) and to account for income
growth to 2016. After applying these adjustments to the $6.3 million value, the VSL is $8.9M.
8 The choice of a discount rate, and its associated conceptual basis, is a topic of ongoing discussion within the
federal government. EPA adopted a 3% discount rate for its base estimate in this case to reflect reliance on a
"social rate of time preference" discounting concept. We have also calculated benefits and costs using a 7% rate
consistent with an "opportunity cost of capital" concept to reflect the time value of resources directed to meet
regulatory requirements. In this case, the benefit and cost estimates were not significantly affected by the
choice of discount rate. Further discussion of this topic appears in EPA's Guidelines for Preparing Economic
Analyses (EPA, 2010).
5-42
-------
Table 5-9. Expected Impact on Estimated Benefits of Premature Mortality Reductions of
Differences Between Factors Used in Developing Applied VSL and Theoretically
Appropriate VSL
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 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, 2000). In developing our primary estimate of the co-
benefits of premature mortality reductions, we have followed this advice and discounted over
the lag period between exposure and premature mortality.
Uncertainties Specific to Premature Mortality Valuation. The economic co-benefits
associated with reductions in the risk of premature mortality are the largest category of
monetized co-benefits of the MATS. In addition, in prior analyses, EPA has identified valuation
of mortality-related benefits as the largest contributor to the range of uncertainty in monetized
benefits (U.S. EPA, 1999).9 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.
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
PM mortality expert elicitation may result in different conclusions about the relative contribution of sources of
uncertainty.
5-43
-------
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
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
5-44
-------
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
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. 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. 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.
5-45
-------
• 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). This
worker must have either a higher workplace risk than the average worker, a lower
risk tolerance than the average worker, or both. However, the risk estimate used in
hedonic studies is generally based on average risk, so the VSL may be upwardly
biased 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.
5.4.4.2 Chronic Bronchitis Valuation
The best available estimate of WTP to avoid a case of CB comes from Viscusi, Magat,
and Huber (1991). The Viscusi, Magat, and Huber study, however, describes a severe case of CB
to the survey respondents. We therefore employ an estimate of WTP to avoid a pollution-
related case of CB, based on adjusting the Viscusi, Magat, and Huber (1991) estimate of the
WTP to avoid a severe case. This is done to account for the likelihood that an average case of
pollution-related CB is not as severe. The adjustment is made by applying the elasticity of WTP
with respect to severity reported in the Krupnick and Cropper (1992) study. Details of this
adjustment procedure are provided in the Benefits Technical Support Document (TSD) for the
Nonroad Diesel rulemaking (Abt Associates, 2003).
We use the mean of a distribution of WTP estimates as the central tendency estimate of
WTP to avoid a pollution-related case of CB in this analysis. The distribution incorporates
uncertainty from three sources: the WTP to avoid a case of severe CB, as described by Viscusi,
Magat, and Huber; the severity level of an average pollution-related case of CB (relative to that
of the case described by Viscusi, Magat, and Huber); and the elasticity of WTP with respect to
severity of the illness. Based on assumptions about the distributions of each of these three
uncertain components, we derive a distribution of WTP to avoid a pollution-related case of CB
by statistical uncertainty analysis techniques. The expected value (i.e., mean) of this
5-46
-------
distribution, which is about $340,000 (2006$), is taken as the central tendency estimate of WTP
to avoid a PM-related case of CB.
5.4.4.3 Non-fatal Myocardial Infarctions Valuation
We were not able to identify a suitable WTP value for reductions in the risk of non-fatal
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 present discounted value in lost
earnings (in 2006$) over 5 years due to a myocardial infarction of $8,774 for someone between
the ages of 25 and 44, $12,932 for someone between the ages of 45 and 54, and $74,746 for
someone between the ages of 55 and 65. The corresponding age-specific estimates of lost
earnings (in 2006$) using a 7% discount rate are $7,855, $11,578, and $66,920, 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:
• Wittels et al. (1990) estimated expected total medical costs of myocardial infarction
over 5 years to be $51,211 (in 1986$) for people who were admitted to the hospital
and survived hospitalization. (There does not appear to be any discounting used.)
Wittels et al. was used to value coronary heart disease in the 812 Retrospective
Analysis of the Clean Air Act. Using the CPI-U for medical care, the Wittels estimate
is $144,111 in year 2006$. 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.
5-47
-------
However, this may include patients who died in the hospital (not included among
our non-fatal myocardial infarction cases), 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 $44,663 in 1997$, or $64,003 in
2006$ for myocardial infarction patients, using statistical prediction (regression)
models to estimate inpatient costs. Only inpatient costs (physician fees and hospital
costs) were included.
• Russell et al. (1998) estimated first-year direct medical costs of treating non-fatal
myocardial infarction of $15,540 (in 1995$) and $1,051 annually thereafter.
Converting to year 2006$, that would be $30,102 for a 5-year period (without
discounting) or $38,113 for a 10-year period.
In summary, the three different studies provided significantly different values (see
Table 5-10).
Table 5-10. Alternative Direct Medical Cost of Illness Estimates for Non-fatal Heart Attacks
Study Direct Medical Costs (2006$) Over an x-Year Period, for x =
Wittels et al. (1990) 144,llla 5
Russell et al. (1998) 30,102b 5
Eisenstein et al. (2001) 64,003b 10
Russell etal. (1998) $38,113b 10
3 Wittels et al. (1990) did not appear to discount costs incurred in future years.
b 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 $65,902, and added it to the 5-year opportunity cost estimate. The resulting
estimates are given in Table 5-11.
5-48
-------
Table 5-11. Estimated Costs Over a 5-Year Period (in 2006$) of a Non-fatal Myocardial
Infarction
Cost, $
Age Group
0-24
25-44
45-54
55-65
>65
Opportunity
0
10,757b
15,855b
91,647b
0
Medical"
84,955
84,955
84,955
84,955
84,955
Total
84,955
95,713
100,811
176,602
84,955
a 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.
5.4.5 Hospital Admissions Valuation
In the absence of estimates of societal WTP to avoid hospital visits/admissions for
specific illnesses, estimates of total cost of illness (total medical costs plus the value of lost
productivity) typically are used as conservative, or lower bound, estimates. These estimates are
biased downward, because they do not include the willingness-to-pay value of avoiding pain
and suffering.
The International Classification of Diseases (ICD-9, 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 (AHRQ 2000). 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 1990 median weekly wage by five and inflated the result to year 2006$ using the
CPI-U "all items." The resulting estimate is $127.93. The total cost-of-illness estimate for an ICD
code-specific hospital stay lasting n days, then, was the mean hospital charge plus $127.93
multiplied by n.
5.4.5.1 Asthma-Related Emergency Room Visits Valuation
To value asthma emergency room 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 room visits in 1987, at a total
cost of $186.5 million (1987$). The average cost per visit that year was $155; in 2006$, that cost
5-49
-------
was $400.88 (using the CPI-U for medical care to adjust to 2006$). The second estimate comes
from Stanford et al. (1999), who reported the cost of an average asthma-related emergency
room visit at $335.14, based on 1996-1997 data. A simple average of the two estimates yields a
(rounded) unit value of $368.
5.4.5.2 Minor Restricted Activity Days Valuation
No studies are reported to have estimated WTP to avoid a minor restricted activity day.
However, one of EPA's contractors, lEc (1994) has 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. The lEc estimate
of WTP to avoid a minor respiratory restricted activity day is $38.37 (1990$), or about $62.04
(2006$).
5-50
-------
Table 5-12. Unit Values for Economic Valuation of Health Endpoints (2006$)a
Central Estimate of Value Per
Statistical Incidence,
Income Level
Health Endpoint
2000
2016
Derivation of Distributions of Estimates
Premature Mortality (Value
of a Statistical Life)
$6,300,000 $8,600,000
Chronic Bronchitis (CB)
$340,000 $470,000
EPA currently recommends a central VSL of $6.3m
(2000$) 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 5B of EPA's current Guidelines for
Preparing Economic Analyses (U.S. EPA, 2000).
The WTP to avoid a case of pollution-related CB is
calculated as where x is the severity of an average
CB case, WTP13 is the WTP for a severe case of CB,
and $ is the parameter relating WTP to severity,
based on the regression results reported in
Krupnick and Cropper (1992). The distribution of
WTP for an average severity-1 eve I case of CB was
generated by Monte Carlo methods, drawing
from each of three distributions: (1) WTP to avoid
a severe case of CB is assigned a 1/9 probability of
being each of the first nine deciles of the
distribution of WTP responses in Viscusi et al.
(1991); (2) the severity of a pollution-related case
of CB (relative to the case described in the Viscusi
study) is assumed to have a triangular
distribution, with the most likely value at severity
level 6.5 and endpoints at 1.0 and 12.0; and (3)
the constant in the elasticity of WTP with respect
to severity is normally distributed with mean =
0.18 and standard deviation = 0.0669 (from
Krupnick and Cropper [1992]). This process and
the rationale for choosing it is described in detail
in the Costs and Benefits of the Clean Air Act,
1990 to 2010 (U.S. EPA, 1999).
(continued)
5-51
-------
Table 5-12. Unit Values for Economic Valuation of Health Endpoints (2006$) (continued)
Central Estimate of Value Per
Statistical Incidence,
Income Level
Health Endpoint
2000
2016
Derivation of Distributions of Estimates
Non-fatal Myocardial
Infarction (heart attack)
3% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
7% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
$79,685
$88,975
$93,897
$167,532
$79,685
$77,769
$87,126
$91,559
$157,477
$77,769
$79,685
$88,975
$93,897
$167,532
$79,685
$77,769
$87,126
$91,559
$157,477
$77,769
No distributional information available. Age-
specific cost-of-illness values reflect lost earnings
and direct medical costs over a 5-year period
following a non-fatal 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:
age of onset: at 3% at 7%
25-44 $8,774 $7,855
45-54 $12,932 11,578
55-65 $74,746 66,920
Direct medical expenses: An average of:
1. Wittels et al. (1990) ($102,658-no
discounting)
2. Russell et al. (1998), 5-year period ($22,331
at 3% discount rate; $21,113 at 7% discount
rate)
Hospital Admissions
Chronic Obstructive
Pulmonary Disease (COPD)
Asthma Admissions
$16,606 $16,606 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 COPD
category illnesses) reported in Agency for
Healthcare Research and Quality (2000)
(www.ahrq.gov).
$8,900 $8,900 No distributional information available. The COI
estimates (lost earnings plus direct medical costs)
are based on ICD-9 code-level information (e.g.,
average hospital care costs, average length of
hospital stay, and weighted share of total asthma
category illnesses) reported in Agency for
Healthcare Research and Quality (2000)
(www.ahrq.gov).
(continued)
5-52
-------
Table 5-12. Unit Values for Economic Valuation of Health Endpoints (2006$) (continued)
Central Estimate of Value Per
Statistical Incidence,
Income Level
Health Endpoint
2000
2016
Derivation of Distributions of Estimates
All Cardiovascular
All respiratory (ages 65+)
All respiratory (ages 0-2)
Emergency Room Visits for
Asthma
$24,668 $24,668 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
(2000) (www.ahrq.gov).
$24,622 $24,622 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 COPD
category illnesses) reported in Agency for
Healthcare Research and Quality, 2000
(www.ahrq.gov).
$10,385 $10,385 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 COPD
category illnesses) reported in Agency for
Healthcare Research and Quality, 2000
(www.ahrq.gov).
$384 $384 No distributional information available. Simple
average of two unit COI values:
(1) $311.55, from Smith et al. (1997) and
(2) $260.67, from Stanford et al. (1999).
(continued)
5-53
-------
Table 5-12. Unit Values for Economic Valuation of Health Endpoints (2006$) (continued)
Central Estimate of Value Per
Statistical Incidence,
Income Level
Health Endpoint
2000
2016
Derivation of Distributions of Estimates
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory
Symptoms (URS)
Lower Respiratory
Symptoms (LRS)
Asthma Exacerbations
$30 $30 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 $43.1.
$16 $19 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 $24.46.
$43 $53 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 $15.6 and $70.8.
(continued)
5-54
-------
Table 5-12. Unit Values for Economic Valuation of Health Endpoints (2006$) (continued)
Central Estimate of Value Per
Statistical Incidence,
Income Level
Health Endpoint
2000
2016
Derivation of Distributions of Estimates
Acute Bronchitis
Work Loss Days (WLDs)
Minor Restricted Activity
Days (MRADs)
$360 $440 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.
Variable (U.S. Variable (U.S. No distribution available. Point estimate is based
median = median = on county-specific median annual wages divided
$130) $130) by 52 and then by 5—to get median daily wage.
U.S. Year 2000 Census, compiled by Geolytics, Inc.
$51 $62 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.
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 $16.00) and be less than that for a
WLD. The triangular distribution acknowledges
that the actual value is likely to be closer to the
point estimate than either extreme.
aValues reported in this table are in 2006$, but we used 2007$ for this analysis. Inflating to 2007$ would increase
the values approximately 2.8% for WTP estimates up to 4.4% for COI estimates.
Although Ostro and Rothschild (1989) statistically linked ozone and minor restricted
activity days, it is likely that most MRADs associated with ozone 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.
5.4.5.3 Growth in WTP Reflecting National Income Growth Over Time
Our analysis accounts for expected growth in real income over time. Economic theory
argues that WTP for most goods (such as environmental protection) will increase if real incomes
5-55
-------
increase. There is substantial empirical evidence that the income elasticity10 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,
2000). 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, 2004b, p. 52)" and that "The same increase should be
assumed for the WTP for serious non-fatal health effects (U.S. EPA-SAB, 2004b, p. 52)," 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, p. 53). 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.
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
10 Income elasticity is a common economic measure equal to the percentage change in WTP for a 1% change in
income.
5-56
-------
found in Kleckner and Neumann (1999). An abbreviated description of the procedure we used
to account for WTP for real income growth between 1990 and 2016 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. 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. We also
expect that the WTP for improved visibility in Class I areas 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. The elasticity values used to adjust estimates of benefits
in 2016 are presented in Table 5-13.
Table 5-13. 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
Visibility 0.90
Derivation of estimates can be found in Kleckner and Neumann (1999) and Chestnut (1997). COI estimates are
assigned an adjustment factor of 1.0.
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
5-57
-------
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 2016, 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.n We used projections of
real GDP (in chained 1996 dollars) provided by Standard and Poor's (2000) for the years 2010 to
2016.12
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 5-14. 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. Note that, for premature mortality,
we applied the income adjustment factor to the present discounted value of the stream of
avoided mortalities occurring over the lag period. Also note that because of a lack of data on
the dependence of COI and income, and a lack of data on projected growth in average wages,
no adjustments are made to benefits based on the COI approach or to work loss days and
worker productivity. This assumption leads us to 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).
11 U.S. Bureau of Economic Analysis, Table 2A (1992$) (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.
12 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.
5-58
-------
Table 5-14. Adjustment Factors Used to Account for Projected Real Income Growth3
Benefit Category 2016
Minor Health Effect 1.06
Severe and Chronic Health Effects 1.19
Premature Mortality 1.16
Visibility 1.41
Based on elasticity values reported in Table 5-13, U.S. Census population projections, and projections of real
GDP per capita.
5.5 Unquantified Health and Welfare Benefits
This analysis is limited by the available data and resources. As such, we are not able to
quantify several welfare benefit categories, as shown in Tables 5-3 and 5-4. This section
provides an overview of what is meant by ecosystem services as well as a description of
visibility benefits, which are typically assessed and monetized in relevant RIAs but that were not
quantified in this benefits analysis. The RIA for the final Cross-State Air Pollution Rule (U.S. EPA,
2011b) provides more information on additional major health and welfare benefit categories
associated with reducing N02 and S02 emissions including: health and ecosystem benefits of
reducing nitrogen and sulfur emissions and deposition; vegetation benefits from reducing
ozone; mercury benefits associated with reducing mercury emissions; and the role of sulfate
deposition in mercury methylation. While we are unable to quantify these benefits, previous
relevant EPA assessments show that these benefits could be substantial (U.S. EPA, 2008a; U.S.
EPA, 2009a; U.S. EPA, 2007; U.S. EPA, 1999, U.S. EPA, 2011b). The omission of these endpoints
from the monetized results should not imply that the impacts are small or unimportant.
5.5.1 Visibility Valuation
Reductions in N02 and S02 emissions along with the secondary formation of PM2.5 would
improve the level of visibility throughout the United States because these suspended particles
and gases degrade visibility by scattering and absorbing light (U.S. EPA, 2009a). Visibility has
direct significance to people's enjoyment of daily activities and their overall sense of wellbeing
(U.S. EPA, 2009a). Individuals value visibility both in the places they live and work, in the places
they travel to for recreational purposes, and at sites of unique public value, such as the Great
Smokey Mountains National Park. This section discusses the measurement of the economic
benefits of improved visibility. As there is no analogous approach for estimating visibility
benefits using the BPT approach, visibility benefits are calculated for the modeled interim policy
scenario only and are not included in the co-benefits estimate of the final policy. However,
5-59
-------
since the magnitude of S02 emission reductions did not significantly change in the visibility
study areas between the interim and final emissions scenarios, we expect the visibility benefit
for the final policy scenario would be similar to that calculated for the interim policy scenario
($1.1 billion in total for the U.S., using 2007$; see Appendix 5C).
Visual air quality (VAQ) is commonly measured as either light extinction, which is
defined as the loss of light per unit of distance in terms of inverse megameters (Mm"1) or the
deciview (dv) metric (Pitchford and Malm, 1993), which is a logarithmic function of extinction.
Extinction and deciviews are 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. Pitchford and Malm characterize a change of one deciview as "a small but
perceptible scenic change under many circumstances." 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 (i.e. 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 5-3 for an illustration of the important factors affecting visibility.
In conjunction with the U.S. National Park Service, the U.S. Forest Service, other Federal
land managers, and State organizations in the U.S., the U.S. EPA has supported visibility
monitoring in national parks and wilderness areas since 1988. The monitoring network known
as IMPROVE (Interagency Monitoring of Protected Visual Environments) now includes 150 sites
that represent almost all of the Class I areas across the country (see Figure 5-4) (U.S. EPA,
2009a).
Annual average visibility conditions (reflecting light extinction due to both
anthropogenic and non-anthropogenic sources) vary regionally across the U.S. (U.S. EPA,
2009a). The rural East generally has higher levels of impairment than remote sites in the West,
with the exception of urban-influenced sites such as San Gorgonio Wilderness (CA) and Point
Reyes National Seashore (CA), which have annual average levels comparable to certain sites in
the Northeast (U.S. EPA, 2004). Higher visibility impairment levels in the East are due to
generally higher concentrations of fine particles, particularly sulfates, and higher average
relative humidity levels. While visibility trends have improved in most Class I areas, the recent
data show that these areas continue to suffer from visibility impairment. In eastern parks,
5-60
-------
average visual range has decreased from 90 miles to 15-25 miles, and in the West, visual range
has decreased from 140 miles to 35-90 miles (U.S. EPA, 2004; U.S. EPA, 1999).
Light from clouds
scattered Into
sight path v
Image-forming
light scattered
out of sight path
Sunlight X
scattered Light reflected
from ground
scattered Into
sight path
Image
light absorbed
Figure 5-3. Important Factors Involved in Seeing a Scenic Vista (Malm, 1999)
5-61
-------
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
Figure 5-4. Mandatory Class I Areas in the U.S.
EPA distinguishes benefits from two categories of visibility changes: residential visibility
and recreational visibility. In both cases economic benefits are believed to 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 based on people's beliefs that the environment ought to exist free of human-induced
haze. Nonuse values may be more important for recreational areas, particularly national parks
and monuments.
Residential visibility benefits are those that occur from visibility changes in urban,
suburban, and rural areas. In previous assessments, EPA used a study on residential visibility
valuation conducted in 1990 (McClelland et al., 1993). Subsequently, EPA designated the
McClelland et al. study as significantly less reliable for regulatory benefit-cost analysis
consistent with SAB advice (U.S. EPA-SAB, 1999). Although a wide range of published, peer-
review literature supports a non-zero value for residential visibility (Brookshire et al., 1982; Rae,
1983; Tolley et al., 1986; Chestnut and Rowe, 1990c; McClelland et al., 1993; Loehman et al.,
1994), the residential visibility benefits have not been calculated in this analysis.
5-62
-------
For recreational visibility, only one existing study provides defensible monetary
estimates of the value of visibility changes in a 1988 survey on recreational visibility value
(Chestnut and Rowe, 1990a; 1990b). Although there are a number of other studies in the
literature, they were conducted in the early 1980s and did not use methods that are considered
defensible by current standards. The Chestnut and Rowe study uses the CV method. There has
been a great deal of controversy and significant development of both theoretical and empirical
knowledge about how to conduct CV surveys in the past decade. In EPA's judgment, the
Chestnut and Rowe study contains many of the elements of a valid CV study and is sufficiently
reliable to serve as the basis for monetary estimates of the benefits of visibility changes in
recreational areas.13 This study serves as an essential input to our estimates of the benefits of
recreational visibility improvements in the primary benefits estimates.
For the purposes of the analysis of the visibility benefits of the modeled interim policy
(Appendix 5C), recreational visibility improvements are defined as those that occur specifically
in federal Class I areas.14 A key distinction between recreational and residential benefits is that
only those people living in residential areas are assumed to receive benefits from residential
visibility, while all households in the United States are assumed to derive some benefit from
improvements in Class I areas. Values are assumed to be higher if the Class I area is located
close to their home.15 The Chestnut and Rowe study measured the demand for visibility in Class
I areas managed by the National Park Service (NPS) in three broad regions of the country:
California, the Southwest, and the Southeast. 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 visibility levels in these photographs were later converted to deciviews for the current
analysis. The survey data collected were used to estimate a WTP equation for improved
visibility. In addition to the visibility change variable, the estimating equation also included
household income as an explanatory variable.
13 In SAB advisory letter indicates that "many members of the Council believe that the Chestnut and Rowe study is
the best available" (EPA-SAB-COUNCIL-ADV-00-002,1999, p. 13). However, the committee did not formally
approve use of these estimates because of concerns about the peer-reviewed status of the study. EPA believes
the study has received adequate review and has been cited in numerous peer-reviewed publications (Chestnut
and Dennis, 1997).
14 The Clean Air Act designates 156 national parks and wilderness areas as Class I areas for visibility protection.
15 For details of the visibility estimates discussed in this chapter, please refer to the Benefits TSD for the Nonroad
Diesel rulemaking (Abt Associates, 2003).
5-63
-------
The Chestnut and Rowe study did not measure values for visibility improvement in Class
I areas outside the three regions. Their study covered 86 of the 156 Class I areas in the United
States. We can infer the value of visibility changes in the other Class I areas by transferring
values of visibility changes at Class I areas in the study regions. A complete description of the
benefits transfer method used to infer values for visibility changes in Class I areas outside the
study regions is provided in the Benefits TSD for the Nonroad Diesel rulemaking (Abt
Associates, 2003).
The Chestnut and Rowe study (Chestnut and Rowe, 1990a; 1990b), although
representing the best available estimates, has a number of limitations. These include the
following:
• The age of the study (late 1980s) will increase the uncertainty about the
correspondence of the estimated values to those that might be provided by current
or future populations.
• 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 Chestnut and
Rowe study 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 visibility at national
parks entirely from residential visibility and health effects.
• It is not clear exactly what visibility improvements the respondents to the Chestnut
and Rowe survey were valuing. The WTP question asked about changes in average
visibility, but the survey respondents were shown photographs of only summertime
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. For the purpose of the benefits analysis for this rule, EPA
assumed that respondents provided values for changes in annual average visibility.
Because most policies will result in a shift in the distribution of visibility (usually
5-64
-------
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.
• The Chestnut and Rowe survey focused on visibility improvements in and around
national parks and wilderness areas. The survey also focused on visibility
improvements of national parks in the southwest United States. Given that national
parks and wilderness areas exhibit unique characteristics, it is not clear whether the
WTP estimate obtained from Chestnut and Rowe can be transferred to other
national parks and wilderness areas, without introducing additional uncertainty.
In general, the survey design and implementation reflect the period in which the survey
was conducted. Since that time, many improvements to the stated preference methodology
have been developed. As future survey efforts are completed, EPA will incorporate values for
visibility improvements reflecting the improved survey designs.
The estimated relationship from the Chestnut and Rowe study is only directly applicable
to the populations represented by survey respondents. EPA used benefits transfer methodology
to extrapolate these results to the population affected by the reductions in precursor emissions
associated with this rule. A general WTP equation for improved visibility (measured in
deciviews) was developed as a function of the baseline level of visibility, the magnitude of the
visibility improvement, and household income. The behavioral parameters of this equation
were taken from analysis of the Chestnut and Rowe data. These parameters were used to
calibrate WTP for the visibility changes resulting from this rule. The method for developing
calibrated WTP functions is based on the approach developed by Smith et al. (2002). Available
evidence indicates that households are willing to pay more for a given visibility improvement as
their income increases (Chestnut, 1997). The benefits estimates here incorporate Chestnut's
estimate that a 1% increase in income is associated with a 0.9% increase in WTP for a given
change in visibility. A more detailed explanation of the visibility benefits methodology is
provided in Appendix I of the PM NAAQS RIA (U.S. EPA, 2006a).
One major source of uncertainty for the visibility benefits estimate is the benefits
transfer process used. Judgments used to choose the functional form and key parameters of
the estimating equation for WTP for the affected population could have significant effects on
5-65
-------
the size 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 results.
In addition, our estimate of visibility benefits of the modeled interim policy in Appendix
5C is incomplete. For example, we anticipate improvement in visibility in residential areas for
which we are currently unable to monetize benefits, such as the Northeastern and Central
regions of the U.S. The value of visibility benefits in areas where we were unable to monetize
benefits could also be substantial. EPA requests public comment on the approach taken here to
quantify the monetary value of changes in visibility in Class I areas.
5.5.2 Ecosystem Services
Ecosystem services can be generally defined as the benefits that individuals and
organizations obtain from ecosystems. 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, 2006b). Figure 5-5 provides the Millennium Ecosystem
Assessment's schematic demonstrating the connections between the categories of ecosystem
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).
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
5-66
-------
ECOSYSTEM
Supporting
NUTRIENT CYCLING
SOIL FORMATION
PRIMARY PRODUCTION
LIFE ON EARTH -
SERVICES
Provisioning
FOOD f
FRESHWATER •
WOOD AND FIBER
FUEL
Regulating
CLIMATE REGULATION |
FLOOD REGULATION
DISEASE REGULATION |
WATCH PURIFICATION
Cultural
AESTHETIC
SPIRITUAL
EDUCATIONAL
RECREATIONAL
BIODIVERSITY
CONSTITUENTS OFWELL-BEING
Security
PERSONAL SAFETY
SECURE HESOURCEACCESS
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
ABIUTYTO HELP OTHERS
Freedom
of choice
and action
OPPORTUNITY TO BE
ABLE TO ACHIEVE
WHAT AN INDIVIDUAL
VALUES DOING
AND BEING
Source: Willeni
Figure 5-5. Linkages Between Categories of Ecosystem Services and Components of Human
Weil-Being from Millennium Ecosystem Assessment (MEA, 2005)
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
reductions in them (U.S. EPA, 2006b). 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, 2006b). 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 5-6.
5-67
-------
I Ecological goods and services
affected by the policy
Planning and problem formulation
n
Goods and services
identified
U duds and
services not
identified
Ecological analysis
Identified
goods and
services not
quantified
Quantified
goods and
services not
monetized
Figure 5-6. Schematic of the Benefits Assessment Process (U.S. EPA, 2006b]
5.5.3 Ecosystem Benefits of Reduced Nitrogen and Sulfur Deposition
5.5.3.1 Science of Deposition
Nitrogen and sulfur emissions 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, 2008b). Sulfur is primarily emitted as S02, and nitrogen can be emitted as NO, N02, or NH3.
Secondary particles are formed from NOX and SOX 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 Figure 5-7.
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.
5-68
-------
S02 Atmospheric
Fate and Transport
Deposition Processes
Acidification
MeHg Production
f • 1
X j
^ ~i r~ i r~ ~i
«. ) v y v J
N02 Atmospheric
Fate and Transport
Deposition Processes
Acidification
Nutrient Enrichment
Aquatic
Terrestrial
Aquatic
Terrestrial
Figure 5-7. Schematics of Ecological Effects of Nitrogen and Sulfur Deposition
The 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 N02 and S02
5-69
-------
emissions will decrease both acid deposition and PM concentrations, but not necessarily in a
linear fashion. (U.S. EPA, 2008b). Sulfuric acid is also deposited on surfaces by dry deposition
and can contribute to environmental effects (U.S. EPA, 2008b).
5.5.3.2 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, 2008b). Major
terrestrial effects include a decline in sensitive tree species, such as red spruce (Picea rubens)
and sugar maple (Acer saccharum) (U.S. EPA, 2008b). Biological effects of acidification in
terrestrial ecosystems are generally linked to aluminum toxicity and decreased ability of plant
roots to take up base cations (U.S. EPA, 2008b). 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, 2008b).
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, 2008b). 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, 2008b).
5.5.3.3 Aquatic Ecosystems
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. 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 factor16, changes in species composition and declines in aquatic species richness
16 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, 2008b).
5-70
-------
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 5-15 provides a summary of the biological effects experienced at various ANC
levels.
Table 5-15. Aquatic Status Categories
Category Label ANC Levels Expected Ecological Effects
Acute
Concern
Severe
Concern
Elevated
Concern
Moderate
Concern
Low
Concern
<0 micro
equivalent per
Liter (u.eq/L)
0-20 u.eq/L
20-50 u.eq/L
50-100 u.eq/L
>100 u.eq/L
Near complete loss offish 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.
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, 2008b). 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 5-8 illustrates those areas of the U.S. where
aquatic ecosystems are at risk from acidification.
5-71
-------
Sersrtivity
Acid Sensitive Mtetare {USGSt
_J States
250
750
1,000
—j fcm
Figure 5-8. Areas Potentially Sensitive to Aquatic Acidification (U.S. EPA, 2008b)
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
are affected by the reductions in available fish populations caused by surface water
acidification.
Inland surface waters support several cultural services, including aesthetic and
educational services and recreational fishing. Recreational fishing in lakes and streams is among
5-72
-------
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 2006.17 For recreation days, consumer surplus value is most commonly
measured using recreation demand, travel cost models.
Another estimate of the overarching ecological 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
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, 2009b).18
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.
5.5.3.4 Terrestrial Ecosystems
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
17 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, nc
result of the emission reductions achieved by this rule.
18 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
5-73
-------
restricts the ability of the plant to take up water and nutrients, especially calcium (U. S. EPA,
2008b). 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,
2008b).
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
(2008b). Figure 5-9 depicts the areas across the U.S. that are potentially sensitive to terrestrial
acidification.
I Area of Higasl Potential Sensitivity
Top Guartito N
I co Quaitiie S
1,000
: km
Figure 5-9. Areas Potentially Sensitive to Terrestrial Acidification (U.S. EPA, 2008b)
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
5-74
-------
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 5-10 shows the distribution of
red spruce (brown) and sugar maple (green) in the eastern U.S.
Figure 5-10. Distribution of Red Spruce (Pink) and Sugar Maple (Green) in the Eastern U.S.
(U.S. EPA, 2008b)
Terrestrial acidification affects several important ecological endpoints, including
declines in habitat for threatened and endangered species (cultural), declines in forest
aesthetics (cultural), declines in forest productivity (provisioning), and increases in forest soil
erosion and reductions in water retention (cultural and regulating).
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). Red spruce is also used in a variety of products
including lumber, pulpwood, poles, plywood, and musical instruments. The total removal of red
5-75
-------
spruce saw timber from timberland in the United States was over 300 million board feet in
2006 (USFS, 2006).
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:
• Spruce-fir moss spider (Microhexura montivaga)—endangered
• Rock gnome lichen (Gymnoderma lineare)—endangered
• 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.
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,
5-76
-------
an essential aesthetic component of most fall color landscapes. A survey of residents in the
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.19
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.
5.5.4 Ecological 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.
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 susceptibility to
19 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.
5-77
-------
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, 2008b). 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,
2008b). 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).
In addition to the role of sulfate deposition on methylation, the technologies installed to
reduce emissions of NOX and S02 associated with this rule would also reduce mercury
emissions. EPA recently commissioned an information collection request that will soon provide
greatly improved power industry mercury emissions estimates that will enable the Agency to
better estimate mercury emissions changes from its air emissions control actions. For this
reason, the Agency did not estimate Hg changes in this rule and will instead wait for these new
data which will be available in the near future. Due to time and resource limitations, we were
unable in any event to model mercury dispersion, deposition, methylation, bioaccumulation in
fish tissue, and human consumption of mercury-contaminated fish that would be needed in
order to estimate the human health benefits from reducing these mercury emissions.
5.5.5 Nitrogen Enrichment
5.5.5.1 Aquatic Enrichment
One of the main adverse ecological effects resulting from N deposition, particularly in
the Mid-At I antic region of the United States, is the effect associated with nutrient enrichment
in estuarine waters. A recent assessment of 141 estuaries nationwide by the 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).
5-78
-------
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 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).
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
5-79
-------
$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).20 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.21
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.
5.5.5.2 Terrestrial Enrichment
Terrestrial enrichment occurs when terrestrial ecosystems receive N loadings in excess
of natural background levels, either through atmospheric deposition or direct application.
Evidence presented in the Integrated Science Assessment (U.S. EPA, 2008b) 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 ecosystem conditions and
indicators. This long time scale also affects the timing of the ecosystem service changes.
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
20 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, nc
result of the emission reductions achieved by this rule.
21 These estimates reflect the total value of the service, not the marginal change in the value of the service as a
5-80
-------
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.22 However, the contribution of N deposition to these losses is still uncertain.
5.5.6 Benefits of Reducing Ozone Effects on Vegetation and Ecosystems
Ozone causes discernible injury to a wide array of vegetation (U.S. EPA, 2006c; Fox and
Mickler, 1996). 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, 2006c). Studies
have demonstrated repeatedly that 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,
2006c; 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, 2006c).
22 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.
5-81
-------
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). Because ozone damage can consist of visible
injury to leaves, it can also reduce the aesthetic value of ornamental vegetation and trees in
urban landscapes, and negatively affects scenic vistas in protected natural areas.
Ozone can produce both acute and chronic injury in sensitive species depending on the
concentration level and the duration of the exposure. Ozone effects also tend to accumulate
over the growing season of the plant, so that even lower concentrations experienced for a
longer duration have the potential to create chronic stress on sensitive vegetation. Not all
plants, however, are equally sensitive to ozone. Much of the variation in sensitivity between
individual plants or whole species is related to the plant's ability to regulate the extent of gas
exchange via leaf stomata (e.g., avoidance of ozone uptake through closure of stomata)
(U.S. EPA, 2006c; Winner, 1994). After injuries have occurred, plants may be capable of
repairing the damage to a limited extent (U.S. EPA, 2006c). 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, 2006c). 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, 2006c, McBride et al., 1985; Miller et al., 1982). It is not yet possible to
predict ecosystem responses to ozone with much certainty; however, considerable knowledge
of potential ecosystem responses has been acquired through long-term observations in highly
damaged forests in the United States (U.S. EPA, 2006c).
5-82
-------
5.5.6.1 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, 2006c). 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, 2007).
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 examines ozone injury to ozone-sensitive plant species
at ground monitoring sites in forestland across the country (excluding woodlots and urban
trees). FIA looks for damage on the foliage of ozone-sensitive forest plant species at each site
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 injury is typically
highest.
Monitoring of ozone injury to plants by the USDA Forest Service has expanded over the
last 10 years from monitoring sites in 10 states in 1994 to nearly 1,000 monitoring sites in 41
states in 2002. The data underlying the indictor in Figure 5-11 are based on averages of all
observations collected in 2002, the latest year for which data are publicly available at the time
the study was conducted, and are broken down by U.S. EPA Regions. 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, 2006c; 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.
5-83
-------
Degree of injury:
None
Low
Moderate
High
Severs
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
16.7
11.1
61.9
21.4
7.1
7.1
3.7
2.4
55.9
18,0
14.4
7.2
4.5
75.3
10.1
7.0
75.6
18.3
6.1
94.9
•3.5
4.0
5.1
85.7
9.5
"13.2
J1.6
100.0
76.3
12.5
8.8
1.3
100.0
'Coverage: 945 monitoring sites,
located in 41 states.
"Totals may not add to 100% due to
rounding.
Data source: USD A Forest Service,
2006
EPA Regions
Figure 5-11. Ozone Injury to Forest Plants in U.S. by EPA Regions, 2002
Assessing the impact of ground-level ozone on forests in the eastern United States
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
5-84
-------
species dominate the forest community, the biomass loss from ozone can be significant.
Significant biomass loss can be defined as a more than 2% annual biomass loss, which would
cause long term ecological harm as the short-term negative effects on seedlings compound to
affect long-term forest health (Heck, 1997).
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, 2007). Other common tree species, such as oak (Quercus sppj and
hickory (Carya spp.), are not as sensitive to ozone. Consequently, with knowledge of the
distribution of sensitive species and the level of ozone at particular locations, it is possible to
estimate a "biomass loss" for each species across their range. As shown in Figure 5-12, current
ambient levels of ozone are associated with significant biomass loss across large geographic
areas (U.S. EPA, 2009b). However, this information is unavailable for this rule.
Biomass (% Loss)
•= 1%
| ltd 3%
| 3 to 6%
1 6 to 9%
I >9% (Max 16%)
Figure 5-12. Estimated Black Cherry, Yellow Poplar, Sugar Maple, Eastern White Pine, Virginia
Pine, Red Maple, and Quaking Aspen Biomass Loss due to Current Ozone Exposure, 2006-
2008 (U.S. EPA, 2009b)
5-85
-------
To estimate the biomass loss for forest ecosystems across the eastern United States, the
biomass loss for each of the seven tree species was calculated using the three-month, 12-hour
W126 exposure metric at each location, along with each tree's individual C-R functions. The
W126 exposure metric was calculated using monitored ozone data from CASTNET and AQS
sites, and a three-year average was used to mitigate the effect of variations in meteorological
and soil moisture conditions. The biomass loss estimate for each species was then multiplied by
its prevalence in the forest community using the U.S. Department of Agriculture (USDA) Forest
Service IV index of tree abundance calculated from Forest Inventory and Analysis (FIA)
measurements (Prasad, 2003). Sources of uncertainty include the ozone-exposure/plant-
response functions, the tree abundance index, and other factors (e.g., soil moisture). Although
these factors were not considered, they can affect ozone damage (Chappelka, 1998).
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 impacted 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 (Felzer et al., 2005).
5.5.6.2 Ozone Effects on Crops and Urban Ornamentals
Laboratory and field experiments have also 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,
2007). 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, 2006c). In addition, economic studies have shown
reduced economic benefits as a result of predicted reductions in crop yields, directly affecting
5-86
-------
the amount and quality of the provisioning service provided by the crops in question,
associated with observed ozone levels (Kopp et al., 1985; Adams et al., 1986; Adams et al.,
1989). According to the Ozone Staff Paper, there has been no evidence that crops are becoming
more tolerant of ozone (U.S. EPA, 2007). Using the Agriculture Simulation Model (AGSIM)
(Taylor, 1994) to calculate the agricultural benefits of reductions in ozone exposure, U.S. EPA
estimated that meeting a W126 standard of 21 ppm-hr would produce monetized benefits of
approximately $160 million to $300 million (inflated to 2006 dollars) (U.S. EPA, 2007).23
Urban ornamentals are an additional vegetation category likely to experience some
degree of negative effects associated with exposure to ambient ozone levels. Because ozone
causes visible foliar injury, the aesthetic value of ornamentals (such as petunia, geranium, and
poinsettia) in urban landscapes would be reduced (U.S. EPA, 2007). Sensitive 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, 2007). In addition, many businesses rely on healthy-looking vegetation for
their livelihoods (e.g., horticulturalists, landscapers, Christmas tree growers, farmers of leafy
crops, etc.) and a variety of ornamental species have been listed as sensitive to ozone (Abt
Associates, 2010). The ornamental landscaping industry is valued at more than $30 billion
(inflated to 2006 dollars) annually, by both private property owners/tenants and by
governmental units responsible for public areas (Abt Associates, 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 the absence of adequate exposure-response functions
and economic damage functions for the potential range of effects relevant to these types of
vegetation, we cannot conduct a quantitative analysis to estimate these effects.
5.5.7 Unquantified SO2 and NO^Related Human Health Benefits
Following an extensive evaluation of health evidence from epidemiologic and laboratory
studies, the Integrated Science Assessment for Sulfur Dioxide 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 S02 on 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
23 These 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.
5-87
-------
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
time constraints.
Epidemiological researchers have associated N02 exposure with adverse health effects
in numerous toxicological, clinical and epidemiological studies, as described in the Integrated
Science Assessment for Oxides of Nitrogen - Health Criteria (Final Report) (U.S. EPA, 2008c). 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"
(ISA, section 5.3.2.1). These epidemiologic and experimental studies encompass a number of
endpoints including [Emergency Department (ED)] 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 (ISA, section 5.4). 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 co-benefits due to time constraints.
5.6 Social Cost of Carbon and Greenhouse Gas Co-Benefits
EPA has assigned a dollar value to reductions in carbon dioxide (C02) emissions using
recent estimates of the "social cost of carbon" (SCC). The SCC is an estimate of the monetized
damages associated with an incremental increase in carbon emissions in a given year. It is
5-88
-------
intended to include (but is not limited to) changes in net agricultural productivity, human
health, property damages from increased flood risk, and the value of ecosystem services due to
climate change. The SCC estimates used in this analysis were developed through an interagency
process that included EPA and other executive branch entities, and concluded in February 2010.
EPA first used these SCC estimates in the benefits analysis for the final joint EPA/DOT
Rulemaking to establish Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate
Average Fuel Economy Standards; see the rule's preamble for discussion about application of
SCC (75 FR 25324; 5/7/10). The SCC Technical Support Document (SCC TSD) provides a
complete discussion of the methods used to develop these SCC estimates.24
The interagency group selected four SCC values for use in regulatory analyses, which we
have applied in this analysis: $5.9, $24.3, $39, and $74.4 per metric ton of C02 emissions25 in
2016, in 2007 dollars. The first three values are based on the average SCC from three integrated
assessment models, at discount rates of 2.5%, 3%, and 5%, respectively. SCCs at several
discount rates are included because the literature shows that the SCC is quite sensitive to
assumptions about the discount rate, and because no consensus exists on the appropriate rate
to use in an intergenerational context. The fourth value is the 95th percentile of the SCC from
all three models at a 3% discount rate. It is included to represent higher-than-expected impacts
from temperature change further out in the tails of the SCC distribution. Low probability, high
impact events are incorporated into all of the SCC values through explicit consideration of their
effects in two of the three models as well as the use of a probability density function for
equilibrium climate sensitivity. Treating climate sensitivity probabilistically results in more high
temperature outcomes, which in turn lead to higher projections of damages.
The SCC increases over time because future emissions are expected to produce larger
incremental damages as physical and economic systems become more stressed in response to
greater climatic change. Note that the interagency group estimated the growth rate of the SCC
directly using the three integrated assessment models rather than assuming a constant annual
24 Docket ID EPA-HQ-OAR-2009-0472-114577, Technical Support Document: Social Cost of Carbon for Regulatory
Impact Analysis Under Executive Order 12866, Interagency Working Group on Social Cost of Carbon, with
participation by Council of Economic Advisers, Council on Environmental Quality, Department of Agriculture,
Department of Commerce, Department of Energy, Department of Transportation, Environmental Protection
Agency, National Economic Council, Office of Energy and Climate Change, Office of Management and Budget,
Office of Science and Technology Policy, and Department of Treasury (February 2010). Also available at
http://www.epa.gov/otaq/climate/regulations.htm
25 Note that upstream and downstream emission changes were not considered for this rule. For example, there
may be changes in greenhouse gas emissions (in particular, methane) due to changes in fossil fuel extraction and
transport in response to this proposal, but those emission changes were not quantified.
5-89
-------
growth rate. This helps to ensure that the estimates are internally consistent with other
modeling assumptions. The SCC estimates for the analysis year of 2016 in 2007$ are provided in
Table 5-16.
Table 5-16. Social Cost of Carbon (SCC) Estimates (per tonne of CO2) for 2016 (in 2007$)a
Discount Rate and Statistic SCC Estimate, $
5% Average 5.9
3% Average 24.3
2.5% Average 39.0
3% 95th percentile 74.4
3 The SCC values are dollar-year and emissions-year specific. SCC values represent only a partial accounting of
climate impacts.
When attempting to assess the incremental economic impacts of carbon dioxide
emissions, the analyst faces a number of serious challenges. A recent report from the National
Academies of Science (NRC 2009) points out that any assessment will suffer from uncertainty,
speculation, and lack of information about (1) future emissions of greenhouse gases, (2) the
effects of past and future emissions on the climate system, (3) the impact of changes in climate
on the physical and biological environment, and (4) the translation of these environmental
impacts into economic damages. As a result, any effort to quantify and monetize the harms
associated with climate change will raise serious questions of science, economics, and ethics
and should be viewed as provisional.
The interagency group noted a number of limitations to the SCC analysis, including the
incomplete way in which the integrated assessment models capture catastrophic and non-
catastrophic impacts, their incomplete treatment of adaptation and technological change,
uncertainty in the extrapolation of damages to high temperatures, and assumptions regarding
risk aversion. Current integrated assessment models do not assign value to all of the important
physical, ecological, and economic impacts of climate change because models understandably
lag behind the most recent research. The limited amount of research linking climate impacts to
economic damages makes the interagency modeling exercise even more difficult. The
interagency group hopes that over time researchers and modelers will work to fill these gaps
and that the SCC estimates used for regulatory analysis by the federal government will continue
to evolve with improvements in modeling. Additional details on these limitations are discussed
intheSCCTSD.
5-90
-------
In light of these limitations, the interagency group has committed to updating the
current estimates as the science and economic understanding of climate change and its impacts
on society improves over time. Specifically, the interagency group has set a preliminary goal of
revisiting the SCC values in the next few years or at such time as substantially updated models
become available, and to continue to support research in this area.
Applying the global SCC estimates shown in Table 5-16 to the estimated reductions in
annual C02 emissions of 15 million metric tons for the policy scenario, we estimate the dollar
value of the climate related co-benefits captured by the models for 2016 using three discount
rates 5%, 3%, and 2.5% rather than 3% and 7%.26 These climate co-benefit estimates are
provided in Table 5-17. The C02 emission reductions associated with the policy scenario were
developed using IPM and result largely from projected increases in electricity generation from
natural gas sources and reductions in coal-fired generation by 2016. Even within the coal
generation fleet, there are likely some modest generation shifts away from the least efficient
units towards units that are more efficient to operate. These C02 emission reductions are net of
any C02 emission increases associated with the energy usage for control technologies required
by the rule.
Table 5-17. Monetized Co-Benefits of CO2 Emissions Reductions in 2016 (in millions of
2007$ja,b,c,d
5%
3%
2.5%
3%
Discount Rate and Statistic
Average
Average
Average
95th percentile
SCC Estimate, $
89
360
590
1,100
3 All estimates have been rounded.
b The SCC values are dollar-year and emissions-year specific.
SCC values represent only a partial accounting of climate impacts.
Three discount rates are used to estimate the dollar value of the climate related co-benefits.
As noted above, there are a number of limitations associated with the SCC and its use to
assess the climate benefits of regulations. Beyond the SCC's incomplete treatment of impacts
associated with C02 emissions, it is important to note that SCC is limited to assessing the
26 See SCC TSD for more information about discount rate selection. Also, it is possible that other benefits or costs
of proposed regulations unrelated to CO2 emissions will be discounted at rates that differ from those used to
develop the SCC estimates.
5-91
-------
climate benefits associated with changes in C02 emissions only. However this rule will have an
impact on the emissions of other pollutants that will affect the climate. These other pollutants
include other greenhouse gases, aerosols and aerosols precursors such as black carbon, organic
carbon, sulfur dioxide and nitrogen oxides, and ozone precursors such as nitrogen oxides and
volatile organic carbon compounds. Changes in these pollutants (both increases and decreases)
can be a direct result of changes in electricity generation, including but not limited to the
changes in S02, NOX, and filterable particulate matter identified in Chapter 3 of the RIA, but can
also result from upstream changes in emissions due to changes in fossil fuel extraction and
transport or downstream emission changes for secondary market impacts (not calculated for
this rule). Reductions in black carbon or ozone precursors would lead to further cooling, but
reductions in the other aerosol species and precursors would lead to warming. Therefore,
changes in non-C02 pollutants could potentially augment or offset the climate benefits
calculated here. These pollutants can act in different ways and on different timescales than
carbon dioxide. For example, aerosols reflect (and in the case of black carbon, absorb) incoming
radiation, whereas greenhouse gases absorb outgoing infrared radiation. These aerosols can
also affect climate indirectly by altering properties of clouds. Black carbon can also deposit on
snow and ice, darkening these surfaces and accelerating melting. In terms of lifetime, while
carbon dioxide emissions can increase concentrations in the atmosphere for hundreds to
thousands of years, many of these other pollutants are short lived and remain in the
atmosphere for short periods of time ranging from days to weeks and can therefore exhibit
large spatial and temporal variability. The climate impacts of these other pollutants can be
complex and have not been calculated for this rule.
5.7 Co-Benefits Results
Applying the impact and valuation functions described previously in this chapter to the
estimated changes in ambient PM yields estimates of the changes in physical damages (e.g.,
premature deaths, hospital admissions). Since the air quality modeling performed for this RIA
does not reflect the changes in emissions of PM2.5 precursors associated with the final
emissions control requirements of the rule, we extrapolate the co-benefits of the final rule from
the co-benefits of the air quality modeled emissions (see Appendices 5A and 5B). From these
modeled co-benefits, we calculate BPT values for S02 and direct PM (carbonaceous and crustal),
separately for Eastern and Western states, following the general methodology described by
Fann et al. (2009). We then apply the BPT values to the final emission changes associated with
the revised policy scenario. Since the geographic distribution of emission changes did not
change dramatically from the modeled emission scenarios to the final policy scenario,
extrapolating co-benefits using the BPT approach reasonably approximates the co-benefits of
5-92
-------
the final policy scenario. However, there is additional uncertainty in the extrapolated benefits
estimates relative to the benefits estimated for the air quality modeled emissions.
This section summarizes the health co-benefits estimated for the final policy scenario in
2016. Co-benefits associated with the modeled air quality changes are described in Appendix
5C. Although extrapolating recreational visibility impacts to the final revised policy scenario is
not possible, we estimate that visibility co-benefits add $1.1 billion to the total monetized
benefits of the modeled interim policy scenario (see Appendix 5C). Visibility benefits are not
included in the co-benefits estimate for the final policy. Table 5-18 presents health impacts
among eastern and western states. Monetized values for both health and welfare endpoints are
presented in Table 5-19. All monetary benefits are in constant-year 2007$.
Not all known health and welfare co-benefits for non-HAP pollutants could be quantified
or monetized in this analysis. The monetized value of these unquantified effects is represented
by adding an unknown "B" to the aggregate total. The estimate of total monetized co-benefits
is thus equal to the subset of monetized PM- and C02-related health and welfare co-benefits
plus B, the sum of the non-monetized health and welfare; this B represents both uncertainty
and a bias in this analysis, as it reflects those co-benefits categories that we are unable quantify
in this analysis.
This assessment estimates that in 2016 MATS will result in between 4,200 and 11,000
PM2.5-related avoided premature deaths annually. The total monetized health and climate co-
benefits of MATS in 2016 are between $37 billion and $90 billion using a 3% discount rate and
between $33 billion and $81 using a 7% discount rate. As shown in Appendix 5C, 95% of the
health co-benefits result from reduced exposure to sulfate particles. Mortality co-benefits
account for approximately 93% to 97% of total monetized co-benefits depending on the PM2.5
estimates used, in part because we are unable to quantify most of the non-health co-benefits.
The next largest benefit is for reductions in chronic illness (CB and non-fatal heart attacks),
although this value is more than an order of magnitude lower than for premature mortality.
Hospital admissions for respiratory and cardiovascular causes, visibility, MRADs and work loss
days account for the majority of the remaining co-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.
Figure 5-13 summarizes an array of PM2.5-related monetized co-benefits estimates
based on alternative epidemiology and expert-derived PM-mortality estimate. A comparison of
the incidence table to the monetary co-benefits table reveals that there is not always a close
5-93
-------
correspondence between the number of incidences avoided for a given endpoint and the
monetary value associated with that endpoint. For example, there are over 100 times more
work loss days than premature mortalities, yet work loss days account for only a very small
fraction of total monetized co-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 Table 5-19.
5-94
-------
Table 5-18. Estimated Reduction in Incidence of Adverse Health Effects of the Mercury and
Air Toxics Standards in 2016 (95% confidence intervals)3'15
Impact
Premature death
Pope et al. (2002) (age >30)
Laden et al. (2006) (age
>25)
Infant (< 1 year)
Chronic bronchitis
Non-fatal heart attacks (age >
18)
Hospital admissions-
respiratory (all ages)
Hospital admissions-
cardiovascular (age > 18)
Emergency room visits for
asthma (age < 18)
Acute bronchitis (age 8-12)
Lower respiratory symptoms
(age 7-14)
Upper respiratory symptoms
(asthmatics age 9- 18)
Asthma exacerbation
(asthmatics 6- 18)
Lost work days (ages 18-65)
Minor restricted-activity days
(ages 18-65)
Eastern U.S.0
4,100
(1,100-7,000)
10,000
(4,800 - 16,000)
19
(-21-59)
2,700
(89 - 5,400)
4,600
(1,200-8,100)
820
(320-1,300)
1,800
(1,200-2,100)
3,000
(1,500-4,500)
6,000
(-1,400 - 13,000)
77,000
(30,000 - 120,000)
58,000
(11,000 - 110,000)
130,000
(4,500 - 430,000)
520,000
(440,000 - 600,000)
3,100,000
(2,500,000 - 3,700,000)
Western U.S.
130
(30 - 220)
320
(140-510)
1
(-1-2)
100
(-1-210)
120
(25-210)
17
(6-27)
42
(27-50)
110
(52-160)
250
(-69 - 560)
3,100
(1,100-5,200)
2,400
(360 - 4,400)
5,200
(-6 - 18,000)
21,000
(18,000-24,000)
120,000
(99,000 - 150,000)
Total
4,200
(1,200-7,200)
11,000
(5,000 - 17,000)
20
(-22-61)
2,800
(88-5,600)
4,700
(1,200-8,300)
830
(330-1,300)
1,800
(1,200-2,200)
3,100
(1,600-4,700)
6,300
(-1,400 - 14,000)
80,000
(31,000-130,000)
60,000
(11,000-110,000)
130,000
(4,500 - 450,000)
540,000
(460,000 - 620,000)
3,200,000
(2,600,000 - 3,800,000)
Estimates rounded to two significant figures; column values will not sum to total value.
The negative estimates for certain endpoints are the result of the weak statistical power of the study used to
calculate these health impacts and do not suggest that increases in air pollution exposure result in decreased
health impacts.
c Includes Texas and those states to the north and east.
5-95
-------
Table 5-19. Estimated Economic Value of Health and Welfare co-benefits of the Mercury and
Air Toxics Standards in 2016 (95% confidence intervals, billions of 2007$)a
Impact
Pollutant
Adult premature death (Pope et al. 2002 PM
3% discount rate
7% discount rate
Adult premature death (Laden
3% discount rate
7% discount rate
Infant premature death
Chronic Bronchitis
Non-fatal heart attacks
3% discount rate
7% discount rate
Hospital admissions-
respiratory
Hospital admissions-
cardiovascular
Emergency room visits for
asthma
Acute bronchitis
Lower respiratory symptoms
Upper respiratory symptoms
Asthma exacerbation
Lost work days
Minor restricted-activity days
CO2-related co-benefits (3%
discount rate)
PM2.5
PM2.5
Eastern U.S.b
mortality estimate)
$33
($2.6 -$99)
$30
($2.3 -$90)
Western U.S.
$1.0
(<$0.01-$3.1)
$0.9
(<$0.01-$2.8)
Total
$34
($2.6 -$100)
$30($2.4-$92)
et al. 2006 PM mortality estimate)
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
C02
$84
($7.4 -$240)
$76
($6.7 -$220)
$0.2
($-0.2 -$0.8)
$1.3
($0.1 -$6.1)
$0.5
($0.1 -$1.3)
$0.4
($0.1 -$1.0)
$0.01
(<$0.01-$0.02)
$0.03
(<$0.01-$0.05)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$0.1
($0.1 -$0.1)
$0.2
($0.1 -$0.3)
$2.6
($0.1 -$7.6)
$2.3
($0.1 -$6.9)
<$0.01
$0.1
(<$0.01-$0.2)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$87
($7.5 -$250)
$78
($6.8 -$230)
$0.2
($-0.2 - $0.8)
$1.4
($0.1 -$6.4)
$0.5
($0.1 -$1.3)
$0.4
($0.1 -$1.0)
$0.01
($0.01 - $0.02)
$0.03
(<$0.01 - $0.05)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$0.1
($0.1 -$0.1)
$0.2
($0.1 -$0.3)
$0.36
(continued)
5-96
-------
Table 5-19. Estimated Economic Value of Health and Welfare co-benefits of the Mercury and
Air Toxics Standards in 2016 (95% confidence intervals, billions of 2007$)a
(continued)
Total Monetized co-benefits (Pope et al
3% discount rate
7% discount rate
Total Monetized Benefits (Laden et al.
3% discount rate
7% discount rate
. 2002 PM2.5 mortality estimate)
$35+B
($2.8 -$110)
$32+B
($2.5 - $98)
2006 PM2.5 mortality estimate)
$87+B
($7.5 -$250)
$78+B
($6.8 -$230)
$1.1+B
($0.03 - $3.4)
$1.0+B
($0.03 -$3.1)
$2.7+B
($0.1 -$7.9)
$2.4+B
($0.1 -$7.2)
$37+B
($3.2 -$110)
$33+B
($2.9 - $100)
$90+ B
($8.0 - $260)
$81+B
($7.3 - $240)
a Estimates rounded to two significant figures. The negative estimates for certain endpoints are the result of the
weak statistical power of the study used to calculate these health impacts and do not suggest that increases in
air pollution exposure result in decreased health impacts. Confidence intervals reflect random sampling error
and not the additional uncertainty associated with co-benefits scaling described above. The net present value of
reduced CO2 emissions are calculated differently than other co-benefits. The same discount rate used to
discount the value of damages from future emissions (SCC at 5, 3, 2.5 percent) is used to calculate net present
value of SCC for internal consistency. This table shows monetized CO2 co-benefits at discount rates at 3 and 7
percent that were calculated using the global average SCC estimate at a 3% discount rate because the
interagency workgroup on this topic deemed this marginal value to be the central value. In Section 5.6 we also
report CO2 co-benefits using discount rates of 5 percent (average), 2.5 percent (average), and 3 percent (95th
percentile).
b Includes Texas and those states to the north and east.
PM2.5 mortality benefits represent a substantial proportion of total monetized co-
benefits (over 90%), and these estimates have the following key assumptions and uncertainties.
1. The PM2.5-related co-benefits were derived through a benefit per-ton approach,
which does not fully reflect local variability in population density, meteorology,
exposure, baseline health incidence rates, or other local factors that might lead to
an over-estimate or under-estimate of the actual co-benefits of controlling PM
precursors. In addition, differences in the distribution of emissions reductions
across states between the modeled scenario and the final rule scenario add
uncertainty to the final benefits estimates.
2. This rule is expected to reduce emissions of S02, NOX, and directly emitted PM2.5.
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 via transported precursors emitted from EGUs may differ
significantly from direct PM2.5 released from diesel engines and other industrial
5-97
-------
sources, but the scientific evidence is not yet sufficient to allow differential effects
estimates by particle type.
3. We assume that the health impact function for fine particles is linear within the
range of ambient concentrations under consideration. Thus, the estimates include
health co-benefits from reducing fine particles in areas with varied concentrations of
PM2.5, including both regions that are in attainment with fine particle standard and
those that do not meet the standard down to the lowest modeled concentrations.
Based on our review of the current body of scientific literature, EPA estimated PM-
related mortality without applying an assumed concentration threshold. EPA's Integrated
Science Assessment for Particulate Matter (U.S. EPA, 2009a), which was reviewed by EPA's
Clean Air Scientific Advisory Committee (U.S. EPA-SAB, 2009a; U.S. EPA-SAB, 2009b), concluded
that the scientific literature consistently finds that a no-threshold log-linear model most
adequately portrays the PM-mortality concentration-response relationship while also
recognizing potential uncertainty about the exact shape of the concentration-response
function. Consistent with this finding, we incorporated a "Lowest Measured Level" (LML)
assessment, which is a method EPA has employed in several recent RIA's including the Cross-
State Air Pollution Rule (U.S. EPA, 2011b).
This approach summarizes the distribution of avoided PM mortality impacts according
to the baseline (i.e. pre-MATS) PM2.5 levels experienced by the population receiving the PM2.5
mortality benefit (Figures 5-14 and 5-15). We identify on this figure the lowest air quality levels
measured in each of the two primary epidemiological studies EPA used to quantify PM-related
mortality. This information allows readers to determine the portion of PM-related premature
deaths avoided occurring at or above the LML of each study; in general, our confidence in the
estimated PM-related premature deaths avoided decreases as we consider air quality levels
further below the LML in the two epidemiological studies. While the LML analysis provides
some insight into the level of uncertainty in the estimated PM mortality co-benefits, EPA does
not view the LML as a threshold and continues to quantify PM-related mortality impacts using a
full range of modeled air quality concentrations. 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, 2010e),
which is provided in Appendix 5E of this RIA. While this figure describes the relationship
between baseline PM2.5 exposure and avoided premature deaths for the modeled air quality
scenario, we expect the distribution of mortality impacts to be fairly similar between the two
cases.
5-98
-------
PM2.5 Benefits estimates derived from 2 epidemiology functions and 11 expert
functions
Figure 5-13. Economic Value of Estimated PM2.5-Related Health co-benefits of the Mercury
and Air Toxics Standards in 2016 According to Epidemiology or Expert-Derived PM Mortality
Risk Estimate3 b
3 Based on the modeled interim baseline, which is approximately equivalent to the final baseline (see Appendix
5A).
b Column total equals sum of PM2.5-related mortality and morbidity co-benefits.
5-99
-------
35%
LML of Laden et al. (2006) study
LM L of Popeetal. (2002) study
0%
14
IS
1 2 3 4 5 6 7 7.5 8 9 10 12
Baseline annual mean PM2 5 level (ng/rn5)
Of the total PM-related deaths avoided:
73% occur among population exposed to PM levels at or above the LML of the Pope etal. study.
I I % occur among population exposed to PM levels at or above the LML of the Laden etal study.
20
Figure 5-14. Percentage of Total PM-Related Mortalities of the Mercury and Air Toxics
Standards in 2016 Avoided by Baseline Air Quality Level3
3 Based on the modeled interim baseline, which is approximately equivalent to the final baseline (see Appendix
5A)
Some proportion of the avoided PM-related impacts we estimate in this analysis occur
among populations exposed at or above the LML of the Laden et al. (2006) study, while a
majority of the impacts occur at or above the LML of the Pope et al. (2002) study (Figure 5-14),
increasing our confidence in the PM-related premature mortality analysis. Based on the
modeled interim baseline which is approximately equivalent to the final baseline (see Appendix
5A), 11% and 73% of the estimated avoided premature deaths occur at or above an annual
mean PM2.5 level of 10 u.g/m3 (the LML of the Laden etal. 2006 study) and 7.5 u.g/m3(the LML
of the Pope et al. 2002 study), respectively. Using these percentages derived from the modeled
interim baseline, Table 5-20 shows the allocation of reduced incidence above and below the
LMLs of Laden et al. (2006) and Pope et al. (2002). As we model avoided premature deaths
among populations exposed to levels of PM2.s, we have lower confidence in levels below the
LML for each study.
5-100
-------
Table 5-20. Estimated Reduction in Incidence of Premature Adult Mortality due to the
Mercury and Air Toxics Standards in 2016 Occurring Above and Below the Lowest Measured
Levels in the Underlying Epidemiology Studies3
Allocation of Reduced Mortality Incidence
Study and
Lowest Measured Level (LML)
Pope et al. (2002), 7.5 u.g/m3
Laden et al. (2006), 10 u.g/m3
Total Reduced
Mortality Incidence
4,200
11,000
Below LML
1,100
9,600
At or Above LML
3,100
1,200
' These estimates are rounded to two significant digits. 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. See Appendix 5E for more
information.
A large fraction of the PM2.5-related benefits associated with this rule occur below the
level of the National Ambient Air Quality Standard (NAAQS) for annual PM2.5 at 15 u.g/m3, which
was set in 2006. It is important to emphasize that NAAQS are not set at a level of zero risk.
Instead, the NAAQS reflect the level determined by the Administrator to be protective of public
health within an adequate margin of safety, taking into consideration effects on susceptible
populations. While benefits occurring below the standard may be less certain than those
occurring above the standard, EPA considers them to be legitimate components of the total
benefits estimate.
5-101
-------
100%
90%
«
in
T3
70%
r
° 50%
a) 40%
30%
I 20%
3
10%
0% 4
LML of Pope et al. (2002) study
z
LML of Laden et al. (2006) study
1 2 3 4 5 6 7 7.5 8 9 10 12 14 16 18 20
Baseline annual mean PM15 level (ng/rnj)
Of the total PM-related deaths avoided:
73% occur among population exposed to PM levels at or above the LML of the Pope etal. study.
I l%occur among population exposed to PM levels atorabovethe LMLofthe ndeneta study.
Figure 5-15. Cumulative Percentage of Total PM-Related Mortalities of the Mercury and Air
Toxics Standards in 2016 Avoided by Baseline Air Quality Level3
a Based on the modeled interim baseline, which is approximately equivalent to the final baseline (see Appendix
5A)
While the LML of each study is important to consider when characterizing and
interpreting the overall level PM2.5-related co-benefits, as discussed earlier in this chapter, EPA
believes that both 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 Laden et al. (2006) analysis of the Harvard Six Cities and the Pope et al. (2002)
analysis of the American Cancer Society cohorts there are innumerable other attributes that
may affect the size of the reported risk 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.
5-102
-------
5.8 Discussion
This analysis demonstrates the significant health and welfare co-benefits of MATS. We
estimate that in 2016 the rule will have reduced the number of PM2.5-related premature deaths
by between 4,200 and 11,000 and produce substantial non-mortality co-benefits. We estimate
the monetized health and climate co-benefits of MATS to be $37 billion to $90 billion at a 3%
discount rate and $33 billion to $81 billion at a 7% discount rate in 2016, depending on the
epidemiological function used to estimate reductions in premature mortality. All estimates are
in 2007$. Health co-benefits comprise approximately 99% of these total monetized co-benefits.
This co-benefits assessment omits several important categories of co-benefits that we were
unable to quantify, including health and ecological co-benefits from reducing exposure to
ozone, ecosystem co-benefits for reducing nitrogen and sulfate deposition, the direct health co-
benefits from reducing exposure to S02 and N02, and reduced visibility impairment in
recreational areas. Inherent in any complex RIA such as this one are multiple sources of
uncertainty. Some of these we characterized through our quantification of statistical error in
the concentration response relationships and our use of the expert elicitation-derived PM2.5
mortality functions. Others are unquantified, including the projection of atmospheric conditions
and source-level emissions, the projection of baseline morbidity rates, incomes and
technological development.
The emissions scenarios for the RIA reflects the Cross-State Air Pollution Rule (CSAPR) as
finalized in July 2011 and the emissions reductions of SOx, NOx, directly emitted PM, and C02
are consistent with application of federal rules, state rules and statutes, and other binding,
enforceable commitments in place by December 2010 for the analysis timeframe27. EPA has
proposed minor modifications to the state level S02 budgets in the Cross State Air Pollution
Rule (CSAPR; see http://www.epa.gov/airtransport/actions.html). These modifications are
expected to result in small changes in the levels of S02 emission reductions expected in a
number of states, with the largest impact expected in Texas. EPA expects that these changes
will slightly reduce the benefits of CSAPR, and will have a small impact on the baseline
emissions for MATS. Because of the change in the baseline S02 emissions for MATS, the MACT
controls may result in slightly larger reductions in S02 and other emissions, and consequently
slightly higher benefits. It is important to note that the total monetized benefits of both rules is
not expected to change significantly, rather, the allocation of the S02 emissions reductions and
27 Consistent with the mercury risk deposition modeling for MATS, EPA did not model non-federally enforceable
mercury-specific emissions reduction rules in the base case or MATS policy case (see preamble Section III. A.).
Note that this approach does not significantly affect SO2 and NOX projections underlying the cost and benefit
results presented in this RIA.
5-103
-------
benefits between the rules is changed, so that MATS accounts for slightly more of the total S02
emissions reductions and benefits, and CSAPR slightly less.
5.9 References
Abbey, D.E., B.L. Hwang, R.J. Burchette, T. Vancuren, and P.K. Mills. 1995. Estimated Long-Term
Ambient Concentrations of PM(10) and Development of Respiratory Symptoms in a
Nonsmoking Population. Archives of Environmental Health 50(2): 139-152.
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 [see comments]. American Journal of Respiratory and Critical
Care Medicine 159(2):373-382.
Abt Associates, Inc. April 2003. Proposed Nonroad Land-based Diesel Engine Rule: Air Quality
Estimation, Selected Health and Welfare Benefits Methods, and Benefit Analysis Results.
Prepared for Office of Air Quality Planning and Standards, U.S. EPA.
Abt Associates, Inc. 2005. U.S. EPA. Urban ornamental plants: sensitivity to ozone and potential
economic losses. Memorandum to Bryan Hubbell and Zachary Pekar.
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.
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.
Agency for Healthcare Research and Quality (AHRQ). 2000. HCUPnet, Healthcare Cost and
Utilization Project.
American Lung Association (ALA). 2010. Trends in Asthma Morbidly and Mortality. American
Lung Association Epidemiology and Statistics Unit, Research and Program Services
Division.
American Lung Association. 1999. Chronic Bronchitis. Available on the Internet at
.
5-104
-------
American Lung Association. 2002. Trends in Asthma Morbidity and Mortality. American Lung
Association, Best Practices and Program Services, Epidemiology and Statistics Unit.
Available on the Internet at .
Banzhaf, S., D. Burtraw, D. Evans, and A. Krupnick. 2006. "Valuation of Natural Resource
Improvements in the Adirondacks." Land Economics 82:445-464.
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.
Brookshire, D.S., Thayer, M.A., Schulze, W.D. & D'Arge, R.C. 1982. "Valuing Public Goods: A
Comparison of Survey and Hedonic Approaches." The American Economic Review.
72(1): 165-177.
Brown, L.H. 2002. Profile of the Annual Fall Foliage Tourist in Vermont: Travel Year 2001. Report
prepared for the Vermont Department of Tourism and Marketing and the Vermont
Tourism Data Center in association with the University of Vermont, Burlington, VT.
Burnett RT, Smith-Doiron M, Stieb D, Raizenne ME, Brook JR, Dales RE, et al. 2001. Association
between ozone and hospitalization for acute respiratory diseases in children less than 2
years of age. Am J Epidemiol 153(5):444-452.
Carnethon, M.R., D. Liao, G.W. Evans, W.E. Cascio, L.E. Chambless, W.D. Rosamond, and G.
Heiss. 2002. Does the Cardiac Autonomic Response to Postural Change Predict Incident
Coronary Heart Disease and Mortality? The Atherosclerosis Risk in Communities Study.
Centers for Disease Control: Wide-ranging OnLine Data for Epidemiologic Research (CDC
Wonder) (data from years 2004-2006), Centers for Disease Control and Prevention
(CDC), U.S. Department of Health and Human Services, Available on the Internet at
.
Centers for Disease Control and Prevention (CDC). 2009. Summary Health Statistics for U.S.
Children: National Health Interview Survey, 2008. Vital and Health Statistics. Series 10,
Number 244. U.S. Department of Health and Human Services. National Center for
Health Statistics. December.
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.
Chen L, Jennison BL, Yang W, Omaye ST. 2000. Elementary school absenteeism and air
pollution. InhalToxicol 12(11):997-1016.
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 NAPAP State of Science and Technology
Report 27.
5-105
-------
Chestnut, L.G. April 15,1997. Draft Memorandum: Methodology for Estimating Values for
Changes in Visibility at National Parks.
Chestnut, L.G., and R.D. Rowe. 1990a. 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., and R.D. Rowe. 1990b. Preservation Values for Visibility Protection at the
National Parks: Draft Final Report. Prepared for Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC, and Air
Quality Management Division, National Park Service, Denver, CO.
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
.
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.
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.
Davidson K, Hallberg A, McCubbin D, Hubbell BJ. 2007. Analysis of PM2.5 Using the
Environmental Benefits Mapping and Analysis Program (BenMAP). J Toxicol Environ
Health 70: 332-346.
De Steiguer, J., Pye, J., Love, C. 1990. Air Pollution Damage to U.S. Forests. Journal of Forestry,
88(8), 17-22.
DeHayes, D.H., P.G. Schaberg, G.J. Hawley, and G.R. Strimbeck. 1999. Acid rain impacts on
calcium nutrition and forest health. Bioscience 49(10):789-800.
Dekker, J.M., R.S. Crow, A.R. Folsom, P.J. Hannan, D. Liao, C.A. Swenne, and E.G. Schouten.
2000. Low Heart Rate Variability in a 2-Minute Rhythm Strip Predicts Risk of Coronary
Heart Disease and Mortality From Several Causes: The ARIC Study. Circulation 2000
102:1239-1244.
5-106
-------
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.
Eisenstein, E.L., L.K. Shaw, K.J. Anstrom, C.L. Nelson, Z. Hakim, V. Hasselblad and D.B. Mark.
2001. Assessing the Clinical and Economic Burden of Coronary Artery Disease: 1986-
1998. Medical Care 39(8):824-35.
Fann, N., C.M. Fulcher, B.J. Hubbell. 2009. The influence of location, source, and emission type
in estimates of the human health benefits of reducing a ton of air pollution. Air dual
Atmos Health 2:169-176.
Felzer, B., J. Reilly, J. Melillo, D Kicklighter, M. Sarogim, C Wang, R. Prinn, Q. Zhuang. 2005.
Future effects of Ozone on Carbon Sequestration and Climate Change Policy using a
Global Biogeochemical Model. Climatic Change (2005) 73: 345-373
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.
Freeman(lll), AM. 1993. The Measurement of Environmental and Resource Values: Theory and
Methods. Washington, DC: Resources for the Future.
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.
Gold, D.R., A. Litonjua, J. Schwartz, E. Lovett, A. Larson, B. Nearing, G. Allen, M. Verrier, R.
Cherry., and R. Verrier. 2000. Ambient Pollution and Heart Rate Variability. Circulation
101(ll):1267-73.
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.
Harrington, W., and P.R. Portney. 1987. Valuing the Benefits of Health and Safety Regulation.
Journal of Urban Economics 22:101-112.
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.
5-107
-------
Hollman, F.W., T.J. Mulder, and J.E. Kalian. January 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.
Hubbell BJ, Fann N, LevyJI. 2009. Methodological Considerations in Developing Local-Scale
Health Impact Assessments: Balancing National, Regional and Local Data. Air dual
Atmos Health doi: 10.1007/sll869-009-0037-z [online 31 March 2009].
Hubbell BJ, Hallberg A, McCubbin D, Post, E. 2005. Health-Related Benefits of Attaining the 8-Hr
Ozone Standard. Environ Health Perspect 113: 73—82.
Hutchinson, J., D. Maynard, and L. Geiser. 1996. Air Quality and Lichens —A Literature Review
Emphasizing the Pacific Northwest, USA. USDA Forest Service, Pacific Northwest Region
Air Resource Management Program, U.S. Forest Service, U.S. Department of Agriculture.
Hutchison, R., and C.E. Kraft. 1994. Hmong Fishing Activity and Fish Consumption. Journal of
Great Lakes Research 20(2):471-487.
Industrial Economics, Inc. 2006. Expanded Expert Judgment Assessment of the Concentration-
Response Relationship Between PM2.s Exposure and Mortality. Prepared for the U.S.
EPA, Office of Air Quality Planning and Standards, September. Available on the Internet
at .
Industrial Economics, Incorporated (lEc). 1994. Memorandum to Jim DeMocker, Office of Air
and Radiation, Office of Policy Analysis and Review, U.S. Environmental Protection
Agency. March 31.
Intergovernmental Panel on Climate change (IPCC). 2007. Climate Change 2007 - Synthesis
Report Contribution of Working Groups I, II and III to the Fourth Assessment Report of
the IPCC.
Ito, K. 2003. Associations of Particulate Matter Components with Daily Mortality and Morbidity
in Detroit, Michigan. In Revised Analyses of Time-Series Studies of Air Pollution and
Health. Special Report. Health Effects Institute, Boston, MA.
Jenkins, D.H., J. Sullivan, and G.S. Amacher. 2002. Valuing high altitude spruce-fir forest
improvements: Importance of forest condition and recreation activity. Journal of Forest
Economics 8:77-99.
Jerrett M, Burnett RT, Pope CA, III, et al. 2009. Long-Term Ozone Exposure and Mortality. N Engl
J Med 360:1085-95.
Joslin, J.D., Kelly, J.M., 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.
5-108
-------
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.
Kleckner, N., and J. Neumann. June 3, 1999. Recommended Approach to Adjusting WTP
Estimates to Reflect Changes in Real Income. Memorandum to Jim Democker, U.S.
EPA/OPAR.
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.
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.O. Sills and K.L.
Abt. Dordrecht, The Netherlands: Kluwer Academic Publishers.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi, Y, et al. 2009. Extended follow-up and
spatial analysis of the American Cancer Society study linking particulate air pollution and
mortality. HEI Research Report, 140, Health Effects Institute, Boston, MA.
Krewski, D., R.T. Burnett, M.S. Goldbert, K. Hoover, J. Siemiatycki, M. Jerrett, M. Abrahamowicz,
and W.H. White. 2000. Reanalysis of the Harvard Six Cities Study and the American
Cancer Society Study of Particulate Air Pollution and Mortality. Special Report to the
Health Effects Institute. Cambridge MA. July.
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, et al. 2000. Public-health
imjpact of outdoor and traffic-related air pollution: A European Assessment. The Lancet
356(9232):795-801.
Kunzli, N., S. Medina, R. Kaiser, P. Quenel, F. Horak Jr, and M. Studnicka. 2001. Assessment of
Deaths Attributable to Air Pollution: Should We Use Risk Estimates Based on Time Series
or on Cohort Studies? American Journal of Epidemiology 153(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.
5-109
-------
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>.
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.
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, doi: 10.1111/J.1539-
6924.2009.01227.x [Online 9 Apr 2009]
Liao, D., J. Cai, W.D. Rosamond, R.W. Barnes, R.G. Hutchinson, E.A. Whitsel, P. Rautaharju, and
G. Heiss. 1997. Cardiac Autonomic Function and Incident Coronary Heart Disease: A
Population-Based Case-Cohort Study. The ARIC Study. Atherosclerosis Risk in
Communities Study. American Journal of Epidemiology 145(8):696-706.
Liao, D., J. Creason, C. Shy, R. Williams, R. Watts, and R. Zweidinger. 1999. Daily Variation of
Particulate Air Pollution and Poor Cardiac Autonomic Control in the Elderly. Environ
Health Perspect 107:521-5.
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.
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.
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.
5-110
-------
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.
Magari, S.R., R. Mauser, J. Schwartz, P.L. Williams, T.J. Smith, and D.C. Christian!. 2001.
Association of Heart rate Variability with Occupational and Environmental Exposure to
Particulate Air Pollution. Circulation 104(9):986-91.
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.
September 1993. Valuing Eastern Visibility: A Field Test of the Contingent Valuation
Method. Prepared for Office of Policy, Planning and Evaluation, U.S. Environmental
Protection Agency. Available on the Internet at <
http://yosemite.epa.gov/ee/epa/eerm.nsf/vwAN/EE-0008-l.pdf/$file/EE-0008-l.pdf>.
Millennium Ecosystem Assessment Board (MEA). 2005. Ecosystems and Human Well-being:
Synthesis. Washington, DC: World Resources Institute.
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).
Moolgavkar SH, Luebeck EG, Anderson EL. 1997. Air pollution and hospital admissions for
respiratory causes in Minneapolis St. Paul and Birmingham. Epidemiology. 8(4):364-370.
Moolgavkar, S.H. 2000. Air Pollution and Hospital Admissions for Diseases of the Circulatory
System in Three U.S. Metropolitan Areas. Journal of the Air and Waste Management
Association 50:1199-1206.
Moolgavkar, S.H. 2003. Air Pollution and Daily Deaths and Hospital Admissions in Los Angeles
and Cook Counties. In Revised Analyses of Time-Series Studies of Air Pollution and
Health. Special Report. Boston, MA: Health Effects Institute.
Mrozek, J.R., and L.O. Taylor. 2002. What Determines the Value of Life? A Meta-Analysis.
Journal of Policy Analysis and Management 21(2):253-270.
National 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.
5-111
-------
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 Oceanic and Atmospheric Administration (NOAA). 2007. Annual Commercial Landing
Statistics. August. Available on the Internet at
-------
Peel, J. L, P. E. Tolbert, M. Klein, et al. 2005. Ambient air pollution and respiratory emergency
department visits. Epidemiology. Vol. 16 (2): 164-74.
Peters, A., D.W. Dockery, J.E. Muller, and M.A. Mittleman. 2001. Increased Particulate Air
Pollution and the Triggering of Myocardial Infarction. Circulation 103:2810-2815.
Peterson, D.E., M.S. Kanarek, M.A. Kuykendall, J.M. Died rich, H.A. Anderson, P.L. Remington,
and T.B. Sheffy. 1994. Fish Consumption Patterns and Blood Mercury Levels in
Wisconsin Chippewa Indians. Archives of Environmental Health 49(l):53-58.
Pitchford, M. and W. Malm. 1993. Development and applications of a standard visual index.
Atmospheric Environment 28(5):1049-1054.
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, 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, C.A., III, M.J. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer, and C.W.
Heath, Jr. 1995. Particulate Air Pollution as a Predictor of Mortality in a Prospective
Study of U.S. Adults. American Journal of Respiratory Critical Care Medicine 151:669-
674.
Pope, C.A., III, R.T. Burnett, G.D. Thurston, M.J. Thun, E.E. Calle, D. Krewski, and J.J. Godleski.
2004. Cardiovascular Mortality and Long-term Exposure to Particulate Air Pollution.
Circulation 109: 71-77.
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, 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.
Prasad, A.M. and Iverson, L.R. 2003. Little's range and FIA importance value database for 135
eastern U.S. tree species. Northeastern Research Station, USDA Forest Service. Available
on the Internet at http://www.fs.fed.us/ne/delaware/4153/global/littlefia/index.html
Pye, J.M. 1988. Impact of ozone on the growth and yield of trees: A review. Journal of
Environmental Quality, 17, 347-360.
Rae, D.A. 1983. Benefits of Visual Air Quality in Cincinnati Results of a Contingent Ranking
Survey. Prepared for Electric Power Research Institute. May.
5-113
-------
Ransom, Michael, and C. Arden Pope. 1992. M.R. Ransom and C.A. Pope, III, Elementary school
absences and PMi0 pollution in Utah Valley. Environ. Res. 58, pp. 204-219.
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.
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
Samet, J.M., S.L. Zeger, F. Dominici, F. Curriero, I. Coursac, D.W. Dockery, J. Schwartz, and A.
Zanobetti. 2000. The National Morbidity, Mortality and Air Pollution Study: Part II:
Morbidity, Mortality and Air Pollution in the United States. Research Report No. 94, Part
II. Health Effects Institute, Cambridge MA. June.
Schwartz J. 1994a. PM(10) Ozone, and Hospital Admissions For the Elderly in Minneapolis St
Paul, Minnesota. Arch Environ Health. 49(5):366-374.
Schwartz J. 1994b. Air Pollution and Hospital Admissions For the Elderly in Detroit, Michigan.
Am J Respir Crit Care Med. 150(3):648-655.
Schwartz J. 1995. Short term fluctuations in air pollution and hospital admissions of the elderly
for respiratory disease. Thorax. 50(5):531-538.
Schwartz, J. 1993. Particulate Air Pollution and Chronic Respiratory Disease.
EnvironmentResearch 62:7-13.
Schwartz, J. 2005. How sensitive is the association between ozone and daily deaths to control
for temperature? Am J Respir Crit Care Med. Vol. 171 (6): 627-31.
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.
5-114
-------
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 Life. Journal of Political Economy
110:1168-1197.
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, 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, V.K., Pattanayak, S.K., and Van Houtven. 2006. Structural benefit transfer: An example
using VSL estimates. Ecological Economics 60(2):361-371.
Smith, V.K., G. Van Houtven, and S.K. Pattanayak. 2002. Benefit Transfer via Preference
Calibration. Land Economics 78:132-152.
Spencer, D.M., and D.F. Holecek. 2007. Basic characteristics of the fall tourism market. Tourism
Management 28:491-504.
Standard and Poor's. 2000. "The U.S. Economy: The 25 Year Focus." Winter. Web site:
http://www2standardandpoors.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. Vol. 160 (1): 211-5.
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.
Taylor R. 1994. Deterministic versus stochastic evaluation of the aggregate economic effects of
price support programs. Agricultural Systems 44: 461-473.
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. U.S. Environmental
Protection Agency Grant #807768-01-0.
5-115
-------
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.
January
Tsuji, H., M.G. Larson, F.J. Venditti, Jr., E.S. Manders, J.C. Evans, C.L. Feldman, D. Levy. 1996.
Impact of Reduced Heart Rate Variability on Risk for Cardiac Events. The Framingham
Heart Study. Circulation 94(ll):2850-2855.
U.S. Bureau of Census. 2000. Population Projections of the United States by Age, Sex, Race,
Hispanic Origin and Nativity: 1999 to 2100. Population Projections Program, Population
Division, U.S. Census Bureau, Washington, DC. Available on the Internet at
.
U.S. Department of Agriculture - Forest Service (USDA Forest Service). 2006. Ozone bioindicator
data. Accessed 2006. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 1999. The Benefits and Costs of the Clean Air
Act 1990 to 2010: EPA Report to Congress. Office of Air and Radiation, Office of Policy,
Washington, DC. November. EPA report no. EPA410-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 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). 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. Ecological Benefits Assessment
Strategic Plan. EPA-240-R-06-001. Office of the Administrator. Washington, DC. October.
Available on the Internet at
.
5-116
-------
U.S. Environmental Protection Agency (U.S. EPA). 2006c. 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). 2007. 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
http://www.epa.gov/ttn/naaqs/standards/ozone/data/2007_07_ozone_staff_paper.pdf
U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis, 2008
National Ambient Air Quality Standards for Ground-level Ozone, Chapter 6. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. March. Available at
.
U.S. Environmental Protection Agency (U.S. EPA). 2008b. Integrated Science Assessment for
Oxides of Nitrogen and Sulfur-Ecological Criteria National (Final Report). National
Center for Environmental Assessment, Research Triangle Park, NC. EPA/600/R-08/139.
December. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2008c. 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). 2009a. 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). 2009b. Risk and Exposure Assessment for
Review of the Secondary National Ambient Air Quality Standards for Oxides of Nitrogen
and Oxides of Sulfur (Final). 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. 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 .
U.S. Environmental Protection Agency (U.S. EPA). 2010c. Quantitative Health Risk Assessment
for Particulate Matter. 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. 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). 2011a. Proposed Regulatory Impact Analysis
(RIA) for the Toxics Rule. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. March. Available on the Internet at
.
U.S. Environmental Protection Agency (U.S. EPA). 2011b. Regulatory Impact Analysis for the
Federal Implementation Plans to Reduce Interstate Transport of Fine Particulate Matter
and Ozone in 27 States; Correction of SIP Approvals for 22 States. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. June. Available on the Internet at
.
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. 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.
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.
5-118
-------
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.
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
<
http://vosemite.epa.gov/sab/sabproduct.nsf/7CCBBFE15CD4C8B185256F17005E3079/$
File/council adv 04004.pdf>.
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). 2009b. Review
of EPA'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). 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
.
5-119
-------
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. Food and Drug Administration (U.S. FDA). 2001. Fish and Fisheries Products Hazards and
Controls Guidance: Chapter 10: Methyl Mercury. Third Edition. June. Available on the
Internet at <
http://www.fda.gov/Food/GuidanceComplianceRegulatorylnformation/GuidanceDocum
ents/Seafood/ucm092041.htm>.
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.
Vedal, S., J. Petkau, R. White, and J. Blair. 1998. Acute Effects of Ambient Inhalable Particles in
Asthmatic and Nonasthmatic Children. American Journal of Respiratory and Critical Care
Medicine 157(4):1034-1043.
Viscusi, V.K., and J.E. Aldy. 2003. The Value of a Statistical Life: A Critical Review of Market
Estimates throughout the World. Journal of Risk and Uncertainty 27(l):5-76.
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.
Wilson, A. M., C. P. Wake, T. Kelly, et al. 2005. Air pollution, weather, and respiratory
emergency room visits in two northern New England cities: an ecological time-series
study. Environ Res. Vol. 97 (3): 312-21.
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.
5-120
-------
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 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. 2008. Population by Single Year of Age CD. CD-ROM. Woods &
Poole Economics, Inc. Washington, D.C.
World Health Organization (WHO). 1977. International Classification of Diseases, 9th revision
(ICD-9). Geneva:World Health Organization.
5-121
-------
APPENDIX 5A
IMPACT OF THE INTERIM POLICY SCENARIO ON EMISSIONS
5A.1 Introduction
This section summarizes the emissions inventories that are used to create emissions
inputs to the air quality modeling performed for this rule. A summary of the emissions
reductions that were modeled for this rule is provided. Note that the emissions processing and
corresponding air quality modeling were used to develop BPT scaling factors for the benefits
calculation as described in this RIA. More information on this approach can be found in
Appendix 5C. The emissions inventories were processed into the form required by the
Community Multi-scale Air Quality (CMAQ) model. CMAQ simulates the numerous physical and
chemical processes involved in the formation, transport, and destruction of ozone, particulate
matter and air toxics.
As part of the analysis for this rulemaking, the modeling system was used to calculate
daily and annual PM2.5 concentrations, 8-hr maximum ozone and visibility impairment. Model
predictions of PM2.5 and ozone are used in a relative sense to estimate scenario-specific, future-
year design values of PM2.5 and ozone. These are combined with monitoring data to estimate
population-level exposures to changes in ambient concentrations for use in estimating health
and welfare effects. In the remainder of this section we provide an overview of (1) the
emissions components of the modeling platform, (2) the development of the 2005 base year
emissions, (3) the development of the future year baseline emissions, and (4) the development
of the future year control case emissions.
5A.2 Overview of Modeling Platform and Emissions Processing Performed
A modeling platform is the collection of the inputs to an air quality model, including the
settings and data used for the model, including emissions data, meteorology, initial conditions,
and boundary conditions. The 2005-based air quality modeling platform used for this RIA
includes 2005 base year emissions and 2005 meteorology for modeling ozone and PM2.5 with
CMAQ. In support of this rule, EPA modeled the air quality in the Eastern and the Western
United States using two separate model runs, each with a horizontal grid resolution of 12 km x
12 km. These 12 km modeling domains were "nested" within a modeling domain covering the
remainder of the lower 48 states and surrounding areas using a grid resolution of 36 x 36 km.
The results from the 36-km modeling were used to provide incoming "boundary" for the 12km
grids. Additional details on the non-emissions portion of the 2005v4.3 modeling platform used
for this RIA are described in the air quality modeling section (Appendix 5B).
5A-1
-------
The 2005-based air quality modeling platform used in support of this RIA is version 4.3
and is referred to as the 2005v4.3 platform. It is an update to the 2005-based platform, version
4.1 (i.e., 2005v4.1) used for the proposal modeling and for the appropriate and necessary
finding. The Technical Support Document "Preparation of Emissions Inventories for the Version
4.1, 2005-based Platform" (see http://www.epa.gov/ttn/chief/emch/index.htmltftoxics)
provides information on the platform used for the proposed version of this rule and for the
appropriate and necessary finding. The 2005v4.3 platform builds upon the 2005-based
platform, version 4.2 which was the version of the platform used for the final Cross-State Air
Pollution Rule and incorporated changes made in response to public comments on the
proposed version of that rule. Detailed documentation about the 2005v4.3 platform emissions
inventories used for this rule can be found in the "Emissions Modeling for the Final Mercury
and Air Toxics Standards Technical Support Document".
5A.3 Development of 2005 Base Year Emissions
Emissions inventory inputs representing the year 2005 were developed to provide a
base year for forecasting future air quality. The emission source sectors and the basis for
current and future-year inventories include Electric Generating Utility point sources, non-EGU
point sources, and the following types of sources with inventories primarily at the county level:
onroad mobile, nonroad mobile, nonpoint, and fires. The specific sectors used for modeling are
listed and defined in detail in the "Emissions Modeling for the Final Mercury and Air Toxics
Standards Technical Support Document". The inventories used include emissions of criteria
pollutants, and for some sectors the pollutants benzene, formaldehyde, acetaldehyde and
methanol are used to speciate VOC into the chemical species needed by CMAQ.
The 2005v4 platform was the initial starting point for the 2005v4.3 platform used for
this modeling. There were two intermediate versions: the version used for the MATS proposal
modeling (2005v4.1), and the version used for the final Cross-State Air Pollution Rule modeling
(2005v4.2). The basis of the 2005v4 platform and subsequent versions is the U.S. inventory is
the 2005 National Emission Inventory (NEI), version 2 from October 6, 2008
(http://www.epa.gov/ttn/chief/net/2005inventory.html). The 2005 NEI v2 includes 2005-
specific data for point and mobile sources, while most nonpoint data were carried forward from
version 3 of the 2002 NEI.
Emissions for point sources were primarily from the 2005 NEI v2 inventory, consisting
mostly of 2005 values with some 2002 emissions values used where 2005 data were not
available. The point sources are split into "ECU" (aka "ptipm") and "Non-EGU" (aka
5A-2
-------
"ptnonipm") sectors for modeling purposes based on the matching of the unit level data in the
NEI units in the National Electric Energy Database System (NEEDS) version 4.10 database. All
units that matched NEEDS were included in the ECU sector so that the future year emissions
could easily be taken from the Integrated Planning Model (IPM) as its outputs are also based on
the NEEDS units. Efforts made to ensure the quality of the 2005 ECU inventory included
ensuring that there were not duplicate emissions (e.g., data pulled forward from earlier
inventories), accounting for plants or units that shutdown prior to 2005, adding estimates for
ethanol plants, and accounting for installed emissions control devices.
The 2005 annual NOX and S02 emissions for sources in the ECU sector are based
primarily on data from EPA's Clean Air Markets Division's Continuous Emissions Monitoring
(CEM) program, with other pollutants estimated using emission factors and the CEM annual
heat input. For EGUs without CEMs, emissions were obtained from the state-submitted data in
the NEI. For the 2005 base year, the annual ECU NEI emissions were allocated to hourly
emissions values needed for modeling based on the 2004, 2005, and 2006 CEM data. The NOX
CEM data were used to create N0x-specific profiles, the S02 data were used to create S02-
specific profiles, and the heat input data were used to allocate all other pollutants. The three
years of data were used to create monthly profiles by state, while the 2005 data were used to
create state-averaged profiles for allocating monthly emissions to daily. These daily values were
input into SMOKE, which utilized state-averaged 2005-based hourly profiles to allocate to
hourly values. This approach to temporal allocation was used for all base and control cases
modeled to provide a temporal consistency between the years modeled without tying the
temporalization to the events of a single year.
For nonpoint sources, the 2002 NEI v2 inventory was augmented with updated oil and
gas exploration emissions for Texas and Oklahoma (for CO, NOX, PM, S02, VOC). These oil and
gas exploration emissions were in addition to oil and gas data previously available in the 2005
v4 platform that includes emissions within the following states: Arizona, Colorado, Montana,
Nevada, New Mexico, North Dakota, Oregon, South Dakota, Utah, and Wyoming.
The commercial marine category 3 (C3) vessel emissions portion of the nonroad sector
used point-based gridded 2005 emissions that reflect the final projections developed for the
category 3 commercial marine Emissions Control Area (ECA) proposal to the International
Maritime Organization (EPA-420-F-10-041, August 2010). These emissions include Canada as
part of the ECA, and were updated using region-specific growth rates and thus contain
Canadian province codes. The state/federal water boundaries were based on a file available
5A-3
-------
from the Mineral Management Service (MMS) that specifies boundaries ranging from three to
ten nautical miles from the coast.
The onroad emissions were primarily based on the version of the Motor Vehicle
Emissions Simulator (MOVES) (http://www.epa.gov/otaq/models/moves/) used for the Tier 3
proposed rule. The first step was to run MOVES to output emission factors for a set of counties
with characteristics representative of the counties within the continental United States. Data
for each representative county included county-specific fuels, vehicle age distribution,
inspection and maintenance programs, temperatures and relative humidity. The emission
factors produced by MOVES were then combined by SMOKE with county-based activity data
(vehicle miles traveled, speed data, and vehicle population) and gridded temperature data to
create hourly, gridded emissions. Additional information on this approach is available in the
"Emissions Modeling for the Final Mercury and Air Toxics Standards Technical Support
Document".
The nonroad emissions utilized the National Mobile Inventory Model (NMIM) to run the
NONROAD model for all states to create county/month emissions, updated from the annual
emissions in the 2005 NEI v2 with some improvements. For this case, NMIM was run using
representing county mode, with improved fuels, an improved toxics emission factor (1,3-
butadiene for 2-stroke snowmobiles), and a small coding change that enabled NONROAD to
process 15% ethanol (E15) fuels.
Other emissions inventories used included average-year county-based inventories for
emissions from wildfires and prescribed burning. These emissions are intended to be
representative of both base and future years and are held constant for each. As a result, post-
processing techniques minimize their impact on the modeling results. The 2005v4.3 platform
utilizes the same 2006 Canadian inventory and a 1999 Mexican inventory as were used since
the v4 platform, as these were the latest available data from these countries.
Once developed, the emissions inventories were processed to provide the hourly,
gridded emissions for the model-species needed by CMAQ. Details on this processing are
further described in the "Emissions Modeling for the Final Mercury and Air Toxics Standards
Technical Support Document". Table 5A-1 provides summaries of the 2005 U.S. emissions
inventories modeled for this rule by sector. Tables 5A-2 through 5A-3 provide state-level
summaries of S02, and PM2.s by sector. Note that the nonroad columns include emissions from
traditional nonroad sources that are found "on-land," along with commercial marine sources.
The nonpoint columns include area fugitive dust, agriculture, and other nonpoint emissions.
5A-4
-------
Table 5A-1. 2005 US Emissions by Sector
Emissions Sector
Agriculture
Area fugitive dust
Average fires
Commercial marine
Category 3 (US)
ECU
Locomotive/marine
Non-EGU point
Nonpoint
Nonroad
Onroad
US Total
Table 5A-2. 2005
State
Alabama
Arizona
Arkansas
California
Colorado
2005 NO*
[tons/yr]
2005 SO2
[tons/yr]
2005 PM2.5
[tons/yr]
2005 PM10 2005 NH3
[tons/yr] [tons/yr]
2005 CO
[tons/yr]
2005 VOC
[tons/yr]
3,251,990
189,428
130,164
3,729,161
1,922,723
2,213,471
1,696,902
2,031,527
8,235,002
20,148,378
Base Year
EGU
460,123
52,733
66,384
601
64,174
49,094
97,485
10,380,883
153,068
2,030,759
1,216,362
196,277
168,480
14,292,410
1,030,391
684,035
10,673
496,877
56,666
433,346
1,079,906
201,406
301,073
4,294,373
8,858,992
796,229
11,628
602,236
59,342
647,873
1,349,639
210,767
369,911
12,906,616 3
SO2 Emissions (tons/year) for States
Non-EGU
66,373
23,966
13,039
33,097
1,550
Nonpoint
52,325
2,571
27,260
77,672
6,810
36,777
21,995
773
158,342
133,962
1,971
144,409
,750,218
8,554,551
11,862
603,788
270,007
3,201,418
7,410,946
20,742,873
41,117,658
81,913,104
1,958,992
4,570
41,089
67,690
1,279,308
7,560,061
2,806,422
3,267,931
16,986,064
by Sector
Nonroad Onroad
5,622
6,151
5,678
40,222
4,897
3,554
3,622
1,918
4,526
2,948
Fires
983
2,888
728
6,735
1,719
Total
588,980
91,931
115,008
162,852
82,098
(continued)
5A-5
-------
Table 5A-2. 2005 Base Year SO2 Emissions (tons/year) for States by Sector (continued)
State
Connecticut
Delaware
District of Columbia
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
EGU
10,356
32,378
1,082
417,321
616,063
0
330,382
878,979
130,264
136,520
502,731
109,875
3,887
283,205
84,234
349,877
101,678
75,047
284,384
19,715
74,955
53,363
51,445
57,044
30,628
180,847
512,231
137,371
1,116,095
Non-EGU
1,831
34,859
686
57,429
52,827
17,151
131,357
86,337
41,010
12,926
25,808
165,705
18,512
34,988
19,620
76,510
24,603
29,892
78,308
11,056
7,910
2,253
3,155
7,639
7,831
58,426
59,433
9,582
115,155
Nonpoint
18,455
1,030
1,559
70,490
56,829
2,915
5,395
59,775
19,832
36,381
34,229
2,378
9,969
40,864
25,261
42,066
14,747
6,796
44,573
2,600
7,659
12,477
7,408
10,726
3,193
125,158
22,020
6,455
19,810
Nonroad
2,557
2,657
414
31,190
9,224
2,304
19,305
9,437
8,838
8,035
6,943
25,451
1,625
9,353
6,524
14,626
10,409
5,930
10,464
3,813
9,199
2,880
789
13,321
3,541
15,666
8,766
5,986
15,425
On road
1,337
486
205
12,388
6,939
902
6,881
4,641
2,036
1,978
3,240
2,902
963
3,016
2,669
8,253
2,934
2,590
4,901
874
1,510
656
746
3,038
1,801
6,258
6,287
533
7,336
Fires
4
6
0
7,018
2,010
3,845
20
24
25
103
364
892
150
32
93
91
631
1,051
186
1,422
105
1,346
38
61
3,450
113
696
66
22
Total
34,540
71,416
3,947
595,836
743,893
27,117
493,339
1,039,194
202,004
195,943
573,315
307,202
35,106
371,458
138,402
491,423
155,002
121,306
422,816
39,480
101,337
72,975
63,580
91,830
50,445
386,468
609,433
159,994
1,273,843
(continued)
5A-6
-------
Table 5A-2. 2005 Base Year SO2 Emissions (tons/year) for States by Sector (continued)
State
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
Table 5A-3. 2005
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
EGU
110,081
12,304
1,002,203
176
218,781
12,215
266,148
534,949
3
34,813
9
220,287
3,409
469,456
180,200
89,874
10,380,883
Base Year
EGU
23,366
7,418
1,688
347
4,342
562
2,169
17
Non-EGU
40,482
9,825
83,375
2,743
31,495
1,702
65,693
223,625
1,511
9,132
902
69,401
24,211
46,710
66,807
22,321
2,030,759
Nonpoint
8,556
9,845
68,349
3,365
13,489
10,347
32,714
115,192
0
3,577
5,385
32,923
7,254
14,589
6,369
6,721
1,216,362
Nonroad
5,015
5,697
11,999
816
7,719
3,412
6,288
34,944
0
2,439
385
10,095
18,810
2,133
7,163
2,674
446,831
On road
3,039
1,790
6,266
254
3,589
623
5,538
16,592
0
1,890
342
4,600
3,343
1,378
3,647
721
168,480
Fires
469
4,896
32
1
646
498
277
1,178
0
1,934
49
399
407
215
70
1,106
49,094
Total
167,642
44,357
1,172,224
7,354
275,719
28,797
376,659
926,480
1,515
53,784
7,073
337,705
57,433
534,481
264,256
123,417
14,292,410
PM2.s Emissions (tons/year) for States by Sector
Non-EGU
19,498
3,940
10,820
21,517
7,116
224
1,810
172
Nonpoint
35,555
21,402
34,744
94,200
25,340
11,460
1,590
589
Nonroad
4,142
4,486
3,803
22,815
3,960
1,740
818
277
On road
5,775
6,920
3,102
26,501
4,377
2,544
922
367
Fires
13,938
37,151
10,315
97,302
24,054
56
87
0
Total
102,273
81,316
64,472
262,682
69,189
16,586
7,397
1,421
(continued)
5A-7
-------
Table 5A-3. 2005 Base Year PM2.5 Emissions (tons/year) for States by Sector (continued)
State
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
EGU
24,217
28,057
0
16,585
34,439
8,898
5,549
19,830
5,599
52
15,417
3,110
11,022
3,262
2,029
6,471
2,398
1,246
3,341
2,586
4,625
5,583
9,648
16,967
6,397
53,572
1,411
412
55,547
Non-EGU
25,193
12,666
2,072
15,155
14,124
6,439
7,387
10,453
32,201
3,783
6,768
2,245
12,926
10,538
10,602
6,966
2,729
2,340
4,095
568
2,588
1,460
4,994
12,665
598
12,847
6,246
8,852
16,263
Nonpoint
52,955
63,133
41,492
74,045
74,443
54,312
138,437
31,245
28,164
15,037
23,323
31,116
47,722
73,990
34,217
76,419
30,096
45,661
9,920
13,316
13,623
50,698
48,540
49,551
41,504
52,348
90,047
58,145
44,607
Nonroad
15,035
6,504
2,140
12,880
6,515
6,969
5,719
4,762
9,440
1,363
3,410
3,293
8,561
8,541
4,133
7,230
2,654
5,848
2,212
907
5,042
1,959
8,607
6,272
4,552
9,847
3,765
3,741
7,565
On road
16,241
12,449
1,402
12,574
7,585
3,468
3,109
5,566
4,288
1,759
5,504
5,913
13,006
6,842
4,195
7,665
1,347
2,620
1,290
1,512
5,963
2,861
11,139
8,939
976
11,785
4,559
3,375
11,058
Fires
99,484
24,082
52,808
277
344
349
1,468
5,155
12,647
2,127
531
1,324
1,283
8,943
14,897
2,636
17,311
1,483
19,018
534
865
48,662
1,601
9,870
934
316
6,644
65,350
454
Total
233,125
146,892
99,914
131,516
137,450
80,436
161,669
77,010
92,339
24,120
54,952
47,001
94,520
112,116
70,074
107,388
56,536
59,198
39,876
19,423
32,707
111,224
84,529
104,264
54,962
140,715
112,672
139,874
135,494
(continued)
5A-8
-------
Table 5A-3. 2005 Base Year PM2.5 Emissions (tons/year) for States by Sector (continued)
State
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
EGU
10
14,455
390
12,856
21,464
0
5,055
37
12,357
2,396
26,377
5,233
8,068
496,877
Non-EGU
256
4,779
2,982
21,912
37,563
1,569
3,595
337
11,455
4,618
5,154
7,967
10,298
433,346
Nonpoint
1,289
26,598
33,678
32,563
194,036
0
14,761
6,943
38,140
45,599
14,778
37,277
31,645
2,110,298
Nonroad
394
3,491
2,910
5,072
21,361
0
1,627
479
5,968
6,697
1,702
6,083
1,455
268,745
On road
577
5,061
1,056
8,514
29,859
0
2,703
605
6,661
6,721
1,930
6,783
1,103
301,073
Fires
14
9,163
7,062
3,934
21,578
0
27,412
696
5,659
4,487
3,050
994
15,686
684,035
Total
2,540
63,548
48,079
84,851
325,861
1,569
55,153
9,098
80,241
70,519
52,991
64,337
68,254
4,294,373
5A.4 Development of Future year baseline Emissions
The future year baseline scenario, also known as the "reference case", represents
predicted emissions including adjustments for known promulgated federal measures for all
sectors as of the year 2017, which is assumed to be representative of 2016. The EGU and
mobile sectors reflect projected economic and fuel usage changes. Emissions from non-EGU
stationary sectors have previously been shown to not be well correlated with economic
forecasts, and therefore economic impacts were not included for non-EGU stationary sources.
Like the 2005 base case, these emissions cases include criteria pollutants and for some sectors,
benzene, formaldehyde, acetaldehyde and methanol from the inventory is used in VOC
speciation. The future year baseline scenario represents predicted emissions in the absence of
any further controls beyond those Federal measures already promulgated. For EGUs, all state
and other programs available at the time of modeling have been included. For mobile sources,
all national measures promulgated at the time of modeling have been included. Additional
details on the future year baseline (i.e., reference case) emissions modeling can be found in the
5A-9
-------
"Emissions Modeling for the Final Mercury and Air Toxics Standards Technical Support
Document".
The future year baseline ECU emissions were obtained using version 4.10 Final of the
Integrated Planning Model (IPM) (http://www.epa.gov/airmarkt/progsregs/epa-
ipm/index.html). The IPM is a multiregional, dynamic, deterministic linear programming model
of the U.S. electric power sector. Version 4.10 Final reflects state rules and consent decrees
through December 1, 2010, information obtained from the 2010 Information Collection
Request (ICR), and information from comments received on the IPM-related Notice of Data
Availability (NODA) published on September 1, 2010. Notably, IPM 4.1 Final included the
addition of over 20 GW of existing Activated Carbon Injection (ACI) for coal-fired EGUs reported
to EPA via the ICR. Additional unit-level updates that identified existing pollution controls (such
as scrubbers) were also made based on the ICR and on comments from the IPM NODA. Units
with S02 or NOX advanced controls (e.g., scrubber, SCR) that were not required to run for
compliance with Title IV, New Source Review (NSR), state settlements, or state-specific rules
were modeled by IPM to either operate those controls or not based on economic efficiency
parameters. The IPM run for future year baseline case modeled with CMAQ assumed that 100%
of the HCI found in the coal was emitted into the atmosphere. However, in the final IPM results
for the rule, neutralization of 75% of the available HCI was included based on recent findings.
Further details on the future year baseline ECU emissions inventory used for this rule
can be found in the IPM v.4.10 Documentation, available at
http://www.epa.gov/airmarkets/progsregs/epa-ipm/transport.html. The future year baseline
modeled in IPM for this rule includes estimates of emissions reductions that will result from the
Cross-State Air Pollution Rule. However, reductions from the Boiler MACT rule were not
represented this modeling because the rule was stayed at the time the modeling was
performed. A complete list of state regulations, NSR settlements, and state settlements
included in the IPM modeling is given in Appendices 3-2, 3-3, and 3-4 beginning on p. 68 of
http://www.epa.gov/airmarkets/progsregs/epa-
ipm/CSAPR/docs/DocSuppv410 FTransport.pdf. For the future year baseline ECU emissions,
the IPM outputs for 2020, which are also representative of the year 2017, were used as part of
the 2017 reference case modeling. These emissions were very similar to the year 2015
emissions output from the same IPM modeling case.
Inventories of onroad mobile emissions for the future year baseline and control cases
were created using the MOVES model with an approach consistent with the 2005 base year. As
with the 2005 emissions, the future year onroad emissions were based on emission factors
5A-10
-------
developed using the Tier 3 version of MOVES processed through the SMOKE-MOVES interface.
Future-year vehicle miles travelled (VMT) were projected from the 2005 NEI v2 VMT using
growth rates from the 2009 Annual Energy Outlook (AEO) data. The VMT for heavy duty diesel
vehicles class 8a and 8b was updated based on data from Oak Ridge National Laboratory. The
future year onroad emissions reflect control program implementation through 2017 and
include the Light-Duty Vehicle Tier 2 Rule, the Onroad Heavy-Duty Rule, the Mobile Source Air
Toxics (MSAT) final rule, and the Renewable Fuel Standard version 2 (RFS2).
Future year nonroad mobile emissions were created using NMIM to run NONROAD in a
consistent manner as was done for 2005, but with future-year equipment population estimates,
fuels, and control programs. The fuels in 2017 are assumed to be E10. Emissions for
locomotives and category 1 and 2 (Cl and C2) commercial marine vessels were derived based
on emissions published in the Final Locomotive Marine Rule, Regulatory Impact Assessment,
Chapter 3 (see http://www.epa.gov/otaq/locomotives.htmW2008final). The future year baseline
nonroad mobile emissions reductions include emissions reductions to locomotives, various
nonroad engines including diesel engines and various marine engine types, fuel sulfur content,
and evaporative emissions standards, including the category 3 marine residual and diesel
fuelled engines and International Maritime Organization standards which include the
establishment of emission control areas for these ships. A summary of the onroad and nonroad
mobile source control programs included in the projected future year baseline is shown in Table
5A-4.
Table 5A-4. Summary of Mobile Source Control Programs Included in the Future Year
Baseline
National Onroad Rules:
Tier 2 rule (Signature date: February 28, 2000)
Onroad heavy-duty rule (February 24, 2009)
Final mobile source air toxics rule (MSAT2) (February 9, 2007)
Renewable fuel standard Version 2 (March 26, 2010)
Light duty greenhouse gas standards (May, 2010)
Corporate Average Fuel Economy (CAFE) standards for 2008-2011
Local Onroad Programs:
National low emission vehicle program (NLEV) (March 2,1998)
Ozone transport commission (OTC) LEV Program (January, 1995)
(continued)
5A-11
-------
Table 5A-4. Summary of Mobile Source Control Programs Included in the Future Year
Baseline (continued)
National Nonroad Controls:
Tier 1 nonroad diesel rule (June 17, 2004)
Phase 1 nonroad SI rule (July 3,1995)
Marine SI rule (October 4,1996)
Nonroad large SI and recreational engine rule (November 8, 2002)
Clean Air Nonroad Diesel Rule—Tier 4 (June 29, 2004)
Locomotive and marine rule (May 6, 2008)
Nonroad SI rule (October 8, 2008)
Aircraft:
Itinerant (ITN) operations at airports adjusted to year 2017
Locomotives:
Locomotive Emissions Final Rulemaking (December 17,1997)
Clean Air nonroad diesel final rule—Tier 4 (June 29, 2004)
Locomotive rule (April 16, 2008)
Locomotive and marine rule (May 6, 2008)
Commercial Marine:
Locomotive and marine rule (May 6, 2008)
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)
Tier 1 Marine Diesel Engines (February 28, 2003)
Category 3 marine diesel engines Clean Air Act and International Maritime Organization standards (April, 30,
2010)
For non-EGU point sources, emissions were projected by including emissions reductions
and increases from a variety of source data. Other than for certain large municipal waste
combustors and airports, non-EGU point source emissions were not grown using economic
growth projections, but rather were held constant at the emissions levels in 2005. Emissions
reductions were applied to non-EGU point source to reflect final federal measures, known plant
closures, and consent decrees. The starting inventories for this rule were the projected
5A-12
-------
emission inventories developed for the 2005v4.2 platform for the final Cross-State Air Pollution
Rule (see http://www.epa.gov/ttn/chief/emch/index.htmltffinal). The most significant updates
to the emission projections for this rule are the addition of future year ethanol, biodiesel and
cellulosic plants that account for increased ethanol production from the Renewable Fuel
Standard Rule that is incorporated into the base case for 2017.
Since aircraft at airports were treated as point emissions sources in the 2005 NEI v2, we
developed future year baseline emissions for airports by applying projection factors based on
activity growth projected by the Federal Aviation Administration Terminal Area Forecast (TAF)
system, published in January 2010 for these sources.
Emissions from stationary nonpoint sources were projected using procedures specific to
individual source categories. Refueling emissions were projected using refueling emissions from
MOVES inventory mode runs. Portable fuel container emissions were projected using estimates
from previous rulemaking inventories compiled by the Office of Transportation and Air Quality
(OTAQ). Emissions of ammonia and dust from animal operations were projected based on
animal population data from the Department of Agriculture and EPA. Residential wood
combustion emissions were projected by replacement of obsolete woodstoves with new
woodstoves and a 1 percent annual increase in fireplaces. Landfill emissions were projected
using MACT controls. Other nonpoint sources were held constant between the 2005 and future
year scenarios.
A summary of all rules and growth assumptions impacting non-EGU stationary sources is
provided in Table 5A-5, along with the affected pollutants. Note that reductions associated with
the Boiler MACT are not included in the future year baseline.
Table 5A-5. Control Strategies and/or Growth Assumptions Included in the Future Year
Baseline for Non-EGU Stationary Sources
MACT rules, national, VOC: national applied by SCC, MACT VOC
Consent decrees and settlements, including refinery consent decrees, and settlements All
for: Alcoa, TX and Premcor (formerly MOTIVA), DE
Municipal waste combustor reductions—plant level PM
Hazardous waste combustion PM
Hospital/medical/infectious waste incinerator regulations NOX, PM, SO2
Large municipal waste combustors—growth applied to specific plants All
(continued)
5A-13
-------
Table 5A-5. Control Strategies and/or Growth Assumptions Included in the Future Year
Baseline for Non-EGU Stationary Sources (continued)
MACT rules, plant-level, VOC: auto plants VOC
MACT rules, plant-level, PM & SO2: lime manufacturing PM, SO2
MACT rules, plant-level, PM: taconite ore PM
Municipal waste landfills: projection factor of 0.25 applied All
Livestock emissions growth from year 2002 to year 2017 NH3, PM
Residential wood combustion growth and change-outs from years 2005 to year 2017 All
Gasoline Stage II growth and control via MOVES from year 2005 to year 2017 VOC
Portable fuel container mobile source air toxics rule 2: inventory growth and control VOC
from year 2005 to year 2017
NESHAP: Portland Cement (09/09/10)—plant level based on industrial sector Hg, NOX, SO2, PM, HCI
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 standards VOC, HAP VOC, NOX
Additional plant and unit closures provided by state, regional, and EPA agencies All
Emission reductions resulting from controls put on specific boiler units (not due to NOX, SO2, HCL
MACT) after 2005, identified through analysis of the control data gathered from the
ICRfrom the ICI boiler NESHAP.
NESHAP: Reciprocating Internal Combustion Engines (RICE). NOX, CO, PM, SO2
RICE controls applied to Phase II WRAP 2018 oil and gas emissions VOC, SO2, NOX, CO
RICE controls applied to 2008 Oklahoma and Texas Oil and gas emissions VOC, SO2, NOX, CO, PM
Ethanol plants that account for increased ethanol due to RFS2 All
State fuel sulfur content rules for fuel oil—effective in 2017, only in Maine, New Jersey, SO2
and New York
In all future year cases, the same emissions were used for Canada and Mexico as were
used in the 2005 base case because appropriate future year emissions for sources in these
countries were not available. Future year emissions need to reflect expected percent
reductions or increases between the base year and the future year to be considered
appropriate for this type of modeling and such emissions were not available.
Table 5A-6 shows a summary of the 2005 and modeled future year baseline emissions
for the lower 48 states. Tables 5A-7 and 5A-8 below provide summaries of S02 and PM2.5 in the
5A-14
-------
2017 baseline for each sector by state. Table 5A-9 shows the future year baseline ECU
emissions by state for all criteria air pollutants.
Table 5A-6. Summary of Modeled Base Case Annual Emissions (tons/year) for 48 States by
Sector: SO2 and PM2.5
Source Sector SO2 Emissions
EGU point
Non-EGU point
Nonpoint
Nonroad
On-road
Average fire
Total SO2, all sources
Source Sector PM2.5 Emissions
EGU point
Non-EGU point
Nonpoint
Nonroad
On-road
Average fire
Total PM2.s, all sources
2005
10,380,883
2,030,759
1,216,362
446,831
168,480
49,094
14,292,410
2005
496,877
433,346
2,110,298
268,745
301,073
684,035
4,294,373
2017
3,281,364
1,534,991
1,125,985
15,759
29,288
49,094
6,036,480
2017
276,430
411,437
1,912,757
150,221
129,416
684,035
3,564,296
Table 5A-7. Future Year Baseline SO2 Emissions (tons/year) for States by Sector
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
EGU
186,084
36,996
92,804
5,346
74,255
3,581
Non-EGU
63,053
24,191
12,160
21,046
1,415
1,833
Nonpoint
52,341
2,467
26,801
67,846
6,210
18,149
Nonroad
146
59
123
3,311
50
100
Onroad
569
724
314
2,087
532
311
Fires
983
2,888
728
6,735
1,719
4
Total
303,177
67,324
132,929
106,370
84,181
23,978
(continued)
5A-15
-------
Table 5A-7. Future Year Baseline SO2 Emissions (tons/year) for States by Sector (continued)
State
Delaware
District of Columbia
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
EGU
2,835
5
117,702
96,712
182
118,217
200,969
85,178
45,740
116,927
142,447
2,564
29,786
15,133
163,168
52,380
34,865
178,143
24,018
70,910
14,140
6,719
9,042
10,211
14,653
71,113
105,344
180,935
Non-EGU
4,770
686
49,082
44,248
17,133
81,683
73,930
22,865
10,288
23,530
129,730
14,285
33,562
17,077
48,697
24,742
24,284
33,757
7,212
6,885
2,132
2,471
6,700
7,813
45,222
58,517
9,915
93,600
Nonpoint
1,018
1,505
70,073
55,946
2,894
5,650
59,771
19,929
36,140
33,852
2,669
2,007
40,642
24,907
42,185
14,635
6,635
44,680
1,875
7,899
12,028
7,284
9,528
2,719
71,060
21,713
5,559
19,777
Nonroad
500
3
1,255
192
23
295
150
86
57
215
1,449
72
494
266
448
220
208
191
25
58
27
21
686
26
659
197
36
373
Onroad
91
38
2,111
1,158
162
1,107
760
324
294
463
447
149
593
565
995
558
396
722
106
202
200
137
757
262
1,466
890
71
1,093
Fires
6
0
7,018
2,010
3,845
20
24
25
103
364
892
150
32
93
91
631
1,051
186
1,422
105
1,346
38
61
3,450
113
696
66
22
Total
9,220
2,237
247,241
200,266
24,240
206,971
335,604
128,407
92,622
175,350
277,634
19,226
105,110
58,041
255,584
93,164
67,440
257,679
34,657
86,058
29,873
16,671
26,774
24,480
133,173
153,125
120,991
295,799
(continued)
5A-16
-------
Table 5A-7. Future Year Baseline SO2 Emissions (tons/year) for States by Sector (continued)
State
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
Table 5A-8. Future
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
EGU
141,433
13,211
126,316
0
103,694
29,711
33,080
249,748
0
34,912
264
51,004
5,569
84,344
50,777
48,198
3,281,364
Non-EGU
27,873
9,790
64,697
2,745
28,536
1,655
59,145
129,667
676
6,599
902
50,387
19,780
32,458
61,080
20,491
1,534,991
Year Baseline PM2.5
EGU
13,154
3,889
2,838
475
3,845
400
434
Non-EGU
17,052
3,809
10,527
20,693
7,037
222
772
Nonpoint
7,731
9,508
67,650
3,338
13,310
10,301
32,624
108,633
0
3,365
5,283
32,439
6,885
14,322
6,260
5,944
1,125,985
Nonroad
49
218
427
33
294
23
154
1,146
0
27
8
275
881
64
122
18
15,759
Emissions (tons/year)
Nonpoint
33,235
20,214
33,486
73,607
19,868
6,838
1,207
Nonroad
2,403
2,674
2,042
14,875
2,350
1,038
383
Onroad
501
361
1,060
85
500
86
757
2,483
0
291
129
849
633
178
633
87
29,288
for States
Onroad
2,217
2,762
1,242
13,492
2,387
1,414
375
Fires
469
4,896
32
1
646
498
277
1,178
0
1,934
49
399
407
215
70
1,106
49,094
by Sector
Fires
13,938
37,151
10,315
97,302
24,054
56
87
Total
178,056
37,985
260,183
6,202
146,980
42,273
126,037
492,855
676
47,128
6,634
135,353
34,156
131,582
118,941
75,844
6,036,480
Total
81,999
70,498
60,450
220,443
59,540
9,968
3,259
(continued)
5A-17
-------
Table 5A-8. Future Year Baseline PM2.5 Emissions (tons/year) for States by Sector (continued)
State
District of Columbia
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
EGU
1
12,723
13,445
36
8,587
22,354
4,298
3,199
12,078
3,093
355
3,969
1,465
8,102
2,598
2,201
7,061
3,870
2,358
2,505
1,130
2,452
3,153
2,331
9,983
5,870
18,920
3,530
Non-EGU
172
24,620
12,105
2,076
13,471
13,570
7,000
6,895
10,353
30,865
3,543
6,382
2,123
11,688
9,867
10,492
6,384
2,562
2,834
4,032
464
2,520
1,442
4,859
12,656
795
12,353
5,695
Nonpoint
536
50,472
59,412
40,288
70,775
72,501
51,684
136,633
26,811
27,082
8,213
18,960
23,729
43,055
68,121
31,474
69,722
28,479
44,904
9,351
8,981
8,559
49,789
44,334
43,398
40,802
47,811
88,862
Nonroad
139
8,100
3,803
1,186
6,885
3,491
3,348
2,872
2,717
5,107
881
1,975
1,914
4,696
4,483
2,337
3,954
1,332
2,967
1,319
576
2,929
1,148
5,032
3,583
2,126
5,302
2,029
On road
154
7,652
4,863
714
4,926
3,380
1,519
1,268
2,059
1,673
750
2,492
2,590
4,949
2,882
1,525
3,059
492
919
857
663
3,244
1,103
6,723
3,521
383
5,013
2,006
Fires
0
99,484
24,082
52,808
277
344
349
1,468
5,155
12,647
2,127
531
1,324
1,283
8,943
14,897
2,636
17,311
1,483
19,018
534
865
48,662
1,601
9,870
934
316
6,644
Total
1,002
203,050
117,711
97,108
104,922
115,640
68,198
152,335
59,173
80,467
15,869
34,310
33,145
73,773
96,893
62,926
92,816
54,048
55,465
37,083
12,348
20,569
105,298
64,879
83,011
50,910
89,715
108,767
(continued)
5A-18
-------
Table 5A-8. Future Year Baseline PM2.5 Emissions (tons/year) for States by Sector (continued)
State
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
Table 5A-9. Future
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
EGU
381
16,727
4
9,997
737
5,053
21,677
1
4,524
67
4,529
1,444
13,602
5,323
5,662
276,430
Non-EGU
8,869
14,874
256
4,527
2,399
21,553
34,648
1,568
3,530
336
10,165
4,421
4,281
7,853
10,225
411,437
Nonpoint
39,503
38,523
1,070
23,430
32,697
28,449
187,604
0
13,978
4,930
32,254
35,706
12,951
27,656
30,812
1,912,757
Nonroad
2,148
4,582
222
1,932
1,339
2,939
11,901
0
963
307
3,507
3,328
1,048
3,161
850
150,221
On road
1,627
4,854
383
1,929
416
3,057
9,289
0
1,318
653
3,446
2,874
762
3,148
392
129,416
Fires
65,350
454
14
9,163
7,062
3,934
21,578
0
27,412
696
5,659
4,487
3,050
994
15,686
684,035
Total
117,877
80,014
1,949
50,978
44,650
64,985
286,698
1,569
51,724
6,989
59,561
52,259
35,695
48,135
63,626
3,564,296
Year Baseline EGU CAP Emissions (tons/year) by State
CO
27,024
16,797
9,925
45,388
9,006
9,180
4,256
67
NOX
64,064
36,971
36,297
20,910
50,879
2,738
2,452
11
VOC
1,524
825
658
1,031
636
139
132
2
S02
186,084
36,996
92,804
5,346
74,255
3,581
2,835
5
NH3
1,472
1,163
560
2,519
398
313
119
3
PM10
16,686
5,038
3,507
580
4,605
431
580
1
PM2.5
13,154
3,889
2,838
475
3,845
400
434
1
(continued)
5A-19
-------
Table 5A-9. Future Year Baseline ECU CAP Emissions (tons/year) by State (continued)
State
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
CO
72,915
16,537
1,532
51,862
30,587
8,316
5,066
37,287
32,626
12,789
13,446
7,128
25,856
9,365
9,704
16,499
5,266
4,691
9,677
5,667
25,831
9,079
19,731
17,367
7,437
33,481
26,165
5,905
NOX
83,174
43,778
613
56,128
106,881
42,698
25,163
71,259
33,509
6,121
17,933
7,991
66,846
36,867
27,319
52,464
20,946
28,898
15,627
4,908
11,178
65,189
21,172
44,141
53,778
93,150
47,454
10,828
voc
2,253
1,293
41
3,091
2,295
791
683
1,604
852
306
533
279
1,497
746
440
1,714
338
542
438
206
823
574
731
1,076
867
2,005
957
203
S02
117,702
96,712
182
118,217
200,969
85,178
45,740
116,927
142,447
2,564
29,786
15,133
163,168
52,380
34,865
178,143
24,018
70,910
14,140
6,719
9,042
10,211
14,653
71,113
105,344
180,935
141,433
13,211
NH3
3,997
903
57
1,437
1,317
452
305
928
1,427
269
301
395
874
460
469
740
198
292
953
207
747
570
1,076
654
383
1,317
1,073
381
PM10
19,098
18,668
38
9,926
33,816
5,735
3,996
16,279
3,677
366
5,322
1,915
11,056
3,034
3,113
9,093
6,117
2,948
3,095
1,234
2,948
3,833
3,248
13,368
6,757
25,688
4,457
446
PM2.5
12,723
13,445
36
8,587
22,354
4,298
3,199
12,078
3,093
355
3,969
1,465
8,102
2,598
2,201
7,061
3,870
2,358
2,505
1,130
2,452
3,153
2,331
9,983
5,870
18,920
3,530
381
(continued)
5A-20
-------
Table 5A-9. Future Year Baseline ECU CAP Emissions (tons/year) by State (continued)
State
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
CO
38,767
1,748
10,305
742
10,693
78,317
601
5,632
1,868
30,205
7,183
15,496
19,247
9,087
873,344
NOX
123,501
456
37,516
14,293
16,982
145,182
73
67,476
458
39,408
14,284
54,247
35,179
71,380
1,930,769
VOC
2,023
44
726
129
862
4,975
15
526
52
821
326
1,320
1,137
970
46,050
S02
126,316
0
103,694
29,711
33,080
249,748
0
34,912
264
51,004
5,569
84,344
50,777
48,198
3,281,364
NH3
1,522
136
515
48
406
5,304
47
279
25
1,115
346
658
649
481
40,259
PM10
22,117
7
14,469
764
6,313
31,404
2
5,843
69
5,404
1,706
18,415
6,503
7,385
371,101
PM2.5
16,727
4
9,997
737
5,053
21,677
1
4,524
67
4,529
1,444
13,602
5,323
5,662
276,430
Note: Emission estimates apply to all fossil Electric Generating Units, including those with capacity < 25MW.
5A.5 Development of Future Year Control Case Emissions for Air Quality Modeling
For the future year control case (i.e., policy case) air quality modeling, the emissions for
all sectors were unchanged from the base case modeling except for those from EGUs. The IPM
model was used to prepare the future year policy case for ECU emissions. The air quality
modeling for MATS relied on ECU emission projections from an interim IPM platform based on
the Cross-state Air Pollution Rule version 4.10_FTransport, and was subsequently updated
during the rulemaking process. The updates made include: updated assumptions regarding the
removal of HCI by alkaline fly ash in subbituminous and lignite coals; an update to the fuel-
based mercury emission factor for petroleum coke, which was corrected based on re-
examination of the 1999 ICR data; updated capital cost for new nuclear capacity and nuclear life
extension costs; corrected variable operating and maintenance cost (VOM) for ACI retrofits;
adjusted coal rank availability for some units, consistent with EIA From 923 (2008); updated
state rules in Washington and Colorado; and numerous unit-level revisions based on comments
received through the notice and comment process. In particular, the policy case modeled with
5A-21
-------
CMAQ did not include the neutralization of 75% of HCI as did the final policy case. Additional
details on the version of IPM used to develop the control case are available in Chapter 3.
The changes in ECU S02, and PM2.5 emissions as a result of the policy case for the lower
48 states are summarized in Table 5A-10. Table 5A-11 shows the CAP emissions for the
modeled MATS control case by State. State-specific difference summaries of ECU S02 and PM2.5
for the sum of the lower 48 states are shown in Tables 5A-12 and 5A-13, respectively.
Table 5A-10.Summary of Emissions Changes for the MATS AQ Modeling in the Lower 48
States
Future Year EGU Emissions
Base case EGU emissions (tons)
Control case EGU emissions (tons)
Reductions to base case in control case (tons)
Percentage reduction of base EGU emissions
SO2
3,281,364
1,866,247
1,415,117
43%
PM2.5
276,430
223,320
53,110
19%
Total Man-Made Emissions3
Total base case emissions (tons)
Total control case emissions (tons)
Percentage reduction of all man-made emissions
6,036,480
4,621,363
23%
3,564,296
3,511,186
1%
a In this table, man-made emissions includes average fires.
Table 5A-11.EGU Emissions Totals for the Modeled MATS Control Case in the Lower 48 States
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
CO
20,873
13,238
9,036
56,360
8,219
8,017
1,312
66,378
NOX
61,863
34,804
35,788
27,159
44,409
2,800
2,527
61,676
voc
1,313
749
642
1,307
582
136
67
2,055
SO2
68,517
23,459
35,112
5,041
19,564
1,400
4,160
64,791
NH3
1,235
921
490
2,548
358
313
93
3,482
PM10
9,734
4,264
1,696
1,057
3,492
439
3,056
16,434
PM2.5
7,844
3,494
1,593
942
2,859
412
1,455
11,377
(continued)
5A-22
-------
Table 5A-11. ECU Emissions Totals for the Modeled MATS Control Case in the Lower 48 States
(continued)
State
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
CO
14,217
1,523
24,365
17,061
7,340
4,683
25,911
28,171
10,992
4,283
5,408
18,792
8,699
8,782
12,249
2,223
4,493
7,178
6,781
8,350
7,987
18,725
15,195
7,266
29,956
26,687
6,002
24,865
NOX
41,006
609
50,655
102,045
41,247
22,136
70,126
31,655
5,683
16,554
7,211
60,982
34,942
20,749
52,755
19,758
28,180
14,382
4,862
7,699
64,922
20,863
35,309
53,267
85,565
44,725
9,671
104,906
voc
1,197
41
2,353
1,872
747
623
1,476
767
302
400
226
1,215
709
410
1,605
264
533
336
232
315
545
699
1,033
858
1,852
892
198
1,645
SO2
78,197
182
103,867
156,781
48,030
22,767
125,430
30,509
1,372
18,091
5,033
82,834
33,214
15,975
95,965
6,399
34,631
6,372
2,102
6,404
9,984
28,174
59,551
23,889
139,208
44,602
3,565
93,606
NH3
790
56
1,050
1,110
410
282
882
1,261
267
211
344
718
430
397
686
133
277
725
232
546
554
1,086
602
371
1,229
970
379
1,349
PM10
11,165
38
7,309
29,683
3,318
2,504
12,544
2,003
342
3,851
1,702
8,261
3,332
1,949
5,216
2,637
2,152
2,626
1,336
2,020
2,961
3,123
8,885
5,940
19,599
2,293
241
17,330
PM2.5
9,742
36
6,588
20,388
2,947
2,263
10,635
1,899
331
3,143
1,267
6,893
2,936
1,720
4,809
1,727
1,828
2,073
1,264
1,583
2,750
2,350
7,988
5,051
15,823
2,056
233
14,080
(continued)
5A-23
-------
Table 5A-11. ECU Emissions Totals for the Modeled MATS Control Case in the Lower 48 States
(continued)
State
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
Table 5A-12. State
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
CO NO*
1,721 443
9,826 37,849
641 14,290
5,551 16,931
71,475 138,086
266 32
4,003 65,286
1,868 458
26,778 37,255
6,334 3,834
13,923 47,836
16,124 32,865
7,516 71,135
707,640 1,789,790
Specific Changes in
Future Year
Baseline SO2
(tons)
186,084
36,996
92,804
5,346
74,255
3,581
2,835
5
voc
43
725
117
723
4,444
7
474
52
707
179
1,263
1,012
932
40,875
Annual ECU
SO2
0
40,901
2,483
42,666
105,958
0
17,007
264
33,704
854
66,857
28,322
28,456
1,866,247
SO2 for the
NH3
134
459
41
334
4,774
21
241
25
748
254
632
578
467
35,493
Lower 48
PM10 PM2.5
7 4
9,627 6,963
260 245
6,721 5,272
25,359 17,601
1 1
4,755 3,896
69 67
5,306 4,506
183 176
14,321 11,572
4,725 3,969
5,946 4,671
281,811 223,320
States
Future Year Policy
Case SO2 EGU SO2 Reduction EGU SO2 Reduction
(tons) (tons) (%)
68,517
23,459
35,112
5,041
19,564
1,400
4,160
0
117,568
13,537
57,692
305
54,690
2,181
-1,324
5
63%
37%
62%
6%
74%
61%
-47%
100%
(continued)
5A-24
-------
Table 5A-12. State Specific Changes in Annual ECU SO2 for the Lower 48 States (continued)
State
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Future Year
Baseline SO2
(tons)
117,702
96,712
182
118,217
200,969
85,178
45,740
116,927
142,447
2,564
29,786
15,133
163,168
52,380
34,865
178,143
24,018
70,910
14,140
6,719
9,042
10,211
14,653
71,113
105,344
180,935
141,433
Future Year Policy
Case SO2
(tons)
64,791
78,197
182
103,867
156,781
48,030
22,767
125,430
30,509
1,372
18,091
5,033
82,834
33,214
15,975
95,965
6,399
34,631
6,372
2,102
6,404
9,984
28,174
59,551
23,889
139,208
44,602
EGU SO2 Reduction
(tons)
52,911
18,515
0
14,350
44,189
37,148
22,973
-8,503
111,938
1,191
11,695
10,100
80,334
19,165
18,890
82,177
17,618
36,279
7,768
4,618
2,638
228
-13,521
11,562
81,455
41,727
96,831
EGU SO2 Reduction
(%)
45%
19%
0%
12%
22%
44%
50%
-7%
79%
46%
39%
67%
49%
37%
54%
46%
73%
51%
55%
69%
29%
2%
-92%
16%
77%
23%
68%
(continued)
5A-25
-------
Table 5A-12. State Specific Changes in Annual ECU SO2 for the Lower 48 States (continued)
State
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
Table 5A-13.State
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Future Year
Baseline SO2
(tons)
13,211
126,316
0
103,694
29,711
33,080
249,748
0
34,912
264
51,004
5,569
84,344
50,777
48,198
3,281,364
Specific Changes
Future Year
Baseline PM2.5
(tons)
13,154
3,889
2,838
475
3,845
400
Future Year Policy
Case SO2 EGU SO2 Reduction EGU SO2 Reduction
(tons) (tons) (%)
3,565
93,606
0
40,901
2,483
42,666
105,958
0
17,007
264
33,704
854
66,857
28,322
28,456
1,866,247
in Annual EGU PPVh.sfor
Future Year Policy
Case PM2.5
(tons)
7,844
3,494
1,593
942
2,859
412
9,646
32,710
0
62,793
27,228
-9,586
143,790
0
17,905
0
17,300
4,716
17,488
22,454
19,742
1,415,117
the Lower 48 States
EGU PM2.5
Reduction
(tons)
5,310
395
1,246
-467
985
-12
73%
26%
N/A
61%
92%
-29%
58%
N/A
51%
0%
34%
85%
21%
44%
41%
EGU PM2.5
Reduction
(%)
40%
10%
44%
-98%
26%
-3%
(continued)
5A-26
-------
Table 5A-13. State Specific Changes in Annual ECU PM2.5for the Lower 48 States (continued)
State
Delaware
District of Columbia
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Future Year
Baseline PM2.5
(tons)
434
1
12,723
13,445
36
8,587
22,354
4,298
3,199
12,078
3,093
355
3,969
1,465
8,102
2,598
2,201
7,061
3,870
2,358
2,505
1,130
2,452
3,153
2,331
9,983
5,870
Future Year Policy
Case PM2.5
(tons)
1,455
0
11,377
9,742
36
6,588
20,388
2,947
2,263
10,635
1,899
331
3,143
1,267
6,893
2,936
1,720
4,809
1,727
1,828
2,073
1,264
1,583
2,750
2,350
7,988
5,051
EGU PM2.5
Reduction
(tons)
-1,021
1
1,346
3,703
0
2,000
1,966
1,351
936
1,443
1,193
24
826
198
1,210
-339
481
2,252
2,143
530
432
-134
868
403
-19
1,995
819
EGU PM2.5
Reduction
(%)
-235%
100%
11%
28%
0%
23%
9%
31%
29%
12%
39%
7%
21%
14%
15%
-13%
22%
32%
55%
22%
17%
-12%
35%
13%
-1%
20%
14%
(continued)
5A-27
-------
Table 5A-13. State Specific Changes in Annual ECU PM2.5for the Lower 48 States (continued)
State
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Tribal
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total
Future Year
Baseline PM2.5
(tons)
18,920
3,530
381
16,727
4
9,997
737
5,053
21,677
1
4,524
67
4,529
1,444
13,602
5,323
5,662
276,430
Future Year Policy
Case PM2.5
(tons)
15,823
2,056
233
14,080
4
6,963
245
5,272
17,601
1
3,896
67
4,506
176
11,572
3,969
4,671
223,320
EGU PM2.5
Reduction
(tons)
3,097
1,474
148
2,646
0
3,033
492
-219
4,077
1
627
0
24
1,268
2,031
1,354
991
53,110
EGU PM2.5
Reduction
(%)
16%
42%
39%
16%
2%
30%
67%
-4%
19%
56%
14%
0%
1%
88%
15%
25%
17%
5A-28
-------
APPENDIX 5B
IMPACT OF THE INTERIM POLICY SCENARIO ON AIR QUALITY
5B.1 Air Quality Modeling Platform
This document describes the air quality modeling performed by EPA in support of the
final National Emissions Standard for Hazardous Air Pollutants (NESHAP) related to electrical
generating utilities. A national scale air quality modeling analysis was performed to estimate
the impact of the sector emissions changes on future year annual and 24-hour PM2.s
concentrations, 8-hr maximum ozone, as well as visibility impairment. Air quality benefits are
estimated with the Community Multi-scale Air Quality (CMAQ) model. CMAQ simulates the
numerous physical and chemical processes involved in the formation, transport, and
destruction of ozone, particulate matter and other air pollutants. In addition to the CMAQ
model, the modeling platform includes the emissions, meteorology, and initial and boundary
condition data which are inputs to this model.
Emissions and air quality modeling decisions are made early in the analytical process.
For this reason, it is important to note that the inventories used in the air quality modeling may
be slightly different than the final utility sector inventories presented in the RIA. However, the
air quality inventories and the final rule inventories are generally consistent, so the air quality
modeling adequately reflects the effects of the rule. Similarly, the projected future year
inventory used for this analysis is generally representative of several years around 2017 such as
2016.
Photochemical grid models use state of the science numerical algorithms to estimate
pollutant formation, transport, and deposition over a variety of spatial scales that range from
urban to continental. Emissions of precursor species are injected into the model where they
react to form secondary species such as ozone and then transport around the modeling domain
before ultimately being removed by deposition or chemical reaction. The 2005-based CMAQ
modeling platform was used as the basis for the air quality modeling for this rule. This platform
represents a structured system of connected modeling-related tools and data that provide a
consistent and transparent basis for assessing the air quality response to projected changes in
emissions. The base year of data used to construct this platform includes emissions and
meteorology for 2005. This modeling platform is described in more detail in the modeling
technical support document for this rule (USEPA, 2011).
5B-1
-------
5.B.1.1 Photochemical Model Background
The Community Multi-scale Air Quality (CMAQ) model v4.7.1 (www.cmaq-model.org) is
a state of the science three-dimensional Eularian "one-atmosphere" photochemical transport
model used to estimate air quality (Appel et al., 2008; Appel et al., 2007; Byun and Schere,
2006). CMAQ simulates the formation and fate of photochemical oxidants, ozone, primary and
secondary PM concentrations, and air toxics over regional and urban spatial scales for given
input sets of meteorological conditions and emissions. 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 (Gery et al., 1989).
5.B.I.2 Model Setup, Application, and Post-Processing
The modeling analyses were performed for a domain covering the continental United
States, as shown in Figure 5B-1. This domain has a parent horizontal grid of 36 km with two
finer-scale 12 km grids over portions of the eastern and western U.S. The model extends
vertically from the surface to 100 millibars (approximately 15 km) using a sigma-pressure
coordinate system. Air quality conditions at the outer boundary of the 36 km domain were
taken from a global model and vary in time and space. The 36 km grid was only used to
establish the incoming air quality concentrations along the boundaries of the 12 km grids. Only
the finer grid data were used in determining the impacts of the emissions changes. Table 5B-1
provides geographic information about the photochemical model domains.
Figure 5B-1. Map of the Photochemical Modeling Domains. The black outer box denotes the
36 km national modeling domain; the red inner box is the 12 km western U.S. grid; and the
blue inner box is the 12 km eastern U.S. grid.
5B-2
-------
Table 5B-1. Geographic Elements of Domains Used in Photochemical Modeling
Photochemical Modeling Configuration
National Grid Western U.S. Fine Grid Eastern U.S. Fine Grid
Map Projection Lambert Conformal Projection
Grid Resolution 36 km 12 km 12 km
Coordinate Center 97 deg W, 40 deg N
True Latitudes 33 deg N and 45 deg N
Dimensions 148x112x14 213x192x14 279x240x14
Vertical extent 14 Layers: Surface to 100 millibar level
The 36 km and both 12 km modeling domains were modeled for the entire year of 2005.
Data from the entire year were utilized when looking at the estimation of PM2.5 and visibility
impacts from the regulation. Data from April through October is used to estimate ozone
impacts.
As part of the analysis for this rulemaking, the modeling system was used to calculate
daily and annual PM2.5 concentrations, 8-hr maximum ozone, and visibility impairment. Model
predictions are used to estimate future-year design values of PM2.5 and ozone. Specifically, we
compare a 2017 reference scenario to a 2017 control scenario. This is done by calculating the
simulated air quality ratios between any particular future year simulation and the 2005 base.
These predicted ratios are then applied to ambient base year design values. The design value
projection methodology used here followed EPA guidance for such analyses (USEPA, 2007).
5.B.I.3 Emissions Input Data
The emissions data used in the base year and future reference and future emissions
adjustment case are based on the 2005 v4.1 platform. Emissions are processed to
photochemical model inputs with the SMOKE emissions modeling system (Houyoux et al.,
2000). The 2017 reference case is intended to represent the emissions associated with growth
and controls in that year projected from the 2005 simulation year. The United States ECU point
source emissions estimates for the future year reference and control case are based on an
Integrated Planning Model (IPM) run for criteria pollutants. Both control and growth factors
were applied to a subset of the 2005 non-EGU point and non-point emissions to create the
2017 reference case. The 2005 v4 platform projection factors were the starting point for most
5B-3
-------
of the 2017 SMOKE-based projections. The estimated total anthropogenic emissions and
emissions for the utility sector used in this modeling assessment are shown in Appendix 5A.
Other North American emissions are based on a 2006 Canadian inventory and 1999 Mexican
inventory. Both inventories are not grown or controlled when used as part of future year
inventories. Global emissions of criteria pollutants are included in the modeling system through
boundary condition inflow. More details on these emissions are available in Appendix 5A.
5B.2 Impacts of Sector on Future Annual PM2.5 Levels
This section summarizes the results of our modeling of annual average PM2.5 air quality
impacts in the future due to reductions in emissions from this sector. Specifically, we compare a
2017 baseline scenario to a 2017 control scenario. The modeling assessment indicates a
decrease up to 1.03 u.g/m3 in annual PM2.5 design values is possible given an area's proximity to
controlled sources. The median reduction in annual PM2.5 design value over all monitor
locations is 0.36 u.g/m3. The change in future year projected design value is shown in Figure
5B-2. Negative changes indicate an improvement in air quality.
An annual PM2.5 design value is the concentration that determines whether a
monitoring site meets the annual NAAQS for PM2.5. The full details involved in calculating an
annual PM2.5 design value are given in appendix N of 40 CFR part 50. Projected air quality
benefits are estimated using procedures outlined by United States Environmental Protection
Agency modeling guidance (USEPA, 2007).
5B-4
-------
Annual PM2.5 Difference (ug'mS)
] >-025to<--010
]> -0.10 to <=•• 0.05
| | > -0.05 to •« 0.05
•0.05
Figure 5B-2. Change in Design Values Between the 2017 Baseline and 2017 Control
Simulations. Negative numbers indicate lower (improved) design values in the control case
compared to the baseline
5B.3 Impacts of Sector on Future 24-hour PM2.5 Levels
This section summarizes the results of our modeling of 24-hr average PM2.5 air quality
impacts in the future due to reductions in emissions from this sector. Specifically, we compare a
2017 baseline scenario to a 2017 control scenario. A decrease up to 1.9 u.g/m3 in 24-hr average
PM2.5 design value at monitor locations in the United States is possible given an area's
proximity to controlled sources and the amount of reduced emissions from those sources. A
median decrease of 0.6 u.g/m3 in 24-hr average PM2.5 design value at monitor locations in the
United States is possible given an area's proximity to controlled sources and the amount of
reduced emissions from those sources. The change in future year projected design value is
shown in Figure 5B-3. Negative changes indicate an improvement in air quality.
A 24-hour PM2.5 design value is the concentration that determines whether a
monitoring site meets the 24-hour NAAQS for PM2.5. The full details involved in calculating a 24-
hour PM2.5 design value are given in appendix N of 40 CFR part 50. Projected air quality benefits
are estimated using procedures outlined by United States Environmental Protection Agency
modeling guidance (USEPA, 2007).
5B-5
-------
Daily PM2.5 Difference (ug/m3)
| <= -1.0ugfm3
^B '-1-0 to'-=-0.80
| > -o.so in «-o.50
| | > -0.50 to «-0.25
^ >-0.25 to <=-0.10
| >-0.10 to <=0.10
~~| >0.10
Figure 5B-3. Change in Design Values Between the 2017 Base Case and 2017 Control
Simulations. Negative numbers indicate lower (improved) design values in the control case
compared to the baseline.
5B.4 Impacts of Sector on Future Visibility Levels
Air quality modeling conducted for this rule was used to project visibility conditions in
138 mandatory Class I federal areas across the U.S. in 2017 (USEPA, 2007). The level of visibility
impairment in an area is based on the light-extinction coefficient and a unitless visibility index,
called a "deciview," which is used in the valuation of visibility. The deciview metric provides a
scale for perceived visual changes over the entire range of conditions, from clear to hazy. Under
many scenic conditions, the average person can generally perceive a change of one deciview.
Higher deciview values are indicative of worse visibility. Thus, an improvement in visibility is a
decrease in deciview value.
The modeling assessment indicates a median visibility improvement of 0.09 deciviews in
annual 20% worst visibility days over all Class I area monitors. An improvement in visibility up to
0.97 deciviews on the 20% worst visibility days at Class I monitor locations in the United States
is possible given an area's proximity to controlled sources and the amount of reduced emissions
5B-6
-------
from these sources. The change in future year projected visibility is shown in Figure 5B-4.
Negative changes indicate an improvement in air quality.
Improvement in Deciviews (dv)
• --= -O 7 dv
• > -0.7 to '- -0.6
O > -0 6 to <- -0 5
• > -0.5 to <- -0 3
O >-03to<=-02
O ' -0.2 to <= -0 1
O > -o 1 to <• o 1
Figure 5B-4. Change in 20% Worst Days Between the 2017 Baseline and 2017 Control
Simulations. Negative numbers indicate lower (improved) visibility expressed in deciviews in
the control case compared to the baseline
5B.5 Impacts of Sector on Future Ozone Levels
This section summarizes the results of our modeling of 8-hr maximum ozone air quality
impacts in the future due to reductions in emissions from this sector. Specifically, we compare a
2017 baseline scenario to a 2017 control scenario. The modeling assessment indicates a
decrease of up to 3.5 ppb in 8-hr averaged ozone design value is possible given an area's
proximity to controlled sources and the amount of reduced emissions from these sources. A
median decrease across all monitors of 0.20 ppb in 8-hr averaged ozone design value is possible
given an area's proximity to controlled sources and the amount of reduced emissions from
these sources. The change in future year projected design value is shown in Figure 5B-5.
Negative changes indicate an improvement in air quality. The full details involved in calculating
5B-7
-------
design value are given in appendix P of 40 CFR part 50. Projected air quality benefits are
estimated using procedures outlined by United States Environmental Protection Agency
modeling guidance (USEPA, 2007).
8-hr Ozone Difference (ppb)
H ---3.5 pi*
^B >-3.5tO<=-2.0
IB -: ^!' - - -1 o
^B T Oto<=-0.2
>-02lo<=0.2
> D 2 to <= 0 A
Figure 5B-5. Change in Design Values Between the 2017 Baseline and 2017 Control
Simulations. Negative numbers indicate lower (improved) design values in the control case
compared to the baseline
5B.6 References
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.
5B-8
-------
Byun, D., Schere, K.L, 2006. Review of the governing equations, computational algorithms, and
other components of the models-3 Community Multiscale Air Quality (CMAQ) modeling
system. Applied Mechanics Reviews 59, 51-77.
Carlton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.D., Sarwar, G., Finder, 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.
Gery, M.W., Whitten, G.Z., Killus, J.P., Dodge, M.C., 1989. A Photochemical Kinetics Mechanism
for Urban and Regional Scale Computer Modeling. Journal of Geophysical Research-
Atmospheres 94, 12925-12956.
Houyoux, M.R., Vukovich, J.M., Coats, C.J., Wheeler, N.J.M., Kasibhatla, P.S., 2000. Emission
inventory development and processing for the Seasonal Model for Regional Air Quality
(SMRAQ) project. Journal of Geophysical Research-Atmospheres 105, 9079-9090.
Nenes, A., Pandis, S.N., Pilinis, C., 1998. ISORROPIA: A new thermodynamic equilibrium model
for multiphase multicomponent inorganic aerosols. Aquatic Geochemistry 4, 123-152.
USEPA, 2007. Guidance on the Use of Models and Other Analyses for Demonstrating
Attainment of Air Quality Goals for Ozone, PM2.s, and Regional Haze, RTP.
USEPA, 2011. Air Quality Modeling Technical Support Document: Final ECU NESHAP (EPA-
454/R-11-009), Research Triangle Park, North Carolina.
5B-9
-------
APPENDIX 5C
HEALTH AND WELFARE CO-BENEFITS OF THE MODELED INTERIM POLICY SCENARIO
In this appendix to the co-benefits chapter we report the estimates of the benefits of
reductions in emissions of S02 and directly emitted PM2.5 based on air quality modeling of an
interim policy scenario.
As noted in Chapter 5 of the RIA, the air quality modeling performed for the RIA does
not reflect the emission changes associated with the final rule requirements. To estimate the
benefits of those emissions changes for the final rule, we developed BPT estimates for S02 and
directly emitted PM2.5 based on air quality modeling of an interim policy scenario. These BPT
values were used to adjust benefits estimates for changes in the emission reductions resulting
from the final policy scenario. This appendix reports the results of the benefits analysis
associated with the modeled interim policy scenario described in Appendix 5A and 5B, along
with the derivation of BPT values used to estimate the health benefits of the final policy
scenario.
As described in the benefits chapter, the chief difference between the modeled and
revised scenarios relates to the magnitude and distribution of S02 emission reductions (Figure
5C-1). In general, the modeled and revised policy cases achieve roughly similar levels of S02
reductions (1.42 versus 1.33 million tons, respectively) with a similar distribution among states.
However, for some states (notably Alabama, Colorado, Louisiana, Michigan, Missouri, North
Dakota, Oklahoma, and Texas), S02 emission reductions were lower for the final case versus the
interim case. By far, the greatest difference in S02 emission reductions was in Michigan where
the final case emission reduction was 70% lower than for the interim case. In a few states
(notably Arkansas, Ohio, and South Carolina), S02 emission reductions were slightly larger for
the final case versus the interim case. Since differences between the interim and final cases are
not concentrated in any particular region of the country and the overall distribution of emission
reductions is similar, we conclude that it is reasonable to apply BPT values derived from the
interim case to the final case. While NOX emissions reductions decreased by 70% between the
interim and final cases (141,000 vs. 46,000 tons), the impact of NOX on PM2.5 concentrations
and mortality is very minor relative to the impact of S02 emission reductions. Therefore,
differences in the magnitude and distribution of NOX emission reductions are likely to have only
a minor effect on results.
5C-1
-------
5C.1 PM2.5-Related Health Impacts and Monetized Benefits of Reductions in Emissions of
SO2 and Directly Emitted PM2.5for the Air Quality Modeled Interim Policy Scenario
Health benefits of the interim policy scenario are calculated using the modeled changes
in PM2.5 concentration described in Appendix 5B, which result from the emission changes
described in Appendix 5A. Concentration changes are input into BenMAP to calculate the
changes in incidence of an array of health endpoints, along with their associated monetary
value. BenMAP is described in more detail in Chapter 5. In addition, more information can be
found at http://www.epa.gov/air/benmap/.
Tables 5C-1 and 5C-2 summarize the PM2.5-related health impacts and monetized
benefits of the air quality modeled interim policy scenario. Not all known PM- and ozone-
related health and welfare 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- and
ozone-related health and welfare benefits plus B, the sum of the nonmonetized health and
welfare benefits; this B represents both uncertainty and a bias in this analysis, as it reflects
those benefits categories that we are unable quantify in this analysis. Figure 5C-2 illustrates the
distribution of avoided PM-related deaths by county across the U.S.
Methods for quantifying recreational visibility are described in Section 5.5.1. Visibility
benefits are calculated for the modeled interim policy scenario only since there is no analogous
approach for estimating visibility benefits using the BPT approach. However, the magnitude of
S02 emission reductions did not significantly change in the visibility study areas between the
interim and final emissions scenarios. Therefore, we expect the visibility benefit for the final
policy scenario would be similar to that calculated for the interim policy scenario ($1.1 billion in
total for the U.S., using 2007$). These benefits are not included in the co-benefits estimate of
the final policy.
5C-2
-------
1GO
140
h^
NJ
O
nd
h-
O
O
(tho
00
O
£» O1
O O
V
70
-20
-40
it
Li
ij
n
Interim scenario
Final policy
Figure 5C-1. Comparison of state-level SO2 emission changes between the interim modeled
scenario and the final policy.
5C-3
-------
Table 5C-1. Estimated Reduction in Incidence of Adverse Health Effects of the Interim
Modeled Mercury and Air Toxics Standards in 2016 (95% confidence intervals)3'6
Impact
Premature Mortality
Pope et al. (2002) (age >30)
Laden et al. (2006) (age
>25)
Infant (< 1 year)
Chronic Bronchitis
Non-fatal heart attacks (age >
18)
Hospital admissions-
respiratory (all ages)
Hospital admissions-
cardiovascular (age > 18)
Emergency room visits for
asthma (age < 18)
Acute bronchitis (age 8-12)
Lower respiratory symptoms
(age 7-14)
Upper respiratory symptoms
(asthmatics age 9-18)
Asthma exacerbation
(asthmatics 6- 18)
Eastern U.S.0
4,200
(1,200-7,300)
11,000
(5,000-17,000)
20
(-22-61)
2,800
(94-5,500)
4,800
(1,200-8,400)
840
(340-1,300)
1,800
(1,200-2,200)
3,100
(1,600-4,700)
6,200
(-1,400-14,000)
80,000
(31,000-130,000)
60,000
(11,000-110,000)
130,000
(4,700-450,000)
Western U.S.
120
(19-220)
310
(110-510)
1
(-1-2)
100
(-12-210)
110
(13-220)
16
(5-27)
41
(26-51)
100
(44-160)
240
(-100-570)
3,000
(860-5,200)
2,300
(130-4,400)
5,000
(-520-17,000)
Total
4,400
(1,200-7,500)
11,000
(5,100-17,000)
20
(-23-63)
2,900
(82-5,800)
4,900
(1,200-8,600)
860
(340-1,400)
1,900
(1,300-2,200)
3,200
(1,700-4,800)
6,500
(-1,500-14,000)
83,000
(32,000-130,000)
62,000
(11,000-110,000)
140,000
(4,200-460,000)
Lost work days (ages 18-65)
Minor restricted-activity days
(ages 18-65)
540,000
(460,000-620,000)
3,200,000
(2,600,000-3,800,000)
20,000
(16,000-24,000)
120,000
(93,000-140,000)
560,000
(470,000-640,000)
3,300,000
(2,700,000-3,900,000)
a Estimates rounded to two significant figures; column values will not sum to total value.
The negative estimates for certain endpoints are the result of the weak statistical power of the study used to
calculate these health impacts and do not suggest that increases in air pollution exposure result in decreased
health impacts.
c Includes Texas and those states to the north and east.
5C-4
-------
Table 5C-2. Estimated Economic Value of Health and Welfare Benefits of the Interim
Modeled Mercury and Air Toxics Standards in 2016 (95% confidence intervals,
billions of 2007$)
Impact
Premature mortality (Pope et al.
3% discount rate
7% discount rate
Premature mortality (Laden et al
3% discount rate
7% discount rate
Infant mortality
Chronic bronchitis
Non-fatal heart attacks
3% discount rate
7% discount rate
Hospital admissions-
respiratory
Hospital admissions-
cardiovascular
Emergency room visits for
asthma
Acute bronchitis
Lower respiratory
symptoms
Upper respiratory
symptoms
Asthma exacerbation
Lost work days
Minor restricted-activity
days
Recreational visibility, Class
1 areas
Social cost of carbon (3%
discount rate, 2016 value)
Pollutant
Eastern U.S.3
Western U.S.
Total
2002 PM mortality estimate)
PM2.5
PM2.5
. 2006 PM
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
C02
$34
(S2.7-S100)
$31
($2.4-$93)
mortality estimate)
$87
($7.7-$250)
$79
($6.9-$230)
$0.2
(S-0.2-S0.8)
$1.4
(S0.1-S6.4)
$0.5
(S0.1-S1.3)
$0.4
(S0.1-S1.0)
$0.01
($0.01— $0.02)
$0.03
(<$0.01-$0.05)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$0.1
(So.i-so.1)
$0.2
(S0.1-S0.3)
$0.9
$1.0
(S0.1-S3.1)
$0.9
(S0.1-S2.8)
$2.5
(S0.2-S7.5)
$2.3
(S0.2-S6.7)
<$0.01
$0.05
(<$0.01-$0.23)
$0.01
(<$0.01-$0.03)
$0.01
(<$0.01-$0.03)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$0.2
$35
($2.8— $100)
$32
($2.5-$96)
$90
($7.9-$260)
$81
($7.1-$240)
$0.2
(S-0.2-S0.8)
$1.4
(S0.1-S6.6)
$0.6
(S0.1-S1.3)
$0.4
(So.i-Si.0)
$0.01
($0.01— $0.02)
$0.03
(<$0.01-$0.06)
<$0.01
<$0.01
<$0.01
<$0.01
<$0.01
$0.1
(So.i-so.i)
$0.2
(S0.1-S0.3)
$1.1
(continued)
5C-5
-------
Table 5C-2. Estimated Economic Value of Health and Welfare Benefits of the Interim
Modeled Mercury and Air Toxics Standards in 2016 (95% confidence intervals,
billions of 2007$) (continued)
Impact
Pollutant
Eastern U.S.
Western U.S.
3% discount rate
7% discount rate
3% discount rate
7% discount rate
Total
Total Monetized benefits (Pope et al. 2002 PM2.s mortality estimates)
$37+B $1.2+B $39+B
($3.7-$110) ($0.2-$3.6) ($4.0-$120)
$34+B $1.1+B $35+B
($3.5-$100) ($0.2-$3.3) ($3.7-$110)
Total Monetized benefits (Laden et al. 2006 PM2.s mortality estimates)
$91+B $2.8+B $93+B
($8.7-$260) ($0.4-$8.0) ($9.1-$270)
$82+B $2.5+B $84+B
($7.9-$240) ($0.4-$7.2) ($8.3-$240)
Includes Texas and those states to the north and east.
PM-related premature deaths avoided
Pope et al. (2002) mortality estimate
Figure 5C-2. Estimated Reduction in Excess PM2.5-Related Premature Deaths Estimated to
Occur in Each County in 2016 as a Result of the Interim Modeled Mercury and Air Toxics
Standards
5C-6
-------
5C.2 Derivation of the BPT Values Used to Calculate the Health Benefits of the Final Policy
Scenario
The health benefits summarized in Tables C-l and C-2 include impacts of changes in
sulfate, nitrate, and direct PM2.5. To quantify the health benefits of the final policy scenario
reported in Chapter 5, we calculate the health benefits per ton of emission reduced, separately
for eastern and western states, and separately for S02, directly emitted carbonaceous PM2.5,
and directly emitted crustal PM2.5. This calculation is shown by Equation 1:
where BPT is the BPT for a particular pollutant i (S02, directly emitted carbonaceous PM2.5, or
directly emitted crustal PM2.5), region j (Eastern U.S. or Western U.S.), health endpoint k (e.g.
adult mortality, infant mortality, etc.), and the interim baseline and policy scenario (denoted as
"1"). As described in Chapter 5, we do not generate BPT values for NOX. Nitrate increases in the
modeled policy scenario were two orders of magnitude smaller than the sulfate decreases and
were not included in the BPT estimates. Including nitrate in the S02 BPT estimate would reduce
the S02 BPT by 1-2%, with a corresponding impact on the total health benefits of the rule.
Furthermore, as described in Appendix 5A, NOX emission changes resulting from this rule were
75% smaller for the final policy scenario relative to the interim modeled policy scenario.
Therefore, excluding the impacts of NOX emission changes is unlikely to materially impact the
final benefit results.
Table 5C-3 reports the economic value of the adult mortality benefits resulting from
reductions in S02, directly emitted carbonaceous PM2.5, and directly emitted crustal PM2.5for
the modeled interim policy scenario, along with the BPT values derived from these benefits.
Only adult mortality benefits are shown here as they contribute 93-97% of the total health
benefits, however, BPT values were calculated and applied separately for each health endpoint.
Since premature mortality is discounted after BPT values are applied to the final emissions, the
values reported in Table 5C-3 are not discounted. Sulfate reductions resulting from S02
emission reductions contribute approximately 95% of the benefits of S02 and directly emitted
PM2.5 combined. In some locations, directly emitted carbonaceous PM2.5 increased slightly in
the Western U.S. for the interim policy scenario relative to the interim baseline, which overall
resulted in negative BPT values for the West. However, since the magnitudes of the emission
and concentration changes are small relative to the changes in S02 emissions and sulfate
concentrations, the resulting increase in premature mortality is only 0.04% of the total health
impact of the rule.
5C-7
-------
Table 5C-3. Estimated Economic Value of Adult Mortality Benefits by Pollutant, in Total and
Per Ton of Emissions Reduced Interim Modeled Mercury and Air Toxics Standard
in 2016 (95% confidence intervals, 2007$)
Pollutant and Source of Adult
Mortality Estimate
SO2 emissions (tons)b
Pope et al. (2002) estimate
Laden et al. (2006) estimate
Carbonaceous PM2.5 emissions (tons) b
Pope et al. (2002) estimate0
Laden et al. (2006) estimate0
Crustal PM2.s emissions (tons) b
Pope et al. (2002) estimate
Laden et al. (2006) estimate
Includes Texas and those states to the
Total Monetized Benefits
(billions)
Eastern U.S.3
1,268,961
$36
($2.9— $110)
$93
($8.2-$270)
5,860
$1.3
(S0.1-S3.9)
$3.3
(S0.3-S9.6)
34,742
$0.6
(<$0.01-$1.9)
$1.6
(S0.1-S4.7)
north and east.
Western U.S.
146,155
$1.2
(S0.1-S3.7)
$3.1
(S0.3-S9.0)
231
<-$0.01
<-$0.01
29,148
$0.1
(<$0.01— $0.2)
$0.1
(<$0.01— $0.4)
BPT (thousands)
Eastern U.S.3
$29
($2.3-$87)
$73
(S6.4-S210)
$220
($17-$670)
$560
($49-$l,600)
$18
(S1.4-S55)
$47
($4.1-$ 140)
Western U.S.
$8.3
(S0.1-S25)
$21
(S1.9-S62)
-$66
(-$450-$210)
-$170
(-$960-$350)
$9.6
(S0.1-S31)
$25
(S2.1-S74)
Emission reductions are reported for the modeled interim policy case, from which the BPT values were
generated.
c Directly emitted carbonaceous PM2.5 increased slightly in some locations in the Western U.S. for the interim
policy scenario relative to the interim baseline, which overall resulted in negative BPT values for the West.
However, since the magnitudes of the emission and concentration changes are small relative to the changes in
SO2 emissions and sulfate concentrations, the resulting increase in premature mortality is only 0.04% of the total
health impact of the rule.
The BPT values reported in Table 5C-3, along with those calculated for the other health
endpoints listed in Table 5C-2, are applied to the final emission changes described in Chapter 3,
resulting in the final benefit values summarized in Chapter 5. This calculation is shown by
Equation 2:
TotalBenefitSijiki =
where 2 refers to the final baseline and policy scenarios.
x BPT
tjik
5C-8
-------
APPENDIX 5D
PM2.5 CO-BENEFITS OF THE FINAL RULE BY STATE
5D.1 Introduction
This appendix describes the distribution of the health-related PM2.5 co-benefits
associated with this rule by state. We describe our approach for allocating the national-level
PM2.5-related mortality and monetized benefits to the state-level. We also summarize the
results of this analysis and describe the limitations and uncertainties associated with our
approach. This rule is expected to achieve PM2.5-related health benefits in all states, resulting
from both emission reductions in that state and reduced transport of PM2.5 between states. A
key limitation of our approach is that it does not account for differences in the distribution of
S02 and direct PM2.5 emission reductions between the modeled interim scenario and the final
policy (see Appendix 5C). PM2.5-related co-benefits may therefore be under- or over-estimated
for certain states.
5D.2 Methods
As described in Appendix 5C, the PM2.5 health co-benefits of the final rule are calculated
using a BPT approach. The BPT values are derived from air quality modeling of an interim
emissions scenario. Since the distribution of the S02 emission reductions in the interim
modeled scenario and the final policy were generally consistent, applying BPT values from the
interim modeled scenario to the final policy reasonably approximates the total monetized
benefits of the final policy. However, this approach requires aggregation of benefits in the
interim scenario to larger spatial scales to account for transport of pollution across state
boundaries. Therefore, the final rule benefits described in Appendix 5C are estimated for the
eastern and western US, the same resolution at which BPT values were generated from the
interim scenario.
Since spatially resolved estimates of the co-benefits are useful for understanding how
the expected benefits of this rule are distributed across the U.S., we developed an
approximating approach for allocating national-level PM2.5 co-benefits estimated for the final
policy to the state level. This approach follows three steps. First, we quantified the state-level
mortality and monetized health co-benefits of the air quality modeled scenario using the
BenMAP software. From these results, we calculated the percentage of national health
benefits occurring in each state. Finally, these percentages were used to scale the national
health benefits of the final policy down to the state level.
5D-1
-------
As another approach, EPA considered scaling the health co-benefits of the final policy
scenario by the percentage of the national total S02 emission change occurring in each state,
since the distribution of emission changes across the U.S. changed between the modeled
interim scenario and the final policy. However, such an approach would not account for the
population in each state which is a main driver for air pollution health impacts, nor would it
account for transport of pollution across state lines. Therefore, EPA judged that scaling
national health co-benefits of the final rule by the state distribution of the co-benefits of the
interim modeled scenario is a more appropriate approach.
5D.3 Limitations and uncertainties
The method described above adds unique uncertainties and limitations beyond those
already described in detail in Chapter 5. A key limitation of this approach is that the
distribution of S02 and direct PM2.5 emissions changed between the modeled interim scenario
and the final policy (see Appendix 5C). Differences in the emission changes would have an
effect on the percentage of health co-benefits occurring in each state for the final policy.
However, our approach necessarily assumes that the state distribution of health co-benefits for
the final policy is equivalent to that of the modeled interim scenario. PM2.5-related health co-
benefits for this rule could therefore be under- or over-estimated for certain states.
5D.4 Results
The reduction in incidence of adult premature PM2.5-related mortality for the final rule
by state is shown in Table 5D-1. Additional non-mortality heath benefits are also expected in
each state but are not included here. The greatest percentage of interim mortality benefits in
any one state is in Texas (10.8%), followed by Florida (6.7%). For the final policy, 460 to 1,200
avoided premature deaths are estimated to be avoided in Texas, and 280 to 750 in Florida,
depending on the concentration-response factor. Although S02 emissions in some states (e.g.
Kentucky, New York, Tennessee) increase between the interim baseline and interim policy
scenario (see Appendix 5A), mortality decreases in these states due to reduced transport of
pollution from other states. All states, therefore, experience health benefits from the interim
scenario and the final rule. Table 5D-2 shows the estimated economic value of health and
welfare benefits by state for the final rule. Approximately $4.0 to $9.7 billion (2007$, 3%
discount rate) in benefits are expected to occur in Texas, and $2.4 to $6.0 billion in Florida,
depending on the concentration-response function used for adult mortality.
5D-2
-------
Table 5D-1. Estimated Reduction in Incidence of Premature Adult Mortality for the Mercury
and Air Toxics Standards in 2016 by Statea'b
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
DC
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
Percent of total
interim benefits
3.34
0.32
2.28
0.13
1.26
0.84
0.30
0.14
6.73
4.53
0.06
5.31
2.64
1.45
1.43
1.97
2.68
0.19
1.99
1.23
3.78
1.34
2.21
3.79
0.07
0.67
0.09
0.23
2.96
0.22
4.11
4.42
0.17
5.19
2.82
0.11
4.91
0.27
3.01
0.25
3.38
10.82
0.20
0.09
2.76
0.28
Final policy benefits
Pope et al. (2002) estimate
140
14
96
6
53
35
13
6
280
190
3
220
110
61
60
83
110
8
84
52
160
57
93
160
3
28
4
10
130
9
170
190
7
220
120
5
210
11
130
11
140
460
8
4
120
12
-adult mortality
Laden et al. (2006) estimate
360
35
250
14
140
90
32
15
730
490
6
570
290
160
160
210
290
20
220
130
410
150
240
410
8
72
10
25
320
24
440
480
19
560
300
12
530
29
330
27
370
1200
22
10
300
31
5D-3
-------
West Virginia 0.89 38 96
Wisconsin 2.06 87 220
Wyoming 0.05 2 6
National Total 4,200 11,000
State level benefits of the final rule are scaled by the distribution of mortality benefits simulated from the
interim scenario described in Appendices 5A, 5B, and 5C.
Estimates rounded to two significant figures; column values will not sum to total value. These estimates do not
include confidence intervals.
5D-4
-------
Table 5D-2. Estimated Economic Value of Health Benefits of the Mercury and Air Toxics
Standard in 2016 by State (billions of 2007$, 3% discount rate)a'b
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
DC
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
Percent of total
interim benefits
3.32
0.32
2.27
0.13
1.28
0.84
0.30
0.14
6.68
4.56
0.06
5.34
2.64
1.45
1.43
1.97
2.67
0.19
1.99
1.23
3.78
1.35
2.20
3.78
0.07
0.67
0.09
0.23
2.96
0.22
4.12
4.41
0.17
5.17
2.81
0.11
4.87
0.27
2.99
0.25
3.38
10.95
0.20
0.09
2.77
0.28
Health benefits (billions of 2007$, 3%
Pope et al. (2002) estimate Laden et
$1.20
$0.12
$0.82
$0.05
$0.46
$0.30
$0.11
$0.05
$2.40
$1.70
$0.02
$1.90
$0.96
$0.52
$0.52
$0.71
$0.97
$0.07
$0.72
$0.45
$1.40
$0.49
$0.80
$1.40
$0.03
$0.24
$0.03
$0.08
$1.10
$0.08
$1.50
$1.60
$0.06
$1.90
$1.00
$0.04
$1.80
$0.10
$1.10
$0.09
$1.20
$4.00
$0.07
$0.03
$1.00
$0.10
discount rate)
al. (2006) estimate
$3.00
$0.29
$2.00
$0.12
$1.10
$0.75
$0.27
$0.12
$6.00
$4.10
$0.05
$4.70
$2.40
$1.30
$1.30
$1.80
$2.40
$0.17
$1.80
$1.10
$3.40
$1.20
$2.00
$3.40
$0.06
$0.60
$0.08
$0.21
$2.60
$0.20
$3.70
$3.90
$0.15
$4.60
$2.50
$0.10
$4.40
$0.24
$2.70
$0.23
$3.00
$9.70
$0.18
$0.08
$2.50
$0.25
5D-5
-------
West Virginia 0.89 $0.32 $0.79
Wisconsin 2.06 $0.75 $1.80
Wyoming 0.05 $0.02 $0.05
National Total0 $36.00 $89.00
State level benefits of the final rule are scaled by the distribution of mortality benefits simulated from the
interim scenario described in Appendices 5A, 5B, and 5C.
Estimates rounded to two significant figures; column values will not sum to total value. These estimates do not
include confidence intervals.
While climate benefits are included in the total co-benefits of this rule as described in Chapter 5, only health
benefits (sum of mortality and morbidity endpoints) are included in the national total here.
5D-6
-------
APPENDIX 5E
TECHNICAL SUPPORT DOCUMENT:
SUMMARY OF EXPERT OPINIONS ON THE EXISTENCE OF A THRESHOLD IN THE
CONCENTRATION-RESPONSE FUNCTION FOR PM2.5-RELATED MORTALITY28
28 U.S. Environmental Protection Agency. 2010. Technical Support Document: Summary of Expert Opinions on the
Existence of a Threshold in the Concentration-Response Function for PM2.5-related Mortality. Research Triangle
Park, NC. June. Available on the Internet at: .
5E-1
-------
Summary of Expert Opinions on the Existence of a Threshold in the
Concentration-Response Function for PM2.s-related Mortality
Technical Support Document (TSD)
June 2010
Compiled by:
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impact Division
Air Benefit-Cost Group
Research Triangle Park, North Carolina
Contents:
A. HES comments on 812 Analysis (2010)
B. American Heart Association Scientific Statement (2010)
C. Integrated Science Assessment for Paniculate Matter (2009)
D. CAS AC comments on PM ISA and REA (2009)
E. Krewski et al. (2009)
F. Schwartz et al. (2008)
G. Expert Elicitation on PM Mortality (2006, 2008)
H. CAS AC comments on PM Staff Paper (2005)
I. HES comments on 812 Analysis (2004)
J. NRC (2002)
5E-2
-------
A. HES Comments on 812 Analysis (2010)
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
.
Pg 2: "The HES generally agrees with other decisions made by the EPA project team with
respect to PM, in particular, the PM mortality effect threshold model, the cessation lag model,
the inclusion of infant mortality estimation, and differential toxicity of PM."
Pg 2: "Further, the HES fully supports EPA's use of a no-threshold model to estimate the
mortality reductions associated with reduced PM exposure."
Pg 6: "The HES also supports the Agency's choice of a no-threshold model for PM-related
effects."
Pg 13: "The HES fully supports EPA's decision to use a no-threshold model to estimate mortality
reductions. This 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. Therefore, there is
no evidence to support a truncation of the CRF."
HES Panel Members
Dr. John Bailar, Chair of the Health Effects Subcommittee, Scholar in Residence, The National
Academies, Washington, DC
Dr. Michelle Bell, Associate Professor, School of Forestry and Environmental Studies, Yale
University, New Haven, CT
Dr. James K. Hammitt, Professor, Department of Health Policy and Management, Harvard
School of Public Health, Boston, MA
Dr. Jonathan Levy, Associate Professor, Department of Environmental Health, Harvard School
of Public Health, Boston, MA
Dr. C. Arden Pope, III Professor, Department of Economics, Brigham Young University,
Provo, UT
Mr. John Fintan Hurley, Research Director, Institute of Occupational Medicine (IOM),
Edinburgh, United Kingdom, UK
Dr. Patrick Kinney, Professor, Department of Environmental Health Sciences, Mailman School
of Public Health, Columbia University, New York, NY
Dr. Michael T. Kleinman, Professor, Department of Medicine, Division of Occupational and
Environmental Medicine, University of California, Irvine, Irvine, C A
Dr. Bart Ostro, Chief, Air Pollution Epidemiology Unit, Office of Environmental Health
Hazard Assessment, California Environmental Protection Agency, Oakland, CA
5E-3
-------
Dr. Rebecca Parkin, Professor and Associate Dean, Environmental and Occupational Health,
School of Public Health and Health Services, The George Washington University Medical
Center, Washington, DC
5E-4
-------
B. Scientific Statement from American Heart Association (2010)
Brook RD, Rajagopalan S, Pope CA 3rd, Brook JR, Bhatnagar A, Diez-Roux AV, Holguin
F, Hong Y, Luepker RV, Mittleman MA, Peters A, Siscovick D, Smith SC Jr, Whitsel L,
Kaufman JD; on behalf of the American Heart Association Council on Epidemiology and
Prevention, Council on the Kidney in Cardiovascular Disease, and Council on Nutrition,
Physical Activity and Metabolism. (2010). "Particulate matter air pollution and
cardiovascular disease: an update to the scientific statement from the American Heart
Association." Circulation. 121: 2331-2378.
Pg 2338: "Finally, there appeared to be no lower-limit threshold below which PMio was not
associated with excess mortality across all regions."
Pg 2350: "There also appears to be a monotonic (eg, linear or log-linear) concentration-response
relationship between PM2.s and mortality risk observed in cohort studies that extends below
present-day regulations of 15 |ig/m3 for mean annual levels, without a discernable "safe"
threshold." (cites Pope 2004, Krewski 2009, and Schwartz 2008)
Pg 2364: "The PM2.5 concentration- cardiovascular risk relationships for both short- and long-
term exposures appear to be monotonic, extending below 15 |ig/m3 (the 2006 annual NAAQS
level) without a discernable "safe" threshold."
Pg 2365: "This updated review by the AHA writing group corroborates and strengthens the
conclusions of the initial scientific statement. In this context, we agree with the concept and
continue to support measures based on scientific evidence, such as the US EPA NAAQS, that
seek to control PM levels to protect the public health. Because the evidence reviewed supports
that there is no safe threshold, it appears that public health benefits would accrue from lowering
PM2.5 concentrations even below present-day annual (15 |ig/m3) and 24-hour (35 |ig/m3)
NAAQS, if feasible, to optimally protect the most susceptible populations."
Pg 2366: "Although numerous insights have greatly enhanced our understanding of the PM-
cardiovascular relationship since the first AHA statement was published, the following list
represents broad strategic avenues for future investigation: ... Determine whether any "safe" PM
threshold concentration exists that eliminates both acute and chronic cardiovascular effects in
healthy and susceptible individuals and at a population level."
Scientific Statement Authors
Dr. Robert D. Brook, MD
Dr. Sanjay Rajagopalan, MD
Dr. C. Arden Pope, PhD
Dr. Jeffrey R. Brook, PhD
Dr. Aruni Bhatnagar, PhD, FAHA
Dr. Ana V. Diez-Roux, MD, PhD, MPH
Dr. Fernando Holguin, MD
5E-5
-------
Dr. Yuling Hong, MD, PhD, FAHA
Dr. Russell V. Luepker, MD, MS, FAHA
Dr. Murray A. Mittleman, MD, DrPH, FAHA
Dr. Annette Peters, PhD
Dr. David Siscovick, MD, MPH, FAHA
Dr. Sidney C. Smith, Jr, MD, FAHA
Dr. Laurie Whitsel, PhD
Dr. Joel D. Kaufman, MD, MPH
5E-6
-------
C. Integrated Science Assessment for Particulate Matter (2009)
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
.
Pg 1-22: "An important consideration in characterizing the public health impacts associated with
exposure to a pollutant is whether the concentration-response relationship is linear across the full
concentration range encountered, or if nonlinear relationships exist along any part of this range.
Of particular interest is the shape of the concentration-response curve at and below the level of
the current standards. The shape of the concentration-response curve varies, depending on the
type of health outcome, underlying biological mechanisms and dose. At the human population
level, however, various sources of variability and uncertainty tend to smooth and "linearize" the
concentration-response function (such as the low data density in the lower concentration range,
possible influence of measurement error, and individual differences in susceptibility to air
pollution health effects). In addition, many chemicals and agents may act by perturbing naturally
occurring background processes that lead to disease, which also linearizes population
concentration-response relationships (Clewell and Crump, 2005, 156359; Crump et al., 1976,
003192; Hoel, 1980, 156555). These attributes of population dose-response may explain why the
available human data at ambient concentrations for some environmental pollutants (e.g., PM, Os,
lead [Pb], ETS, radiation) do not exhibit evident thresholds for health effects, even though likely
mechanisms include nonlinear processes for some key events. These attributes of human
population dose-response relationships have been extensively discussed in the broader
epidemiologic literature (Rothman and Greenland, 1998, 086599)."
Pg 2-16: "In addition, cardiovascular hospital admission and mortality studies that examined the
PMio concentration-response relationship found evidence of a log-linear no-threshold
relationship between PM exposure and cardiovascular-related morbidity (Section 6.2) and
mortality (Section 6.5)."
Pg 2-25: "2.4.3. PM Concentration-Response Relationship
An important consideration in characterizing the PM-morbidity and mortality association is
whether the concentration-response relationship is linear across the full concentration range that
is encountered or if there are concentration ranges where there are departures from linearity (i.e.,
nonlinearity). In this ISA studies have been identified that attempt to characterize the shape of
the concentration-response curve along with possible PM "thresholds" (i.e., levels which PM
concentrations must exceed in order to elicit a health response). The epidemiologic studies
evaluated that examined the shape of the concentration-response curve and the potential presence
of a threshold have focused on cardiovascular hospital admissions and ED visits and mortality
associated with short-term exposure to PMio and mortality associated with long-term exposure to
PM2.5.
"A limited number of studies have been identified that examined the shape of the PM
cardiovascular hospital admission and ED visit concentration-response relationship. Of these
studies, some conducted an exploratory analysis during model selection to determine if a linear
5E-7
-------
curve most adequately represented the concentration-response relationship; whereas, only one
study conducted an extensive analysis to examine the shape of the concentration-response curve
at different concentrations (Section 6.2.10.10). Overall, the limited evidence from the studies
evaluated supports the use of a no-threshold, log-linear model, which is consistent with the
observations made in studies that examined the PM-mortality relationship.
"Although multiple studies have previously examined the PM-mortality concentration-response
relationship and whether a threshold exists, more complex statistical analyses continue to be
developed to analyze this association. Using a variety of methods and models, most of the
studies evaluated support the use of a no-threshold, log-linear model; however, one study did
observe heterogeneity in the shape of the concentration-response curve across cities (Section
6.5). Overall, the studies evaluated further support the use of a no-threshold log-linear model, but
additional issues such as the influence of heterogeneity in estimates between cities, and the effect
of seasonal and regional differences in PM on the concentration-response relationship still
require further investigation.
"In addition to examining the concentration-response relationship between short-term exposure
to PM and mortality, Schwartz et al. (2008, 156963) conducted an analysis of the shape of the
concentration-response relationship associated with long-term exposure to PM. Using a variety
of statistical methods, the concentration-response curve was found to be indistinguishable from
linear, and, therefore, little evidence was observed to suggest that a threshold exists in the
association between long-term exposure to PM2.s and the risk of death (Section 7.6)."
Pg 6-75: "6.2.10.10. Concentration Response
The concentration-response relationship has been extensively analyzed primarily through studies
that examined the relationship between PM and mortality. These studies, which have focused on
short- and long-term exposures to PM have consistently found no evidence for deviations from
linearity or a safe threshold (Daniels et al., 2004, 087343; Samoli et al., 2005, 087436; Schwartz,
2004, 078998; Schwartz et al., 2008, 156963) (Sections 6.5.2.7 and 7.1.4). Although on a more
limited basis, studies that have examined PM effects on cardiovascular hospital admissions and
ED visits have also analyzed the PM concentration-response relationship, and contributed to the
overall body of evidence which suggests a log-linear, no-threshold PM concentration-response
relationship.
"The results from the three multicity studies discussed above support no-threshold log-linear
models, but issues such as the possible influence of exposure error and heterogeneity of shapes
across cities remain to be resolved. Also, given the pattern of seasonal and regional differences
in PM risk estimates depicted in recent multicity study results (e.g., Peng et al., 2005, 087463),
the very concept of a concentration-response relationship estimated across cities and for all-year
data may not be very informative."
Pg 6-197: "6.5.2.7'. Investigation of Concentration-Response Relationship
The results from large multicity studies reviewed in the 2004 PM AQCD (U.S. EPA, 2004,
056905) suggested that strong evidence did not exist for a clear threshold for PM mortality
effects. However, as discussed in the 2004 PM AQCD (U.S. EPA, 2004, 056905), there are
several challenges in determining and interpreting the shape of PM-mortality concentration-
5E-8
-------
response functions and the presence of a threshold, including: (1) limited range of available
concentration levels (i.e., sparse data at the low and high end); (2) heterogeneity of susceptible
populations; and (3) investigate the PM-mortality concentration-response relationship.
"Daniels et al. (2004, 087343) evaluated three concentration-response models: (1) log-linear
models (i.e., the most commonly used approach, from which the majority of risk estimates are
derived); (2) spline models that allow data to fit possibly non-linear relationship; and (3)
threshold models, using PMio data in 20 cities from the 1987-1994 NMMAPS data. They
reported that the spline model, combined across the cities, showed a linear relation without
indicating a threshold for the relative risks of death for all-causes and for cardiovascular-
respiratory causes in relation to PMio, but "the other cause" deaths (i.e., all cause minus
cardiovascular-respiratory) showed an apparent threshold at around 50 (J,g/m3 PM10, as shown in
Figure 6-35. For all-cause and cardio-respiratory deaths, based on the Akaike's Information
Criterion (AIC), a log-linear model without threshold was preferred to the threshold model and
to the spline model.
"The FIEI review committee commented that interpretation of these results required caution,
because (1) the measurement error could obscure any threshold; (2) the city-specific
concentration-response curves exhibited a variety of shapes; and (3) the use of AIC to choose
among the models might not be appropriate due to the fact it was not designed to assess scientific
theories of etiology. Note, however, that there has been no etiologically credible reason
suggested thus far to choose one model over others for aggregate outcomes. Thus, at least
statistically, the result of Daniels et al. (2004, 087343) suggests that the log-linear model is
appropriate in describing the relationship between PMIO and mortality.
"The Schwartz (2004, 078998) analysis of PMio and mortality in 14 U.S. cities, described in
Section 6.5.2.1, also examined the shape of the concentration-response relationship by including
indicator variables for days when concentrations were between 15 and 25 ug/m3, between 25 and
34 ug/m3, between 35 and 44 ug/m3, and 45 ug/m3 and above. In the model, days with
concentrations below 15 ug/m3 served as the reference level. This model was fit using the single
stage method, combining strata across all cities in the case-crossover design. Figure 6-36 shows
the resulting relationship, which does not provide sufficient evidence to suggest that a threshold
exists. The authors did not examine city-to-city variation in the concentration-response
relationship in this study.
and mortality in 22 European cities (and BS in 15 of the cities) participating in the
APHEA project. In nine of the 22 cities, PMIO levels were estimated using a regression model
relating co-located PMIO to BS or TSP. They used regression spline models with two knots (30
and 50 ug/m3) and then combined the individual city estimates of the splines across cities. The
investigators concluded that the association between PM and mortality in these cities could be
adequately estimated using the log-linear model. However, in an ancillary analysis of the
concentration-response curves for the largest cities in each of the three distinct geographic areas
(western, southern, and eastern European cities): London, England; Athens, Greece; and Cracow,
Poland, Samoli et al. (2005, 087436) observed a difference in the shape of the concentration-
response curve across cities. Thus, while the combined curves (Figure 6-37) appear to support
5E-9
-------
no-threshold relationships between PMio and mortality, the heterogeneity of the shapes across
cities makes it difficult to interpret the biological relevance of the shape of the combined curves.
"The results from the three multicity studies discussed above support no-threshold log-linear
models, but issues such as the possible influence of exposure error and heterogeneity of shapes
across cities remain to be resolved. Also, given the pattern of seasonal and regional differences
in PM risk estimates depicted in recent multicity study results (e.g., Peng et al., 2005, 087463),
the very concept of a concentration-response relationship estimated across cities and for all-year
data may not be very informative."
Authors of ISA
Dr. Lindsay Wichers Stanek (PM Team Leader)—National Center for Environmental
Assessment (NCEA), U.S. Environmental Protection Agency (U.S. EPA), Research Triangle
Park, NC
Dr. Jeffrey Arnold—NCEA, U.S. EPA, Research Triangle Park, NC (now at Institute for Water
Resources, U.S. Army Corps of Engineers, Washington, D.C)
Dr. Christal Bowman—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. James S. Brown—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Barbara Buckley—NCEA, U.S. EPA, Research Triangle Park, NC
Mr. Allen Davis—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Jean-Jacques Dubois—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Steven J. Dutton—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Tara Greaver—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Erin Hines—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Douglas Johns—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Ellen Kirrane—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Dennis Kotchmar—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Thomas Long—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Thomas Luben—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research
Fellow to NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Kristopher Novak—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Joseph Pinto—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—NCEA, US EPA, Research Triangle Park, NC
Dr. Mary Ross—NCEA, U.S. EPA, Research Triangle Park, NC
Mr. Jason Sacks—NCEA, U.S. EPA, Research Triangle Park, NC
5E-10
-------
Dr. Timothy J. Sullivan—E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. David Svendsgaard—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Lisa Vinikoor—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. William Wilson—NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Lori White— NCEA, U.S. EPA, Research Triangle Park, NC (now at National Institute for
Environmental Health Sciences, Research Triangle Park, NC)
Dr. Christy Avery—University of North Carolina, Chapel Hill, NC
Dr. Kathleen Belanger —Center for Perinatal, Pediatric and Environmental Epidemiology,
Yale University, New Haven, CT
Dr. Michelle Bell—School of Forestry & Environmental Studies, Yale University, New Haven,
CT
Dr. William D. Bennett—Center for Environmental Medicine, Asthma and Lung Biology,
University of North Carolina, Chapel Hill, NC
Dr. Matthew J. Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Leland B. Deck— Stratus Consulting, Inc., Washington, DC
Dr. Janneane F. Gent—Center for Perinatal, Pediatric and Environmental Epidemiology, Yale
University, New Haven, CT
Dr. Yuh-Chin Tony Huang—Department of Medicine, Division of Pulmonary Medicine, Duke
University Medical Center, Durham, NC
Dr. Kazuhiko Ito—Nelson Institute of Environmental Medicine, NYU School of Medicine,
Tuxedo, NY
Mr. Marc Jackson—Integrated Laboratory Systems, Inc., Research Triangle Park, NC
Dr. Michael Kleinman—Department of Community and Environmental Medicine, University
of California, Irvine
Dr. Sergey Napelenok—National Exposure Research Laboratory, U.S. EPA, Research Triangle
Park, NC
Dr. Marc Pitchford—National Oceanic and Atmospheric Administration, Las Vegas, NV
Dr. Les Recio—Genetic Toxicology Division, Integrated Laboratory Systems, Inc., Research
Triangle Park, NC
Dr. David Quincy Rich—Department of Epidemiology, University of Medicine and Dentistry
of New Jersey, Piscataway, NJ
Dr. Timothy Sullivan— E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. George Thurston—Department of Environmental Medicine, NYU, Tuxedo, NY
Dr. Gregory Wellenius—Cardiovascular Epidemiology Research Unit, Beth Israel Deaconess
Medical Center, Boston, MA
5E-11
-------
Dr. Eric Whitsel—Departments of Epidemiology and Medicine, University of North Carolina,
Chapel Hill, NC
Peer Reviewers
Dr. Sara Dubowsky Adar, Department of Epidemiology, University of Washington, Seattle,
WA
Mr. Chad Bailey, Office of Transportation and Air Quality, Ann Arbor, MI
Mr. Richard Baldauf, Office of Transportation and Air Quality, Ann Arbor, MI
Dr. Prakash Bhave, National Exposure Research Laboratory, U.S. EPA, Research Triangle
Park, NC
Mr. George Bowker, Office of Atmospheric Programs, U.S. EPA, Washington, D.C.
Dr. Judith Chow, Division of Atmospheric Sciences, Desert Research Institute, Reno, NV
Dr. Dan Costa, U.S. EPA, Research Triangle Park, NC
Dr. Ila Cote, NCEA, U.S. EPA, Research Triangle Park, NC
Dr. Robert Devlin, National Health and Environmental Effects Research Laboratory, U.S. EPA,
Research Triangle Park, NC
Dr. David DeMarini, National Health and Environmental Effects Research Laboratory, U.S.
EPA, Research Triangle Park, NC
Dr. Neil Donahue, Department of Chemical Engineering, Carnegie Mellon University,
Pittsburgh, PA
Dr. Aimen Farraj, National Health and Environmental Effects Research Laboratory, U.S. EPA,
Research Triangle Park, NC
Dr. Mark Frampton, Department of Environmental Medicine, University of Rochester Medical
Center, Rochester, NY
Mr. Neil Frank, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Mr. Tyler Fox, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Dr. Jim Gauderman, Department of Environmental Medicine, Department of Preventive
Medicine, University of Southern California, Los Angeles, CA
Dr. Barbara Glenn, National Center for Environmental Research, U.S. EPA, Washington, D.C.
Dr. Terry Gordon, School of Medicine, New York University, Tuxedo, NY
Mr. Tim Hanley, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Dr. Jack Harkema, Department of Pathobiology and Diagnostic Investigation, Michigan State
University, East Lansing, MI
Ms. Beth Hassett-Sipple, Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC
5E-12
-------
Dr. Amy Herring, Department of Biostatistics, University of North Carolina, Chapel Hill, NC
Dr. Israel Jirak, Department of Meteorology, Embry-Riddle Aeronautical University, Prescott,
AZ
Dr. Mike Kleeman, Department of Civil and Environmental Engineering, University of
California, Davis, CA
Dr. Petros Koutrakis, Exposure, Epidemiology and Risk Program, Harvard School of Public
Health, Boston, MA
Dr. Sagar Krupa, Department of Plant Pathology, University of Minnesota, St. Paul, MN
Mr. John Langstaff, Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC
Dr. Meredith Lassiter, Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC
Mr. Phil Lorang, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Dr. Karen Martin, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Ms. Connie Meacham, NCEA, U.S. EPA, Research Triangle Park, NC
Mr. Tom Pace, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Dr. Jennifer Peel, Department of Environmental and Radiological Health Sciences, College of
Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO
Dr. Zackary Pekar, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Mr. Rob Pinder, National Exposure Research Laboratory, U.S. EPA, Research Triangle Park,
NC
Mr. Norm Possiel, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Dr. Sanjay Rajagopalan, Division of Cardiovascular Medicine, Ohio State University,
Columbus, OH
Dr. Pradeep Rajan, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Mr. Venkatesh Rao, Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC
Ms. Joann Rice, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Mr. Harvey Richmond, Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC
5E-13
-------
Ms. Victoria Sandiford, Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC
Dr. Stefanie Sarnat, Department of Environmental and Occupational Health, Emory University,
Atlanta, GA
Dr. Frances Silverman, Gage Occupational and Environmental Health, University of Toronto,
Toronto, ON
Mr. Steven Silverman, Office of General Council, U.S. EPA, Washington, D.C.
Dr. Barbara Turpin, Department of Environmental Sciences, Rutgers University, New
Brunswick, NJ
Dr. Robert Vanderpool, National Exposure Research Laboratory, U.S. EPA, Research Triangle
Park, NC
Dr. John Vandenberg (Director)—NCEA-RTP Division, U.S. EPA, Research Triangle Park,
NC
Dr. Alan Vette, National Exposure Research Laboratory, U.S. EPA, Research Triangle Park,
NC
Ms. Debra Walsh (Deputy Director)—NCEA-RTP Division, U.S. EPA, Research Triangle
Park, NC
Mr. Tim Watkins, National Exposure Research Laboratory, U.S. EPA, Research Triangle Park,
NC
Dr. Christopher Weaver, NCEA, U.S. EPA, Research Triangle Park, NC
Mr. Lewis Weinstock, Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC
Ms. Karen Wesson, Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle
Park, NC
Dr. Jason West, Department of Environmental Sciences and Engineering, University of North
Carolina, Chapel Hill, NC
Mr. Ronald Williams, National Exposure Research Laboratory, U.S. EPA, Research Triangle
Park, NC
Dr. George Woodall, NCEA, U.S. EPA, Research Triangle Park, NC
Dr. A nt on el hi Zanobetti, Department of Environmental Health, Harvard School of Public
Health, Boston, MA
5E-14
-------
D. CASAC comments on PM ISA and REA (2009)
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2009.
Review of EPA's Integrated Science Assessment for Particulate Matter (First External
Review Draft, December 2008). EPA-COUNCIL-09-008. May. Available on the Internet
at
.
Pg 9: "There is an appropriate discussion of the time-series studies, but this section needs to
have an explicit finding that the evidence supports a relationship between PM and mortality that
is seen in these studies. This conclusion should be followed by the discussion of statistical
methodology and the identification of any threshold that may exist."
U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2009.
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
.
Pg 6: "On the issue of cut-points raised on 3-18, the authors should be prepared to offer a
scientifically cogent reason for selection of a specific cut-point, and not simply try different cut-
points to see what effect this has on the analysis. The draft ISA was clear that there is little
evidence for a population threshold in the C-R function."
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2009. Review
of Integrated Science Assessment for Particulate Matter (Second External Review Draft, July 2009).
EPA-CASAC-10-001. November. Available on the Internet at
.
Pg 2: "The paragraph on lines 22-30 of page 2-37 is not clearly written. Twice in succession it
states that the use of a no-threshold log-linear model is supported, but then cites other studies
that suggest otherwise. It would be good to revise this paragraph to more clearly state - well, I'm
not sure what. Probably that more research is needed."
CASAC Panel Members
Dr. Jonathan M. Samet, Professor and Chair, Department of Preventive Medicine, University of
Southern California, Los Angeles, CA
Dr. Joseph Brain, Philip Drinker Professor of Environmental Physiology, Department of Environmental
Health, Harvard School of Public Health, Harvard University, Boston, MA
Dr. Ellis B. Cowling, University Distinguished Professor At-Large Emeritus, Colleges of Natural
Resources and Agriculture and Life Sciences, North Carolina State University, Raleigh, NC
Dr. James Crapo, Professor of Medicine, Department of Medicine, National Jewish Medical and
Research Center, Denver, CO
5E-15
-------
Dr. H. Christopher Frey, Professor, Department of Civil, Construction and Environmental Engineering,
College of Engineering, North Carolina State University, Raleigh, NC
Dr. Armistead (Ted) Russell, Professor, Department of Civil and Environmental Engineering, Georgia
Institute of Technology, Atlanta, GA
Dr. Lowell Ashbaugh, Associate Research Ecologist, Crocker Nuclear Lab, University of California,
Davis, Davis, CA
Prof. Ed Avol, Professor, Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA
Dr. Wayne Cascio, Professor, Medicine, Cardiology, Brody School of Medicine at East Carolina
University, Greenville, NC
Dr. David Grantz, Director, Botany and Plant Sciences and Air Pollution Research Center, Riverside
Campus and Kearney Agricultural Center, University of California, Parlier, CA
Dr. Joseph Helble, Dean and Professor, Thayer School of Engineering, Dartmouth College, Hanover,
NH
Dr. Rogene Henderson, Senior Scientist Emeritus, Lovelace Respiratory Research Institute,
Albuquerque, NM
Dr. Philip Hopke, Bayard D. Clarkson Distinguished Professor, Department of Chemical Engineering,
Clarkson University, Potsdam, NY
Dr. Morton Lippmann, Professor, Nelson Institute of Environmental Medicine, New York University
School of Medicine, Tuxedo, NY
Dr. Helen Suh Macintosh, Associate Professor, Environmental Health, School of Public Health,
Harvard University, Boston, MA
Dr. William Malm, Research Physicist, National Park Service Air Resources Division, Cooperative
Institute for Research in the Atmosphere, Colorado State University, Fort Collins, CO
Mr. Charles Thomas (Tom) Moore, Jr., Air Quality Program Manager, Western Governors'
Association, Cooperative Institute for Research in the Atmosphere, Colorado State University, Fort
Collins, CO
Dr. Robert F. Phalen, Professor, Department of Community & Environmental Medicine; Director, Air
Pollution Health Effects Laboratory; Professor of Occupational & Environmental Health, Center for
Occupation & Environment Health, College of Medicine, University of California Irvine, Irvine, CA
Dr. Kent Pinkerton, Professor, Regents of the University of California, Center for Health and the
Environment, University of California, Davis, CA
Mr. Richard L. Poirot, Environmental Analyst, Air Pollution Control Division, Department of
Environmental Conservation, Vermont Agency of Natural Resources, Waterbury, VT
Dr. Frank Speizer, Edward Kass Professor of Medicine, Channing Laboratory, Harvard Medical School,
Boston, MA
Dr. Sverre Vedal, Professor, Department of Environmental and Occupational Health Sciences, School of
Public Health and Community Medicine, University of Washington, Seattle, WA
Dr. Donna Kenski, Data Analysis Director, Lake Michigan Air Directors Consortium, Rosemont, IL
Dr. Kathy Weathers, Senior Scientist, Gary Institute of Ecosystem Studies, Millbrook, NY
5E-16
-------
E. Krewski et al. (2009)
Krewski, Daniel, Michael Jerrett, Richard T. Burnett, Renjun Ma, Edward Hughes, Yuanli
Shi, Michelle C. Turner, C. Arden Pope III, George Thurston, Eugenia E. Calle, and
Michael J. Thun with Bernie Beckerman, Pat DeLuca, Norm Finkelstein, Kaz Ito, D.K.
Moore, K. Bruce Newbold, Tim Ramsay, Zev Ross, Hwashin Shin, and Barbara
Tempalski. (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.
Pg 119: [About Pope et al. (2002)] "Each 10-ug/m3 increase in longterm average ambient PIVb.s
concentrations was associated with approximately a 4%, 6%, or 8% increase in risk of death
from all causes, cardiopulmonary disease, and lung cancer, respectively. There was no evidence
of a threshold exposure level within the range of observed PM2.5 concentrations. "
Krewski (2009). Letter from Dr. Daniel Krewski to HEI's Dr. Kate Adams (dated July July
7, 2009) regarding "EPA queries regarding HEI Report 140". Dr. Adams then forwarded
the letter on July 10, 2009 to EPA's Beth Hassett-Sipple. (letter placed in docket #EPA-
HQ-OAR-2007-0492).
Pg 4: "6. The Health Review Committee commented that the Updated Analysis completed by
Pope et al. 2002 reported "no evidence of a threshold exposure level within the range of
observed PM2.5 concentrations" (p. 119). In the Extended Follow-Up study, did the analyses
provide continued support for a no-threshold response or was there evidence of a threshold?
"Response: As noted above, the HEI Health Review Committee commented on the lack of
evidence for a threshold exposure level in Pope et al. (2002) with follow-up through the year
1998. The present report, which included follow-up through the year 2000, also does not appear
to demonstrate the existence of a threshold in the exposure-response function within the range of
observed PM2.5 concentrations."
HEI Health Review Committee Members
Dr. Homer A. Boushey, MD, Chair, Professor of Medicine, Department of Medicine,
University of California-San Francisco
Dr. Ben Armstrong, Reader, in Epidemiological Statistics, Department of Public Health and
Policy, London School of Hygiene and Tropical Medicine, United Kingdom
Dr. Michael Brauer, ScD, Professor, School of Environmental Health, University of British
Columbia, Canada
Dr. Bert Brunekreef, PhD, Professor of Environmental Epidemiology, Institute of Risk
Assessment Sciences, University of Utrecht, The Netherlands
Dr. Mark W. Frampton, MD, Professor of Medicine & Environmental Medicine, University of
Rochester Medical Center, Rochester, NY
Dr. Stephanie London, MD, PhD, Senior Investigator, Epidemiology Branch, National Institute
of Environmental Health Sciences
5E-17
-------
Dr. William N. Rom, MD, MPH, Sol and Judith Bergstein Professor of Medicine and
Environmental Medicine and Director of Pulmonary and Critical Care Medicine, New York
University Medical Center
Dr. Armistead Russell, Georgia Power Distinguished Professor of Environmental Engineering,
School of Civil and Environmental Engineering, Georgia Institute of Technology
Dr. Lianne Sheppard, PhD, Professor, Department of Biostatistics, University of Washington
5E-18
-------
F. Schwartz et al. (2008)
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.
Pg 67: "A key finding of this study is that there is little evidence for a threshold in the
association between exposure to fine particles and the risk of death on follow-up, which
continues well below the U.S. EPA standard of 15 ug/m3."
Pg 68: "In conclusion, penalized spline smoothing and model averaging represent reasonable,
feasible approaches to addressing questions of the shape of the exposure-response curve, and can
provide valuable information to decisionmakers. In this example, both approaches are consistent,
and suggest that the association of particles with mortality has no threshold down to close to
background levels."
5E-19
-------
G. Expert Elicitation on PM-Mortality (2006, 2008)
Industrial Economics, Inc., 2006. Expanded Expert Judgment Assessment of the
Concentration-Response Relationship Between PM2.s Exposure and Mortality. Prepared
for the U.S.EPA, Office of Air Quality Planning and Standards, September. Available on
the Internet at .
Pg v: "Each expert was given the option to integrate their judgments about the likelihood of a
causal relationship and/or threshold in the C-R function into his distribution or to provide a
distribution "conditional on" one or both of these factors."
Pg vii: "Only one of 12 experts explicitly incorporated a threshold into his C-R function.3 The
rest believed there was a lack of empirical and/or theoretical support for a population threshold.
However, three other experts gave differing effect estimate distributions above and below some
cut-off concentration. The adjustments these experts made to median estimates and/or
uncertainty at lower PM2 5 concentrations were modest."
"3 Expert K indicated that he was 50 percent sure that a threshold existed. If there
were a threshold, he thought that there was an 80 percent chance that it would be
less than or equal to 5 ug/m3, and a 20 percent chance that it would fall between 5
and 10 ug/m3."
Pg ix: "Compared to the pilot study, experts in this study were in general more confident in a
causal relationship, less likely to incorporate thresholds, and reported higher mortality effect
estimates. The differences in results compared with the pilot appear to reflect the influence of
new research on the interpretation of the key epidemiological studies that were the focus of both
elicitation studies, more than the influence of changes to the structure of the protocol."
Pg3-25: "3.1.8 THRESHOLDS
The protocol asked experts for their judgments regarding whether a threshold exists in the PM2.5
mortality C-R function. The protocol focused on assessing expert judgments regarding theory
and evidential support for a population threshold (i.e., the concentration below which no member
of the study population would experience an increased risk of death).32 If an expert wished to
incorporate a threshold in his characterization of the concentration-response relationship, the
team then asked the expert to specify the threshold PM2.5 concentration probabilistically,
incorporating his uncertainty about the true threshold level.
"From a theoretical and conceptual standpoint, all experts generally believed that individuals
exhibit thresholds for PM-related mortality. However, 11 of them discounted the idea of a
population threshold in the C-R function on a theoretical and/or empirical basis. Seven of these
experts noted that theoretically one would be unlikely to observe a population threshold due to
the variation in susceptibility at any given time in the study population resulting from
combinations of genetic, environmental, and socioeconomic factors.33 All 11 thought that there
was insufficient empirical support for a population threshold in the C-R function. In addition,
two experts (E and L) cited analyses of the ACS cohort data in Pope et al. (2002) and another (J)
cited Krewski et al. (2000a & b) as supportive of a linear relationship in the study range.
5E-20
-------
"Seven of the experts favored epidemiological studies as ideally the best means of addressing the
population threshold issue, because they are best able to evaluate the full range of susceptible
individuals at environmentally relevant exposure levels. However, those who favored
epidemiologic studies generally acknowledged that definitive studies addressing thresholds
would be difficult or impossible to conduct, because they would need to include a very large and
diverse population with wide variation in exposure and a long follow-up period. Furthermore,
two experts (B and I) cited studies documenting difficulties in detecting a threshold using
epidemiological studies (Cakmak et al. 1999, and Brauer et al., 2002, respectively). The experts
generally thought that clinical and toxicological studies are best suited for researching
mechanisms and for addressing thresholds in very narrowly defined groups. One expert, B,
thought that a better understanding of the detailed biological mechanism is critical to addressing
the question of a threshold.
"One expert, K, believed it was possible to make a conceptual argument for a population
threshold. He drew an analogy with smoking, indicating that among heavy smokers, only a
proportion of them gets lung cancer or demonstrates an accelerated decline in lung function. He
thought that the idea that there is no level that is biologically safe is fundamentally at odds with
toxicological theory. He did not think that a population threshold was detectable in the currently
available epidemiologic studies. He indicated that some of the cohort studies showed greater
uncertainty in the shape of the C-R function at lower levels, which could be indicative of a
threshold.
"Expert K chose to incorporate a threshold into his C-R function. He indicated that he was 50
percent sure that a threshold existed. If there were a threshold, he thought that there was an 80
percent chance that it would be less than or equal to 5 ug/m3, and a 20 percent chance that it
would fall between 5 and 10 ug/m3."
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. Scl Technol, 42(7):2268-2274.
Pg 2271: "Eight experts thought the true C-R function relating mortality to changes in annual
average PM2.5 was log-linear across the entire study range (In(mortality) ) P * PM). Four experts
(B, F, K, and L) specified a "piecewise" log-linear function, with different P coefficients for PM
concentrations above and below an expert-specified break point. This approach allowed them to
express increased uncertainty in mortality effects seen at lower concentrations in major
epidemiological studies. Expert K thought the relationship would be log-linear above a
threshold."
Pg 2271: "Expert K also applied a threshold, T, to his function, which he described
probabilistically. He specified P(T > 0) = 0.5. Given T > 0, he indicated P(T < 5 ug/m3) = 0.8
and P(5 ug/m3 < T < 10 ug/m3) = 0.2. Figure 3 does not include the impact of applying expert
K's threshold, as the size of the reduction in benefits will depend on the distribution of baseline
PM levels in a benefits analysis."
5E-21
-------
Experts:
Dr. Doug W. Dockery, Harvard School of Public Health
Dr. Kazuhiko Ito, Nelson Institute of Environmental Medicine, NYU School of Medicine,
Tuxedo, NY
Dr. Dan Krewski, University of Ottawa
Dr. Nino Kiinzli, University of Southern California Keck School of Medicine
Dr. Morton Lippmann, Professor, Nelson Institute of Environmental Medicine, New York University
School of Medicine, Tuxedo, NY
Dr. Joe Mauderly, Lovelace Respiratory Research Institute
Dr. Bart Ostro, Chief, Air Pollution Epidemiology Unit, Office of Environmental Health
Hazard Assessment, California Environmental Protection Agency, Oakland, CA
Dr. Arden Pope, Professor, Department of Economics, Brigham Young University, Provo, UT
Dr. Richard Schlesinger, Pace University
Dr. Joel Schwartz, Harvard School of Public Health
Dr. George Thurston—Department of Environmental Medicine, NYU, Tuxedo, NY
Dr. Mark Utell, University of Rochester School of Medicine and Dentistry
5E-22
-------
H. CASAC comments on PM Staff Paper (2005)
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2005.
EPA's Review of the National Ambient Air Quality Standards for Particulate Matter
(Second Draft PM Staff Paper, January 2005). EPA-SAB-CASAC-05-007. June. Available
on the Internet at
-------
Dr. Paul J. Lioy, Associate Director and Professor, Environmental and Occupational Health Sciences
Institute, UMDNJ - Robert Wood Johnson Medical School, NJ
Dr. Morton Lippmann, Professor, Nelson Institute of Environmental Medicine, New York
University School of Medicine, Tuxedo, NY
Dr. Joe Mauderly, Vice President, Senior Scientist, and Director, National Environmental
Respiratory Center, Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Roger O. McClellan, Consultant, Albuquerque, NM
Dr. Frederick J. Miller, Consultant, Cary, NC
Dr. Gunter Oberdorster, Professor of Toxicology, Department of Environmental Medicine, School
of Medicine and Dentistry, University of Rochester, Rochester, NY
Mr. Richard L. Poirot, Environmental Analyst, Air Pollution Control Division, Department of
Environmental Conservation, Vermont Agency of Natural Resources, Waterbury, VT
Dr. Robert D. Rowe, President, Stratus Consulting, Inc., Boulder, CO
Dr. Jonathan M. Samet, Professor and Chair, Department of Epidemiology, Bloomberg School of
Public Health, Johns Hopkins University, Baltimore, MD
Dr. Frank Speizer, Edward Kass Professor of Medicine, Channing Laboratory, Harvard Medical
School, Boston, MA
Dr. Sverre Vedal, Professor of Medicine, School of Public Health and Community Medicine
University of Washington, Seattle, WA
Mr. Ronald White, Research Scientist, Epidemiology, Bloomberg School of Public Health, Johns
Hopkins University, Baltimore, MD
Dr. Warren H. White, Visiting Professor, Crocker Nuclear Laboratory, University of California -Davis,
Davis, CA
Dr. George T. Wolff, Principal Scientist, General Motors Corporation, Detroit, MI
Dr. Barbara Zielinska, Research Professor, Division of Atmospheric Science, Desert Research
Institute, Reno, NV
5E-24
-------
I. HES Comments on 812 Analysis (2004)
U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2004.
Advisory on Plans for Health Effects Analysis in the Analytical Plan for EPA's Second
Prospective Analysis - Benefits and Costs of the Clean Air Act, 1990-2020. 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
.
Pg 20: "The Subcommittee agrees that the whole range of uncertainties, such as the questions of
causality, shape of C-R functions and thresholds, relative toxicity, years of life lost, cessation lag
structure, cause of death, biologic pathways, or susceptibilities may be viewed differently for
acute effects versus long-term effects.
"For the studies of long-term exposure, the HES notes that Krewski et al. (2000) have conducted
the most careful work on this issue. They report that the associations between PM2.5 and both all-
cause and cardiopulmonary mortality were near linear within the relevant ranges, with no
apparent threshold. Graphical analyses of these studies (Dockery et al., 1993, Figure 3 and
Krewski et al., 2000, page 162) also suggest a continuum of effects down to lower levels.
Therefore, it is reasonable for EPA to assume a no threshold model down to, at least, the low end
of the concentrations reported in the studies."
HES Panel Members
Dr. Bart Ostro, California Office of Environmental Health Hazard Assessment (OEHHA),
Oakland, CA
Mr. John Fintan Hurley, Institute of Occupational Medicine (IOM), Edinburgh, Scotland
Dr. Patrick Kinney, Columbia University, New York, NY
Dr. Michael Kleinman, University of California, Irvine, CA
Dr. Nino Kiinzli, University of Southern California, Los Angeles, CA
Dr. Morton Lippmann, New York University School of Medicine, Tuxedo, NY Dr. Rebecca
Parkin, The George Washington University, Washington, DC
Dr. Trudy Cameron, University of Oregon, Eugene, OR
Dr. David T. Allen, University of Texas, Austin, TX
Ms. Lauraine Chestnut, Stratus Consulting Inc., Boulder, CO
Dr. Lawrence Goulder, Stanford University, Stanford, CA
Dr. James Hammitt, Harvard University, Boston, MA
Dr. F. Reed Johnson, Research Triangle Institute, Research Triangle Park, NC
Dr. Charles Kolstad, University of California, Santa Barbara, CA
Dr. Lester B. Lave, Carnegie Mellon University, Pittsburgh, PA
5E-25
-------
Dr. Virginia McConnell, Resources for the Future, Washington, DC
Dr. V. Kerry Smith, North Carolina State University, Raleigh, NC
Other Panel Members
Dr. John Evans, Harvard University, Portsmouth, NH Dr. Dale Hattis, Clark University,
Worcester, MA Dr. D. Warner North, NorthWorks Inc., Belmont, CA Dr. Thomas S. Wallsten,
University of Maryland, College Park, MD
5E-26
-------
J. NRC - Committee on Estimating the Health Risk Reduction Benefits of Proposed Air
Pollution Regulations (2002)
National Research Council (NRC). 2002. Estimating the Public Health Benefits of
Proposed Air Pollution Regulations. Washington, DC: The National Academies Press.
Pg 109: "Linearity and Thresholds
"The shape of the concentration-response functions may influence the overall estimate of
benefits. The shape is particularly important for lower ambient air pollution concentrations to
which a large portion of the population is exposed. For this reason, the impact of the existence of
a threshold may be considerable.
"In epidemiological studies, air pollution concentrations are usually measured and modeled as
continuous variables. Thus, it may be feasible to test linearity and the existence of thresholds,
depending on the study design. In time-series studies with the large number of repeated
measurements, linearity and thresholds have been formally addressed with reasonable statistical
power. For pollutants such as PMio and PM2.5, there is no evidence for any departure of linearity
in the observed range of exposure, nor any indication of a threshold. For example, examination
of the mortality effects of short-term exposure to PMio in 88 cities indicates that the
concentration-response functions are not due to the high concentrations and that the slopes of
these functions do not appear to increase at higher concentrations (Samet et al. 2000). Many
other mortality studies have examined the shape of the concentration-response function and
indicated that a linear (nonthreshold) model fit the data well (Pope 2000). Furthermore, studies
conducted in cities with very low ambient pollution concentrations have similar effects per unit
change in concentration as those studies conducted in cities with higher concentrations. Again,
this finding suggests a fairly linear concentration-response function over the observed range of
exposures.
"Regarding the studies of long-term exposure, Krewski et al. (2000) found that the assumption of a linear
concentration-response function for mortality outcomes was not unreasonable. However, the statistical
power to assess the shape of these functions is weakest at the upper and lower end of the observed
exposure ranges. Most of the studies examining the effects of long-term exposure on morbidity compare
subjects living in a small number of communities (Dockery et al. 1996; Ackermmann-Liebrich 1997;
Braun-Fahrlander et al. 1997). Because the number of long-term effects studies are few and the number of
communities studied is relatively small (8 to 24), the ability to test formally the absence or existence of a
no-effect threshold is not feasible. However, even if thresholds exist, they may not be at the same
concentration for all health outcomes.
"A review of the time-series and cohort studies may lead to the conclusion that although a threshold is not
apparent at commonly observed concentrations, one may exist at lower levels. An important point to
acknowledge regarding thresholds is that for health benefits analysis a key threshold is the population
threshold (the lowest of the individual thresholds). However, the population threshold would be very
difficult to observe empirically through epidemiology, because epidemiology integrates information from
very large groups of people (thousands). Air pollution regulations affect even larger groups of people
(millions). It is reasonable to assume that among such large groups susceptibility to air pollution health
effects varies considerably across individuals and depends on a large set of underlying factors, including
5E-27
-------
genetic makeup, age, exposure measurement error, preexisting disease, and simultaneous exposures from
smoking and occupational hazards. This variation in individual susceptibilities and the resulting
distribution of individual thresholds underlies the concentration-response function observed in
epidemiology. Thus, until biologically based models of the distribution of individual thresholds are
developed, it may be productive to assume that the population concentration-response function is
continuous and to focus on finding evidence of changes in its slope as one approaches lower
concentrations.
EPA's Use of Thresholds
"In EPA's benefits analyses, threshold issues were discussed and interpreted. For the PM and ozone
National Ambient Air Quality Standards (NAAQS), EPA investigated the effects of a potential threshold
or reference value below which health consequences were assumed to be zero (EPA 1997). Specifically,
the high-end benefits estimate assumed a 12-microgram per cubic meter (jig/m3) mean threshold for
mortality associated with long-term exposure to PM25. The low-end benefits estimate assumed a 15-
(ig/m3 threshold for all PM-related health effects. The studies, however, included concentrations as low as
7.5 (ig/m3. For the Tier 2 rule and the HD engine and diesel-fuel rule, no threshold was assumed (EPA
1999, 2000). EPA in these analyses acknowledged that there was no evidence for a threshold for PM.
"Several points should be noted regarding the threshold assumptions. If a threshold is assumed where one
was not apparent in the original study, then the data should be refit and a new curve generated with the
assumption of a zero slope over a segment of the concentration-response function that was originally
found to be positively sloped. The assumption of a zero slope over a portion of the curve will force the
slope in the remaining segment of the positively sloped concentration-response function to be greater than
was indicated in the original study. A new concentration-response function was not generated for EPA's
benefits analysis for the PM and ozone NAAQS for which threshold assumptions were made. The
generation of the steeper slope in the remaining portion of the concentration-response function may fully
offset the effect of assuming a threshold. These aspects of assuming a threshold in a benefits analysis
where one was not indicated in the original study should be conveyed to the reader. The committee notes
that the treatment of thresholds should be evaluated in a consistent and transparent framework by using
different explicit assumptions in the formal uncertainty analyses (see Chapter 5)."
Pg 117: "Although the assumption of no thresholds in the most recent EPA benefits analyses was
appropriate, EPA should evaluate threshold assumptions in a consistent and transparent framework using
several alternative assumptions in the formal uncertainty analysis."
Pg 136: "Two additional illustrative examples are thresholds for adverse effects and lag structures.- EPA
considers implausible any threshold for mortality in the particulate matter (PM) exposure ranges under
consideration (EPA 1999a, p. 3-8). Although the agency conducts sensitivity analyses incorporating
thresholds, it provides no judgment as to their relative plausibility. In a probabilistic uncertainty analysis,
EPA could assign appropriate weights to various threshold models. For PM-related mortality in the Tier 2
analysis, the committee expects that this approach would have resulted in only a slight widening of the
probability distribution for avoided mortality and a slight reduction in the mean of that distribution, thus
reflecting EPA's views about the implausibility of thresholds. The committee finds that such formal
incorporation of EPA's expert judgments about the plausibility of thresholds into its primary analysis
would have been an improvement.
"Uncertainty about thresholds is a special aspect of uncertainty about the shape of concentration-response
functions. Typically, EPA and authors of epidemiological studies assume that these functions are linear
on some scale. Often, the scale is a logarithmic transformation of the risk or rate of the health outcome,
5E-28
-------
but when a rate or risk is low, a linear function on the logarithmic scale is approximately linear on the
scale of the rate or risk itself. Increasingly, epidemiological investigators are employing analytic methods
that permit the estimation of nonlinear shapes for concentration-response functions (Greenland et al.
1999). As a consequence, EPA will need to be prepared to incorporate nonlinear concentration-response
functions from epidemiological studies into the agency's health benefits analyses. Any source of error or
bias that can distort an epidemiological association can also distort the shape of an estimated
concentration -response function, as can variation in individual susceptibility (Hattis and Burmaster 1994;
Hattisetal. 2001)."
Pg 137: "In principle, many components of the health benefits model need realistic probabilistic models
(see Table 5-1 for a listing of such components), in addition to concentration-response thresholds and
time lags between exposure and response. For example, additional features of the concentration-response
function—such as projection of the results from the study population to the target populations (which may
have etiologically relevant characteristics outside the range seen in the study population) and the
projection of baseline frequencies of morbidity and mortality into the future—must be characterized
probabilistically. Other uncertainties that might affect the probability distributions are the estimations of
population exposure (or even concentration) from emissions, estimates of emissions themselves, and the
relative toxicity of various classes of particles. Similarly, many aspects of the analysis of the impact of
regulation on ambient concentrations and on population exposure involve considerable uncertainty and,
therefore, may be beneficially modeled in this way. Depending on the analytic approach used, joint
probability distributions will have to be specified to incorporate correlations between model components
that are structurally dependent upon each other, or the analysis will have to be conducted in a sequential
fashion that follows the model for the data-generating process.
"EPA should explore alternative options for incorporating expert judgment into its probabilistic
uncertainty analyses. The agency possesses considerable internal expertise, which should be employed as
fully as possible. Outside experts should also be consulted as needed, individually or in panels. In all
cases, when expert judgment is used in the construction of a model component, the experts should be
identified and the rationales and empirical bases for their judgments should be made available."
NRC members
Dr. JOHN C. BAILAR, III (Chair), (emeritus) University of Chicago, Chicago, Illinois
Dr. HUGH ROSS ANDERSON, University of London, London, England
Dr. MAUREEN L. CROPPER, University of Maryland, College Park
Dr. JOHN S. EVANS, Harvard University, Boston, Massachusetts
Dr. DALE B. HATTIS, Clark University, Worcester, Massachusetts
Dr. ROGENE F. HENDERSON, Lovelace Respiratory Research Institute, Albuquerque, New Mexico
Dr. PATRICK L. KINNEY, Columbia University, New York, New York
Dr. NINO KUNZLI, University of Basel, Basel, Switzerland; as of September 2002, University of
Southern California, Los Angeles
Dr. BART D. OSTRO, California Environmental Protection Agency, Oakland
Dr. CHARLES POOLE, University of North Carolina, Chapel Hill
Dr. KIRK R. SMITH, University of California, Berkeley
Dr. PETER A. VALBERG, Gradient Corporation, Cambridge, Massachusetts
5E-29
-------
Dr. SCOTT L. ZEGER, Johns Hopkins University, Baltimore, Maryland
5E-30
-------
CHAPTER 6
EMPLOYMENT AND ECONOMIC IMPACT ANALYSIS
In this chapter, we estimate select employment effects of the rule, for both the
regulated industry (electric power industry) and the environmental control industry.
6.1 Employment Impacts for the MATS
In addition to addressing the costs and benefits of the MATS, EPA has estimated certain
impacts of this on employment, which are presented in this section.1 While a standalone
analysis of employment impacts is not included in a standard cost-benefit analysis, such an
analysis is of particular concern in the current economic climate of sustained unemployment.
Executive Order 13563, states, "Our regulatory system must protect public health, welfare,
safety and our environment while promoting economic growth, innovation, competiveness, and
job creation " (emphasis added). Therefore, and consistent with recent efforts to characterize
the employment effects of economically significant rules, the Agency has provided this analysis
to inform the discussion of labor demand and employment impacts.
The analysis includes two sets of estimates. The first involves the employment impacts
on the regulated industry over time. The second involves certain short-term and on-going
employment impacts (increase in labor demand) associated with the construction of needed
pollution control equipment, and other activities, to comply with the regulation. EPA estimates
that the net employment effect on the regulated industry will range from -15,000 to +30,000
jobs, with a central estimate of +8,000. This aggregate figure includes potential job losses from
increased costs as well as potential job increases as a result of additional hiring for compliance.
In the pollution control sector, EPA estimates an increase of 46,000 job-years. EPA also provides
a qualitative discussion of other potential employment effects, including both increases and
decreases. Because of the uncertainties involved, these sets of estimates should not be added
in an attempt to characterize the overall employment effect.
The Agency has not quantified the rule's effects on all labor in other sectors not
regulated by the MATS, or the effects induced by changes in workers' incomes. What follows is
an overview of the various ways that environmental regulation can affect employment,
followed by a discussion of the estimated impacts of this rule. EPA continues to explore the
relevant theoretical and empirical literature, which continues to evolve, and to seek public
1 See the employment impacts appendix included in this RIA.
6-1
-------
comments in order to ensure that such estimates are as accurate, useful, and informative as
possible.
From an economic perspective, labor is an input into producing goods and services. If
regulation leads to more labor being used to produce a given amount of output, the additional
labor is reflected by an increase in the cost of production.2 When an increase in employment
occurs as a result of a regulation, it is a cost to firms. Moreover, when the economy is at full
employment, we would not expect an environmental regulation to have an impact on overall
employment because labor is being shifted from one sector to another. On the other hand, in
periods of high unemployment, employment effects (both positive and negative) are possible.
For example, an increase in labor demand due to regulation may result in a short-term net
increase in overall employment due to the potential hiring of previously unemployed workers
by the regulated sector to help meet new requirements (e.g., to install new equipment) or by
the pollution control sector to produce new abatement capital. When significant numbers of
workers are unemployed, the opportunity costs associated with displacing jobs in other sectors
are likely to be smaller. And, in general, if a regulation imposes high costs and does not
increase the demand for labor, it may lead to a decrease in employment.
To provide a partial picture of the employment consequences of this rule, EPA
investigates the expected consequences for the regulated sector and for the pollution control
sector. First, the analysis uses the results of Morgenstern, Pizer, and Shih (2002) to estimate
the effects of the regulation on the regulated industry, the electric power industry in this case.
This approach has been used by EPA previously in recent Regulatory Impact Analyses. Second,
EPA uses information derived from engineering studies and projections of pollution controls
from the power sector modeling to generate estimates of employment impacts to the pollution
control sector.
Section 6.2 discusses the estimates of the employment consequences in the electricity
sectors, using the Morgenstern, et al. approach. Section 6.3 estimates the employment
consequences in the pollution control sector.
2 It should be noted that if more labor must be used to produce a given amount of output, then this implies a
decrease in labor productivity. A decrease in labor productivity will cause a short-run aggregate supply curve to
shift to the left, and businesses will produce less, all other things being equal.
6-2
-------
6.2 Employment Impacts Primarily on the Regulated Industry: Morgenstern, Pizer, and
Shih (2002)
EPA examined possible employment effects within the electric utility sector using a
peer-reviewed, published study that explores historical relationships between industrial
employment and environmental regulations (Morgenstern, Pizer, and Shih, 2002). For context,
in 2007, the electric power generation, transmission and distribution sector (NAICS 2211) had
approximately 510,000 paid employees (according to the 2007 Economic Census). Estimates
from Morgenstern et. al. study have been applied in recent RIAs to derive the employment
effects of new regulations within the regulated industry. (See, for example, the Regulatory
Impact Analyses for the proposed MATS and final CSAPR regulations). With certain
qualifications, we believe that this study is relevant to this employment analysis, as it was for
the MATS proposal, since the pollution control strategies or measures that form the basis of the
cost inputs in the Morgenstern et al. analysis are primarily add-on or end-of-line pollution
controls, in general covering more than 70% of the abatement expenditures in most years and
industries analyzed as shown in Table 6-1. The analysis of control strategies presented in
Chapter 3 of this RIA are composed entirely of add-on or end-of-line pollution controls. Thus,
the cost inputs in the Morgenstern et al. analysis are consistent with the cost inputs that enter
into this analysis of employment impacts within the regulated industry for MATS. It should be
noted that the electric utility sector is less labor-intensive than the industries examined by
Morgenstern et al. (2002). To this extent, it is possible that the positive employment impact
estimates are high.
Table 6-1. Percent of Abatement Expenditures in Different PACE Studies from Add-On or
End-of-Line Control Measures3
Percent
Industry
Pulp and Paper
Plastics
Petroleum Refining
Iron and Steel
1979
84
85
72
96
1983
80
88
57
93
1988
61
75
63
94
1991
47
67
61
92
3 U.S. Bureau of the Census. Pollution Abatement Costs and Expenditures. Washington, DC: U.S. Government
Printing Office, various years. The pulp and paper industry is defined by SIC 2611 & 2621, plastics by SICE 282,
petroleum refining by SIC 2911, and iron and steel by SIC 332. For pulp and paper and iron and steel industries in
1983 the data is partially estimated based on non reported data due to disclosure reasons. The 1984 (1989) data
for plastics (pulp and paper) is used instead of 1983 (1988) due to a lack of reported data in the original year.
6-3
-------
Determining the direction of employment effects in the regulated industry is challenging
due to competing effects. A regulation that imposes costs may, for that reason, have an
adverse effect on employment, but if a regulation leads to the hiring of additional workers, it
may, for that reason, have a positive effect on employment. The fundamental insight of
Morgenstern, et al. is that environmental regulations can be understood as requiring regulated
firms to add a new output (environmental quality) to their product mixes. Although legally
compelled to satisfy this new demand, regulated firms have to finance this additional
production with the proceeds of sales of their other (market) products. Satisfying this new
demand requires additional inputs, including labor, and may alter the relative proportions of
labor and capital used by regulated firms in their production processes. Thus, Morgenstern et
al. decompose the overall effect of a regulation on employment into the following three
subcomponents:
• The "Demand Effect": higher production costs raise market prices, reducing
consumption (and production), thereby reducing demand for labor within the
regulated industry (an unambiguously negative effect);
• The "Cost Effect": Assuming that the capital/labor ratio in the production
process is held fixed, as production costs increase, plants use more of all inputs,
including labor, to maintain a given level of output. For example, in order to
reduce pollutant emissions while holding output levels constant, regulated firms
may require additional labor (an unambiguously positive effect);
• The "Factor-Shift Effect": Regulated firms' production technologies may be more
or less labor intensive after complying with a regulation (i.e., more/less labor is
required per dollar of output). "Environmental activities may be more labor
intensive than conventional production," meaning that "the amount of labor per
dollar of output will rise." However, activities may, instead, be less labor
intensive because "cleaner operations could involve automation and less
employment, for example." (p. 416) (ambiguous effect)
Decomposing the overall employment impact of environmental regulation into three
subcomponents clarifies the conceptual relationship between environmental regulation and
employment in regulated sectors, and permitted Morgenstern, et al. to provide an empirical
estimate of the net impact. For present purposes, the net effect is of particular interest, and is
the focus of our analysis.
The demand effect is expected to have an unambiguously negative effect on
employment, the cost effect to have an unambiguously positive effect on employment, and the
factor-shift effect to have an ambiguous effect on employment. Without more information with
6-4
-------
respect to the magnitudes of these three competing effects, it is not possible to predict the net
employment effect in the regulated sector.
Using plant-level Census information between the years 1979 and 1991, Morgenstern et
al. estimate the size of each effect for four highly polluting and regulated industries (petroleum,
plastic material, pulp and paper, and steel). On average across the four industries, each
additional $1 million ($1987) spending on pollution abatement results in a (statistically
insignificant) net increase of 1.55 (+/- 2.24) jobs. As a result, the authors conclude that
increases in pollution abatement expenditures can have positive effects on employment and do
not necessarily cause economically significant employment changes. The conclusion is similar to
Berman and Bui (2001), who found that increased air quality regulation in Los Angeles did not
cause in large employment changes.
Ideally, the EPA would first apply the methodology of Morgenstern et al. to current
pollution expenditure and market data for the regulated firms to identify the relationship
between abatement costs and employment, then use this relationship to extrapolate the effect
of new projected abatement costs on these firms. Unfortunately, current firm-level abatement
cost and market characteristics are not available. Therefore, the EPA has used the estimated
relationship from the Morgenstern et al. data to extrapolate the employment impact of the
new projected abatement costs without accounting for the industry and firm differences.
Since the Morgenstern, et al. parameter estimates are expressed in jobs per million
($1987)4 of environmental compliance expenditures, their study offers a transparent and simple
way to transfer estimates for other employment analysis. For each of the three job effects
outlined above, EPA used the Morgenstern et. al. four industry average parameters and
standard errors along with the estimated private compliance costs to provide a range (based on
the 95th percentile of results) of employment effects in the electricity sector associated with
the rule. By applying these estimates to annualized cost for the final rule for the electric power
sector as shown in Chapter 3 of this RIA ($9.60 billion in 2007$), the Agency estimated each
effect. The results are:
• Demand effect: -39,000 to +2,000 jobs in the directly affected sector with a
central estimate of-18,000;
The Morgenstern et al. analysis uses "production worker" as defined in the US Census Bureau's Annual Survey of
Manufactures (ASM) in order to define a job. This definition can be found on the Internet at
http://www.census.gov/manufacturing/asm/definitions/index.html
6-5
-------
• Cost effect: +4,000 to +21,000 jobs in the directly affected sector with a central
estimate of +12,000; and
• Factor-shift effect: +200 to +27,000 jobs in the directly affected sector with a
central estimate of+14,000.
EPA estimates the net employment effect to range from -15,000 to +30,000 jobs in the
directly affected sector with a central estimate of +8,000.5'6 EPA recognizes there will be other
employment effects that are not considered in the Morgenstern et al. study. For example,
employment in pollution control industries may increase as firms purchase more pollution
control equipment and services to meet the rule's requirements. EPA does provide such an
estimate of employment change later in this section in a separate analysis.
A defensible methodology for evaluating the employment impacts beyond the pollution
control and regulated sectors is not yet available, though as noted before, net effects on
employment are expected to be at or very close to zero for the economy overall under full
employment. Attempts to estimate such effects usually rely on input-output methodologies
that hold technologies and the proportion of various inputs constant over time, making them
inappropriate for estimating long run impacts of regulation.
6.2.1 Limitations
The Morgenstern et al. approach to employment analysis has the advantage of carefully
controlling for many possibly confounding effects in order to separate the effect of changes in
regulatory costs on employment. Although the Morgenstern et al. paper provides information
about the potential job effects of environmental protection programs, however, there are
several caveats associated with using those estimates to analyze the final rule. First, the
Morgenstern et al. estimates presented in Table 6-2 and used in EPA's analysis represent the
weighted average parameter estimates for a set of manufacturing industries (pulp and paper,
plastics, petroleum, and steel). Unfortunately this set of industries does not overlap directly
with the electric utility sector. Second, relying on Morgenstern et al. implicitly assumes that the
employment estimates derived from 1979-1991 data are still applicable. Third, the
methodology used in Morgenstern et al. assumes that regulations affect plants in proportion to
5 Since Morgenstern's analysis reports environmental expenditures in $1987, we make an inflation adjustment the
IPM costs using the ratio of the annual consumer price index, U.S. city, all items reported by the U.S. Bureau of
Labor Statistics: CPI1987/ CPI2007 = (113.6/207.3) = 0.55.
6 Net employment effect = 1.55x $9,600 million x 0.55. Given the 95% confidence interval for this effect, this
estimated net result is not statistically different from zero.
6-6
-------
their total costs. In other words, each additional dollar of regulatory burden affects a plant by
an amount equal to that plant's total costs relative to the aggregate industry costs. By
transferring the estimates, EPA assumes a similar distribution of regulatory costs by plant size
and that the regulatory burden does not disproportionately fall on smaller or larger plants. EPA
also assumes that the net employment impact can be linearly extrapolated from the abatement
cost (i.e., that every million 1987 dollars generates a central estimate of 1.55 jobs). Fourth, the
Morgenstern et al. analysis makes particular assumptions about the role of imports and the
effect of previous regulation on plant closures. While imports are not an issue for MATS, the
stringency of the current regulation is expected to result in a number of power plant closures
due to early retirement of coal-fired ECU capacity in 2015, as indicated in Chapter 3 of this RIA.
Finally, the Morgenstern et.al. methodology does not examine the effects of regulation
on employment in sectors related to, but outside of the regulated sector. However, it does
suggest that the relationship between the employment impact in any sector and increased
costs due to regulation is ambiguous.
Table 6-2. Employment Impacts Within the Regulated Industry Using Peer-Reviewed Study
Estimates using Morgenstern et al. (2002)
Estimates Using Morgenstern, et. al (2002)
Demand Effect Cost Effect Factor Shift Effect Net Effect
Change in Full-Time Jobs per Million -3.56 2.42 2.68 1.55
Dollars of Environmental Expenditure3
Standard Error 2.03 0.83 1.35 2.24
EPA Estimate for Ruleb -39,000 to+2,000 +4,000 to+21,000 +200 to+27,000 -15,000 to+30,000°
3 Expressed in 1987 dollars. Adjustment of dollars from 2007 to 1987 is accomplished through use of the annual
Consumer Price Index - All Urban Consumers, found on the Internet at
ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt. U.S. Department of Labor, Bureau of Labor Statistics.
Washington, D.C.
According to the 2007 Economic Census, the electric power generation, transmission and distribution sector
(NAICS 2211) had approximately 510,000 paid employees. Both the midpoint and range for each effect are
reported in the last row of the table.
CEPA has used this study to estimate the mean net employment impact of this rule, and provided the 95%
confidence interval results to reflect the high degree of uncertainty regarding the effect on employment within
the regulated industry. The confidence interval includes zero indicating we are uncertain as to the sign of the
effect, but the interval itself does reveal information on the magnitude of the effect.
6-7
-------
6.3 Employment Impacts of the MATS-Pollution Control Sector Approach by 20157
Regulations set in motion new orders for pollution control equipment and services.
New categories of employment have been created in the process of implementing
environmental regulations. When a regulation is promulgated, one typical response of industry
is to order pollution control equipment and services in order to comply with the regulation
when it becomes effective, while closure of plants that choose not to comply occurs after the
compliance date. With such a response by industry as a basis, this section presents estimates
for short term employment needed to design, construct, and install the control equipment in
the three or four year period leading up to the compliance date. Environmental regulation may
increase revenue and employment in the environmental technology industry. While these
increases represent gains for that industry, they are costs to the regulated industries required
to install the equipment. As with any pool of labor, the gross size of the labor pool does not
reflect the net impact on overall employment after adjusting for shifts in other sectors.
Regulated firms hire workers to design and build pollution controls. Once the
equipment is installed, regulated firms hire workers to operate and maintain the pollution
control equipment - much like they hire workers to produce more output. Of course, these
firms may also reassign existing employees to do these activities. Environmental regulations
also support employment in many basic industries. In addition to the increase in employment
in the pollution control industry (to fill increased orders for pollution control equipment placed
by the regulated sector), environmental regulations also support employment in industries that
provide intermediate goods to the pollution control industry. For example, an investment in
capital expenditures to reduce air pollution involves the purchase of abatement equipment.
The equipment manufacturers, in turn, order steel, tanks, vessels, blowers, pumps, and
chemicals to manufacture and install the equipment. A study by Bezdek, Wendling, and
DiPernab (2008) found that "investments in environmental protection create jobs and displace
jobs, but the net effect on employment is positive."8 The majority of the jobs associated with
added pollution controls (e.g., boilermakers, general construction workers, etc.) will provide
domestic employment opportunities, but some goods and services demanded and/or provided
to the pollution control industry (e.g., steel, cement, etc.) are internationally traded goods.
7 EPA expects that the installation of retrofit control equipment in response to the requirements of this proposal
will primarily take place within 3 years of the effective date of the final rule, but there may be a possibility that
some installations may occur within 4 years of the effective date.
8 Environmental protection, the economy, and jobs: National and regional analyses, Roger H. Bezdek, Robert M.
Wendling and Paula DiPerna, Journal of Environmental Management Volume 86, Issue 1, January 2008, Pages 63-
79.
6-8
-------
The focus of this part of the employment analysis is on short-term employment related
to the compliance actions of the affected entities. This analysis estimates the employment
impacts due to the increased demand for pollution control equipment in response to MATS.9
Results indicate that the MATS has the potential to result in a net increase of labor in these
industries, driven by the high demand for new pollution controls. Overall, the results of the
pollution control sector approach indicate that the MATS could support an increase of about
46,000 job-years10 by 2015.
6.3.1 Overall Approach and Methodology for Pollution Control Sector Approach
EPA developed estimates of the potential employment changes for the Pollution Control
Sector using a bottom-up engineering based methodology combined with macroeconomic data
on industrial output and productivity, to estimate employment impacts. The approach relies
heavily on the projections and costing analysis from the IPM model, which uses industry specific
data and assumptions to derive compliance costs and energy impacts (See Chapter 3). Central
to the approach are prior EPA studies on similar issues, and in particular, data and information
from extensive engineering studies that the Agency has commissioned.11 The analysis develops
employment estimates by relying on IPM projections from the MATS analysis for the specific
types of pollution control technologies expected to be installed to comply with the rule.12 More
specifically, the analysis includes estimates for the labor needed to design, manufacture and
install the needed pollution control equipment over the 3 to 4 years leading up to compliance
in 2015.
For construction labor, the labor needs are derived from an update to a 2002 EPA
resource analysis for building various pollution controls (FGD - Flue Gas Desulfurization or
scrubbers, SCR- selective catalytic reduction, ACI - activated carbon injection, DSI - dry sorbent
injection, and FF - Fabric Filters) and are further classified into different labor categories. These
categories include boilermakers, engineers and a catch-all "other" installation labor. For the
inputs needed (e.g., steel), the updated 2002 resource study was used to determine the steel
9 For more detail on methodology, approach, and assumptions, see Appendix 6A.
10 Numbers of job years are not the same as numbers of individual jobs, but represents the amount of work that
can be performed by the equivalent of one full-time individual for a year (or FTE). For example, 25 job years may
be equivalent to five full-time workers for five years, twenty-five full-time workers for one year, or one full-time
worker for twenty-five years.
11 Engineering and Economic Factors Affecting the Installation of Control Technologies for Multipollutant Strategies
EPA-600/R-02/073 (2002) and Engineering and Economic Factors Affecting the Installation of Control
Technologies - An Update (2011).
12 Detailed results from IPM for the MATS can be found in Chapter 3 of the RIA.
6-9
-------
demand for each MW of additional pollution control, combined with labor productivity data
from the Economic Census and BLS for relevant industries. More detail on methodology,
assumptions, and data sources can be found in Appendix 6A for this RIA. Projections from IPM
were used to estimate the incremental retrofit capacities projected in response to the final rule.
These additional pollution controls are shown in Table 6-3, and reflect the added pollution
controls needed to meet the requirements of the rule. Additional information on the power
sector impacts can be found in Chapter 3 of the RIA.
Table 6-3. Increased Pollution Control Installations due to MATS, by 2015 (GW)
Retrofit Type IPM Projected Additional Pollution Control
FGD 17
ACI 99
DSI 44
FF 102
6.3.2 Summary of Employment Estimates from Pollution Control Sector Approach
Table 6-4 shows the results of employment impacts resulting from the additional
demand for the aforementioned pollution controls. The results indicate that MATS could
support or create roughly 46,000 one-time job-years of increased cost of direct labor, driven by
the need to design and build the pollution control retrofits.
Table 6-4. Employment Effects Using the Pollution Control Sector Approach for the MATS
(in Job-Years)13
Employment Incremental Employment
One-Time Employment Changes for Construction
1. Boilermakers 20,000
2. Engineers 5,000
3. General Construction 21,000
Total 46,000
13 Numbers are rounded to nearest thousand. MATS is not anticipated to result in any notable new capacity in
response to the rule, and thus is not considered as part of this analysis.
6-10
-------
6.3.3 Other Employment Impacts of MA TS
In addition to the employment impacts estimated for the regulated sector and pollution
control sectors, there are likely to be other employment impacts associated with MATS. These
include changes resulting from labor needed to operate the needed pollution controls,
increased demand for materials used in pollution control operation, shifts in demand for fuel in
response to the rule, changes in employment resulting from additional coal retirements, and
changes in other industries due to changes in the price of electricity and natural gas.14 The EPA
has provided estimates of some of these effects below, which are discussed in more detail in
Appendix 6A. The most notable of those that the Agency is unable to estimate are the impacts
on employment as a result of the increase in electricity and other energy prices in the economy.
Nor is the Agency able to quantify all the employment changes in industries that support and
supply the pollution control industry. Because of this inability to estimate all the important
employment impacts, EPA neither sums the impacts that the Agency is able to estimate for
these other employment impacts or make any inferences of whether there is a net gain or loss
of employment across these categories. A summary of the other employment impacts can be
found in Table 6-5, with additional detail provided in Appendix 6A.
Table 6-5. Other Employment Impacts of MATS (in Job-Years)
Employment Impacts from Increased Demand for Pollution Control Operating Inputs
Lime (FGD) 280
Activated Carbon (ACI) 460
Trona (DSI) 3,130
Baghouse material (FF) 20
Employment Impacts from Pollution Control Operation 4,320
Employment Impacts from Retirements of Existing Coal Capacity (2,500)
Employment Impacts from Changes in Coal Demand (430)
Employment Impacts from Changes in Natural Gas Demand 670
Note: See Appendix 6A for more detail.
6.4 Summary of Employment Impacts
The employment approaches used by EPA rely on different analytical techniques and are
applied to different industries during different time periods, and they use different units of
14 The employment approaches used by EPA rely on different analytical techniques and are applied to different
industries during different time periods, and they use different units of analysis. These estimates should not be
summed because of the different metrics, length and methods of analysis. The Morgenstern estimates are used
for the ongoing employment impacts for the regulated entities (the electric power sector).
6-11
-------
analysis. These estimates should not be summed because of the different metrics, length and
methods of analysis. The Morgenstern estimates are used for the ongoing employment
impacts for the regulated entities (the electric power sector). The short term estimates for
employment needed to design, construct, and install the control equipment in the three or four
year period leading up to the compliance date are also provided. Finally some of the other
types of employment impacts that will be ongoing are estimated.
In Table 6-6, we show the employment impacts of the MATS as estimated by the
pollution control sector approach and by the Morgenstern approach.
Table 6-6. Estimated Employment Impact Table for the MATS
One Time (Construction
Annual (Reoccurring) During Compliance Period)
Pollution Control Sector approach3
Net Effect on Electric Utility Sector Employment
from Morgenstern et al. approach0
Not Applicable
8,000b
-15, 000 to +30,000d
46,000
Not Applicable
aThese one-time impacts on employment are estimated in terms of job-years. These employment estimates should
not be summed because of the different metrics, length and methods of analysis.
bThis estimate is not statistically different from zero.
°These annual or recurring employment impacts are estimated in terms of production workers as defined by the
US Census Bureau's Annual Survey of Manufacturers (ASM).
d95% confidence interval
6.5 Potential Effect of Electricity Price Increase on Economy-Wide Production Costs
As with any input into production, the new price of electricity, reflecting the costs of
MATS, will be absorbed in some fraction by industries that use electricity in their operations.
Firms can respond to price changes by making changes to their processes, raising their prices,
reducing production, etc. However, electricity expenditures are only a modest component of
overall economic activity.
On an expenditures-weighted basis, electricity comprises only 0.75% of total production
expenditures across all sectors in 2002 (BEA, 2007b, 2007c).15 As reported in Chapter 3, the
Retail Electricity Price Model forecasts a 3.1% increase in the contiguous U.S. electricity price in
2015 (see Table 3-12) as a result of MATS. Therefore, the upper estimate of the initial increase
15 The BEA's benchmark I/O summary-level data includes information on the share of expenditures by industry
spent on 135 commodity categories for 133 different sectors. These data provide a "comprehensive picture of
inner workings of the economy" (Stewart et al., 2007). For more detail, see BEA 2007a and 2009.
6-12
-------
in production costs across all sectors from direct electricity expenses is 0.023% ( =[0.031 *
0.0075]*100% ).16 This 0.023% increase in average production expenditures represents a
credible upper estimate of the average direct effect of higher electricity expenses because it
assumes that production, consumption, and input levels do not change in the economy.17 In
reality, we know that producers and consumers can often use less electricity-intensive
substitute goods and services to avoid a significant portion of these costs even in the short-run,
which would mitigate these production cost increases. We also know that producers of
intermediate goods and services that adjust to higher electricity prices can also make changes
that lead to price adjustments for final goods and services sectors (as discussed below, indirect
electricity price effects are not included in this illustrative estimate). Taking into account the
fact that these numbers represent an upper estimate of initial production costs from the direct
increase in electricity expenditures, EPA does not expect that increases in average production
expenses from direct electricity price changes in this range are sufficient to cause significant
shifts in overall economic activity outside the electricity sector and its major input markets.
Note that this per unit percent cost increase does not reflect other potential economic effects
of this rule. For example, the increased expenditures on pollution abatement equipment could
create more demand for labor in those industries. Alternatively, as producers switch away from
electricity-intensive inputs, the demand for other inputs may increase, changing the cost of
production for those factors of production.
This estimate has a number of limitations. First, as mentioned above, it reflects an
upper estimate on the initial change to the average production cost of goods and services from
direct electricity expenses because the calculation used to estimate these changes assumes
that production, consumption and input use will not change in response to higher electricity
prices. Second, as mentioned above, this analysis also does not account for the effect that
higher electricity prices have on the factors used by other sectors (e.g. the cost of components
and other inputs). For sectors that use both electricity and other energy-intensive inputs, the
effect of higher electricity prices on input prices can be important but applying the BEA data to
account for the indirect effect of electricity price pass-through on factors relies even more
16 Note that we are only performing simple calculations for upper estimate increases in per unit production costs
as a direct consequence of higher electricity prices. A modeling approach would require assumptions about
behavioral response to price changes, and we are assuming for this analysis that there is no behavioral response
to higher electricity prices.
17This means that all other inputs, including capital, labor, and materials are assumed to be fixed when generating
an upper estimate per unit production expense from direct electricity prices for the industries included in this
analysis.
6-13
-------
heavily on assumptions about the inability of sectors to change their factor mix in response to
relative factor prices changes. Third, important differences across sectors, regions and
consumer classes may be masked by the nationwide estimated average expenditure changes.
However, because the Retail Electricity Price Model does not estimate price changes for
different customer classes, and because the BEA data does not provide a regional
decomposition of the economic accounts suitable to calculating regional upper estimates on
per unit production expenses, regional and consumer class price differences cannot be
calculated. Similarly, there are sectors that will have a meaningfully higher or lower maximum
increase in average production expenditures within the context of the national average.
Fourth, the share of electricity used may have changed since 2002. In general, electricity
consumption per dollar of gross domestic product fell from 2002 through 2009, but electricity
expenditures relative to gross domestic product rose slightly over this time (EIA, 2011; BEA
2011). Not accounting for this change over time in expenditures on electricity may lead to a
slight underestimate of the increase in average production expenditures, averaged across the
entire economy, as a result of this rule.
While there are several caveats to this approach, this calculation suggests that
electricity prices under MATS are not expected to have a large impact on production costs for
the economy as a whole. Initial production cost impacts of less than 0.023% from direct
electricity expenditures are unlikely to lead to significant impacts on the overall economy and
would fall within the normal variability range of input price variation observed by producers in
the past. This is consistent with the overall history of the implementation of the Clean Air Act
(Jaffeetal., 1995).
This upper estimate of average initial production cost increases from direct electricity
expenditures cannot be used to estimate changes in employment as a result of the regulation,
either nationally or for individual sectors. First, as noted above, these calculations do not
account for the ability of the real economy to adjust to changes in price through input
substitution, technological innovation, or other means. It is necessary to account for changes in
production, consumption, and input use to estimate the change in total employment. Second,
this approach does not account for changes in consumer and producer behavior as they adjust
the quantity of goods and services supplied or demanded in all of the markets affected by the
regulation. Changes in employment (both increases and decreases) in downstream sectors will
reflect the balance of all of these interactions.
An evaluation of the employment impacts beyond the pollution control and regulated
sectors is not yet available, though as noted before, net effects on employment are expected to
6-14
-------
be at or very close to zero for the economy overall under full employment. In the case of this
rule, labor may be a complement or a substitute to electricity in production, depending on the
sector. It is also the case that environmental regulation may increase labor productivity by
improving health, which may increase employment (via an increase in overall economic
productivity, see the discussion in Chapter 5). Attempts to estimate such economy-wide effects
by holding technologies and the proportion of various inputs constant overtime are
inappropriate for estimating long run impacts of regulation and an inaccurate representation of
the behavior of real-world firms.
6.6 Estimating Social Cost and Economic Impacts
In the Transport Rule proposed in the summer of 2010 and in other rulemakings, EPA
used a different model to estimate the social cost and economic impacts of the regulatory
approach than the model applied in this RIA. That model, EPA's EMPAX, is a CGE model that
dynamically cascades the cost of a regulation through the entire economy. Since that rule was
proposed, a different version of EMPAX was used to estimate the social cost of the Clean Air Act
in a new EPA report entitled "The Benefits and Costs of the Clean Air Act from 1990 to 2020"
(EPA, 2010, herein referred to as the Section 812 report). This version of EMPAX accounts for
the benefits of reducing pollution on labor productivity and on the demand for health care,
which significantly influenced the model's estimates of the social cost and economic impacts of
the Clean Air Act relative to an analysis using EMPAX in which these benefits-related effects
were not accounted for. In December 2010, in its review of the 812 Report EPA's Science
Advisory Board (SAB) found that "The inclusion of benefit-side effects (reductions in mortality,
morbidity, and health-care expenditures) in a computable general equilibrium (CGE) model
represents a significant step forward in benefit-cost analysis" (SAB, 2010). A description of the
changes to the model and implications are described in detail in chapter 8 of the Section 812
report. EPA has determined that it needs to update the EMPAX model version used for RIAs to
account for these beneficial effects of reducing pollution prior to its use in any additional
regulatory analysis. The EMPAX model version used for the Section 812 report cannot be used
for this rulemaking because it contains energy and economic data that are consistent with the
multi-year timeframes and energy scenarios of the 812 study but not with the single target
year, analysis timeframe, and energy scenario most appropriate for this current rulemaking
analysis. For example, much of the energy data in the EMPAX model employed in the Section
812 report is from the Energy Information Administration's Annual Energy Outlook 2005. With
these impacts of reducing pollution on labor productivity and the demand for health care now
in the process of being incorporated into the model, the SAB's perspective on the desirability of
accounting for these effects in the CGE analysis for the 812 study, and the typical practice by
6-15
-------
EPA's Office of Air and Radiation of having analyses within RIAs to be consistent in design with
those included in the most recent available Section 812 report, EPA will not use EMPAX for this
RIA.
6.7 References
Berman, E., and L. T. M. Bui. 2001. Environmental Regulation and Labor Demand: Evidence from
the South Coast Air Basin.II Journal of Public Economics 79(2):265-295.
Bezdek, Roger H., Robert M. Wendlingand Paula DiPerna, Environmental protection, the
economy, and jobs: National and regional analyses. Journal of Environmental
Management Volume 86, Issue 1. January 2008, Pages 63-79.
Environmental Business International (EBI), Inc., San Diego, CA. Environmental Business Journal,
monthly (copyright), http://www.ebiusa.com/
Jaffe, Adam B., Steven R. Peterson, Paul R. Portney, Robert N. Stavins. (1995). "Environmental
Regulation and the Competitiveness of U.S. Manufacturing: What Does the Evidence Tell
Us?" Journal of Economic Literature, 33(1), pp. 132-163
Morgenstern, R. D., W. A. Pizer, and J. S. Shih. 2002. Jobs versus the Environment: An Industry-
Level Perspective.il Journal of Environmental Economics and Management 43(3):412-
436.
Stewart, Ricky L., Jessica Brede Stone and Mary L. Streitweisser. 2007. "U.S. Benchmark Input-
Output Accounts, 2002." Survey of Current Business. October, 2007.
US Bureau of Economic Analysis (BEA). 2007a. Appendix B.-Classification of Value Added and
Final Uses in the 2002 Benchmark Input-Output Accounts. Available in: 2002 Summary
Tables, 2002 Benchmark Input-Output Data. Retrieved from
http://www.bea.gov/industry/io benchmark.htm#2002data.
US Bureau of Economic Analysis (BEA). 2007b. Commodity-by-lndustry Direct Requirements
after Redefinitions, 2002. Available in: 2002 Summary Tables, 2002 Benchmark Input-
Output Data. Retrieved from
http://www.bea.gov/industry/io benchmark.htm#2002data.
US Bureau of Economic Analysis (BEA). 2007c. "The Use of Commodities by Industries after
Redefinitions, 2002." Available in: 2002 Summary Tables, 2002 Benchmark Input-Output
Data. Retrieved from http://www.bea.gov/industry/io bench ma rk.htm#2002data.
US Bureau of Economic Analysis (BEA). 2009. Concepts and Methods of the U.S. Input-Output
Accounts. Retrieved from: http://www.bea.gov/papers/pdf/IOmanual_092906.pdf.
6-16
-------
US Bureau of Economic Analysis (BEA). 2011. Gross Domestic Product. In National Income and
Product Account Tables (Table 1.1.5). Retrieved From
http://www.bea.gOV//national/nipaweb/DownSS2.asp.
US Bureau of the Census. 2007 Economic Census, Washington, DC: U.S. Government Printing
Office, http://www.census.gov/econ/census07/.
US Bureau of the Census. Pollution Abatement Costs and Expenditures Washington, DC: U.S.
Government Printing Office, various years.
US Bureau of Labor Statistics, United States Department of Labor. —Industry Labor Productivity
and Cost Data Tables, Annual Percent Changes. (2010).
US Energy Information Administration (EIA), Annual Energy Review 2010. —Coal Mining
Productivity By State and Mine Type.H
US Energy Information Agency (EIA). 2011. Electricity End Use, 1949-2010. In Annual Energy
Review 2010 (Table 8.9). Retrieved From
http://www.eia.gov/totalenergy/data/annual/index.cfmtfelectricitv
US Environmental Protection Agency. Engineering and Economic Factors Affecting the
Installation of Control Technologies for Multipollutant Strategies, EPA-600/R-02/073
(2002).
6-17
-------
APPENDIX 6A
EMPLOYMENT ESTIMATES OF DIRECT LABOR IN RESPONSE TO THE MERCURY AND AIR TOXICS
STANDARDS IN 2015
This appendix presents the short-term employment estimates of the Mercury and Air
Toxics Standards (MATS), henceforth referred to as the final MATS. The focus of the
employment analysis in this study is only on the first order employment impacts related to the
compliance actions of the affected coal-fired entities within the power sector.18 It does not
include the ripple effects of those impacts on the broader economy (i.e., the "multiplier"
effect), nor does it include the wider economy-wide effects of the changes to energy markets
(such as higher electricity prices).19 Moreover, this study provides only a static snapshot of the
impacts for 2015 and does not account for the dynamic adjustments of the affected entities as
they adapt to the final MATS, such as those arising from technological innovation and learning-
by-doing. This analysis is also independent of other techniques used by the U.S. EPA to
estimate certain employment effects of particular regulations.
The estimates of the employment impacts are divided into several categories: job gains
due to the increased demand for pollution control equipment; job losses due to retirements of
coal capacity; and job shifts due to changes in demand for fuels. The various employment
metrics can also be distinguished by one-time employment changes (e.g., pollution control
construction), and ongoing employment changes (e.g., fuel use changes and pollution control
operation or coal retirements). Results indicate that the final MATS has the potential to
provide significant short-term employment opportunities, primarily driven by the high demand
for new pollution control equipment. The employment gains related to the new pollution
controls are likely to be tempered by some losses due to certain coal retirements, although, as
discussed below, some of these workers who lose their jobs due to plant retirements could find
replacement employment operating the new pollution controls at nearby units. Finally, job
losses due to reduced coal demand are expected to be offset by job gains due to increased
natural gas demand, resulting in very small positive (i.e., less than three hundred) net change in
employment due to fuel demand changes. Overall, the preliminary results indicate that the
final MATS could support a net of slightly over 46,000 one-time job-years and a net of about
6,000 ongoing job-years in 2015. These results are summarized in Table 6A-1 below.
18 This analysis does not include potential employment effects resulting from projected impacts on oil/gas-fired
units.
19 For more detail on the economic impacts of the proposed rule, see Chapter 3 of the Regulatory Impact Analysis
accompanying final MATS.
6A-1
-------
Table 6A-1: Net Employment Changes for 2015 (job-years)a'b
New Pollution Control Equipment
Retirements of Generating Units
Changes in Fuel Use
Net Effect
One Time
46,120
-
-
46,120
Ongoing
8,210
(2,500)
240
5,950
Total job years of labor for controls projected to be installed by 2015. MATS is not anticipated to result in any
notable new capacity in response to the rule, and thus is not considered as part of this analysis.
b Totals may not add due to rounding.
The job-years estimated here is a snapshot of the first order employment effect of the
final MATS in 2015. While there is no temporal dimension to this study, some of these jobs are
likely to be spread over several years, and some will last longer. Most of the construction
related labor demand, for example, is expected to provide a short-term, temporary boost to
employment that could last two or three years, along with any "multiplier effects" (i.e.,
secondary employment supported in upstream sectors) that are not included in these job-year
estimates. Most of the operational labor needs and labor shifts resulting from fuel changes are
likely to be longer term. Thus, in terms of the impacts of the final MATS on economy-wide
employment over time, this analysis shows there could be a significant temporary increase to
employment levels starting well before 2015, which would likely recede thereafter as the
construction phases for the needed pollution controls wind down. Over time, the operational
jobs will continue to provide a small boost to employment over "business as usual" baseline
employment levels. Note that this synopsis does not account for other employment impacts of
the final MATS, such as those resulting from higher energy prices.
6A.1 Overall Approach
The estimates for the near-term employment effects of the final MATS utilize studies
conducted by EPA on engineering and resource requirements for various compliance activities,
such as installing pollution control equipment, switching fuels, or ceasing plant operations as
they become uneconomic. Some of the information used here was obtained from a 2002 EPA
engineering study for multi-pollutant control strategies.20 This study was also the basis for the
employment analysis for the proposed MATS. For the final MATS, EPA has undertaken a
separate study to update the 2002 analysis in order to refine and update the Agency's
understanding of the resource requirements (labor and materials) of various compliance
Engineering and Economic Factors Affecting the Installation of Control Technologies for Multi-pollutant
Strategies. EPA-600/R-02/073 (October, 2002).
6A-2
-------
activities, including estimates for newer pollution control equipment that were not included in
the 2002 analysis.21 For example, the 2002 study focused on pollution control technologies
that directly address S02 and NOX emissions, while the updated study also includes pollution
control technologies that reduce mercury emissions and hazardous air pollutants as well.
Collectively, these studies are referred to as EPA pollution control studies in this appendix. This
employment analysis is based on information from the updated study where available, as well
as data from the original 2002 study, where updated information was unavailable.
The basic approach involved using power sector projections and various energy market
implications under the final MATS from modeling using EPA's data and assumptions with the
Integrated Planning Model (IPM), along with data from secondary sources, to estimate the first
order employment impacts for 2015. Throughout this analysis, incremental employment is
measured in job-years, since there is no temporal dimension to this analysis.22 Also, this
appendix does not include estimates of total employment impacts over time, though there is a
distinction between short-term construction related labor needs and more long-term
operational labor needs for new pollution controls (though these operational labor
requirements are also measured in 2015 job-years only).
6A. 1.2 Employment Changes due to New Pollution Control Equipment
EPA's IPM projections for the final MATS policy case were used to estimate the
incremental pollution control demand. These are shown in Table 6A-2 below:23 Note that the
capacity estimates shown in Table 6A-2 do not include EPA's projections for ESP and FGD
upgrades on existing units. Because the engineering studies used in this employment analysis
do not include resource estimates for these technologies, EPA chose not to analyze the
employment impacts for these technologies. This exclusion is likely to understate the total
employment impacts for the final MATS.
21 Engineering and Economic Factors Affecting the Installation of Control Technologies - An Update. Andover
Technology Partners and ICF International. October 20, 2011.
22 A job-year is defined as the amount of work that can be performed by the equivalent of one full-time individual
for one year (or FTE).
23 According to IPM, there is some overlap between the different types of pollution control equipment demand at
individual facilities. To the extent that there could be some efficiency gains at plants installing multiple controls
due to economies of scale, the job estimates presented here could overstate the impacts.
6A-3
-------
Table 6A-2: Increased Pollution Control Demand due to the final MATS, 2015 (GW)
Pollution Control Type
IPM Projected Additional Pollution Control
Scrubbers (FGD)
Activated Carbon Injection (ACI)
Dry Sorbent Injection (DSI)
Fabric Filter (FF)25
17
99
44
102
The employment impacts due to increased pollution control demand are divided into
three categories, one of which is associated with the construction and installation labor
requirements, while the remaining two are associated with the resources required for the
ongoing operation of the pollution control equipment. The labor needed for constructing and
installing these controls are for construction-related sectors, such as boilermakers, engineers,
and other installation labor. The two categories of labor needs for ongoing resource
requirements include employment in sectors that supply resources needed to run these
pollution controls (such as reagents); and utility sector jobs to operate the control equipment.
The following sections discuss the approach for each:
• For the construction labor estimates, per-unit labor needs were taken from the
pollution control studies, which included man-hours required per M W for each of the
control technologies listed above. The total installation labor was then sub-divided into
different labor categories, such as boilermakers, engineers and a catch-all "other
installation labor", using estimated shares of the different labor types in the EPA
pollution control studies.
• For the longer term labor associated with operating the pollution controls, per-unit
estimates of the main resources needed for the particular types of equipment (see Table
6A-3 below for a list of the resources) are also taken from EPA's pollution control
In addition to the scrubber capacity shown in Table 2, EPA also projects an additional 2.6 GW of dry scrubbers on
units burning waste coal and pet. coke and have existing baghouses. This capacity is not included in the table
above, however, employment impacts associated with these controls are included in this appendix as discussed
below.
25 This number includes the total incremental fabric filters, as reported in Chapter 3. For the purpose of estimating
the construction jobs from fabric filters this appendix uses an incremental capacity of 84 GW (i.e., incremental
fabric filters that are standalone, or installed with DSI or ACI+Toxecon). To avoid double counting, the remaining
18 GW of fabric filters that represent those installed on units with dry scrubbers are excluded from the
employment analysis, under the assumption that the labor estimates for dry scrubbers include resources required
for both the scrubber as well as the fabric filter that goes with it.
6A-4
-------
studies. Resources needed for FF (such as the filter bags) were estimated from the
incremental Variable Operation and Maintenance (VOM) costs from EPA's IPM modeling
results.26 These were then multiplied by the incremental GW for each pollution control
to obtain the total (physical) quantity of resources needed. Total tonnage for each
resource was then converted to dollars of increased economic output for these
resources using price estimates developed by Sargent & Lundy for EPA's IPM Base Case
v4.10 modeling assumptions (see notes at the end of Table 6A-3 below). Finally, the
labor productivity for each particular sector was used to estimate the number of job-
years these could create in 2015. Labor productivities for each sector were adjusted to
account for increased worker productivity in 2015. Data for baseline worker
productivity and corresponding growth rates to account for future productivities came
from the Economic Census and the Bureau of Labor Statistics (BLS) estimates.27
• The final employment vector estimated was for the utility sector labor needed to
operate these pollution control equipment. This estimate was based on the incremental
Fixed Operation and Maintenance (FOM) costs from EPA's IPM modeling results,
excluding costs due to retirements. Thus, this study assumes that the FOM costs are a
reasonable proxy for the payroll costs that are part of the FOM costs in EPA's modeling
(FOM costs are defined as the operating and maintenance costs incurred by the utility,
such as those for payroll, irrespective of whether the equipment is operated). The FOM
costs were then translated into employment based on estimates of payroll per worker
for the power sector taken from the 2007 Economic Census and BLS estimates.28
26 Because FF requirements are not endogenously determined in IPM, it required a different approach than the
other controls.
27 Total value of shipments in 2007 and total employees were taken from 2007 Economic Census, Statistics by
Industry for Mining and Manufacturing sectors. The average annual growth rate of labor productivity was taken
from the Bureau of Labor Statistics. Average growth rate calculated for years 1992-2007, applied to 2007
productivity to determine 2015 estimates of productivity. See the Detailed Methodology section at the end for
more details about the data used for these calculations.
28 Same sources as other productivity estimates (2007 Economic Census and BLS), however, uses employees and
total payroll rather than revenue or value of shipments.
6A-5
-------
Table 6A-3: Estimated Pollution Control Resource Needs (Quantity and Prices Used)
Lime, FGD (tons)
Activated Carbon, ACI (tons)
Trona, DSI (tons)
Amount of Resource Used
1,490,391
184,771
10,667,613
Price Used ($/unit)
$9529
$1,500
$150
Price Sources:
• Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies Mercury Control Cost
Development Methodology FINAL", March 2011, Project 12301-009, Perrin Quarles Associates, Inc.
• Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies Dry Sorbent Injection
Cost Development Methodology FINAL", August 2010, Project 12301-007, Perrin Quarles Associates, Inc.
• Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies SDA FGD Cost
Development Methodology FINAL", August 2010, Project 12301-007, Perrin Quarles Associates, Inc.
6A. 1.3 Results
Table 6A-4 presents the estimated employment impacts in 2015 resulting from the
additional pollution controls needed to meet the final MATS requirements. According to this
analysis, these investments could provide the opportunity to support about 54,500 job-years to
design, construct, and operate the needed pollution control equipment in 2015. Note, some of
these jobs are expected to start before, and continue beyond 2015 (such as the resource
related job-years), but this analysis only provides a snapshot for 2015.
For FGD this study uses the price for Lime (Dry FGD) which is significantly greater than the Limestone (Wet FGD)
price. This price was used because EPA's modeling indicates most of the incremental FGD units are likely to be dry
scrubbers.
6A-6
-------
Table 6A-4: Jobs Due to Pollution Control Equipment under the final MATS (Job-years in
2015)
Jobs for Construction Incremental Employment
1. Boilermakers 20190
2. Engineers 5/060
3. General Construction 20 870
Sub-Total: 46jl2o
Jobs for Operation
Jobs from Increased Operating Resource Use
1. Lime (FGD) 280
2. Activated Carbon (ACI) 4gQ
3. Trona (DSI) 3 130
4. Baghouse material (FF) 20
Sub-Total: 3 gg0
Jobs for Pollution Control Operation 4 320
Total Labor: 54 330
Note: Totals may not add due to rounding
The number of job-years estimated for pollution control installation (i.e., "Jobs for
Construction" in Table 6A-4 above) is driven in large part by the demand for new FFs used in
EPA's modeling. As shown in Table 6A-2, up to 84 GW of new FF capacity is projected to come
online in 2015 due to the final MATS that are relevant for this employment analysis. The
demand for new FFs is estimated to contribute nearly 70 percent of the new employment
resulting from the installation of pollution controls. Moreover, of the labor needed due to
increased resource use, the Trona required for DSI is estimated to support higher number of
jobs than the other resources. This is because the DSI technology requires significantly higher
quantities of reagents than the other pollution controls, based on EPA's engineering estimates
and pollution control studies. The second highest resource-related employment gains would
likely come from the activated carbon needed in response to the final MATS.
Of the roughly 54,500 job-years estimated in Table 6A-4, about 4,300 job-years, or
about 8 percent, are estimated to occur within the utility sector for labor needed to operate
the pollution controls (referred to as "Jobs for Pollution Control Operation" in Table 6A-4). The
rest of the labor demand will benefit the pollution control industry and other economic sectors.
The increased demand for resources and chemicals needed to operate the pollution controls
will result in increased employment in sectors such as mining, chemicals, and other
6A-7
-------
manufacturing sectors. The majority of these first order employment effects, however, are
likely to benefit construction-related sectors, such as construction, boilermaker, heavy
engineering, and other heavy construction sectors, resulting from the construction and
installation of the new pollution controls at affected sources throughout the country.
6A.1.3.1 Employment Changes due to Coal Retirements
Employment changes due to incremental coal plant retirements were estimated by first
identifying the retiring coal units30 from EPA's modeling results (for the base and the final MATS
policy cases). EPA projects roughly 4.7 GW of additional coal retirements by 2015 with the final
MATS in place.31
In order to convert the retired coal capacity into potential employment losses, it was
assumed that changes in the operating costs for the retired coal units can be used as a proxy for
payroll expenditures and the lost economic output due to coal retirements. Thus, the changes
in the FOM costs for these particular retiring units were derived using EPA's IPM modeling
results, and converted to lost jobs using data from the Economic Census and BLS output/worker
estimates for the utility sector.32 Employment losses due to plant retirements will not only
affect those that are directly working at the plant (i.e., plant operators), but would also affect
administrative and other "back-office" workers for those utilities and their support
organizations. This appendix assumes that the FOM costs related to retiring plants are a good
proxy for these types of job losses.
Table 6A-5: Annual Job Losses due to Coal Capacity Retirements for 2015
FOM Decrease from Retirements (million) $288
Workers Per Million$ in payroll 8.7
Workers lost due to retirements (job-year): 2,500
30 Oil and gas steam unit emissions requirements, and potential retirements, were not directly included in EPA's
IPM modeling under the MATS policy scenario. An analysis of these units was conducted separately, and to the
extent that there may be some retirements of oil and gas units, then the estimates of potential job losses due to
retirements provided here will understate the employment losses.
31 Retirement estimates are based on IPM System Summary Reports from EPA's modeling runs. Where applicable,
data from IPM parsed outputs were adjusted to account for partial retirements reported in the parsed outputs.
32 The same specific sources as cited before, however, used workers and total payroll.
6A-8
-------
Results indicate there could potentially be about 2,500 job losses (measured in job-years
for 2015, but any net job losses under this category are likely to be permanent), due to coal
retirements. However, two mitigating factors could reduce the negative employment impacts
due to retirements. First, many of the retiring units are at plants that are likely to have other
units operating under the policy scenario. In such cases, some of the excess labor pool at the
retiring units could well be absorbed at other units within the same firm.
Second, as Table 6A-4 indicates, utilities are expected to have the need to fill about
4,300 additional job slots to operate the pollution controls needed to meet the requirements of
the final MATS. If workers with experience at existing coal facilities become available through
plant retirements, some of these workers could be absorbed in operating these new pollution
controls.
6A.1.3.2 Employment Changes due to Changes in Fuel Use
Employment impacts due to projected fuel use changes (coal and natural gas production
shifts) were estimated using EPA's modeling results. First, employment losses due to
reductions in coal demand were estimated using an approach similar to EPA's coal employment
analyses under Title IV of the Clean Air Act Amendments.33 Using this approach, EPA's
projected coal demand changes (in short tons) for various coal supplying regions were
converted to job-years using EIA data on regional coal mining productivity (in short tons per
employee hour), using 2008 labor productivity estimates.34'35
Results of the coal employment impacts of the final MATS are presented in Table 6A-6
below.
33 Impacts of the Acid Rain Program on Coal Industry Employment. EPA 430-R-01-002
March 2001.
34 From US Energy Information Administration (EIA) Annual Energy Review, Coal Mining Productivity Data. Used
2008.
35 Unlike the labor productivity estimates for various equipment resources which were forecasted to 2015 using
BLS average growth rates, we used the most recent historical productivity estimates for fuel sectors. In general,
labor productivity for the fuel sectors (both coal and natural gas) showed a significantly higher degree of
variability in recent years than the manufacturing sectors, which would have introduced a high degree of
uncertainty in forecasting productivity growth rates for future years.
6A-9
-------
Table 6A-6: Annual Employment Impacts Due To Changes in Coal Use for 2015
Coal by Region
Appalachia
Interior
West
Waste Coal
Net Total
Change in Coal Demand
(MM Tons)
(11.8)
19.9
(17.3)
(0.7)
(9.9)
Labor Productivity
2.91
4.81
19.91
5.96
--
Job-year Change
(1,950)
1,990
(420)
(60)
(430)
Notes: Used US national coal productivity for waste coal
Totals may not add due to rounding
For natural gas production, labor productivity per unit of natural gas was unavailable,
unlike coal labor productivities used above. Most secondary data sources (such as Census and
EIA) provide estimates for the combined oil and gas extraction sector. This appendix thus uses
an adjusted labor productivity estimate for the combined oil and gas sector that accounts for
the relative contributions of oil and natural gas in the total sector output (in terms of the value
of energy output in MMBtu). This estimate of labor productivity is then used with the
incremental natural gas demand for the final MATS to estimate the job-years for 2015.
Table 6A-7: Annual Employment Impact due to Changes in Fuel Use (2015)
Fuel Type Employment
Coal Job Years Lost (430)
Natural Gas
Incremental Natural Gas Use (MMBtu) 175,786,505
Labor Productivity (MMBtu/job-year) 261,840
Job-years gained 670
Net Employment Effects of Fuel use changes 240
Note: Totals may not add due to rounding
Thus, about 430 job losses in the coal mining sector are likely to be offset by about 670
job gains in the natural gas production related sectors, for a net effect of about 240 job-year
gains due to the changes in fuel use. The changes in coal mining employment is driven by a
significant increase in demand for Interior coal which leads to about the same amount of job
gains as is lost due to the decreased demand for Appalachian coal (see Table 6A-6 above). This,
6A-10
-------
coupled with the fact that there is likely to be some job gains due to increased demand for
natural gas, results in a small net job gain due to fuel use changes for the final MATS.
6A.2 Results Summary
Overall, the final MATS is expected to provide an increase to short-term employment
resulting from substantial investments in new pollution control equipment. For 2015, the
results indicate the final MATS could support or create around 46,000 job-years driven by the
need to design and construct the needed equipment. While there could be some employment
losses due to coal retirements that will likely have a negative effect on some utilities and the
coal mining sector, employment gains in pollution control operation activities and the natural
gas sector are likely to offset some of those losses. As previously discussed, this assessment
does not account for the long-run economy-wide effects of the final MATS.
6A.3 Detailed Methodology
This section provides more details on the data and approaches used to estimate the
employment impacts discussed above. The section also details the sources for individual data
elements.
6A.3.1 Pollution Control Equipment Labor
6A.3.1.1 Installation Labor
Table 6A-8: Installation Labor Requirement36
Pollution
Control Type
FGD37
ACI
DSI
Incremental GW
Installed
17
99
44
Man-
hours/MW
1730
10
55
Boilermakers (%)
40
50
50
Engineers (%)
20
17
17
Others (%)
40
33
33
36 See Chapter 3 for more detail.
37 EPA also projects 2.6 GW of dry scrubbers on waste coal and pet. coke units with existing baghouses, which are
not shown in this table. Employment impacts from these units, however, are included in the pollution control
construction figures, calculated using the capital cost for these controls ($220.7 MM) and estimated manhours/$
capital cost (0.00598) developed from the same example dry scrubbers used to find the manhours/MW of
capacity.
6A-11
-------
FF
102
780
45
48
Source: Engineering and Economic Factors affecting the Installation of Control Technologies - An Update, Andover
Technology Partners, 2011
Installation labor is estimated by using the incremental GW installed for each pollution
control type from EPA's modeling using IPM. This was then converted into total man-hours
needed for installation using estimates of man-hours/MW primarily from EPA's 2011 update on
pollution control technology,. Total man-hours for each pollution control type were then
converted into man-years assuming 2,080 working hours per year.
Total man-years for each pollution control type were then broken down into various
sectors using the percentages, shown in Table 6A-8. These percentages were estimated from
the 2002 study, updated from the 2011 study, where applicable.
6A.3.1.2 Operating Resource Labor
Table 6A-9: Resources Needed for Operation
Pollution
Control
Type
FGD
ACI
DSI
FF
Incremental
Total GW
17
99
44
102
Resource
(Units in
parenthesis)
Lime
(Tons/MWh)
Activated
Carbon
(Tons/MWh)
Trona
(Tons/MWh)
Bag-house
Resources
Usage Price
Estimates ($/unit)
0.013 95
0.00025 1,500
0.033 150
*Resource Labor
determined Using
VOM cost for FFs
Industry Assumed
for Productivity
Calculations
Lime Manufacturing
Other Chemical
Product
Manufacturing
Potash Soda and
Borate Mineral
Mining
Plastics Material
and Resin
Manufacturing
Productivity*
2.0
1.6
2.0
0.6
*Workers/$Million in Output, Forecasted to 2015
Sources: Usage:
• Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies Mercury Control Cost
Development Methodology FINAL", March 2011, Project 12301-009, Perrin Quarles Associates, Inc.
• Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies Dry Sorbent Injection
Cost Development Methodology FINAL", August 2010, Project 12301-007, Perrin Quarles Associates, Inc.
• Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies SDA FGD Cost
This number includes the total incremental fabric filters. For the purpose of estimating the construction jobs
from fabric filters, this analysis uses an incremental capacity of 84 GW. The remaining 18 GW represent fabric
filters installed with dry FGD units, which are excluded because the labor estimates for dry scrubbers includes the
labor for fabric filters that is installed in conjunction.
6A-12
-------
Development Methodology FINAL", August 2010, Project 12301-007, Perrin Quarles Associates, Inc.
Labor related to resources used in operating pollution control equipment was estimated
using the total incremental GW of pollution control capacity which was first converted to total
MWh of incremental capacity assuming 85 percent capacity factor. For each pollution control
type, the next step involved choosing the primary operating resource. This approach is
consistent with prior EPA's analyses on similar topics. The next step involved estimating the
resource needs by each control type, generally in tons of material using the resource usage
estimates as shown in Table 6A-9. Using the total usage for each pollution control input (in
tons) and associated average prices, total expenditure by each resource type was then
calculated. This total expenditure was then converted to labor using workers per $Million in
total output for the industry associated with producing each respective input material.39
6A.3.1.3 Operating Labor
Table 6A-10: Operating Labor Assumptions
Incremental FOM from IPM Parsed ($ Billion) 2.20
FF and other Capital Costs Included in FOM ($ Billion) 1.70
Remaining FOM used to find O&M Labor ($ Million) 496.5
Productivity* 8.7
*Workers per $Million in Payroll for Electricity Generating Sector, Forecast to 2015
Sources: Productivity from 2007 Economic Census and Growth Rate from BLS.
Labor requirement to operate the controls is estimated for all equipment types
combined, using the incremental FOM costs from IPM. The IPM incremental FOM cost estimate
included capital costs for fabric filters and scrubber improvement costs, which were first
subtracted to obtain the true FOM costs ($496.5 million). Resulting FOM cost estimate was
then converted to labor needs using the workers/$ Million in total payroll for the Electric
Generating Sector.
6A.3.2 Retirement Labor
Table 6A-11: Inputs to Labor from Retirements
39 Fabric filters follow a different pattern. Instead of a resource usage estimate, we used the VOM cost for FFs and
converted this to jobs using the workers per million dollars output for the relevant manufacturing industry sector.
6A-13
-------
Capacity of Incremental Retirements in SSR (MW)
4.7
O&M Decrease scaled to SSR Retirements ($MM) (To account for Partial Retirements)
288.2
Workers Per $Million in payroll, forecast to 2015
8.7
Sources: Productivity from 2007 Economic Census and Growth Rate from BLS.
Retirement labor was estimated by first identifying the retiring units from EPA's
modeling using IPM parsed outputs (using incremental retirements in the policy case). The next
step involved estimating the capacity of incremental retirements as well as the change in the
FOM costs due to these retirements. Because of the discrepancies between partial retirements
in EPA's parsed outputs and System Summary Reports (SSR), FOM costs were scaled
proportionately to reflect the lower SSR-based estimates, as shown in Table 6A-11 above. FOM
cost decreases were then converted to job-years lost due to retirements using workers per
$Million in payroll.
6A.3.3 fuel Use Labor
Table 6A-12: Inputs to Labor for Fuel Use
Coal by Region
Appalachia
Interior
West
Waste Coal
Natural Gas
2015 Incremental Fuel USE (Tons)
-11,770,000
19,870,000
-17,260,000
-700,000
EIA Total Natural Gas Production 2007 (TCP)
EIA Total Crude Oil Production 2007 (Barrels)
EIA Natural Gas Heat Content 2007 (Btu/cf)
EIA Petroleum Heat Content
(MMBtu/Barrel)
Total Crude Oil and Natural Gas Production (MMBtu)
Economic Census 2007 Oil and Gas Extraction Employees
MMBtu per Man-year for Oil
and Gas Extraction
2008 Short Tons/Employee hour
2.9
4.8
19.9
6.0
24.664
1,848,450,000
1,027
6.151
36,699,744,000
140,160
261,842
6A-14
-------
Incremental Natural Gas from IPM (TCP) 0.171
Incremental Natural Gas from IPM (Converted to MMBtu) 175,786,505
*Workers per $Million in Payroll for Electricity Generating Sector, Forecast to 2015
Note: Heat Contents from EIA are assumed to be for fuels used in Electric Power Sector
Sources: Short Tons per hour from US EIA, Coal Industry Annual. Total Production from 2009 EIA Annual Energy
Review. Heat Contents from EIA, Heat Content of Natural Gas Consumed and 2009 Annual Energy Review.
Employment Data from 2007 Economic Census.
Fuel use related employment impacts were estimated by using IPM results for
incremental changes in coal and natural gas use (policy case over the base case). For coal,
estimates of coal use in tons by region from IPM were used in conjunction with labor
productivity estimates from the EIA for each region (in short tons/ employee hour), to calculate
the change in job-hours needed. These were then converted to job-years, assuming 2,080
working hours per year. As discussed above, because of the high variability in coal mining labor
productivity in recent years, no attempt was made to forecast coal (and natural gas, for
consistency) productivities, instead the most recent historical estimates were used in this
appendix (which was the 2008 labor productivity for coal).
For natural gas, the first step was estimating labor productivity since such information
was not available directly from any reliable source. EIA production data from the Annual
Energy Review for natural gas and crude oil (in TCF and barrels, respectively), along with EIA
heat content estimates were used to find total crude oil and natural gas production in MMBtu
for 2007. Labor productivity in MMBtu per job-year for the Oil and Gas Extraction sector was
then estimated using data from the Census on oil and gas extraction employment. Then, the
incremental natural gas demand from EPA's IPM modeling results (in TCF) was converted to
MMBtu of natural gas demand using EIA data on natural gas heat content. This was then used
with the labor productivity estimated above to calculate the total job-years needed for
increased natural gas demand for the final MATS.
6A-15
-------
6A.4 References
Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for ARC Technologies
Mercury Control Cost Development Methodology FINAL", March 2011, Project 12301-
009, Perrin Quarles Associates, Inc.
Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies Dry
Sorbent Injection Cost Development Methodology FINAL", August 2010, Project 12301-
007, Perrin Quarles Associates, Inc.
Sargent & Lundy, "IPM Model - Revisions to Cost and Performance for APC Technologies SDA
FGD Cost Development Methodology FINAL", August 2010, Project 12301-007, Perrin
Quarles Associates, Inc.
United States Department of Labor. Bureau of Labor Statistics. Industry Labor Productivity and
Cost Data Tables, Annual Percent Changes. 2010.
US Census Bureau. 2007 and 2002 Economic Census, 2000 Annual Survey of Manufacturers,
Manufacturing and Mining: Detailed Statistics by Industry for the US and Utilities:
Summary Statistics for the US.
US Energy Information Administration (EIA). Annual Energy Review 2009, Assorted Data for
Coal, Natural Gas, and Petroleum. 2009.
US Energy Information Administration (EIA). Natural Gas 2009, Heat Content of Natural Gas
Consumed 2009.
US EPA. Clear Skies Act, Technical Package, Section D. "Projected Impacts on Generation and
Fuel Use". 2003.
US EPA. Clean Air Interstate Rule, Technical Support Document. "Boilermaker Labor Analysis
and Installation Timing." OAR-2003-0053 (March, 2005).
US EPA. "Impacts of the Acid Rain Program on Coal Industry Employment." EPA430-R-01-002
(March, 2001).
US EPA, Office of Research and Development. "Engineering and Economic Factors Affecting the
Installation of Control Technologies for Multi-pollutant Strategies." EPA-600/R-02/073
(October, 2002).
US EPA, Engineering and Economic Factors Affecting the Installation of Control Technologies:
An Update. Prepared by Andover Technology Partners and ICF International. October,
2011.
6A-16
-------
CHAPTER 7
STATUTORY AND EXECUTIVE ORDER ANALYSES
7.1 Introduction
This chapter presents discussion and analyses relating to Executive Orders and statutory
requirements relevant for the final Mercury and Air Toxics Standards (MATS). We discuss the
analysis conducted to comply with Executive Order (EO) 12866 and the Paperwork Reduction
Act (PRA) as well as potential impacts to affected small entities required by the Regulatory
Flexibility Act (RFA), as amended by the Small Business Regulatory Enforcement Fairness Act
(SBREFA). We also describe the analysis conducted to meet the requirements of the Unfunded
Mandates Reform Act of 1995 (UMRA) assessing the impact of the final rule on state, local and
tribal governments and the private sector. In addition, we address the requirements of EO
13132: Federalism; EO 13175: Consultation and Coordination with Indian Tribal Governments;
EO 13045: Protection of Children from Environmental Health and Safety Risks; EO 13211:
Actions that Significantly Affect Energy Supply, Distribution, or Use; the National Technology
Transfer and Advancement Act; EO 12898: Federal Actions to Address Environmental Justice in
Minority Populations and Low-Income Populations; and the Congressional Review Act.
7.2 Executive Order 12866: Regulatory Planning and Review and Executive Order 13563,
Improving Regulation and Regulatory Review
Under Executive Order (EO) 12866 (58 FR 51,735, 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 or adversely affect in a material way the economy, a sector of
the economy, productivity, competition, jobs, the environment, public health or safety, or
state, local, or tribal governments or communities. Accordingly, the EPA submitted this action
to the Office of Management and Budget (OMB) for review under Executive Orders 12866 and
13563 any changes in response to OMB recommendations have been documented in the
docket for this action. In addition, EPA prepared this Regulatory Impact Analysis (RIA) of the
potential costs and benefits associated with this action.
When estimating the human health benefits and compliance costs detailed in this RIA,
the EPA applied methods and assumptions consistent with the state-of-the-science for human
health impact assessment, economics and air quality analysis. The EPA applied its best
professional judgment in performing this analysis and believes that these estimates provide a
reasonable indication of the expected benefits and costs to the nation of this rulemaking. This
RIA describes in detail the empirical basis for the EPA's assumptions and characterizes the
7-1
-------
various sources of uncertainties affecting the estimates below. In doing what is laid out above
in this paragraph, the EPA adheres to EO 13563, "Improving Regulation and Regulatory
Review," (76 FR 3821; January 18, 2011), which is a supplement to EO 12866.
In addition to estimating costs and benefits, EO 13563 focuses on the importance of a
"regulatory system [that]...promote[s] predictability and reduce[s] uncertainty" and that
"identify[ies] and use[s] the best, most innovative, and least burdensome tools for achieving
regulatory ends." In addition, EO 13563 states that "[i]n developing regulatory actions and
identifying appropriate approaches, each agency shall attempt to promote such coordination,
simplification, and harmonization. Each agency shall also seek to identify, as appropriate,
means to achieve regulatory goals that are designed to promote innovation." We recognize that
the utility sector faces a variety of requirements, including ones under CAA section 110(a)(2)(D)
dealing with the interstate transport of emissions contributing to ozone and PM air quality
problems, with coal combustion wastes, and with the implementation of CWA section 316(b).
In developing today's final rule, the EPA recognizes that it needs to approach these rulemakings
in ways that allow the industry to make practical investment decisions that minimize costs in
complying with all of the final rules, while still achieving the fundamentally important
environmental and public health benefits that underlie the rulemakings.
A summary of the monetized costs, benefits, and net benefits for the final rule at
discount rates of 3 percent and 7 percent is the Executive Summary and Chapter 8 of this RIA.
7.3 Paperwork Reduction Act
The information collection requirements in this rule have been submitted for approval
to the Office of Management and Budget (OMB) under the Paperwork Reduction Act, 44 U.S.C.
3501 et seq. The information collection requirements are not enforceable until OMB approves
them.
The information requirements are based on notification, record keeping, and reporting
requirements in the NESHAP General Provisions (40 CFR part 63, subpart A), which are
mandatory for all owners and operators subject to national emission standards. These
recordkeeping and reporting requirements are specifically authorized by CAA section 114 (42
U.S.C. 7414). All information submitted to the EPA pursuant to the recordkeeping and reporting
requirements for which a claim of confidentiality is made is safeguarded according to Agency
policies set forth in 40 CFR Part 2, subpart B. This final rule requires maintenance inspections of
the control devices but would not require any notifications or reports beyond those required by
7-2
-------
the General Provisions. The recordkeeping provisions require only the specific information
needed to determine compliance.
The annual monitoring, reporting, and recordkeeping burden for this collection
(averaged over the first 3 years after the effective date of the standards) is estimated to be
$158 million. This includes 698,907 labor hours per year at a total labor cost of $49 million per
year, and total non-labor capital costs of $108 million per year. This estimate includes initial and
annual performance tests, semiannual excess emission reports, developing a monitoring plan,
notifications, and recordkeeping. Initial capital expenses to purchase monitoring equipment for
affected units are estimated at a cost of $231 million. This includes 504,629 labor hours at a
total labor cost of $35 million for planning, selection, purchase, installation, configuration, and
certification of the new systems and total non-labor capital costs of $196 million. All burden
estimates are in 2007 dollars and represent the most cost effective monitoring approach for
affected facilities.
An Agency may not conduct or sponsor, and a person is not required to respond to, a
collection of information unless it displays a currently valid OMB control number. The OMB
control numbers for the EPA's regulations are listed in 40 CFR Part 9. When this ICR is approved
by OMB, the Agency will publish a technical amendment to 40 CFR Part 9 in the Federal Register
to display the OMB control number for the approved information collection requirements
contained in this final rule.
7.4 Final Regulatory Flexibility Analysis
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 today's rule on small entities, small entity is
defined as: (1) a small business that is an electric utility producing 4 billion kilowatt-hours or
less as defined by NAICS codes 221122 (fossil fuel-fired electric utility steam generating units)
and 921150 (fossil fuel-fired electric utility steam generating units in Indian country); (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.
7-3
-------
Pursuant to section 603 of the RFA, the EPA prepared an initial regulatory flexibility
analysis (IRFA) for the proposed rule and convened a Small Business Advocacy Review Panel to
obtain advice and recommendations of representatives of the regulated small entities. A
detailed discussion of the Panel's advice and recommendations is found in the Panel Report
(EPA-HQ-OAR-2009-0234-2921). A summary of the Panel's recommendations is presented at
76 FR 24975.
As required by section 604 of the RFA, we also prepared a final regulatory flexibility
analysis (FRFA) for today's final rule. The FRFA addresses the issues raised by public comments
on the IRFA, which was part of the proposal of this rule. The FRFA is summarized below and in
the preamble.
7.4.1 Reasons Why Action Is Being Taken
In 2000, the EPA made a finding that it was appropriate and necessary to regulate coal-
and oil-fired electric utility steam generating units (EGUs) under Clean Air Act (CAA) section 112
and listed EGUs pursuant to CAA section 112(c). On March 29, 2005 (70 FR 15,994), the EPA
published a final rule (2005 Action) that removed EGUs from the list of sources for which
regulation under CAA section 112 was required. That rule was published in conjunction with a
rule requiring reductions in emissions of mercury from EGUs pursuant to CAA section 111, i.e.,
CAMR, May 18, 2005, 70 FR 28606). The Section 112(n) Revision Rule was vacated on February
8, 2008, by the U.S. Court of Appeals for the District of Columbia Circuit. As a result of that
vacatur, CAMR was also vacated and EGUs remain on the list of sources that must be regulated
under CAA section 112. This action provides the EPA's final NESHAP for EGUs.
7.4.2 Statement of Objectives and Legal Basis for Final Rules
MATS will protect air quality and promote public health by reducing emissions of HAP.
In the December 2000 regulatory determination, the EPA made a finding that it was
appropriate and necessary to regulate EGUs under CAA section 112. The February 2008 vacatur
of the 2005 Action reverted the status the rule to the December 2000 regulatory
determination. CAA section 112(n)(l)(A) and the 2000 determination do not differentiate
between EGUs located at major versus area sources of HAP. Thus, the NESHAP for EGUs will
regulate units at both major and area sources. Major sources of HAP are those that have the
potential to emit at least 10 tons per year (tpy) of any one HAP or at least 25 tpy of any
combination of HAP. Area sources are any stationary sources of HAP that are not major
sources.
7-4
-------
7.4.3 Summary of Issues Raised during the Public Comment Process on the IRFA
The EPA received a number of comments related to the Regulatory Flexibility Act during
the public comment process. A consolidated version of the comments received is reproduced
below. These comments can also be found in their entirety in the response to comment
document in the docket.
Comment: Several commenters expressed concern with the SBAR panel. Some believe Small
Entity Representatives (SERs) were not provided with regulatory alternatives including
descriptions of significant regulatory options, differing timetables, or simplifications of
compliance and reporting requirements, and subsequently were not presented with an
opportunity to respond. One commenter believes the EPA's formal SBAR Panel notification and
subsequent information provided by the EPA to the Panel did not include information on the
potential impacts of the rule as required by section 609(b)(l). Additional
commenters suggested that the EPA's rulemaking schedule put pressure on the SBAR Panel
through the abbreviated preparation for the Panel. Commenters also expressed concerns that
the EPA did not provide participants more than cursory background information on which to
base their comments. One commenter stated that the EPA did not provide deliberative
materials, including draft proposed rules or discussions of regulatory alternatives, to the SBAR
Panel members. One commenter stated the SBAR Panel Report does not meet the statutory
obligation to recommend less burdensome alternatives. The commenter suggested the EPA
panel members declined to make recommendations that went further than consideration or
investigation of broad regulatory alternatives, with the exception of those recommendations in
which the EPA rejected alternative interpretations of the CAA section 112 and relevant court
cases. Two stated that the EPA did not respond to the concerns of the small business
community, the SBA, or OMB, ignoring concerns expressed by the SER panelists. One
commenter believes the EPA failed to convene required meetings and hearings with affected
parties as required by law for small business entities. One commenter stated that the SERs'
input is very important because more than 90 percent of public power utility systems meet the
definition and qualify as small businesses under the SBREFA.
Response: The RFA requires that SBAR Panels collect advice and recommendations from SERs
on the issues related to:
• the number and description of the small entities to which the proposed rule will
apply;
7-5
-------
• the projected reporting, record keeping and other compliance requirements of the
proposed rule;
• duplication, overlap or conflict between the proposed rule and other federal rules;
and
• alternatives to the proposed rule that accomplish the stated statutory objectives and
minimize any significant economic impact on small entities.
The RFA does not require a covered agency to create or assemble information for SERs or for
the government panel members. While section 609(b)(4) requires that the government Panel
members review any material the covered agency has prepared in connection with the RFA, the
law does not prescribe the materials to be reviewed. The EPA's policy, as reflected in its RFA
guidance, is to provide as much information as possible, given time and resource constraints, to
enable an informed Panel discussion. In this rulemaking, because of a court-ordered deadline,
the EPA was unable to hold a pre-panel meeting but still provided SERs with the
information available at the time, held a standard Panel Outreach meeting to collect verbal
advice and recommendations from SERs, and provided the standard 14-day written comment
period to SERs. The EPA received substantial input from the SERs, and the Panel report
describes recommendations made by the Panel on measures the Administrator should consider
that would minimize the economic impact of the proposed rule on small entities. The EPA
complied with the RFA.
Comment: One commenter requested that the EPA work with utilities such that new
regulations are as flexible and cost efficient as possible.
Response: In developing the final rule, the EPA has considered all information provided prior to,
as well as in response to, the proposed rule. The EPA has endeavored to make the final
regulations flexible and cost efficient while adhering to the requirements of the CAA.
Comment: One commenter was concerned about the ability of small entities or nonprofit
utilities such as those owned and/or operated by rural electric co-op utilities, and municipal
utilities to comply with the proposed standards within three years. The commenter believes
that the EPA disregarded the SER panelists who explained that under these current economic
conditions they have constraints on their ability to raise capital for the construction of control
projects and to acquire the necessary resources in order to meet a three-year compliance
deadline. Two commenters expressed concern that smaller utilities and those in rural areas will
be unable to get vendors to respond to their requests for proposals, because they will be able
to make more money serving larger utilities.
7-6
-------
Response: The preamble to the proposed rule (76FR 25054, May 3, 2011) provides a detailed
discussion of how the EPA determined compliance times for the proposed (and final) rule. The
EPA has provided pursuant to section 112(i)(3)(A) the maximum three-year period for sources
to come into compliance. Sources may also seek a one-year extension of the compliance period
from their title V permitting authority if the source needs that time to install controls. CAA
section 112(i)(3)(B). If the situation described by commenters (i.e., where small entities or
nonprofit utilities constraints on ability to raise capital for construction of control projects and
to acquire necessary resources) results in the source needing additional time to install controls,
they would be in a position to request the one-year extension. The EPA discusses in more detail
in section VII of this preamble how the agency plans to address those units that are still unable
to comply within the statutorily mandated period.
Comment: Several commenters believe the EPA did not adequately consider the
disproportionately large impact on smaller generating units. The commenters note the
diseconomies in scale for pollution controls for such units. One commenter noted the rule will
create a more serious compliance hurdle for small communities that depend on coal-fired
generation to meet their base load demand. The commenter notes that by not subcategorizing
units, the EPA is dictating a fuel switch due to the disproportionately high cost on small
communities. The other commenter believes the MACT and NSPS standards are unachievable
by going too far without really considering the impacts on small municipal units, as public
powers is critical to communities, jobs, economic viability and electric reliability. A generating
and transmissions electric cooperative which qualifies as a small entity believes the rule will
ultimately result in increased electricity costs to its members and will negatively impact the
economies of the primarily rural areas that they serve. Another commenter believes there is no
legal or factual basis for creating subcategories or weaker standards for state, tribal, or
municipal governments or small entities that are operating obsolete units, particularly given the
current market situation and applicable equitable factors. The commenter suggests both the
EPA's and SBA's analyses focus exclusively on the effects on entities causing HAP emissions and
primarily on those operating obsolete EGUs, and fail to consider either impacts on downwind
businesses and governments or the positive impacts on small entities and governments owning
and operating competing, clean and modern EGUs.
Response: The EPA disagrees with the commenters' belief that the impacts on smaller
generating units were not adequately considered when developing the rule. The EPA
determined the number of potentially impacted small entities and assessed the potential
impact of the proposed action on small entities, including municipal units. A similar assessment
7-7
-------
was conducted in support of the final action. Specifically, the EPA estimated the incremental
net annualized compliance cost, which is a function of the change in capital and operating
costs, fuel costs, and change in revenue. The projected compliance cost was considered relative
to the projected revenue from generation. Thus, the EPA's analysis accounts not only for the
additional costs these entities face resulting from compliance, but also the impact of higher
electricity prices. The EPA evaluated suggestions from SERs, including subcategorization
recommendations. In the preamble to the proposed rule, the EPA explains that, normally, any
basis for subcategorizing must be related to an effect on emissions, rather than some
difference which does not affect emissions performance. The EPA does not see a distinction
between emissions from smaller generating units versus larger units. The EPA acknowledges
the comment that there is no legal or factual basis for creating subcategories or weaker
standards for state, tribal, or municipal governments or small entities that are operating
obsolete units.
Comment: One commenter notes that the EPA recognizes LEEs in the rule such that they should
receive less onerous monitoring requirements; however, the EPA does not recognize that small
and LEEs also need and merit more flexible and achievable pollution control requirements. The
commenter notes that the capital costs for emissions control at small utility units is
disproportionately high due to inefficiencies in Hg removal, space constraints for control
technology retrofits, and the fact that small units have fewer rate base customers across which
to spread these costs. The commenter cites the Michigan Department of Environmental Quality
report titled "Michigan's Mercury Electric Utility Workgroup, Final Report on Mercury Emissions
from Coal-Fired Power Plants," (June 2005). The commenter notes that the EPA has addressed
such concerns previously, citing the RIA for the 1997 8-hour ozone standard. The commenter
also suggests smaller utility systems generally have less capital to invest in pollution control
than larger, investor-owned systems, due to statutory inability to borrow from the private
capital markets, statutory debt ceilings, limited bonding capacity, borrowing limitations related
to fiscal strain posed by other, non-environmental factors, and other limitations.
Response: The EPA acknowledges that the rule contains reduced monitoring requirements for
existing units that qualify as LEEs. Although the EPA does not believe that reduced pollution
control requirements are warranted for LEEs, including small entity LEEs, we believe that
flexible and achievable pollution control requirements are promoted through alternative
standards, alternative compliance options, and emissions averaging as a means of
demonstrating compliance with the standards for existing EGUs.
7-8
-------
Comment: One commenter believes that the EPA should develop more limited monitoring
requirements for small EGUs. The commenter notes small entities do not possess the monetary
resources, manpower, or technical expertise needed to operate cutting-edge monitoring
techniques such as Hg CEMS and PM CEMs. The commenter notes the EPA could have
identified monitoring alternatives to the SER panel for consideration.
Response: The EPA provided monitoring alternatives to using PM CEMS, HCI CEMS, and Hg
CEMS in its proposed standards and in this final rule. The continuous compliance alternatives
are available to all affected sources, including small entities. As alternatives to the use of PM
CEMS and HCI CEMS, sources are allowed to conduct additional performance testing. Sorbent
trap monitoring is allowed in lieu of Hg CEMS.
Comment: Several commenters believe the EPA has not sufficiently complied with the
requirements of the RFA or adequately considered the impact this rulemaking would have on
small entities. One commenter believes the EPA has not engaged in meaningful outreach and
consultation with small entities and therefore recommends that the EPA seek to revise the
court-ordered deadlines to which this rulemaking is subject, re-convene the SBAR panel,
prepare a new initial regulatory flexibility analysis (IRFA), and issue it for additional public
comment prior to final rulemaking. The commenter believes the IRFA does not sufficiently
consider impacts on small entities as identified in the SBAR Panel Report. The commenter
believes it is not apparent that the EPA considered the recommendations of the Panel. The
commenter believes the description of significant alternatives in the IRFA is almost entirely
quoted from the SBAR Panel Report, which the commenter does not believe is an adequate
substitute for the EPA's own analysis of alternatives. The commenter also notes the EPA does
not discuss the potential impacts of its decisions on small entities or the impacts of possible
flexibilities. Where the EPA does consider regulatory alternatives in principle, the commenter
believes it does not provide sufficient support for its decisions to understand on what basis the
EPA rejected alternatives that may or may not have reduced burden on small entities while
meeting the stated objectives of the rule. Additionally, the commenter notes that the EPA did
not evaluate the economic or environmental impacts of significant alternatives to the proposed
rule. One commenter believes that the EPA's stated reasons for declining to specify or analyze
an area source standard are inadequate under the RFA. The commenter believes the EPA must
give serious consideration to regulatory alternatives that accomplish the stated objectives of
the CAA while minimizing any significant economic impacts on small entities and that the EPA
has a duty to specify and analyze this option or to more clearly state its policy reasons for
excluding serious consideration of a separate standard for area sources. A commenter believes
7-9
-------
the EPA did not fully consider the subcategorization of sources such as boilers designed to burn
lignite coals versus other fossil fuels, especially in regard to non-Hg metal and acid gas
emissions. The commenter references the SBAR Panel Report suggestion provided in the
preamble of the proposed rule that the EPA consider developing an area source vs. major
source distinction for the source category and the EPA's response. Another commenter is
concerned that the recommendations made by the SER participants were ignored and not
discussed in the rulemaking. Specifically, the commenter notes the EPA did not discuss
subcategorizing by age, type of plant, fuel, physical space constraints or useful anticipated life
of the plant. Nor did the EPA establish GACT for smaller emitters to alleviate regulatory costs
and operational difficulties. A commenter believes it is likely that different numerical or work
practice standards are appropriate for area sources of HAP.
Response: The EPA disagrees with one commenter's assertion that the agency has not complied
with the requirements of the RFA. The EPA complied with both the letter and spirit of the RFA,
notwithstanding the constraints of the court-ordered deadline. For example, the EPA notified
the Chief Counsel for Advocacy of the SBA of its intent to convene a Panel; compiled a list of
SERs for the Panel to consult with; and convened the Panel. The Panel met with SERs to collect
their advice and recommendations; reviewed the EPA materials; and drafted a report of Panel
findings. The EPA further disagrees with the commenter's assertion that the EPA's IRFA does
not sufficiently consider impacts on small entities. The EPA's IRFA, which is included in chapter
10 of the RIA for the proposed rule, addresses the statutorily required elements of an IRFA such
as, the economic impact of the proposed rule on small entities and the Panel's findings.
The EPA disagrees with the comment that recommendations made by the SERs and not
considered or discussed in the proposed rulemaking such as recommendations regarding
subcategorization and separate GACT standards for area sources. The preamble to the
proposed standards includes a detailed discussion of how the EPA determined which
subcategories and sources would be regulated (76 FR 25036-25037, May 3, 2011). In that
discussion, the EPA explains the rationale for its proposed subcategories based on five unit
design types. In addition, the EPA acknowledges the subcategorization suggestions from the
SERs and explains its reasons for not subcategorizing on those bases. The preamble to the
proposed standards also includes a discussion of the SERs' suggestion that area source EGUs be
distinguished from major-source EGUs and the EPA's reasons for not making that distinction (76
FR 25020-25021, May 3, 2011).
The EPA also disagrees with the suggestion that the Agency pursue an extension of the timeline
for final rulemaking such that the SBAR Panel can be reconvened and a new IRFA can be
7-10
-------
prepared and released for public comment prior to the final rulemaking. The EPA entered into a
Consent Decree to resolve litigation alleging that the EPA failed to perform a non-discretionary
duty to promulgate CAA section 112(d) standards for EGUs. American Nurses Ass'n v. EPA, 08-
2198 (D.D.C.). That Decree required the EPA to sign the final MATS rule by November 16, 2011,
unless the agency sought to extend the deadline consistent with the requirements of the
modification provision of the Consent Decree. If plaintiffs in the American Nurses litigation
objected to an extension request, which the EPA believes would have been likely based on their
comments on the proposed rule, the Agency would have had to file a motion with the Court
seeking an extension of the deadline. Consistent with governing case law, the Agency would
have been required to demonstrate in its motion for extension that it was impossible to finalize
the rule by the deadline provided in the Consent Decree. See Sierra Club v. Jackson, Civil Action
No. 01-1537 (D.D.C.) (Opinion of the Court denying EPA's motion to extend a consent decree
deadline). The EPA negotiated a 30-day extension and was able to complete the rule by
December 16, 2011; accordingly, the Agency had no basis for seeking a further extension of
time.
A detailed description of the changes made to the rule since proposal, including those made as
a result of feedback received during the public comment process can be found in sections VI
(NESHAP) and X (NSPS) in the preamble. Changes explained in the identified sections include
those related to applicability; subcategorization; work practices; periods of startup, shutdown,
and malfunction; initial testing and compliance; continuous compliance; and notification,
recordkeeping, and reporting.
7.4.4 Description and Estimate of the Affected Small Entities
For the purposes of assessing the impacts of MATS on small entities, a small entity is
defined as:
(1) A small business according to the Small Business Administration size standards by
the North American Industry Classification System (NAICS) category of the owning
entity. The range of small business size standards for electric utilities is 4 billion kilowatt
hours (kWh) of production or less;
(2) A small government jurisdiction that is a government of a city, county, town, district,
or special district with a population of less than 50,000; and
(3) A small organization that is any not for profit enterprise that is independently owned
and operated and is not dominant in its field.
7-11
-------
The EPA examined the potential economic impacts to small entities associated with this
rulemaking based on assumptions of how the affected entities will install control technologies
in compliance with MATS. The SBREFA analysis does not examine potential indirect economic
impacts associated with this rule, such as employment effects in industries providing fuel and
pollution control equipment, or the potential effects of electricity price increases on industries
and households.
The EPA used Velocity Suite's Ventyx data as a basis for identifying plant ownership and
compiling the list of potentially affected small entities. The Ventyx dataset contains detailed
ownership and corporate affiliation information. The analysis focused only on those EGUs
affected by the rule, which includes units burning coal, oil, petroleum coke, or coal refuse as
the primary fuel, and excludes any combustion turbine units or EGUs burning natural gas. Also,
because the rule does not affect combustion units with an equivalent electricity generating
capacity up to 25 megawatts (MW), small entities that do not own at least one combustion unit
with a capacity greater than 25 MW were removed from the dataset. For the affected units
remaining, boiler and generator capacity, heat input, generation, and emissions data were
aggregated by owner and then by parent company. Entities with more than 4 billion kWh of
annual electricity generation were removed from the list, as were municipal owned entities
serving a population greater than 50,000. For cooperatives, investor-owned utilities, and
subdivisions that generate less than 4 billion kWh of electricity annually but which may be part
of a large entity, additional research on power sales, operating revenues, and other business
activities was performed to make a final determination regarding size. Finally, small entities for
which the EPA's modeling with the Integrated Planning Model (IPM) does not project
generation in 2015 in the base case were omitted from the analysis because they are not
projected to be operating and, thus, are not projected to face the costs of compliance with the
rule. After omitting entities for the reasons above, the EPA identified a total of 82 potentially
affected small entities that are affiliated with 102 electric generating units.
7.4.5 Compliance Cost Impacts
This section presents the methodology and results for estimating the impact of MATS on
small entities in 2015 based on the following endpoints:
• annual economic impacts of MATS on small entities and
7-12
-------
• ratio of small entity compliance cost impacts to revenues from electricity
generation.1
7.4.5.1 Methodology for Estimating Impacts of MATS on Small Entities
EPA estimated compliance costs of MATS as follows:
^-Compliance = A C-operating+Capital + A Cpue| - A R
where C represents a component of cost as labeled, and A R represents the value of change in
electricity generation, calculated as the difference in revenues between the base case and
MATS.
Based on this formula, compliance costs for a given small entity could either be positive
or negative (i.e., cost savings) based on their compliance choices and market conditions. Under
MATS, some units will forgo some level of electricity generation (and, thus, revenues) to comply
and this impact will be lessened on those entities by the projected increase in electricity prices
under the MATS scenario (which raises their revenues from the remainder of their sales). On
the other hand, some units may increase electricity generation, and coupled with the increase
in electricity prices, will see an increase in electricity revenues resulting in lower net compliance
costs. If entities are able to increase revenue more than an increase in retrofit and fuel costs,
ultimately they will have negative net compliance costs (or savings). Because this analysis
evaluates the total costs as a sum of the costs associated with compliance choices as well as
changes in electricity revenues, it captures savings or gains such as those described. As a result,
what EPA describes as a cost is really more of a measure of the net economic impact of the rule
on small entities.
For this analysis, EPA used unit-level IPM parsed outputs - from modeling runs
conducted with EPA's base case v4.10_MATS assumptions - to estimate costs based on the
parameters above. These impacts were then summed for each small entity, adjusting for
ownership share.2 Net impact estimates were based on the following: changes in operating and
capital costs, driven mainly by retrofit installations or upgrades, change in fuel costs, and
1 This methodology for estimating small entity impacts has been used in recent EPA rulemakings such as the CSAPR
promulgated by EPA in July, 2011.
2 Unit-level cost impacts are adjusted for ownership shares for individual small entities, so as not to overestimate
burden on each company. If an individual unit is owned by multiple small entities, total costs for that unit to
meet MATS obligations are distributed across all owners based on the percentage of the unit owned by each
company. Ownership percentage was estimated based on the Ventyx database.
7-13
-------
change in electricity generation revenues under MATS relative to the base case. These
individual components of compliance cost were estimated as follows:
(1) Operating and capital costs: Using the IPM parsed outputs for the base case and
MATS policy case, EPA identified units that installed one or more pollution
control technologies under the rule. The equations for calculating operating and
capital costs were adopted from technology assumptions used in EPA's version
of IPM (version 4.10). The model calculates the capital cost (in $/MW); the fixed
operation and maintenance (O&M) cost (in $/MW-year); and the variable O&M
cost (in S/MWh).
(2) Fuel costs: Fuel costs were estimated by multiplying fuel input (in million British
thermal units, MMBtu) by region and fuel prices ($/MMBtu) from EPA's
modeling with IPM. The incremental fuel expenditures under MATS were then
estimated by taking the difference in fuel costs between MATS and the base
case.
(3) Value of electricity generated: EPA estimated the value of electricity generated
by multiplying the electricity generation from EPA's IPM modeling results with
the regionally-adjusted retail electricity price ($/MWh), for all entities except
those categorized as "Private" in Ventyx. For private entities, EPA used
wholesale electricity price instead of retail electricity price because most of the
private entities are independent power producers (IPP). IPPs sell their electricity
to wholesale purchasers and do not own transmission facilities and, thus, their
revenue was estimated based on wholesale electricity prices.
7.4.5.2 Results
The number of potentially affected small entities by ownership type and potential
impacts of MATS are summarized in Table 7-1. All costs are presented in 2007 dollars. EPA
estimated the annualized net compliance cost to small entities to be approximately $106
million in 2015.
7-14
-------
Table 7-1. Projected Impact of MATS on Small Entities in 2015
EGU Ownership
Type
Co-Op
IOU
Municipal
Sub-division
Private
Total
Number of
Potentially
Affected
Entities
19
8
42
9
4
82
Number of
Entities
Projected to
Withdrawal!
Affected Units
as Uneconomic
0
0
0
0
3
3
Total Net
Compliance
Costs (2007$
millions)
-29.7
33.0
49.7
44.8
8.4
106
Number of Small
Entities with
Compliance Cost >
1% of Generation
Revenues
9
7
16
4
4
40
Number of
Small Entities
with
Compliance
Cost > 3% of
Generation
Revenues
8
5
15
3
4
35
Notes: The total number of entities with costs greater than 1 percent or 3 percent of revenues includes only
entities experiencing positive costs. About 23 of the 82 total potentially affected small entities are estimated to
have cost savings under MATS (see text above for an explanation).
Definitions of ownership types are based on those provided by Ventyx's Energy Velocity.
Co-op (Cooperative): non-profit, customer-owned electric companies that generate and/or distribute electric
power.
IOU (Investor-Owned Utility): Includes Investor Owned assets (e.g., a marketer, independent power producer,
financial entity) and electric companies owned by stockholders, etc.
Municipal: A municipal utility, responsible for power supply and distribution in a small region, such as a city.
Sub-division: Political Subdivision Utility is a county, municipality, school district, hospital district, or any other
political subdivision that is not classified as a municipality under state law.
Private: Similar to investor-owned, but ownership shares are not openly traded on the stock markets.
Source: ICF International analysis based on IPM modeling results
EPA assessed the economic and financial impacts of the final rule using the ratio of
compliance costs to the value of revenues from electricity generation, and our results focus on
those entities for which this measure could be greater than 1 percent or 3 percent. Of the 82
small entities identified, EPA's analysis shows 40 entities may experience compliance costs
greater than 1 percent of base generation revenues in 2015, and 35 may experience
compliance costs greater than 3 percent of base revenues.3 Also, all generating capacity at 3
small entities is projected to be uneconomic to maintain. In this analysis, the cost of
withdrawing a unit as uneconomic is estimated as the base case profit that is forgone by not
operating under the policy case. Because 35 of the 82 total entities, or more than 40 percent,
are estimated to incur compliance cost greater than 3 percent of base revenues, EPA has
One percent and three percent of generation revenue criteria based on: "EPA's Action Development Process:
Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the Small Business Regulatory
Enforcement Fairness Act." OPEI Regulatory Development Series. November 2006. This can be found on the
Internet at http://www.epa.gov/sbrefa/documents/rfaguidancell-00-06.pdf.
7-15
-------
concluded that it cannot certify that there will be no SISNOSE for this rule. Results for small
entities discussed here, however, do not account for the reality that electricity markets are
regulated in parts of the country. Entities operating in regulated or cost-of-service markets
should be able to recover all of their costs of compliance through rate adjustments.
Note that the estimated costs for small entities are significantly lower than those
estimated by EPA for the MATS proposal (which were $379 million). This is driven by a small
group of units (less than 6 percent) which were projected to be uneconomic to operate under
the proposal (and hence incurred lost profits due to lost electricity revenues), but are now
projected to continue their operations under MATS. In addition, EPA's modeling indicates one
unit that would have operated at a low capacity factor under the base case would find it
economical to increase its generation significantly under MATS to meet electricity demand in its
region. Excluding this unit, the total cost impacts across all entities would be roughly $175
million. Changes in compliance behavior for this small group of units, in particular the one unit
which operates at a higher capacity factor, has a substantial impact on total costs for the entire
group as their increased generation revenues offsets a large portion of the compliance costs.
The separate components of annualized costs to small entities under MATS are
summarized in Table 7-2. The most significant components of incremental costs to these
entities are increased capital and operating costs for retrofits, followed by changes in electricity
revenues (i.e., cost savings).
7-16
-------
Table 7-2. Incremental Annualized Costs under MATS Summarized by Ownership Group and
Cost Category in 2015 (2007$ millions)
EGU Ownership
Type
Co-Op
IOU
Municipal
Sub-division
Private
Total
Capitak
Operating
Costs ($MM)
A
161.5
39.3
76.4
73.9
5.5
356
Fuel Costs
($MM)
B
86.4
0
1.9
2.2
0
91
Change in
Electricity
Revenue ($MM)
C
277.5
6.3
28.7
31.3
-2.9
341
Total
=A+B-C
-29.7
33.0
49.7
44.8
8.4
106
Note: Totals may not add due to rounding.
Source: ICF International analysis based on IPM modeling results
Capital and operating costs increase across all ownership types, but the direction of
changes in electricity revenues vary among ownership types. All ownership types, with the
exception of private entities, experience a net gain in electricity revenues under the MATS,
unlike projections from EPA's modeling during the proposal, where only municipals benefitted
from higher electricity revenues. The change in electricity revenue takes into account both the
profit lost from units that do not operate under the policy case and the difference in revenue
for operating units under the policy case. According to EPA's modeling, an estimated 274 MW
of capacity owned by small entities is considered uneconomic to operate under the policy case,
resulting in a net loss of $13 million (in 2007$) in profits. On the other hand, many operating
units actually increase their electricity revenue due to higher electricity prices under MATS. In
addition, as mentioned above, EPA's modeling indicates one unit finds it economical to increase
its capacity factor significantly under the policy case which results in significantly higher
revenues offsetting the costs.
7.4.6 Description of Steps to Minimize Impacts on Small Entities
Consistent with the requirements of the RFA and SBREFA, the EPA has taken steps to
minimize the significant economic impact on small entities. Because this rule does not affect
units with a generating capacity of less than 25 MW, small entities that do not own at least one
generating unit with a capacity greater than 25 MW are not subject to the rule. According to
the EPA's analysis, among the coal- and oil-fired EGUs (i.e., excluding combined cycle gas
turbines and gas combustion turbines) about 26 potentially small entities only own EGUs with a
7-17
-------
capacity less than or equal to 25 MW, and none of those entities are subject to the final rule
based on the statutory definition of potentially regulated units.
For units affected by the proposed rule, the EPA considered a number of comments
received, both during the Small Business Advocacy Review (SBAR) Panel and the public
comment period. While none of the alternatives adopted are specifically applied to small
entities, the EPA believes these modifications will make compliance less onerous for all
regulated units, including those owned by small entities.
7.4.6.1 Work Practice Standards
Consistent with Sierra Club v. EPA, the EPA proposed numerical emission standards that
would apply at all times, including during periods of startup and shutdown. After reviewing
comments and other data regarding the nature of these periods of operation, the EPA is
finalizing a work practice standard for periods of start up and shut down. The EPA is also
finalizing work practice standards for organic HAP from all subcategories of EGUs. The EPA has
chosen to finalize work practice standards because the significant majority of data for
measured organic HAP emissions from EGUs are below the detection levels of the EPA test
methods, and, as such, the Agency considers it impracticable to reliably measure emissions
from these units. Descriptions of the work practice requirements for startup and shutdown, as
well as organic HAP, can be found in Section VI.D-E. of the preamble.
7.4.6.2 Continuous Compliance and Notification, Record-keeping, and Reporting
The final rule greatly simplifies the continuous compliance requirements and provides
two basic approaches for most situations: use of continuous monitoring and periodic testing.
The frequency of periodic testing has been decreased from monthly in the proposal to quarterly
in the final rule. In addition to simplifying compliance, the EPA believes these changes
considerably reduce the overall burden associated with recordkeeping and reporting. These
changes to the final rule are described in more detail in Section VI.G-H. of the preamble.
7.4.6.3 Subcategorization
The Small Entity Representatives on the SBAR Panel were generally supportive of
subcategorization and suggested a number of additional subcategories the EPA should consider
when developing the final rule. While it was not practicable to adopt the proposed
subcategories, the EPA maintained the existing subcategories and split the "liquid oil-fired
units" subcategory into two individual subcategories - continental and non-continental units.
7-18
-------
7.4.6.4 MACT Floor Calculations
As recommended by the EPA SBAR Panel representative, the EPA established the MACT
floors using all the available ICR data that was received to the maximum extent possible
consistent with the CAA requirements. The Agency believes this approach reasonably ensures
that the emission limits selected as the MACT floors adequately represent the level of emissions
actually achieved by the average of the units in the top 12 percent, considering operational
variability of those units. Additionally, following proposal, the EPA reviewed and revised the
procedure intended to account for the contribution of measurement imprecision to data
variability in establishing effective emissions limits.
7.4.6.5 Alternatives Not Adopted
The EPA chose not to adopt several of the suggestions posed either during the SBAR
Panel or public comment period. The EPA did not propose a percent reduction standard as an
alternative to the concentration-based MACT floor. The percent reduction format for Hg and
other HAP emissions would not have addressed the EPA's desire to promote, and give credit
for, coal preparation practices that remove Hg and other HAP before firing. Also, to account for
the coal preparation practices, sources would be required to track the HAP concentrations in
coal from the mine to the stack, and not just before and after the control device(s), and such an
approach would be difficult to implement and enforce. Furthermore, the EPA does not believe
the percent reduction standard is in line with the Court's interpretation of the Clean Air Act
section 112 requirements. Even if we believed it was appropriate to establish a percent
reduction standard, we do not have the data necessary to establish percent reduction
standards for HAP, as explained further in the response to comments document.
The EPA chose not to establish GACT standards for area sources for a number of
reasons. The data show that similar HAP emissions and control technologies are found on both
major and area sources greater than 25 MWe, and some large units are synthetic area sources.
In fact, because of the significant number of well-controlled EGUs of all sizes, we believe it
would be difficult to make a distinction between MACT and GACT. Moreover, the EPA believes
the standards for area source EGUs should reflect MACT, rather than GACT, because there is no
essential difference between area source and major source EGUs with respect to emissions of
HAP.
The EPA chose not to exercise its discretionary authority to establish health-based
emission standards for HCL and other HAP acid gases. Given the limitations of the currently
available information (e.g., the HAP mix where EGUs are located, and the cumulative impacts of
7-19
-------
respiratory irritants from nearby sources), the environmental effects of HCI and the other acid
gas HAP, and the significant co-benefits from reductions in criteria pollutants the EPA
determined that setting a conventional MACT standard for HCI and the other acid gas HAP was
the appropriate course of action.
As required by section 212 of SBREFA, the EPA also is preparing a Small Entity
Compliance Guide to help small entities comply with this rule. Small entities will be able to
obtain a copy of the Small Entity Compliance guide at the following Web site:
http://www.epa.gov/airquality/powerplanttoxics/actions.html.
7.5 Unfunded Mandates Reform Act (UMRA) Analysis
Title II of the UMRA of 1995 (Public Law 104-4)(UMRA) establishes requirements for
federal agencies to assess the effects of their regulatory actions on state, local, and tribal
governments and the private sector. Under Section 202 of the UMRA, 2 U.S.C. 1532, EPA
generally must prepare a written statement, including a cost-benefit analysis, for any proposed
or final rule that "includes any Federal mandate that may result in the expenditure by State,
local, and tribal governments, in the aggregate, or by the private sector, of $100,000,000 or
more ... in any one year." A "Federal mandate" is defined under Section 421(6), 2 U.S.C. 658(6),
to include a "Federal intergovernmental mandate" and a "Federal private sector mandate." A
"Federal intergovernmental mandate," in turn, is defined to include a regulation that "would
impose an enforceable duty upon State, Local, or tribal governments," Section 421(5)(A)(i), 2
U.S.C. 658(5)(A)(i), except for, among other things, a duty that is "a condition of Federal
assistance," Section 421(5)(A)(i)(l). A "Federal private sector mandate" includes a regulation
that "would impose an enforceable duty upon the private sector," with certain exceptions,
Section 421(7)(A), 2 U.S.C. 658(7)(A).
Before promulgating an EPA rule for which a written statement is needed under Section
202 of the UMRA, Section 205, 2 U.S.C. 1535, of the UMRA generally requires EPA to identify
and consider a reasonable number of regulatory alternatives and adopt the least costly, most
cost-effective, or least burdensome alternative that achieves the objectives of the rule.
Moreover, section 205 allows EPA to adopt an alternative other than the least costly, most cost-
effective or least burdensome alternative if the Administrator publishes an explanation why
that alternative was not adopted.
In a manner consistent with the intergovernmental consultation provisions of Section
204 of the UMRA, EPA carried out consultations with the governmental entities affected by this
rule. EPA held meetings with states and tribal representatives in which the Agency presented its
7-20
-------
plan to develop a proposal and provided opportunities for participants to provide input as part
of the rulemaking process. EPA has also analyzed the economic impacts of MATS on
government entities and this section presents the results of that analysis. The UMRA analysis
does not examine potential indirect economic impacts associated with the rule, such as
employment effects in industries providing fuel and pollution control equipment, or the
potential effects of electricity price increases on industries and households.
7.5.1 Identification of Affected Government Entities
Using Ventyx data, EPA identified state- and municipality-owned utilities and
subdivisions that would be affected by this rule. EPA then used IPM parsed outputs (based on
EPA modeling assumptions) to associate these entities with individual generating units. The
analysis focused only on EGUs affected by MATS, which includes units burning coal, oil,
petroleum coke, or waste coal as the primary fuel, and excludes any combustion turbine units.
Entities that did not own at least one unit with a generating capacity of greater than 25 MW
were also removed from the dataset because of their exemption from the rule. Finally,
government entities for which EPA's modeling does not project generation in 2015 under the
base case were also exempted from this analysis, because they are not projected to operate
and are thus not projected to face compliance costs with this rule. Based on this, EPA identified
96 state, municipal, and sub-divisions affiliated with 172 electric generating units that are
potentially affected by MATS.
7.5.2 Compliance Cost Impacts
After identifying the potentially affected government entities, EPA estimated the impact
of MATS in 2015 based on the following:
• total impacts of compliance on government entities and
• ratio of government entity impacts to revenues from electricity generation.
7.5.2.1 Methodology for Estimating Impacts MATS on Government Entities
EPA estimated compliance costs of MATS as follows:
^Compliance = & Coperating+Capital + ^Fuel ~ A R
7-21
-------
where C represents a component of cost as labeled, and A R represents the retail value of
change in electricity generation, calculated as the difference in projected revenues between the
base case and MATS.
Based on this formula, compliance costs for a given government entity could either be
positive or negative (i.e., cost savings) based on their compliance choices and market
conditions. Under MATS, some units will forgo some level of electricity generation (and thus
revenues) to comply and this impact will be lessened on those entities by the projected
increase in electricity prices under MATS. On the other hand, some units may increase
electricity generation, and coupled with the increase in electricity prices, will see an increase in
electricity revenues resulting in lower net compliance costs. If entities are able to increase
revenue more than an increase in retrofit and fuel costs, ultimately they will have negative net
compliance costs (or savings). Because this analysis evaluates the total costs as a sum of the
costs associated with compliance choices as well as changes in electricity revenues, it captures
savings or gains such as those described. As a result, what EPA describes as a cost is really more
of a measure of the net economic impact of the rule on government entities.
For this analysis, EPA used unit-level data from IPM runs conducted with EPA's modeling
assumptions to estimate costs based on the parameters above. These impacts were then
aggregated for each government entity, adjusting for ownership share. Compliance cost
estimates were based on the following: changes in capital and operating costs, change in fuel
costs, and change in electricity generation revenues under MATS relative to the base case.
These components of compliance cost were estimated as follows:
(1) Capital and operating costs: Using EPA's modeling results for the base case and the
MATS policy case, EPA identified units that install control technology under this rule and
the technologies installed. The equations for calculating operating and capital costs
were adopted from EPA's version of IPM (version 4.10_MATS). The model calculates the
capital cost (in $/MW); the fixed operation and maintenance (O&M) cost (in
$/MW-year); and the variable O&M cost (in $/MWh)
(2) Fuel costs: Fuel costs were estimated by multiplying fuel input (MMBtu) by region and
fuel prices ($/MMBtu) from EPA's modeling. The change in fuel expenditures under
MATS was then estimated by taking the difference in fuel costs between MATS and the
base case.
7-22
-------
(3) Value of electricity generated: EPA estimated the value of electricity generated by
multiplying the estimated electricity generation from EPA's IPM modeling results with
the regional-adjusted retail electricity prices ($/MWh).
7.5.2.2 Results
As was done for the small entities analysis, EPA assessed the economic and financial
impacts of the rule using the ratio of compliance costs to the value of revenues from electricity
generation, and our results focus on those entities for which this measure could be greater than
1 percent or 3 percent of base revenues. EPA projects that 42 government entities will have
compliance costs greater than 1 percent of base generation revenue in 2015 and 32 may
experience compliance costs greater than 3 percent of base revenues. Overall, 6 units owned
by government entities are projected to be uneconomic to maintain.
The separate components of the annualized costs to government entities under MATS
are summarized in Table 7-3 below. The most significant components of incremental costs to
these entities are the increased capital and operating costs, followed by increases in electricity
revenues (i.e., a cost saving).
Table 7-3. Incremental Annualized Costs under MATS Summarized by Ownership Group and
Cost Category (2007$ millions) in 2015
EGU
Ownership
Type
Sub-Division
State
Municipal
Total
Capital Costs +
Operating
Costs($MM)
A
128.0
65.9
516.3
710
Fuel Costs
($MM)
B
50.7
1.2
45.4
97
Change in
Revenue
($MM)
C
106.4
32.7
374.3
513
Total
=A+B-C
72.3
34.4
187.4
294
Note: Totals may not add due to rounding.
Definitions of ownership types are based on those provided by Ventyx's Energy Velocity.
Municipal: A municipal utility, responsible for power supply and distribution in a small region, such as a city.
Sub-division: Political Subdivision Utility is a county, municipality, school district, hospital district, or any other
political subdivision that is not classified as a municipality under state law.
Source: ICF International analysis based on IPM modeling results
The number of potentially affected government entities by ownership type and
potential impacts of MATS are summarized in Table 7-4. All costs are reported in 2007$
millions. EPA estimated the annualized net compliance cost to government entities to be
approximately $294 million in 2015.
7-23
-------
Table 7-4. Summary of Potential Impacts on Government Entities under MATS in 2015
EGU
Ownership
Type
Sub-Division
State
Municipal
Total
Number of
Potentially
Affected
Entities
11
5
80
96
Number of
Entities
Withdrawing all
Affected units
0
0
0
0
Total Net Costs
of MACT
compliance ($
MM)
72.3
34.4
187.4
294
Number of
Government Entities
with Compliance Cost
> 1% of Generation
Revenues
5
4
33
42
Number of Government
Entities with Compliance
Cost > 3% of Generation
Revenues
4
3
25
32
Note: The total number of entities with costs greater than 1 percent or 3 percent of revenues includes only entities
experiencing positive costs. About 30 of the 96 total potentially affected government entities are estimated to
have cost savings under the MACT policy case (see text above for an explanation).
Source: ICF International analysis based on IPM modeling results
Capital and operating costs increase over all ownership types. All ownership types,
however, also experience a net gain in electricity revenue, mainly due to higher electricity
prices under the policy case. As described in the small entity analysis, the change in electricity
revenue takes into account both the profit lost from units that do not operate under the policy
case and the difference in revenue for operating units under the policy case. According to
EPA's modeling, an estimated 757 MW of electricity generation is estimated to be uneconomic
to operate under the policy case, accounting for about $20 million in lost profits. On the other
hand, many operating units actually increase their electricity revenue due to higher electricity
prices under the MATS policy scenario.
7.6 Executive Order 13132, Federalism
Under EO 13132, the EPA may not issue an action that has federalism implications, that
imposes substantial direct compliance costs, and that is not required by statute, unless the
Federal government provides the funds necessary to pay the direct compliance costs incurred
by state and local governments, or the EPA consults with state and local officials early in the
process of developing the final action.
The EPA has concluded that this action may have federalism implications, because it
may impose substantial direct compliance costs on state or local governments, and the Federal
government will not provide the funds necessary to pay those costs. Accordingly, the EPA
provides the following federalism summary impact statement as required by section 6(b) of EO
13132.
7-24
-------
Based on estimates in the RIA, provided in the docket, the final rule may have
federalism implications because the rule may impose approximately $294 million in annual
direct compliance costs on an estimated 96 state or local governments. Specifically, we
estimate that there are 80 municipalities, 5 states, and 11 political subdivisions (i.e., a public
district with territorial boundaries embracing an area wider than a single municipality and
frequently covering more than one county for the purpose of generating, transmitting and
distributing electric energy) that may be directly impacted by this final rule. Responses to the
EPA's 2010 ICR were used to estimate the nationwide number of potentially impacted state or
local governments. As previously explained, this 2010 survey was submitted to all coal- and oil-
fired EGUs listed in the 2007 version of DOE/EIA's "Annual Electric Generator Report," and
"Power Plant Operations Report."
The EPA consulted with state and local officials in the process of developing the rule to
permit them to have meaningful and timely input into its development. The EPA met with 10
national organizations representing state and local elected officials to provide general
background on the rule, answer questions, and solicit input.
7.7 Executive Order 13175, Consultation and Coordination with Indian Tribal
Governments
EPA has concluded that this action may have tribal implications. The EPA offered
consultation with tribal officials early in the regulation development process to permit them an
opportunity to have meaningful and timely input. Consultation letters were sent to 584 tribal
leaders and provided information regarding the EPA's development of this rule and offered
consultation. Three consultation meetings were held: December 7, 2010, with the Upper Sioux
Community of Minnesota; December 13, 2010, with the Moapa Band of Paiutes, Forest County
Potawatomi, Standing Rock Sioux Tribal Council, and Fond du Lac Band of Chippewa; January 5,
2011, with the Forest County Potawatomi and a representative from the National Tribal Air
Association. In these meetings, the EPA presented the authority under the CAA used to develop
these rules and an overview of the industry and the industrial processes that have the potential
for regulation. Tribes expressed concerns about the impact of EGUs on Indian country.
Specifically, they were concerned about potential Hg deposition and the impact on the water
resources of the Tribes, with particular concern about the impact on subsistence lifestyles for
fishing communities, the cultural impact of impaired water quality for ceremonial purposes,
and the economic impact on tourism. In light of these concerns, the Tribes expressed interest in
an expedited implementation of the rule. Other concerns expressed by Tribes related to how
the Agency would consider variability in setting the standards and the use of tribal-specific fish
7-25
-------
consumption data from the Tribes in our assessments. They were not supportive of using work
practice standards as part of the rule and asked the Agency to consider going beyond the MACT
floor to offer more protection for the tribal communities.
In addition to these consultations, the EPA also conducted outreach on this rule through
presentations at the National Tribal Forum in Milwaukee, Wl; phone calls with the National
Tribal Air Association; and a webinar for Tribes on the proposed rule. The EPA specifically
requested tribal data that could support the appropriate and necessary analyses and the RIA for
this rule. In addition, the EPA held individual consultations with the Navajo Nation on October
12, 2011; as well as the Gila River Indian Community, Ak-Chin Indian Community, and the Hopi
Nation on October 14, 2011. These Tribes expressed concerns about the impact of the rule on
the Navajo Generating Station (NGS), the impact on the cost of the water allotted to the Tribes
from the Central Arizona Project (CAP), the impact on tribal revenues from the coal mining
operations (i.e., assumptions about reduced mining if NGS were to retire one or more units),
and the impacts on employment of tribal members at both the NGS and the mine. More
specific comments can be found in the docket.
7.8 Protection of Children from Environmental Health and Safety Risks
This final rule is subject to EO 13045 (62 FR 19885, April 23,1997) because it is an
economically significant regulatory action as defined by EO 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
the standards on children.
Although this final rule is based on technology performance, the standards are designed
to protect against hazards to public health with an adequate margin of safety as described in
the preamble. The protection offered by this rule may be particularly important for children,
especially the developing fetus. As referenced in Chapter 4 of this RIA, "Mercury and Other HAP
Benefits Analysis," children are more vulnerable than adults to many HAP emitted by EGUs due
to differential behavior patterns and physiology. These unique susceptibilities were carefully
considered in a number of different ways in the analyses associated with this rulemaking, and
are summarized in the RIA.
7.9 Statement of Energy Effects
Our analysis to comply with EO 13211 (Statement of Energy Effects) can be found in
Section 3.16 of this RIA.
7-26
-------
7.10 National Technology Transfer and Advancement Act
Section 12(d) of the National Technology Transfer and Advancement Act (NTTAA) of
1995 (Public Law No. 104-113; 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, 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.
This rulemaking involves technical standards. The EPA cites the following standards in
the final rule: EPA Methods 1, 2, 2A, 2C, 2F, 2G, 3A, 3B, 4, 5, 5D, 17, 19, 23, 26, 26A, 29, SOB of
40 CFR Part 60 and Method 320 of 40 CFR Part 63. Consistent with the NTTAA, the EPA
conducted searches to identify voluntary consensus standards in addition to these EPA
methods. No applicable voluntary consensus standards were identified for EPA Methods 2F, 2G,
5D, and 19. The search and review results have been documented and are placed in the docket
for the proposed rule.
The three voluntary consensus standards described below were identified as acceptable
alternatives to EPA test methods for the purposes of the final rule.
The voluntary consensus standard American National Standards Institute (ANSI) /
American Society of Mechanical Engineers (ASME) PTC 19-10-1981, "Flue and Exhaust Gas
Analyses [Part 10, Instruments and Apparatus]" is cited in the final rule for its manual method
for measuring the 02, C02, and CO content of exhaust gas. This part of ANSI/ASME PTC 19-10-
1981 is an acceptable alternative to Method 3B.
The voluntary consensus standard ASTM D6348-03 (Reapproved 2010), "Standards Test
Method for Determination of Gaseous Compounds by Extractive Direct Interface Fourier
Transform (FTIR) Spectroscopy" is acceptable as an alternative to Method 320 and is cited in
the final rule, but with several conditions: (1) The test plan preparation and implementation in
the Annexes to ASTM D-6348-03, Sections Al through AS are mandatory; and (2) In ASTM
D6348-03 Annex A5 (Analyte Spiking Technique), the percent (%) R must be determined for
each target analyte (Equation A5.5). In order for the test data to be acceptable for a compound,
%R must be 70 % > R < 130%. If the %R value does not meet this criterion for a target
compound, the test data are not acceptable for that compound and the test must be repeated
for that analyte (i.e., the sampling and/or analytical procedure should be adjusted before a
7-27
-------
retest). The %R value for each compound must be reported in the test report, and all field
measurements must be corrected with the calculated %R value for that compound by using the
following equation: Reported Result = (Measured Concentration in the Stack x 100) / % R.
The voluntary consensus standard ASTM D6784-02, "Standard Test Method for
Elemental, Oxidized, Particle-Bound and Total Mercury in Flue Gas Generated from Coal-Fired
Stationary Sources (Ontario Hydro Method)," is an acceptable alternative to use of EPA Method
29 for Hg only or Method SOB for the purpose of conducting relative accuracy tests of mercury
continuous monitoring systems under this final rule. Because of the limitations of this method
in terms of total sampling volume, it is not appropriate for use in performance testing under
this rule. In addition to the voluntary consensus standards the EPA used in the final rule, the
search for emissions measurement procedures identified 16 other voluntary consensus
standards. The EPA determined that 14 of these 16 standards identified for measuring
emissions of the HAP or surrogates subject to emission standards in the final rule were
impractical alternatives to EPA test methods for the purposes of this final rule. Therefore, the
EPA does not intend to adopt these standards for this purpose. The reasons for this
determination for the 14 methods are discussed below, and the remaining 2 methods are
discussed later in this section.
The voluntary consensus standard ASTM D3154-00, "Standard Method for Average
Velocity in a Duct (Pitot Tube Method)," is impractical as an alternative to EPA Methods 1, 2,
3B, and 4 for the purposes of this rulemaking because the standard appears to lack in quality
control and quality assurance requirements. Specifically, ASTM D3154-00 does not include the
following: (1) proof that openings of standard pitot tube have not plugged during the test; (2) if
differential pressure gauges other than inclined manometers (e.g., magnehelic gauges) are
used, their calibration must be checked after each test series; and (3) the frequency and validity
range for calibration of the temperature sensors.
The voluntary consensus standard ASTM D3464-96 (Reapproved 2001), "Standard Test
Method Average Velocity in a Duct Using a Thermal Anemometer," is impractical as an
alternative to EPA Method 2 for the purposes of this rule primarily because applicability
specifications are not clearly defined, e.g., range of gas composition, temperature limits. Also,
the lack of supporting quality assurance data for the calibration procedures and specifications,
and certain variability issues that are not adequately addressed by the standard limit the EPA's
ability to make a definitive comparison of the method in these areas.
7-28
-------
The voluntary consensus standard ISO 10780:1994, "Stationary Source Emissions-
Measurement of Velocity and Volume Flowrate of Gas Streams in Ducts," is impractical as an
alternative to EPA Method 2 in this rule. The standard recommends the use of an L-shaped
pitot, which historically has not been recommended by the EPA. The EPA specifies the S-type
design which has large openings that are less likely to plug up with dust.
The voluntary consensus standard, CAN/CSA Z223.2-M86 (1999), "Method for the
Continuous Measurement of Oxygen, Carbon Dioxide, Carbon Monoxide, Sulphur Dioxide, and
Oxides of Nitrogen in Enclosed Combustion Flue Gas Streams," is unacceptable as a substitute
for EPA Method 3A because it does not include quantitative specifications for measurement
system performance, most notably the calibration procedures and instrument performance
characteristics. The instrument performance characteristics that are provided are non-
mandatory and also do not provide the same level of quality assurance as the EPA methods. For
example, the zero and span/calibration drift is only checked weekly, whereas the EPA methods
require drift checks after each run.
Two very similar voluntary consensus standards, ASTM D5835-95 (Reapproved 2001),
"Standard Practice for Sampling Stationary Source Emissions for Automated Determination of
Gas Concentration," and ISO 10396:1993, "Stationary Source Emissions: Sampling for the
Automated Determination of Gas Concentrations," are impractical alternatives to EPA Method
3A for the purposes of this final rule because they lack in detail and quality assurance/quality
control requirements. Specifically, these two standards do not include the following: (1)
sensitivity of the method; (2) acceptable levels of analyzer calibration error; (3) acceptable
levels of sampling system bias; (4) zero drift and calibration drift limits, time span, and required
testing frequency; (5) a method to test the interference response of the analyzer; (6)
procedures to determine the minimum sampling time per run and minimum measurement
time; and (7) specifications for data recorders, in terms of resolution (all types) and recording
intervals (digital and analog recorders, only).
The voluntary consensus standard ISO 12039:2001, "Stationary Source Emissions-
Determination of Carbon Monoxide, Carbon Dioxide, and Oxygen—Automated Methods," is not
acceptable as an alternative to EPA Method 3A. This ISO standard is similar to EPA Method 3A,
but is missing some key features. In terms of sampling, the hardware required by ISO
12039:2001 does not include a 3-way calibration valve assembly or equivalent to block the
sample gas flow while calibration gases are introduced. In its calibration procedures, ISO
12039:2001 only specifies a two-point calibration while EPA Method 3A specifies a three-point
calibration. Also, ISO 12039:2001 does not specify performance criteria for calibration error,
7-29
-------
calibration drift, or sampling system bias tests as in the EPA method, although checks of these
quality control features are required by the ISO standard.
The voluntary consensus standard ASTM D6522-00, "Standard Test Method for the
Determination of Nitrogen Oxides, Carbon Monoxide, and Oxygen Concentrations in Emissions
from Natural Gas-Fired Reciprocating Engines, Combustion Turbines, Boilers and Process
Heaters Using Portable Analyzers" is not an acceptable alternative to EPA Method 3A for
measuring CO and 02 concentrations for this final rule as the method is designed for
application to sources firing natural gas.
The voluntary consensus standard ASME PTC-38-80 R85 (1985), "Determination of the
Concentration of Particulate Matter in Gas Streams," is not acceptable as an alternative for EPA
Method 5 because ASTM PTC-38-80 is not specific about equipment requirements, and instead
presents the options available and the pros and cons of each option. The key specific
differences between ASME PTC-38-80 and the EPA methods are that the ASME standard: (1)
allows in-stack filter placement as compared to the out-of-stack filter placement in EPA
Methods 5 and 17; (2) allows many different types of nozzles, pitots, and filtering equipment;
(3) does not specify a filter weighing protocol or a minimum allowable filter weight fluctuation
as in the EPA methods; and (4) allows filter paper to be only 99 percent efficient, as compared
to the 99.95 percent efficiency required by the EPA methods.
The voluntary consensus standard ASTM D3685/D3685M-98, "Test Methods for
Sampling and Determination of Particulate Matter in Stack Gases," is similar to EPA Methods 5
and 17, but is lacking in the following areas that are needed to produce quality, representative
particulate data: (1) requirement that the filter holder temperature should be between 120oC
and 134oC, and not just "above the acid dew-point;" (2) detailed specifications for measuring
and monitoring the filter holder temperature during sampling; (3) procedures similar to EPA
Methods 1, 2, 3, and 4, that are required by EPA Method 5; (4) technical guidance for
performing the Method 5 sampling procedures, e.g., maintaining and monitoring sampling train
operating temperatures, specific leak check guidelines and procedures, and use of reagent
blanks for determining and subtracting background contamination; and (5) detailed equipment
and/or operational requirements, e.g., component exchange leak checks, use of glass cyclones
for heavy particulate loading and/or water droplets, operating under a negative stack pressure,
exchanging particulate loaded filters, sampling preparation and implementation guidance,
sample recovery guidance, data reduction guidance, and particulate sample calculations input.
7-30
-------
The voluntary consensus standard ISO 9096:1992, "Determination of Concentration and
Mass Flow Rate of Particulate Matter in Gas Carrying Ducts - Manual Gravimetric Method," is
not acceptable as an alternative for EPA Method 5. Although sections of ISO 9096 incorporate
EPA Methods 1, 2, and 5 to some degree, this ISO standard is not equivalent to EPA Method 5
for collection of PM. The standard ISO 9096 does not provide applicable technical guidance for
performing many of the integral procedures specified in Methods 1, 2, and 5. Major
performance and operational details are lacking or nonexistent and detailed quality
assurance/quality control guidance for the sampling operations required to produce quality,
representative particulate data (e.g., guidance for maintaining and monitoring train operating
temperatures, specific leak check guidelines and procedures, and sample preparation and
recovery procedures) are not provided by the standard, as in EPA Method 5. Also, details of
equipment and/or operational requirements, such as those specified in EPA Method 5, are not
included in the ISO standard, e.g., stack gas moisture measurements, data reduction guidance,
and particulate sample calculations.
The voluntary consensus standard CAN/CSA Z223.1-M1977, "Method for the
Determination of Particulate Mass Flows in Enclosed Gas Streams," is not acceptable as an
alternative for EPA Method 5. Detailed technical procedures and quality control measures that
are required in EPA Methods 1, 2, 3, and 4 are not included in CAN/CSA Z223.1. Second,
CAN/CSA Z223.1 does not include the EPA Method 5 filter weighing requirement to repeat
weighing every 6 hours until a constant weight is achieved. Third, EPA Method 5 requires the
filter weight to be reported to the nearest 0.1 milligram (mg), while CAN/CSA Z223.1 requires
reporting only to the nearest 0.5 mg. Also, CAN/CSA Z223.1 allows the use of a standard pitot
for velocity measurement when plugging of the tube opening is not expected to be a problem.
The EPA Method 5 requires an S-shaped pitot.
The voluntary consensus standard EN 1911-1,2,3 (1998), "Stationary Source Emissions-
Manual Method of Determination of HCI-Part 1: Sampling of Gases Ratified European Text-Part
2: Gaseous Compounds Absorption Ratified European Text-Part 3: Adsorption Solutions Analysis
and Calculation Ratified European Text," is impractical as an alternative to EPA Methods 26 and
26A. Part 3 of this standard cannot be considered equivalent to EPA Method 26 or 26A because
the sample absorbing solution (water) would be expected to capture both HCI and chlorine gas,
if present, without the ability to distinguish between the two. The EPA Methods 26 and 26A use
an acidified absorbing solution to first separate HCI and chlorine gas so that they can be
selectively absorbed, analyzed, and reported separately. In addition, in EN 1911 the absorption
7-31
-------
efficiency for chlorine gas would be expected to vary as the pH of the water changed during
sampling.
The voluntary consensus standard EN 13211 (1998), is not acceptable as an alternative
to the Hg portion of EPA Method 29 primarily because it is not validated for use with impingers,
as in the EPA method, although the method describes procedures for the use of impingers. This
European standard is validated for the use of fritted bubblers only and requires the use of a
side (split) stream arrangement for isokinetic sampling because of the low sampling rate of the
bubblers (up to 3 liters per minute, maximum). Also, only two bubblers (or impingers) are
required by EN 13211, whereas EPA Method 29 require the use of six impingers. In addition, EN
13211 does not include many of the quality control procedures of EPA Method 29, especially
for the use and calibration of temperature sensors and controllers, sampling train assembly and
disassembly, and filter weighing.
Two of the 16 voluntary consensus standards identified in this search were not available
at the time the review was conducted for the purposes of the final rule because they are under
development by a voluntary consensus body: ASME/BSR MFC 13M, "Flow Measurement by
Velocity Traverse," for EPA Method 2 (and possibly 1); and ASME/BSR MFC 12M, "Flow in
Closed Conduits Using Multiport Averaging Pitot Primary Flowmeters," for EPA Method 2.
Finally, in addition to the three voluntary consensus standards identified as acceptable
alternatives to EPA methods required in the final rule, the EPA is also specifying four voluntary
consensus standards in the rule for use in sampling and analysis of liquid oil samples for
moisture content. These standards are: ASTM D95-05 (Reapproved 2010), "Standard Test
Method for Water in Petroleum Products and Bituminous Materials by Distillation", ASTM
D4006-11, "Standard Test Method for Water in Crude Oil by Distillation", ASTM D4177-95
(Reapproved 2010), "Standard Practice for Automatic Sampling of Petroleum and Petroleum
Products, and ASTM D4057-06 (Reapproved 2011), "Standard Practice for Manual Sampling of
Petroleum and Petroleum Products."
Table 5, section 4.1.1.5 of appendix A, and section 3.1.2 of appendix B to subpart
UUUUU, 40 CFR Part 63, list the EPA testing methods included in the final rule. Under section
63.7(f) and section 63.8(f) of subpart A of the General Provisions, a source may apply to the EPA
for permission to use alternative test methods or alternative monitoring requirements in place
of any of the EPA testing methods, performance specifications, or procedures specified.
7-32
-------
7.11 Environmental Justice
7.11.1 EnvironmentalJustice Impacts
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 (EJ) part of their
mission by identifying and addressing, as appropriate, disproportionately high and adverse
human health or environmental effects of their programs, policies, and activities on minority
populations and low-income populations in the U.S.
The EPA has determined that this final rule will not have disproportionately high and
adverse human health or environmental effects on minority, low income, or indigenous
populations because it increases the level of environmental protection for all affected
populations.
This final rule establishes national emission standards for new and existing EGUs that
combust coal and oil. The EPA estimates that there are approximately 1,400 units located at
575 facilities covered by this final rule.
This final rule will reduce emissions of all the listed HAP that come from EGUs. This
includes metals (Hg, As, Be, Cd, Cr, Pb, Mn, Ni, and Se), organics (POM, acetaldehyde, acrolein,
benzene, dioxins, ethylene dichloride, formaldehyde, and PCB), and acid gases (HCI and HF). At
sufficient levels of exposure, these pollutants can cause a range of health effects including
cancer; irritation of the lungs, skin, and mucous membranes; effects on the central nervous
system such as memory and IQ loss and learning disabilities; damage to the kidneys; and other
acute health disorders.
The final rule will also result in substantial reductions of criteria pollutants such as CO,
PM, and S02. Sulfur dioxide is a precursor pollutant that is often transformed into fine PM
(PM2.5) in the atmosphere. Reducing direct emissions of PM2.5 and S02 will, as a result, reduce
concentrations of PM2.5 in the atmosphere. These reductions in PM2.5 will provide large health
benefits, such as reducing the risk of premature mortality for adults, chronic and acute
bronchitis, childhood asthma attacks, and hospitalizations for other respiratory and
cardiovascular diseases. (For more details on the health effects of metals, organics, and PM2.5,
please refer to Chapters 4 and 5 of this RIA.) This final rule will also have a small effect on
electricity and natural gas prices but has the potential to affect the cost structure of the utility
industry and could lead to shifts in how and where electricity is generated.
7-33
-------
Today's final rule is one of a group of regulatory actions that the EPA will take over the
next several years to respond to statutory and judicial mandates that will reduce exposure to
HAP and PM2.s, as well as to other pollutants, from EGUs and other sources. In addition, the
EPA will pursue energy efficiency improvements throughout the economy, along with other
Federal agencies, states and other groups. This will contribute to additional environmental and
public health improvements while lowering the costs of realizing those improvements.
Together, these rules and actions will have substantial and long-term effects on both the U.S.
power industry and on communities currently breathing dirty air. Therefore, we anticipate
significant interest in these actions from EJ communities, as well as many others.
7.11.1.1 Key EJ Aspects of the Rule
This is an air toxics rule; therefore, it does not permit emissions trading among sources.
Instead, this final rule will place a limit on the rates of Hg and other HAP emitted from each
affected ECU. As a result, emissions of Hg and other HAP such as HCI will be substantially
reduced in the vast majority of states. In some states, however, there may be small increases in
Hg and other HAP emissions due to shifts in electricity generation from EGUs with higher
emission rates to EGUs with already low emission rates. Hydrogen chloride emissions are
projected to increase at a small number of sources but that does not lead to any increased
emissions at the state level.
The primary risk analysis to support the finding that this final rule is both appropriate
and necessary includes an analysis of the effects of Hg from EGUs on people who rely on
freshwater fish they catch as a regular and frequent part of their diet. These groups are
characterized as subsistence level fishing populations or fishers. A significant portion of the
data in this analysis came from published studies of EJ communities where people frequently
consume locally-caught freshwater fish. These communities included: (1) White and black
populations (including female and poor strata) surveyed in South Carolina; (2) Hispanic,
Vietnamese and Laotian populations surveyed in California; and (3) Great Lakes tribal
populations (Chippewa and Ojibwe) active on ceded territories around the Great Lakes. These
data were used to help estimate risks to similar populations beyond the areas where the study
data was collected. For example, while the Vietnamese and Laotian survey data were collected
in California, given the ethnic (heritage) nature of these high fish consumption rates, we
assumed that they could also be associated with members of these ethnic groups living
elsewhere in the U.S. Therefore, the high-end consumption rates referenced in the California
study for these ethnic groups were used to model risk at watersheds elsewhere in the U.S. As a
result of this approach, the specific fish consumption patterns of several different EJ groups are
7-34
-------
fundamental to the EPA's assessment of both the underlying risks that make this final rule
appropriate and necessary, and of the analysis of the benefits of reducing exposure to Hg and
the other hazardous air pollutants.
The EPA's analysis of risks from consumption of Hg-contaminated fish is contained in
Chapter 4 of this RIA. The effects of this final rule on the health risks from Hg and other HAP are
presented in the preamble and in the RIA for this rule.
7.11.1.2 Potential Environmental and Public Health Impacts to Vulnerable Populations
The EPA has conducted several analyses that provide additional insight on the potential
effects of this rule on EJ communities. These include: (1) The socio-economic distribution of
people living close to affected EGUs who may be exposed to pollution from these sources; and
(2) an analysis of the distribution of health effects expected from the reductions in PM2.s that
will result from implementation of this final rule ("co-benefits").
Socio-economic distribution. As part of the analysis for this final rule, the EPA reviewed
the aggregate demographic makeup of the communities near EGUs covered by this final rule.
Although this analysis gives some indication of populations that may be exposed to levels of
pollution that cause concern, it does NOT identify the demographic characteristics of the most
highly affected individuals or communities. EGUs usually have very tall emission stacks; this
tends to disperse the pollutants emitted from these stacks fairly far from the source. In
addition, several of the pollutants emitted by these sources, such as a common form of
mercury and S02, are known to travel long distances and contribute to adverse impacts on the
environment and human health hundreds or even thousands of miles from where they were
emitted (in the case of elemental mercury, globally).
The proximity-to-the-source review is included in the analysis for this final rule because
some EGUs emit enough hazardous air pollutants such as Nickel or Chromium (VI) to cause
elevated lifetime cancer risks greater than 1 in a million in nearby communities. In addition, the
EPA's analysis indicates that there are localized areas with elevated levels of Hg deposition
around most U.S. EGUs.4
The analysis of demographic data used proximity-to-the-source as a surrogate for
exposure to identify those populations considered to be living near affected sources, such that
they have notable exposures to current hazardous air pollutant emissions from these sources.
The demographic data for this analysis were extracted from the 2000 census data which were
See Excess Local Deposition TSD for more detail.
7-35
-------
provided to the EPA by the US Census Bureau. Distributions by race are based on demographic
information at the census block level, and all other demographic groups are based on the
extrapolation of census block group level data to the census block level. The socio-demographic
parameters used in the analysis included the following categories: Racial (White, African
American, Native American, Other or Multiracial, and All Other Races); Ethnicity (Hispanic); and
Other (Number of people below the poverty line, Number of people with ages between 0 and
18, Number of people greater than or equal to 65, Number of people with no high school
diploma).
In determining the aggregate demographic makeup of the communities near affected
sources, the EPA focused on those census blocks within three miles of affected sources and
determined the demographic composition (e.g., race, income, etc.) of these census blocks and
compared them to the corresponding compositions nationally. The radius of three miles (or
approximately 5 kilometers) is consistent with other demographic analyses focused on areas
around potential sources. In addition, air quality modeling experience has shown that the area
within three miles of an individual source of emissions can generally be considered the area
with the highest ambient air levels of the primary pollutants being emitted for most sources,
both in absolute terms and relative to the contribution of other sources (assuming there are
other sources in the area, as is typical in urban areas). While facility processes and fugitive
emissions may have more localized impacts, the EPA acknowledges that because of various
stack heights there is the potential for dispersion beyond 3 miles. To the extent that any
minority, low income, or indigenous subpopulation is disproportionately impacted by the
current emissions as a result of the proximity of their homes to these sources, that
subpopulation also stands to see increased environmental and health benefit from the
emissions reductions called for by this rule.
The results of EPA's demographic analysis for coal fired EGUs are shown in Table 7-5.
The data indicate that affected sources are located in areas where the minority share of
the population living within a three mile buffer is higher than the national average by 12
percentage points or 48%. For these same areas, the percent of the population below the
poverty line is also higher than the national average by 4 percentage points or 31%. These
results are presented in more detail in the "Review of Proximity Analysis," February 2011, a
copy of which is available in the docket.
PM2.s (co-benefits) analysis. As mentioned above, many of the steps EGUs take to
reduce their emissions of air toxics as required by this final rule will also reduce emissions of
7-36
-------
PM and S02. As a result, this final rule will reduce concentrations of PM2.5 in the atmosphere.
Exposure to PM2.5 can cause or contribute to adverse health effects, such as asthma and heart
disease, that significantly affect many minority, low-income, and tribal individuals and their
communities. Fine PM (PM2.5) is particularly (but not exclusively) harmful to children, the
elderly, and people with existing heart and lung diseases, including asthma. Exposure can cause
premature death and trigger heart attacks, asthma attacks in children and adults with asthma,
chronic and acute bronchitis, and emergency room visits and hospitalizations, as well as milder
illnesses that keep children home from school and adults home from work. Missing work due to
illness or the illness of a child is a particular problem for people who have jobs that do not
provide paid sick days. Low-wage employees also risk losing their jobs if they are absent too
often, even if it is due to their own illness or the illness of a child or other relative. Finally, many
individuals in these communities lack access to high quality health care to treat these types of
illnesses. Due to all these factors, many minority and low-income communities are particularly
susceptible to the health effects of PM2.5 and receive a variety of benefits from reducing it.
We estimate that in 2016 the annual PM related benefits of the final rule for adults
include approximately 4,200 to 11,000 fewer premature mortalities, 2,800 fewer cases of
chronic bronchitis, 4,800 fewer non-fatal heart attacks, 2,600 fewer hospitalizations (for
respiratory and cardiovascular disease combined), 3.2 million fewer days of restricted activity
due to respiratory illness and approximately 540,000 fewer lost work days. We also estimate
substantial health improvements for children in the form of 130,000 fewer asthma attacks,
3,100 fewer emergency room visits due to asthma, 6,300 fewer cases of acute bronchitis, and
approximately 140,000 fewer cases of upper and lower respiratory illness.
We also examined the level of PM2.5 mortality risks prior to the implementation of the
rule according to race, income, and educational attainment. We then estimated the change in
PM2.5 mortality risk as a result of this final rule among people living in the counties with the
highest (top 5 percent) PM2.5 mortality risk in 2005. We then compared the change in risk
among the people living in these "high-risk" counties with people living in all other counties.
In 2005, people living in the highest risk counties and in the poorest counties were
estimated to be at substantially higher risk of PM2.5- related death than people living in the
other 95 percent of counties. This was true regardless of race; the difference between the
groups of counties for each race was large while the differences among races in both groups of
counties were very small. In contrast, the analysis found that people with less than high school
education were predicted to have significantly greater risk from PM2.5 mortality than people
with a greater than high school education. This was true both for the highest-risk counties and
7-37
-------
for the other counties. In summary, the analysis indicates that in 2005, educational status, living
in one of the poorest counties, and living in a high-risk county are associated with higher
estimated PM2.5 mortality risk while race is not.
Our analysis predicts that this final rule will likely significantly reduce the risk of PM2.5-
related premature mortality among all populations of different races living throughout the U.S.
compared to both 2005 and 2016 pre-rule (i.e., base case) levels. The analysis indicates that
people living in counties with the highest rates (top 5 percent) of PM2.5 mortality risk in 2005
receive the largest reduction in mortality risk after this rule takes effect. We also estimate that
people living in the poorest 5 percent of the counties will experience a larger reduction in PM2.5
mortality risk when compared to all other counties. More information can be found below in
section 7.11.3.
The EPA estimates that the benefits of the final rule are likely distributed among races,
income levels, and levels of education fairly evenly, although there is insufficient data to
generate different concentration response functions for each demographic group. However,
the analysis does indicate that this final rule in conjunction with the implementation of existing
or final rules (e.g., the Cross-State Air Pollution Rule) may help reduce the disparity in risk
between those in the highest-risk counties and the other 95 percent of counties for all races
and educational levels.
7-38
-------
Table 7-5. Comparative Summary of the Demographics within 5 Kilometers (3 Miles) of the Affected Sources (population in
millions)3
•vj
ID
Near source total (3 mi)
% of near source total
National total
% of national total
Population White
13.9 8.78
63%
285 215
75%
African
American
2.51
18%
35.0
12%
Native
American
0.10
1%
2.49
1%
Other or
Multiracial
2.52
18%
33.3
12%
Minority13
5.13
37%
70.8
25%
Hispanic
or
Latino"
2.86
21%
39.1
14%
Age
0-17
3.37
24%
77.4
27%
Age
65+
1.65
12%
35.4
12%
No High
School
Diploma
2.20
16%
36.7
13%
Below
Poverty
Line
2.43
17%
37.1
13%
Sources: The demographics are from the U.S. Census Bureau, 2000. Information on the facilities is from U.S. EPA.
a Racial and ethnic categories overlap and cannot be summed.
b The "Minority" population is the overall population (in the first row) minus white population (in the second row).
c The Census Bureau defines "Hispanic or Latino" as an ethnicity rather than a racial category, Hispanics or Latinos may belong to any race.
-------
7.11.1.3 Meaningful Public Participation
The EPA defines "environmental justice" to include meaningful involvement of all
people regardless of race, color, national origin, or income with respect to the development,
implementation, and enforcement of environmental laws, regulations, and policies. To promote
meaningful involvement, the EPA publicized the rulemaking via newsletters, EJ listserves, and
the internet, including the Office of Policy's (OP) Rulemaking Gateway Web site
(http://yosemite.epa.gov/opei/RuleGate.nsf/). During the comment period, the EPA discussed
the proposed rule via a conference call with communities, conducted a community-oriented
webinar on the proposed rule, and posted the webinar presentation on- line. The EPA also held
three public hearings to receive additional input on the proposal.
Once this rule is finalized, affected EGUs will need to update their Title V operating
permits to reflect their new emission limits, any other new applicable requirements, and the
associated monitoring and recordkeeping from this rule. The Title V permitting process provides
that when most permits are reopened (for example, to incorporate new applicable
requirements) or renewed, there must be opportunity for public review and comments. In
addition, after the public review process, the EPA has an opportunity to review the proposed
permit and object to its issuance if it does not meet CAA requirements.
7.11.1.4 Summary
This final rule strictly limits the emissions rate of Hg and other HAP from every affected
ECU. The EPA's analysis indicates substantial health benefits, including for vulnerable
populations, from reductions in PM2.s.
The EPA's analysis also indicates reductions in risks for individuals, including for
members of minority populations, who eat fish frequently from U.S. lakes and rivers and who
live near affected sources. Based on all the available information, the EPA has determined that
this final rule will not have disproportionately high and adverse human health or environmental
effects on minority, low income, or indigenous populations. The EPA is providing multiple
opportunities for EJ communities to both learn about and comment on this rule and welcomes
their participation.
7.11.2 Analysis of High Risk Sub-Populations
In addition to the previously described assessment of EJ impacts, EPA is providing a
qualitative assessment of sub-populations with particularly high potential risks of mercury
exposure due to high rates of fish consumption. These populations overlap in many cases with
7-40
-------
traditional EJ populations and would benefit from mercury reductions resulting from this rule.
This section describes the available information on consumption rates for subpopulations with
high fish consumption, and shows their locations in the U.S. Because of their high rates offish
consumption, reductions in mercury occurring in waterbodies where these populations catch
fish will have a larger IQ benefit for these populations relative to the general fish consuming
population.
Based on a detailed review of the literature, EPA identified several high-risk sub-
populations (Moya, 2004; Burger, 2002, Shilling et al., 2010, Dellinger, 2004). The analysis of
potentially high-risk groups focuses on six subpopulations:
• low-income African-American recreational/subsistence fishers in the Southeast
region5
• low-income white recreational/subsistence fishers in the Southeast region
• low-income female recreational/subsistence fishers
• Hispanic subsistence fishers
• Laotian subsistence fishers
• Chippewa/Ojibwe Tribe members in the Great Lakes area
These specific subpopulations were selected based on published empirical evidence of
particularly high self-caught freshwater fish consumption rates among these groups. Evidence
for the first three groups is based on a study by Burger (2002), which collected survey data from
a random sample of participants in the Palmetto Sportsmen's Classic in Columbia, SC. Of 458
respondents, 39 were black, 415 were white, and 149 were female. The sample size for the
black population is relatively small, which increases uncertainty, particularly in higher percentile
consumption rate values provided for this group. In this study, results are also split out for poor
respondents (0-20K$ annual income). These consumption rates are relatively high, particularly
for the higher percentiles. This observation forms the basis for our decision to assess a number
of the subsistence populations only for watersheds located in US Census tracts containing
members of source populations below the poverty line for the white and black populations.
5The low-income designation is based on Census 2000 estimates of populations living in poverty. The Southeast for
purposes of this analysis comprises Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North
Carolina, South Carolina, Tennessee, Virginia, and West Virginia.
7-41
-------
Evidence for the Hispanic and Laotian groups is based on a study by Shilling et al. (2010).
This study looks at subsistence fishing activity among ethnic groups associated with more
urbanized areas near the Sacramento and San Joaquin rivers in the Central Valley in CA. The
authors note that many of these ethnic groups relied on fishing in origin countries and bring
that practice here (e.g., Cambodian, Vietnamese and Mexican). The authors also note that fish
consumption rates reported here for specific ethnic groups (specifically Southeast Asian) are
generally in-line with rates seen in WA and OR studies. For the Chippewa population, we use
results from a study by Dellinger (2004), which gathered data on self-reported fish consumption
rates by Tribes in the Great Lakes area. Because fishing activity is highly variable across Tribes
(and closely associated with heritage cultural practices) we have not extrapolated fishing
behavior outside of the areas ceded to the Tribes covered in the study (regions in the vicinity of
the Great Lakes). The terms "subsistence" and "recreational" fishing are based on the
terminology used in these published studies to describe the population of interest. In general,
subsistence fishers are individuals whose primary objective in fishing is to acquire food for
household consumption. For recreational fishers, the primary objective is to enjoy the outdoor
activity; however, fish consumption is also often an objective.
Table 7-6. Reported Distributions of Self-Caught Freshwater Fish Consumption Rates
Among Selected Potentially High-Risk Subpopulations
Population
Self-Caught
Sample
Size
Freshwater Fish Consumption Rate
(g/day)
Mean
(Median)
90th (95th)
Percentile
Study
Low-income African-American
recreational/subsistence fishers in
Southeast
39
171(137)
446 (557)
Burger (2002)
Low-income white recreational/
subsistence fishers in Southeast
Low-income female recreational/
subsistence fishers
Hispanic subsistence fishers
Laotian subsistence fishers
Great Lakes tribal groups
a .
415
149
45
54
822
38.8
39.1
25.8
(15.3)
(11.6)
(19.1)
47.2 (17)
60 (
113b)
93 (129)
123 (173)
98a (155.9)
144.8a (265.8)
136.2a(213.1)a
Burger (2002)
Burger (2002)
Shilling etal. (2010)
Shilling etal. (2010)
Dellinger (2004)
deviation reported in study.
Standard deviation in parentheses, rather than median.
7-42
-------
Using county-level growth projections, there were an estimated 3.09 million low-income
African Americans in census tracts that have (1) at least one HUC-12 within 20 miles with a
mercury fish tissue concentration estimate and (2) at least 25 African-American inhabitants
living below the poverty level, and 3.56 million are projected to reside in these areas in 2016.
The geographic distribution of the expected 2016 population is shown in Figure 7-1. The total
low-income (below the poverty level) White population in the southeastern states was 3.26
million for 2005 and is projected to be 3.58 million in 2016. The geographic distribution of this
population for 2016 is shown in Figure 7-2. The total modeled low-income female population
was 18.4 million for 2005 and is projected to be 20.1 million for 2016. The geographic
distribution of the population modeled for 2016 is shown in Figure 7-3. The total modeled
Hispanic population was 19.6 million for 2005 and is projected to be 27.2 million in 2016. The
geographic distribution of the population modeled for 2016 is shown in Figure 7-4. The total
modeled Laotian population was 80,000 for 2005 and projected to be 137,500 in 2016. The
geographic distribution of the population modeled for 2016 is shown in Figure 7-5. The total
modeled Chippewa population used to simulate the distribution of IQ loss was 23,900 for 2005
and is projected to be 29,500 for 2016. The geographic distribution of the population modeled
for 2016 is shown in Figure 7-6.
7-43
-------
MODELED AFRICAN-AMERICAN POPULATION
BELOW THE POVERTY LEVEL BY CENSUS TRACT 2016 :
Afr Am below pov pop 2016
25-100
^B 101-500
^B 501 - 1000
^B 1001-4536
Census tract
• Excludes tracts that:
(1) have fewer than 25 low-income African-American
inhabitants in 2016 or
(2) are more than 20 miles from the nearest HUC-12
with at least one freshwater fish tissue mercury sample.
Figure 7-1. Projected African-American Population Below the Poverty Level by Census
Tract in the Southeast for 2016
7-44
-------
MODELED WHITE POPULATION
BELOW THE POVERTY LEVEL BY CENSUS TRACT 2016 '
white below pov pop 2016
25 - 200
HI 201 - so°
^B 501 -1000
^H 1001 • 7302
Census trad
* Excludes tracts that:
(1) have fewer than 25 low-income white inhabitants
in 2016 or
(2) are more than 20 miles from the nearest HUC-12
with at least one freshwater fish tissue mercury sample.
Figure 7-2. Projected White Population Below the Poverty Level by Census Tract in the
Southeast for 2016
7-45
-------
MODELED FEMALE POPULATION
BELOW THE POVERTY LEVEL BY CENSUS TRACT 2016 *
tot female below pov 2016
25 - 200
^H 201- 600
^H 501-1000
j^B 1001 -10727
Census tract
* Excludes tracts that:
(1) have fewer than 25 low-income female inhabitants
in 2016 or
(2) are more than 20 miles from the nearest HUC-12
with at least one freshwater fish tissue mercury sample.
Figure 7-3. Projected Female Population Below the Poverty Level by Census Tract for 2016
7-46
-------
MODELED HISPANIC POPULATION BY CENSUS TRACT 2016
* Excludes tracts that:
(1) have fewer than 25 Hispanic inhabitants in 2016 or
(2) are more than 20 miles from the nearest HUC-12
with at least one freshwater fish tissue mercury sample.
Figure 7-4. Modeled Hispanic Population by Census Tract for 2016
7-47
-------
MODELED LAOTIAN POPULATION BY CENSUS TRACT 2016 '
* Excludes tracts that:
(1) have fewer than 25 Laotian inhabitants in 2016 or
(2) are more than 20 miles from the nearest HUC-12
with at least one freshwater fish tissue mercury sample.
7-48
-------
MODELED CHIPPEWA POPULATION BY CENSUS TRACT 2016 '
* Excludes tracts that:
(1) have fewer than 25 Chippewa inhabitants in 2016 or
(2) are more than 20 miles from the nearest HUC-12
with at least one freshwater fish tissue mercury sample.
Figure 7-6. Modeled Chippewa Population by Census Tract in the Great Lakes Area for
2016.
7.11.3 Characterizing the Distribution of Health Impacts across Populations
EPA is developing new approaches and metrics to improve its characterization of the
impacts of EPA rules on different populations. This analysis reflects one such approach, which
explores two principal questions regarding the distribution of PM2.5-related benefits resulting
from the implementation of MATS:
1. What is the baseline distribution of PM2.5-related mortality risk for adults according
to the race, income and education of the population?
7-49
-------
2. How does MATS change the distribution of PM2.5 mortality risk among populations of
different races—particularly among those populations at greatest risk in the
baseline?6
In this analysis we estimated that PM2.5 mortality risk from the modeled scenarios is not
distributed equally throughout the U.S., or among populations of different levels of educational
attainment—though the level of PM2.5 mortality risk appears to be shared fairly equally among
populations of different races. We estimate that the air quality and PM2.5-related mortality risk
improvements achieved by MATS are relatively equally distributed among minority populations,
and that the rule reduces PM2.5 mortality risk the most among those populations at greatest
risk in the 2005 baseline we selected for this analysis. We note that while the methods used for
this analysis have been employed in recent EPA Regulatory Impact Assessments (EPA, 2011)
and are drawn from techniques described in the peer reviewed literature (Fann et al. 2011b)
EPA will continue to modify these approaches based on evaluation of the methods.
7.11.3.1 Methodology
The methods used here to describe the distribution of PM2.5 mortality impacts are
consistent with the approach used in the proposed MATS RIA (U.S. EPA, 2011a) and the final
CSAPR RIA (U.S. EPA, 2011b). As a first step, we estimate the level of PM2.5-related mortality risk
in each county in the continental U.S. based on 2005 air quality levels, which provides a
baseline distribution of risk which we use to identify populations with initial higher and lower
baseline PM2.5-related mortality risk. This portion of the analysis follows an approach
described elsewhere (Fann et al. 2011a, Fann et al. 2011b), wherein modeled 2005 PM2.5 levels
are used to calculate the proportion of all-cause mortality risk attributable to total PM2.5 levels
in each county in the Continental U.S. Within each county we estimate the level of all-cause
PM2.5 mortality risks for adult populations as well as the level of PM2.5 mortality risk according
to the race, income and educational attainment of the population.
Our approach to calculating the distribution of PM2.5 mortality risk across the population
is generally consistent with the benefits analysis conducted for the modeled scenario described
in Appendix 5C with two exceptions: the PM2.5 mortality risk coefficients used to quantify
impacts and the baseline mortality rates used to calculate mortality impacts (a detailed
discussion of how both the mortality risk coefficients and baseline incidence rates are used to
estimate the incidence of PM2.5-related deaths may be found in the Chapter 5 of the RIA). We
6 In this analysis we assess the change in risk among populations of different race, income 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.
7-50
-------
substitute risk estimates drawn from the Krewski et al. (2009) extended analysis of the ACS
cohort. In particular, we applied the all-cause mortality risk estimate random effects Cox model
that controls for 44 individual and 7 ecological covariates, using average exposure levels for
1999-2000 over 116 U.S. cities (Krewski et al. 2009) (RR=1.06, 95% confidence intervals 1.04—
1.08 per 10u.g/m3 increase in PM2.5).This mean relative risk estimate is identical to the Pope et
al. (2002) risk estimate applied for the benefits analysis (though the standard error around the
mean RR estimate is slightly narrower).
Within both this and other analyses of the ACS cohort (see Krewski et al. 2000),
educational attainment has been found to be inversely related to the risk of all-cause mortality.
That is, populations with lower levels of education (in particular, < grade 12) are more
vulnerable to PM2.5-related mortality. Krewski and colleagues note that "...the level of
education attainment may likely indicate the effects of complex and multifactorial
socioeconomic processes on mortality...," factors that we would like to account for in this EJ
assessment. When estimating PM mortality impacts among populations according to level of
education, we applied PM2.5 mortality risk coefficients modified by educational attainment: less
than grade 12 (RR = 1.082, 95% confidence intervals 1.024—1.144 per 10 u.g/m3 change), grade
12 (RR = 1.072, 95% confidence intervals 1.020—1.127 per 10 u.g/m3 change), and greater than
grade 12 (RR = 1.055, 95% confidence intervals 1.018—1.094 per 10 u.g/m3 change). The Pope
et al. (2002) study does not provide education-stratified RR estimates. The principal reason we
applied risk estimates from the Krewski et al. (2009) study was to ensure that the risk
coefficients used to estimate all-cause mortality risk and education-modified mortality risk
were drawn from a consistent modeling framework.
The other key difference between this distributional analysis and the benefits analysis
for this rule relates to the baseline mortality rates. As described in Chapter 5 of this RIA, we
calculate PM2.5-related mortality risk relative to baseline mortality rates in each county.
Traditionally, for benefits analysis, we have applied county-level age- and sex-stratified baseline
mortality rates when calculating mortality impacts (Abt, 2010). To calculate PM2.5 impacts by
race, we incorporated race-specific (stratified by White/Black/Asian/Native American) baseline
mortality rates. This approach improves our ability to characterize the relationship between
race and PM2.5-related mortality however, we do not have a differential concentration-
response function as we do for education, and as a result, we are not able to capture the full
impacts of race in modifying the benefits associated with reductions in PM2.5
The result of this analysis is a distribution of PM2.5 mortality risk estimates by county,
stratified by each of the three population variables (race, income and educational attainment).
7-51
-------
We have less confidence in county-level estimates of mortality than the national or even state
estimates, however, the modeling down to the county level can be considered reasonable
because the estimates are based on 12km air quality modeling estimates of PM2.5, county level
baseline mortality rates, and a concentration-response function that is derived from county
level data. We next identified the counties at or above the median and upper 95th percentile of
the PM2.5 mortality risk distribution. We selected this percentile cut-off to capture the very
highest levels of PM2.5 mortality risk. The second step of the analysis was to repeat the
sequence above by estimating PM2.5 mortality risk in 2016 prior to, and after, the
implementation of MATS.
7.11.3.2 Results
We estimated the change in PM2.5 mortality risk in 2016 among populations living in
those counties at the upper 95th percentile of the mortality risk in 2005. We then compared
the change in risk among these populations living in high-risk counties with populations living in
all other counties (Tables 7-17 through 7-9).
Table 7-17. Estimated Change in the Percentage of All Deaths Attributable to PM2.5 Before
and After Implementation of MATS by 2016 for Each Populations, Stratified by Race
Race
Native
Year Asian Black American White
Among populations at greater risk
2016 (pre-MATS Rule) 4.3% 4.4% 4.4% 4.5%
2016 (post-MATS Rule) 4.1% 4.1% 4.2% 4.3%
Among all other populations
2016 (pre-MATS Rule) 3.2% 3.1% 3.1% 3.3%
2016 (post-MATS Rule) 3% 2.9% 2.9% 3.1%
7-52
-------
Table 7-8. Estimated Change in the Percentage of All Deaths Attributable to PM2.5 Before
and After Implementation of MATS by 2016 for Each Population, Stratified by
Race and Poverty Level
Race
Year
Asian
Black
Native American
White
Among populations living in counties with the largest number of individuals living below the poverty line
2016 (pre-MATS) 3.6% 3.5% 3.6% 3.6%
2016 (post-MATS) 3% 3.4% 3% 3.5%
Among all other populations
2016 (pre-MATS) 3.2% 3.2% 3.2% 3.3%
2016 (post-MATS) 3% 2.9% 3% 3.1%
Table 7-9. Estimated Change in the Percentage of All Deaths Attributable to PM2.5 Before
and After the Implementation of MATS by 2016 for Each Population, Stratified by
Educational Attainment
Year
Among all other populations
2016 (pre-MATS)
2016 (post-MATS)
Race
< Grade 12
= Grade 12
4.5%
4.2%
4%
3.8%
> Grade 12
Among populations at greater risk
2016 (pre-MATS)
2016 (post-MATS)
6.2%
5.9%
5.5%
5.3%
4.3%
4.1%
3.1%
2.9%
Table 7-7, shows the estimated level of PM2.5 mortality risk among populations of
different races according to whether those populations live in counties identified as "greater
risk" counties or "all other counties." As described above, we define "greater risk" counties as
those at or above the 95th percentile of the estimated PM2.5 mortality risk in 2005, and "all
other counties" as those with estimated PM2.5 mortality risk below this level. The results of this
7-53
-------
analysis suggest that the PM2.5 mortality risk among these populations at "greater risk" falls
with implementation of the 2016 MATS. These results also suggest that all populations,
irrespective of race, may receive an estimated reduction in PM2.5 mortality risk. However, limits
to data resolution prevent us from delineating the PM2.5 mortality risk according to population
race with confidence.
Table 7-8 illustrates the estimated change in the level of PM2.5 mortality risk among
populations living in those counties that meet two criteria: (1) the county is at the upper 95th
percentile of mortality risk in 2005; (2) the county is at the upper 95th percentile in terms of the
number of individuals living below the poverty line. We also estimate the change in PM2.5 risk
among all other counties. The analysis suggests that people living in the highest mortality risk
and poorest counties may experience a larger improvement in PM2.5 mortality risk than those
living in lower risk counties containing a smaller number of individuals living below the poverty
line.
Table 7-9 summarizes the estimated change in PM2.5 mortality risk among populations
who have attained three alternate levels of education—less than high school, high school and
greater than high school. As described above, we apply education-stratified PM2.5 mortality risk
coefficients for this analysis. These results indicate that populations with less than a high school
education are at higher risk of PM2.5 mortality, irrespective of whether these populations live in
"greater risk" counties, according to the definition described above. We estimate that with the
implementation of MATS, all populations see their PM2.5 mortality risk fall, regardless of
educational attainment.
7.12 Congressional Review Act
The Congressional Review Act, 5 U.S.C. 801 et seq., as added by the Small Business
Regulatory Enforcement Fairness Act of 1996, generally provides that before a rule may take
effect, the agency promulgating the rule must submit a rule report, which includes a copy of the
rule, to each House of the Congress and to the Comptroller General of the United States. EPA
will submit a report containing this rule and other required information to the U.S. Senate, the
U.S. House of Representatives, and the Comptroller General of the United States prior to
publication of the rule in the Federal Register. A major rule cannot take effect until 60 days
after it is published in the Federal Register. This action is a "major rule" as defined by 5 U.S.C.
804(2). This rule will be effective 60 days after publication.
7-54
-------
7.13 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
.
Bullard RD, Mohai P, Wright B, Saha R, et al. Toxic Waste and Race at Twenty 1987-2007. United
Church of Christ. March, 2007.
Burger, J. (2002). Daily consumption of wild fish and game: Exposures of high end
recreationalists, International Journal of Environmental Health Research, 12:4, p. 343-
354.
Dellinger, JA (2004). Exposure assessment and initial intervention regarding fish consumption of
tribal members in the Upper Great Lakes Region in the United States. Environmental
Research 95 (2004) p. 325-340.
Fann N, Lamson AD, Anenberg SC, Wesson K, Risley D, Hubbell B. 2011a. Estimating the national
public health burden associated with exposure to ambient PM2.s and ozone. Risk
Analysis, in press.
Fann N, Roman HA, Fulcher CM, Gentile MA, Hubbell BJ, Wesson K, et al. 2011b. Maximizing
Health Benefits and Minimizing Inequality: Incorporating Local-Scale Data in the Design
and Evaluation of Air Quality Policies. Risk Analysis 31:908-922; doi:10.1111/j.l539-
6924.2011.01629.x.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi, Y, et al. 2009. Extended follow-up and
spatial analysis of the American Cancer Society study linking particulate air pollution and
mortality. HEI Research Report, 140, Health Effects Institute, Boston, MA.
Krewski, D., R.T. Burnett, M.S. Goldbert, K. Hoover, J. Siemiatycki, M. Jerrett, M. Abrahamowicz,
and W.H. White. 2000. Reanalysis of the Harvard Six Cities Study and the American
Cancer Society Study of Particulate Air Pollution and Mortality. Special Report to the
Health Effects Institute. Cambridge MA. July.
Mennis J. "Using Geographic Information Systems to Create and Analyze Statistical Surfaces of
Populations and Risk for Environmental Justice Analysis." Social Science Quarterly,
2002;83(l):281-297.
Mohai P, Saha R. "Reassessing Racial and Socio-economic Disparities in Environmental Justice
Research." Demography. 2006;43(2): 383-399.
Moya, J. 2004. Overview of fish consumption rates in the United States. Human and Ecological
Risk Assessment: An International Journal 10, no. 6: 1195-1211.
7-55
-------
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.
Shilling, Fraser, Aubrey White, Lucas Lippert, Mark Lubell (2010). Contaminated fish
consumption in California's Central Valley Delta. Environmental Research 110, p. 334-
344.
Woods & Poole Economics, Inc. 2008. Population by Single Year of Age CD. CD-ROM. Woods &
Poole Economics, Inc.
U.S. GAO (Government Accountability Office). "Demographics of People Living Near Waste
Facilities." Washington DC: Government Printing Office; 1995.
U.S. Environmental Protection Agency (U.S. EPA). 2011a. Proposed Regulatory Impact Analysis
(RIA) for the Mercury and Air Toxics Rule. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. January.
U.S. Environmental Protection Agency (U.S. EPA). 2011b. Final Regulatory Impact Analysis (RIA)
for the Transport Rule. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. June.
7-56
-------
CHAPTER 8
COMPARISON OF BENEFITS AND COSTS
8.1 Comparison of Benefits and Costs
The estimated costs to implement the final MATS Rule, as described earlier in this
document, are approximately $9.6 billion annually for 2016 (2007 dollars). Thus, the net
benefits (benefits minus costs) of the program in 2016 are approximately $27 to 80 +B billion or
$24 to 71 +B billion annually (2007 dollars, based on a discount rate of 3 percent and 7 percent
for the benefits, respectively and rounded to two significant figures). (B represents the sum of
all unquantified benefits and disbenefits of the regulation.) Therefore, implementation of this
rule is expected, based purely on economic efficiency criteria, to provide society with a
significant net gain in social welfare, even given the limited set of health and environmental
effects we were able to quantify. Addition of health endpoints other than IQ loss to children
exposed to mercury from recreationally caught freshwater fish and acidification-, and
eutrophication-related impacts would likely increase the net benefits of the rule. Table 8-1
presents a summary of the benefits, costs, and net benefits of the final MATS Rule.
As with any complex analysis of this scope, there are several uncertainties inherent in
the final estimate of benefits and costs that are described fully in Chapters 3, 4 and 5.
8-1
-------
Table 8-1. Summary of EPA's Estimates of Annualized3 Benefits, Costs, and Net Benefits of
the Final MATS in 2016b (billions of 2007$)
Estimate Estimate
Description (3% Discount Rate) (7% Discount Rate)
Costs0 $9.6 $9.6
Benefitsd'e'f $37 to $90 + B $33 to $81 + B
Net benefits (benefits-costs)8 $27 to $80+B $24 to $71+B
a All estimates presented in this report represent annualized estimates of the benefits and costs of the final MATS
in 2016 rather than the net present value of a stream of benefits and costs in these particular years of analysis.
Estimates rounded to two significant figures and represent annualized benefits and costs anticipated for the
year 2016.
Total social costs are approximated by the compliance costs. Compliance costs consist of IPM projections,
monitoring/reporting/record-keeping costs, and oil-fired fleet analysis costs. For a complete discussion of these
costs refer to Chapter 3. Costs are estimated using a 6.15% discount rate
d Total benefits are composed primarily of monetized PM-related health benefits. The reduction in premature
fatalities each year accounts for over 90% of total monetized benefits. Benefits in this table are nationwide and
are associated with directly emitted PM2.5 and SO2 reductions. The estimate of social benefits also includes CO2-
related benefits calculated using the social cost of carbon, discussed further in chapter 5.
e Not all possible benefits or disbenefits are quantified and monetized in this analysis. B is the sum of all
unquantified benefits and disbenefits. 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. Estimates
here are subject to uncertainties discussed further in the body of the document. Potential benefit categories
that have not been quantified and monetized are listed in Table ES-4.
Mortality risk valuation assumes discounting over the SAB-recommended 20-year segmented lag structure.
Results reflect the use of 3% and 7% discount rates consistent with EPA and OMB guidelines for preparing
economic analyses (EPA, 2000; OMB, 2003).
8 Net benefits are rounded to two significant figures. Columnar totals may not sum due to rounding
8-2
-------
8.2 References
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.
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.
U.S. Environmental Protection Agency (EPA). December 2010. Guidelines for Preparing
Economic Analyses. EPA 240-R-10-001.
U.S. Office of Management and Budget (OMB). 2003. Circular A-4 Guidance to Federal Agencies
on Preparation of Regulatory Analysis.
Woodruff, T.J., J. Grille, and K.C. Schoendorf. 1997. "The Relationship Between Selected Causes
of Postneonatal Infant Mortality and Particulate Infant Mortality and Particulate Air
Pollution in the United States." Environmental Health Perspectives 105(6):608-612.
8-3
-------
United States Office of Air Quality Planning and Standards Publication No. EP A-452/R-11-011
Environmental Protection Health and Environmental Impacts Division [December 2011]
Agency Research Triangle Park, NC
------- |