United States
Environmental Protection
Agency
Office of Water
Washington, DC 20460
EPA-821 -R-19-011
November 1, 2019
oERA Benefit and Cost Analysis for
Proposed Revisions to the
Effluent Limitations
Guidelines and Standards for
the Steam Electric Power
Generating Point Source
Category
-------
United Slates
Environmental Protection
Ag«ncy
Benefit and Cost Analysis for Proposed
Revisions to the Effluent Limitations
Guidelines and Standards for the Steam
Electric Power Generating Point Source
Category
EPA-821-R-19-011
November 2019
U.S. Environmental Protection Agency
Office of Water (4303T)
Engineering and Analysis Division
1200 Pennsylvania Avenue, NW
Washington, DC 20460
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Acknowledgements and Disclaimer
This report was prepared by the U.S. Environmental Protection Agency. Neither the United States
Government nor any of its employees, contractors, subcontractors, or their employees make any warrant,
expressed or implied, or assume any legal liability or responsibility for any third party's use of or the results
of such use of any information, apparatus, product, or process discussed in this report, or represents that its
use by such party would not infringe on privately owned rights.
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Table of Contents
Table of Contents
Table of Contents i
List of Figures vi
List of Tables vii
Abbreviations xi
Executive Summary 1
1 Introduction 1-1
1.1 Steam Electric Power Plants 1-2
1.2 Baseline and Regulatory Options Analyzed 1-2
1.3 Analytic Framework 1-4
1.3.1 Constant Prices 1-5
1.3.2 Discount Rate and Year 1-5
1.3.3 Period of Analysis 1-5
1.3.4 Annualization of future costs and benefits 1-6
1.3.5 Direction of Environmental Changes and Benefits 1-6
1.3.6 Population and Income Growth 1-7
1.4 Organization of the Benefit and Cost Analysis Report 1-7
2 Benefits Overview 2-1
2.1 Human Health Impacts Associated with Changes in Surface Water Quality 2-4
2.1.1 Drinking Water 2-4
2.1.2 Fish Consumption 2-6
2.1.3 Complementary Measure of Human Health Impacts 2-7
2.2 Ecological Impacts Associated with Changes in Surface Water Quality 2-7
2.2.1 Changes in Surface Water Quality 2-8
2.2.2 Impacts on Threatened and Endangered Species 2-9
2.2.3 Changes in Sediment Contamination 2-9
2.3 Economic Productivity 2-10
2.3.1 Marketability of Coal Ash for Beneficial Use 2-10
2.3.2 Water Supply and Use 2-11
2.3.3 Sedimentation Changes in Navigational Waterways 2-12
2.3.4 Commercial Fisheries 2-13
2.3.5 Tourism 2-13
2.3.6 Property Values 2-13
2.4 Changes in Air Pollution 2-14
EPA-821-R-19-011
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Table of Contents
2.5 Reduced Water Withdrawals 2-16
2.6 Summary of Benefits Categories 2-16
3 Water Quality Effects of Regulatory Options 3-1
3.1 Waters Affected by Steam Electric Power Plant Discharges 3-1
3.2 Changes in Pollutant Loadings 3-2
3.2.1 Timing of ELG Implementation 3-2
3.2.2 Results 3-5
3.3 Water Quality Downstream from Steam Electric Power Plants 3-7
3.4 Overall Water Quality Changes 3-8
3.4.1 WQI Data Sources 3-8
3.4.2 WQI Calculation 3-10
3.4.3 Baseline WQI 3-10
3.4.4 Estimated Changes in Water Quality (AWQI) from the Regulatory Options 3-11
3.5 Limitations and Uncertainty 3-12
4 Human Health Benefits from Changes in Pollutant Exposure via Drinking Water Pathways.... 4-1
4.1 Background 4-1
4.2 Overview of the Analysis 4-2
4.3 Analysis Steps 4-5
4.3.1 Step 1: Modeling Bromide Concentrations in Surface Water 4-5
4.3.2 Step 2: Modeling Changes in Trihalomethanes in Treated Water Supplies 4-5
4.3.3 Step 3: Quantifying Population Exposure and Health Effects 4-11
4.3.4 Step 4: Quantifying the Monetary Value of Benefits 4-16
4.4 Results of Analysis of Human Health Benefits from Estimated Changes in Bromide Discharges
Analysis 4-16
4.5 Additional Measures of Human Health Effects from Exposure to Steam Electric Pollutants via
Drinking Water Pathway 4-21
4.6 Limitations and Uncertainties 4-22
5 Human Health Effects from Changes in Pollutant Exposure via Fish Ingestion Pathway 5-1
5.1 Affected Population 5-2
5.2 Pollutant Exposure from Fish Consumption 5-4
5.2.1 Fish Tissue Pollutant Concentrations 5-4
5.2.2 Average Daily Dose 5-4
5.3 Health Effects in Children from Changes in Lead Exposure 5-6
5.3.1 Methods 5-7
EPA-821-R-19-011
ii
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Table of Contents
5.3.2 Results 5-9
5.4 Heath Effects in Children from Changes in Mercury Exposure 5-10
5.4.1 Methods 5-10
5.4.2 Results 5-11
5.5 Estimated Changes in Cancer Cases from Arsenic Exposure 5-12
5.6 Total Monetary Values of Estimated Changes in Human Health Effects 5-12
5.7 Additional Measures of Potential Changes in Human Health Effects 5-12
5.8 Limitations and Uncertainties 5-13
6 Nonmarket Benefits from Water Quality Changes 6-1
6.1 Linking Changes in Water Quality to Valuation 6-1
6.2 Total WTP for Water Quality Changes 6-1
6.3 Limitations and Uncertainties 6-4
7 Impacts and Benefits to Threatened and Endangered Species 7-1
7.1 Introduction 7-1
7.2 Baseline Status of Freshwater Fish Species 7-1
7.3 T&E Species Affected by the Regulatory Options 7-2
7.3.1 Identifying T&E Species Potentially Affected by the Regulatory Options 7-2
7.3.2 Estimating Effects of the Proposed Rule on T&E Species 7-4
7.4 Limitations and Uncertainties 7-6
8 Air-Related Benefits 8-1
8.1 Data and Methodology 8-2
8.1.1 Changes in Air Emissions 8-2
8.1.2 XO\ and SO 8-4
8.1.3 CO 8-5
8.2 Results 8-8
8.3 Limitations and Uncertainties 8-10
9 Changes in Water Withdrawals 9-1
9.1 Methods 9-1
9.2 Results 9-1
9.3 Limitations and Uncertainties 9-2
10 Estimated Changes in Dredging Costs 10-1
10.1 Methods 10-1
10.1.1 Estimated Changes in Navigational Dredging Costs 10-1
EPA-821-R-19-011
iii
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Table of Contents
10.1.2Estimated Changes in Reservoir Dredging Costs 10-2
10.2 Limitation and Uncertainty 10-3
11 Summary of Estimated Total Monetized Benefits 11-1
12 Summary of Total Social Costs 12-1
12.1 Overview of Costs Analysis Framework 12-1
12.2 Key Findings for Regulatory Options 12-2
13 Benefits and Social Costs 13-1
13.1 Comparison of Benefits and Costs by Option 13-1
13.2 Analysis of Incremental Benefits and Costs 13-1
14 Environmental Justice 14-1
14.1 Socio-economic Characteristics of Populations Residing in Proximity to Steam Electric Power Plants
14-1
14.2 Distribution of Human Health Impacts and Benefits 14-4
14.2.1 Socio-economic Characteristics of Populations Impacted by Changes in Exposure to Pollutants
via Drinking Water Pathway 14-4
14.2.2 Socio-economic Characteristics of Populations Impacted by Changes in Exposure to Pollutants
via Fish Ingestion Pathway 14-7
14.3 EJ Analysis Findings 14-12
14.4 Limitations and Uncertainties 14-12
15 Cited References 15-1
Appendix A Changes to Benefits Methodology since 2015 Rule Analysis A-l
Appendix B WQI Calculation and Regional Subindices B-l
B.l WQI Calculation B-l
B.2 Regional Subindices B-4
Appendix C Additional Details on Modeling Change in Bladder Cancer Incidence from Change in
TTHM Exposure C-l
C. 1 Details on Life Table Approach C-l
C. 1.1 Health Impact Function C-l
C. 1.2 Health Risk Model C-2
C. 1.3 Detailed Input Data C-7
C.2 Detailed Results from Analysis C-l 1
C.3 Temporal Distribution of Benefits C-13
C.4 Sensitivity Analysis Results C-14
C.4.1 Sensitivity to bromide loads C-14
EPA-821-R-19-011
iv
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Table of Contents
C.4.2 Sensitivity to relationship between bromide and TTHM changes C-15
Appendix D Derivation of Ambient Water and Fish Tissue Concentrations in Receiving and
Downstream Reaches D-l
D.l Toxics D-l
D. 1.1 Estimating Water Concentrations in each Reach D-l
D.1.2 Estimating Fish Tissue Concentrations in each Reach D-2
D.2 Nutrients and Suspended Sediment D-3
Appendix E Georeferencing Surface Water Intakes to the Medium-resolution Stream Network....E-l
Appendix F Estimation of Exposed Population for Fish Ingestion Pathway F-l
Appendix G Sensitivity Analysis for IQ Point-based Human Health Effects G-l
G. 1 Health Effects in Children from Changes in Lead Exposure G-l
G.2 Heath Effects in Children from Changes in Mercury Exposure G-2
Appendix H Methodology for Estimating WTP for Water Quality Changes H-l
Appendix I Uncertainty Associated with Estimating the Social Cost of Carbon 1-1
1.1 Overview of Methodology Used to Develop Interim Domestic SC-CO2 Estimates 1-1
1.2 Treatment of Uncertainty in Interim Domestic SC-CO2 Estimates 1-2
1.3 Forgone Global Climate Benefits 1-5
EPA-821-R-19-011
v
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
List of Figures
List of Figures
Figure 2-1: Summary of Benefits Resulting from the Regulatory Options 2-3
Figure 4-1: Overview of Analysis of Human Health Benefits of Altering Bromide Discharges 4-4
Figure 4-2: Modeled Relationship between Changes in Bromide Concentration and Changes in TTHM
Concentrations based on Median Values in Regli et al. (2015) 4-10
Figure 4-3: Estimated Number of Bladder Cancer Cases Avoided under the Regulatory Options 4-17
Figure 4-4: Estimated Number of Cancer Deaths Avoided under the Regulatory Options 4-18
Figure 4-5: Contributions of Individual Steam Electric Power Plants to Total Annualized Benefits of Changes
in Bromide Discharges under the Regulatory Options (3 Percent Discount Rate) 4-20
Appendix Figures
Figure C-l: Estimated Relationships between Lifetime Bladder Cancer Risk and TTHM Concentrations in
Drinking Water C-l
Figure C-2: Cumulative Annual Value of Cancer Morbidity Avoided, 2021-2121 (2018$ undiscounted)...C-13
Figure C-3: Cumulative Annual Value of Mortality Avoided, 2021-2121 (2018$ undiscounted) C-14
Figure C-4: Modeled Sensitivity Analysis Relationship between Changes in Bromide Concentration and
Changes in TTHM Concentrations based on Regli et al. (2015) C-16
Figure E-l: PWS Intakes Review Subset E-2
Figure F-l. Illustration of Intersection of Census Block Groups and COMIDs F-l
EPA-821-R-19-011
vi
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
List of Tables
List of Tables
Table 1-1: Regulatory Options 1-3
Table 2-1: Estimated Annual Pollutant Loadings and Changes in Loadings for Baseline and Regulatory
Options 2-1
Table 2-2: Drinking Water Maximum Contaminant Levels and Goal for Selected Pollutants in Steam Electric
FGD Wastewater or Bottom Ash Transport Water Discharges 2-4
Table 2-3: Estimated Welfare Effects of Changes in Pollutant Discharges from Steam Electric Power Plants 2-
17
Table 3-1: Strahler Stream Order Designation for Reaches Receiving Steam Electric Power Plant Discharges
3-1
Table 3-2: Implementation Schedule by Wastestream and Regulatory Option 3-4
Table 3-3: Annual Average Changesin Total Pollutant Loading in 2021-2047 for Selected Pollutants in Steam
Electric Power Plant Discharges, Relative to Baseline (lb/year) 3-6
Table 3-4: Water Quality Data used in Calculating WQI for the Baseline and Regulatory Options 3-10
Table 3-5: Estimated Percentage of Potentially Affected Inland Reach Miles by WQI Classification: Baseline
Scenario 3-11
Table 3-6: Ranges of Estimated Water Quality Changes for Regulatory Options 3-12
Table 3-7: Limitations and Uncertainties in Estimating Environmental Effects of Regulatory Options 3-12
Table 4-1: Estimated Reaches, Surface Water Intakes, Public Water Systems, and Populations Potentially
Affected by Bromide Discharges from Steam Electric Power Plants 4-6
Table 4-2: Estimated Distribution of Changes in Source Water and PWS-level Bromide Concentrations by
Regulatory Option 4-8
Table 4-3: Estimated Increments of Change in TTHM Levels (j^ig/L) as a Function of Change in Bromide
Levels ((ig/L) 4-9
Table 4-4: Distribution of Estimated Changes in TTHM Concentration by the Number of PWS and
Population Served 4-10
Table 4-5: Summary of Data Sources Used in Lifetime Health Risk Model 4-14
Table 4-6: Summary of Sex- and Age-specific Mortality and Bladder Cancer Incidence Rates 4-15
Table 4-7: Estimated Bromide-related Bladder Cancer Mortality and Morbidity Monetized Benefits 4-18
Table 4-8: Limitations and Uncertainties in the Analysis of Human Health Benefits from Changes in Bromide
Discharges 4-22
Table 5-1: Summary of Potentially Affected Population Living within 50 Miles of Affected Reaches
(baseline, as of 2016) 5-4
Table 5-2: Summary of Group-specific Consumption Rates for Fish Tissue Consumption Risk Analysis .... 5-5
Table 5-3: Value of an IQ Point (2018$) based on Expected Reductions in Lifetime Earnings 5-9
Table 5-4: Estimated Monetary Value of Changes in IQ Points for Children Exposed to Lead 5-9
EPA-821-R-19-011 vii
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs List of Tables
Table 5-5: Estimated Monetary Values from Changes in IQ Points for Infants from Mercury Exposure 5-11
Table 5-6: Total Monetary Values of Changes in Human Health Outcomes Associated with Fish Consumption
for Regulatory Options (millions of 2018$) 5-12
Table 5-7: Estimated Number of Reaches Exceeding Human Health Criteria for Steam Electric Pollutants 5-13
Table 5-8: Limitations and Uncertainties in the Analysis of Human Health Effects 5-14
Table 6-1: Estimated Household Willingness-to-Pay for Water Quality Changes 6-3
Table 6-2: Estimated Total Annualized Willingness-to-Pay for Water Quality Changes Compared to Baseline
(Millions 2018$) 6-4
Table 6-3: Limitations and Uncertainties in the Analysis of Nonmarket Water Quality Benefits 6-4
Table 7-1: T&E Species with High Vulnerability Habitat Occurring within Waterbodies Affected by Steam
Electric Power Plants 7-4
Table 7-2: T&E Species Whose Habitat May Benefit from the Regulatory Options 7-5
Table 7-3: T&E Species Whose Habitat May be Adversely Affected by the Regulatory Options 7-5
Table 7-4: Limitations and Uncertainties in the Analysis of T&E Species Benefits 7-6
Table 8-1: IPM Run Years 8-2
Table 8-2: Estimated Changes in Air Pollutant Emissions due to Increase in Power Requirements and
Trucking at Steam Electric Power Plants 2021-2047, Relative to Baseline 8-3
Table 8-3: Estimated Changes in Annual Air Pollutant Emissions due to Changes in Electricity Generation
Profile, Relative to Baseline 8-3
Table 8-4: Estimated Net Changes in Air Pollutant Emissions due to Changes in Power Requirements,
Trucking, and Electricity Generation Profile, Relative to Baseline 8-4
Table 8-5: National Benefits per Ton Estimates for NOx and SO2 Emissions (2018$/ton) from the Benefits
per Ton Analysis Reported by U.S. EPA (2018d) 8-5
Table 8-6: Interim Domestic Social Cost of Carbon Values (2018$/metric tonne CO2) 8-7
Table 8-7: Estimated Domestic Climate Benefits from Changes in CO2 Emissions for Selected Years
(millions; 2018$) 8-8
Table 8-8: Estimated Total Annualized Domestic Climate Benefits from Changes in CO2 Emissions (Millions;
2018$) 8-9
Table 8-9: Extrapolated Annualized Domestic Climate Benefits from Changes in CO2 Emissions (Millions;
2018$) 8-10
Table 8-10: Limitations and Uncertainties in Analysis of Air-related Benefits 8-11
Table 9-1: Industry-level Total Changes in Water Withdrawals (Surface Water and Aquifers) 9-1
Table 9-2: Estimated Annualized Benefits from Increased Groundwater Withdrawals (Millions; 2018$) 9-2
Table 9-3: Limitations and Uncertainties in Analysis of Changes in Groundwater Withdrawals 9-2
Table 10-1: Estimated Annualized Dredging Costs at Affected Reaches under the Baseline (Millions of
2018$) 10-2
EPA-821-R-19-011
viii
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs List of Tables
Table 10-2: Estimated Annualized Changes in Navigational Dredging Costs (Thousands of 2018$) 10-2
Table 10-3: Estimated Annualized Reservoir Dredging Costs under Baseline (Millions 2018$) 10-2
Table 10-4: Estimated Total Annualized Changes in Reservoir Dredging Costs (2018$) 10-3
Table 10-5: Limitations and Uncertainties in Analysis of Changes in Dredging Costs 10-3
Table 11-1: Summary of Estimated Total Annualized Benefits at 3 Percent (Millions; 2018$) 11-2
Table 11-2: Summary of Estimated Total Annualized Benefits at 7 Percent (Millions; 2018$) 11-3
Table 12-1: Summary of Estimated Annualized Costs (Millions; $2018) 12-2
Table 12-2: Time Profile of Costs to Society (Millions; $2018) 12-3
Table 13-1: Total Estimated Annualized Benefits and Costs by Regulatory Option and Discount Rate
(Millions; 2018$) 13-1
Table 13-2: Estimated Incremental Net Benefit Analysis (Millions; 2018$) 13-2
Table 14-1: Socio-economic Characteristics of Communities Living in Proximity to Steam Electric Power
Plants, Compared to National Average 14-3
Table 14-2: Socio-economic Characteristics of Affected Communities, Compared to State Average 14-4
Table 14-3: Socio-economic Characteristics of Affected Counties, Compared to State Average 14-5
Table 14-4: Socio-economic Characteristics of Affected Tribal Areas, Compared to State Average 14-6
Table 14-5: Characteristics of Population Potentially Exposed to Lead from Steam Electric Power Plants via
Consumption of Self-caught Fish 14-7
Table 14-6: Estimated Distribution of Baseline IQ Point Decrements by Pollutant (2021 to 2047) 14-8
Table 14-7: Distribution of Changes in IQ Point Relative to the Baseline, by Pollutant (2021 to 2047) 14-9
Table 14-8: Estimated Distribution of Baseline IQ Point Decrements by Pollutant and Fishing Mode (2021 to
2047) 14-10
Table 14-9: Estimated Distribution of Changes in IQ Point Decrements Relative to the Baseline by Fishing
Mode, and Pollutant (2021 to 2047) 14-11
Table 14-10: Limitations and Uncertainties in EJ Analysis 14-13
Appendix Tables
Table A-l: Changes to Benefits Analysis Since 2015 Final Rule A-l
Table B-l: Freshwater Water Quality Subindices B-2
Table B-2: Freshwater Water Quality Subindex for Toxics B-3
Table B-3: TSS Subindex Curve Parameters, by Ecoregion B-4
Table B-4: TN Subindex Curve Parameters, by Ecoregion B-6
Table B-5: TP Subindex Curve Parameters, by Ecoregion B-8
Table C-l: Health Risk Model Variable Definitions C-6
EPA-821-R-19-011
ix
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
List of Tables
Table C-2: Summary of Sex- and Age-specific Bladder Cancer Stage Distribution and Relative Survival ...C-8
Table C-3: Number of Adverse Health Effects Avoided Over Time Starting from 2021 C-12
Table C-4: Distribution of Estimated Changes in TTHM Concentration, Number of PWS and Populations. ..C-
16
Table C-5: Sensitivity of Estimated Bromide-related Benefits of Regulatory Option 4 C-17
Table D-l: Assumed Background Fish Tissue Concentrations, based on 10th percentile D-3
Table D-2: Imputed and Validated Fish Tissue Concentrations by Regulatory Option D-3
Table E-l: Summary of Intakes Potentially Affected by Steam Electric Power Plant E-l
Table G-l: Value of an IQ Point (2018$) based on Expected Reductions in Lifetime Earnings G-l
Table G-2: Estimated Monetary Value of Changes in IQ Losses for Children Exposed to Lead G-l
Table G-3: Estimated Monetary Values from Changes in IQ Losses for Infants from Mercury Exposure.... G-2
Table H-l: Independent Variable Assignments for Surface Water Quality Meta-Analysis H-5
EPA-821-R-19-011
x
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Abbreviations
Abbreviations
ACE
Affordable Clean Energy
ACS
American Community Survey
ADD
Average daily dose
AFSC
Alaska Fisheries Science Center
ASMFC
Atlantic States Marine Fisheries Commission
As
Arsenic
BA
Bottom ash
BAT
Best available technology economically achievable
BCA
Benefit-cost-analysis
BEA
Bureau of Economic Analysis
BenMAP
Environmental Benefits Mapping and Analysis Program
BLS
Bureau of Labor Statistics
BMP
Best management practices
BOD
Biochemical oxygen demand
BPT
Best practicable control technology currently available
BW
Body weight
C&D
Construction and development
CBG
Census Block Group
CCI
Construction Cost Index
CCME
Canadian Council of Ministers of the Environment
CCR
Coal combustion residuals
CDC
Center for Disease Control
CFR
Code of Federal Regulations
C02
Carbon dioxide
COD
Chemical oxygen demand
COI
Cost-of-illness
COPD
Chronic obstructive pulmonary disease
CPI
Consumer Price Index
CSF
Cancer slope factor
CVD
Cardiovascular disease
CWA
Clean Water Act
D-FATE
Downstream Fate and Transport Equations
DBP
Disinfection byproduct
DBPR
Disinfectants and Disinfection Byproduct Rule
DCN
Document Control Number
DHHS
Department of Health and Human Services
DO
Dissolved oxygen
DOE
Department of Energy
E2RF1
Enhanced River File 1
EA
Environmental Assessment
ECI
Employment Cost Index
EGU
Electricity Generating Unit
EJ
Environmental justice
ELGs
Effluent limitations guidelines and standards
EPA-821-R-19-011
xi
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Abbreviations
EO
Executive Order
EPA
United States Environmental Protection Agency
EROM
Enhanced Runoff Method
ESA
Endangered Species Act
FC
Fecal coliform
FCA
Fish consumption advisories
FGD
Flue gas desulfurization
FR
Federal Register
GDP
Gross Domestic Product
GHG
Greenhouse gas
GIS
Geographic Information System
HAA
Haloacetic acids
Hg
Mercury
HISA
Highly influential scientific assessment
HRTR
High Residence Time Reduction
HUC
Hydrologic unit code
IAM
Integrated assessment model
ICR
Information Collection Rule
IEUBK
Integrated Exposure, Uptake, and Biokinetics
IPCC
Intergovernmental Panel on Climate Change
IPM
Integrated Planning Model
ISA
Integrated science assessment
IRIS
Integrated Risk Information System
IQ
Intelligence quotient
LADD
Lifetime average daily dose
LML
Lowest measured level
LRTR
Low Residence Time Reduction
MCL
Maximum contaminant level
MCLG
Maximum contaminant level goal
MGD
Million gallons per day
MRM
Meta-regression model
MWTP
Marginal willingness-to-pay
NCHS
National Center for Health Statistics
NEFSC
Northeast Fisheries Science Center
NERC
North American Electric Reliability Corporation
NHD
National Hydrography Dataset
NLCD
National Land Cover Dataset
NMFS
National Marine Fisheries Service
NO A A
National Oceanic and Atmospheric Administration
NOx
Nitrogen oxides
NPDES
National Pollutant Discharge Elimination System
NRWQC
National Recommended Water Quality Criteria
NWIS
National Water Information System
OAQPS
Office of Air Quality Planning and Standards
O&M
Operation and maintenance
OMB
Office of Management and Budget
EPA-821-R-19-011
xii
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Abbreviations
OSHA Occupational Safety and Health Administration
OTSA Oklahoma Tribal Statistical Area
PAM Below the poverty level and minority
Pb Lead
PbB Blood lead concentration
PIFSC Pacific Islands Fisheries Science Center
PM Particulate matter
POM Below the poverty level or minority
ppm parts per million
PSES Pretreatment Standards for Existing Sources
PWS Public water system
QA Quality assurance
QC Quality control
RIA Regulatory Impact Analysis
RSEI Risk-Screening Environmental Indicators
SAB Science Advisory Board
SBREFA Small Business Regulatory Enforcement Fairness Act
SCC Social cost of carbon
SC-CO2 Domestic social cost of carbon
SDWIS Safe Drinking Water Information System
Se Selenium
SEER Surveillance, Epidemiology, and End Results
SEFSC Southeast Fisheries Science Center
SO2 Sulfur dioxide
SPARROW SPAtially Referenced Regressions On Watershed attributes
STORET STOrage and RETrieval Data Warehouse
SWAT Surface Water Analytical Tool
SWFSC Southwest Fisheries Science Center
T&E Threatened and endangered
TDD Technical Development Document
TDS Total dissolved solids
TN Total nitrogen
TP Total phosphorus
TRI Toxics Release Inventory
TSD Technical support document
TSS Total suspended solids
TTHM Total trihalomethanes
TWTP Total willingness-to-pay
U.S. FWS United States Fish and Wildlife Service
USGS United States Geological Survey
VIP Voluntary Incentive Program
VOC Volatile organic compounds
VSL Value of a statistical life
WBD Watershed Boundary Dataset
WQ Water quality
WQI Water quality index
EPA-821-R-19-011 xiii
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Abbreviations
WQI-BL Baseline water quality index
WQI-PC Post-compliance water quality index
WQL Water quality ladder
WTP Willingness-to-pay
EPA-821-R-19-011
xiv
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Executive Summary
Executive Summary
The U.S. Environmental Protection Agency (EPA) is proposing a regulation that would revise the technology-
based effluent limitations guidelines and standards (ELGs) for the steam electric power generating point
source category, 40 CFR part 423, which the EPA promulgated in November 2015 (80 FR 67838). The
regulatory options would revise certain best available technology (BAT) effluent limitations and pretreatment
standards for existing sources (PSES) for two wastestreams: flue gas desulfurization (FGD) wastewater and
bottom ash transport water.
Regulatory Options
The EPA analyzed four regulatory options, summarized in Table ES-1. The baseline for the analyses reflects
ELG requirements (in absence of any new final EPA action).1 The Agency calculated the difference between
the baseline and the regulatory options to determine the net incremental effect (as positive or negative change)
of the regulatory options. The EPA is proposing Option 2. The incremental effects between the baseline and
Option 2 are very small and are expected to yield very small benefits (positive or negative).
Table ES-1: Regulatory Options
Wastestream
Subcategory
Technology Basis for BAT/PSES Regulatory Options3
2015 Rule
(Baseline)
Option 1
Option 2
Option 3
Option 4
FGD
Wastewater
NAb
Chemical
Precipitation
+ HRTR
Biological
Treatment
Chemical
Precipitation
Chemical
Precipitation
+ LRTR
Biological
Treatment
Chemical
Precipitation
+ LRTR
Biological
Treatment
Membrane
Filtration
High FGD Flow Facilities:
Plant-level scrubber purge
flow >4 MGD
NS
NS
Chemical
Precipitation
Chemical
Precipitation
Chemical
Precipitation
Low Utilization Boilers: All
units have net generation <
876,000 MWh
NS
NS
Chemical
Precipitation
NS
NS
Boilers retiring by 2028
NS
Surface
Impoundment
Surface
Impoundment
Surface
Impoundment
Surface
Impoundment
FGD Wastewater Voluntary Incentives
Program (Direct Dischargers Only)
Chemical
Precipitation
+ Evaporation
Membrane
Filtration
Membrane
Filtration
Membrane
Filtration
NA
Bottom Ash
Transport
Water
NAb
Dry Handling/
Closed loop
Dry Handling
or High
Recycle Rate
Systems
Dry Handling
or High
Recycle Rate
Systems
Dry Handling
or High
Recycle Rate
Systems
Dry Handling
or High
Recycle Rate
Systems
Low Utilization Boilers: All
units have net generation <
876,000 MWh
NS
NS
Surface
Impoundment
+ BMP Plan
NS
NS
Boilers retiring by 2028
NS
Surface
Impoundment
Surface
Impoundment
Surface
Impoundment
Surface
Impoundment
This includes the 2015 rule as well as the September 2017 postponement rule which delayed the earliest compliance date for the
ELGs applicable to FGD wastewater and bottom ash transport water.
EPA-821-R-19-011
ES-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Executive Summary
Table ES-1: Regulatory Options
Wastestream
Subcategory
Technology Basis for BAT/PSES Regulatory Options3
2015 Rule
(Baseline)
Option 1
Option 2
Option 3
Option 4
Abbreviations: BMP = Best Management Practice; HRTR = High Residence Time Reduction; LRTR = Low Residence Time Reduction;
NS = Not subcategorized; NA = Not applicable
a. See Supplemental TDD for a description of these technologies.
b. The 2015 rule subcategorized units with nameplate capacity 50 MW or less and the EPA is not revising requirements for these
units in this proposal.
Source: U.S. EPA Analysis, 2019
Benefits of Regulatory Options
The EPA estimated the potential social welfare effects of the regulatory options and, where possible,
quantified and monetized the benefits (see Chapters 3 through 11 for details of the methodology and results).
Table ES-2 and Table ES-3 summarize the benefits that the EPA quantified and monetized using 3 percent
and 7 percent discounts, respectively. In the tables, positive values indicate improvements in social welfare,
relative to the baseline, whereas negative values reflect forgone benefits of the regulatory options, i. e., social
welfare losses. In general, the estimated effects of implementing the regulatory options are small compared to
those estimated in 2015 (see U.S. EPA, 2015a).
The EPA quantified but did not monetize other welfare effects of the regulatory options, including expected
changes of pollutant concentrations in excess of human health-based NRWQC limits, and discusses other
potential welfare effects qualitatively, including impacts to commercial fisheries or changes in the
marketability of coal ash for beneficial use; the EPA evaluated these effects qualitatively in Chapter 2.
Table ES-2: Summary of Total Annualized Benefits at 3 Percent (Millions; 2018$)
Benefit Category
Option V
Option 2a
Option 3a
Option 4a
Low
Mid
High
Low
Mid
High
Low
Mid
High
Low
Mid
High
Human Health
-$0.7
$34.8
$39.7
$82.8
Changes in IQ losses in children
from exposure to leadb
<$0.0
<$0.0
<$0.0
<$0.0
Changes in IQ losses in children
from exposure to mercury
-$0.3
-$2.8
-$2.9
-$1.5
Changes in cancer risk from
DBPs in drinking water
-$0.4
$37.6
$42.6
$84.3
Ecological Conditions and
Recreational Uses Changes
-$10.0
-$12.5
-$55.5
$11.8
$16.7
$65.6
$16.3
$22.5
$90.7
$19.8
$27.3
$110.2
Use and nonuse values for
water quality changes
-$10.0
-$12.5
-$55.5
$11.8
$16.7
$65.6
$16.3
$22.5
$90.7
$19.8
$27.3
$110.2
Market and Productivity
-$0.1
-$0.1
-$0.1
-$0.2
-$0.2
-$0.2
-$0.1
-$0.1
-$0.1
$0.6
$0.6
$0.7
Changes in dredging costs
-$0.1
-$0.1
-$0.1
-$0.1
-$0.2
-$0.2
-$0.1
-$0.1
-$0.1
$0.6
$0.6
$0.7
Reduced water withdrawals'5
$0.0
<$0.0
$0.0
$0.0
Air-related effects
-$30.3
-$31.6
-$20.9
-$4.8
Changes in C02 air emissions0
-$30.3
-$31.6
-$20.9
-$4.8
Total"
-$41.0
-$43.6
-$86.6
$14.8
$19.6
$68.5
$35.1
$41.3
$109.4
$98.4
$105.9
$188.9
EPA-821-R-19-011
ES-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Executive Summary
Table ES-2: Summary of Total Annualized Benefits at 3 Percent (Millions; 2018$)
Benefit Category
Option la
Option 2a
Option 3a
Option 4a
Low
Mid
High
Low
Mid
High
Low
Mid
High
Low
Mid
High
a. Negative values represent forgone benefits and positive values represent realized benefits.
b. "<$0.0" indicates that monetary values are greater than -$0.1 million but less than $0.00 million.
c. The EPA estimated the air-related benefits for Option 2 using the IPM sensitivity analysis scenario that includes the ACE rule in
the baseline (IPM-ACE). EPA extrapolated estimates for Options 1 and 3 air-related benefits from the estimate for Option 2 that is
based on IPM-ACE outputs. The values for Option 4 air-related benefits were estimated using the IPM analysis scenario that does
not include the ACE rule in the baseline.
d. Values for individual benefit categories may not sum to the total due to independent rounding.
Source: U.S. EPA Analysis, 2019
Table ES-3: Summary of Total Annualized Benefits at 7 Percent
(Millions; 2018$)
Benefit Category
Option la
Option 2a
Option 3a
Option 4a
Low
Mid
High
Low
Mid
High
Low
Mid
High
Low
Mid
High
Human Health
-$0.3
$23.6
$26.9
$54.0
Changes in IQ losses in children
from exposure to leadb
<$0.0
<$0.0
<$0.0
<$0.0
Changes in IQ losses in children
from exposure to mercuryb
-$0.1
-$0.6
-$0.6
-$0.3
Changes in cancer risk from
DBPs in drinking water
-$0.2
$24.2
$27.5
$54.3
Ecological Conditions and
Recreational Uses Changes
-$8.6
-$10.9
-$48.1
$10.1
$14.3
$56.1
$14.0
$19.4
$77.8
$17.0
$23.6
$94.6
Use and nonuse values for
water quality changes
-$8.6
-$10.9
-$48.1
$10.1
$14.3
$56.1
$14.0
$19.4
$77.8
$17.0
$23.6
$94.6
Market and Productivity
-$0.1
-$0.1
-$0.1
-$0.1
-$0.2
-$0.2
$0.0
-$0.1
-$0.1
$0.5
$0.5
$0.7
Changes in dredging costs
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.2
$0.0
-$0.1
-$0.1
$0.5
$0.5
$0.7
Reduced water withdrawals'5
$0.0
<$0.0
$0.0
$0.0
Air-related Effects
-$4.8
-$5.2
-$3.7
-$0.9
Changes in C02 air emissions0
-$4.8
-$5.2
-$3.7
-$0.9
Total"
-$13.7
-$16.0
-$53.3
$28.4
$32.6
$74.4
$37.1
$42.5
$100.9
$70.6
$77.2
$148.4
a. Negative values represent forgone benefits and positive values represent realized benefits.
b. "<$0.0" indicates that monetary values are greater than -$0.1 million but less than $0.00 million.
c. The EPA estimated the air-related benefits for Option 2 using the IPM sensitivity analysis scenario that includes the ACE rule in the
baseline (IPM-ACE). EPA extrapolated estimates for Options 1 and 3 air-related benefits from the estimate for Option 2 that is based
on IPM-ACE outputs. The values for Option 4 air-related benefits were estimated using the IPM analysis scenario that does not
include the ACE rule in the baseline.
d. Values for individual benefit categories may not sum to the total due to independent rounding.
Source: U.S. EPA Analysis, 2019
Social Costs of Regulatory Options
Table ES-5 presents the incremental costs attributable to the regulatory options, calculated as the difference
between each option and the baseline. The regulatory options generally result in cost savings across the four
options and discount rates, with the exception of Option 4 which results in additional costs at 3 percent
discount rate. Chapter 12 describes the social cost analysis. The compliance costs of the regulatory options
are detailed in the Regulatory Impact Analysis (RIA) document.
EPA-821-R-19-011
ES-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Executive Summary
Comparison of Benefits and Social Costs of Regulatory Options
In accordance with the requirements of Executive Order 12866: Regulatory Planning and Review and
Executive Order 13563: Improving Regulation and Regulatory Review, the EPA compared the benefits and
costs of each regulatory option. Table ES-5 presents the incremental monetized benefits and incremental
social costs attributable to the regulatory options, calculated as the difference between each option and the
baseline.
Table ES-5: Total Annualized Benefits and Social Costs by Regulatory
Option and Discount Rate (Millions; 2018$)
Regulatory Option
Total Monetized Benefits
Total Costs
Low
Mid
High
3% Discount Rate
Option 1
-$41.0
-$43.6
-$86.6
-$130.6
Option 2
$14.8
$19.6
$68.5
-$136.3
Option 3
$35.1
$41.3
$109.4
-$90.1
Option 4
$98.4
$105.9
$188.9
$11.9
7% Discount Rate
Option 1
-$13.7
-$16.0
-$53.3
-$154.0
Option 2
$28.4
$32.6
$74.4
-$166.2
Option 3
$37.1
$42.5
$100.9
-$119.5
Option 4
$70.6
$77.2
$148.4
-$27.3
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
ES-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
1 Introduction
The EPA is proposing a regulation that would revise the technology-based ELGs for the steam electric power
generating point source category, 40 CFRpart 423, which the EPA promulgated in November 2015 (80 FR
67838). The proposed rule would revise certain effluent limitations based on BAT and pretreatment standards
for existing sources for two wastestreams: FGD wastewater and bottom ash transport water (BA).
This document presents an analysis of the social benefits and social costs of the regulatory options, including
the proposed option (Option 2), and complements other analyses the EPA conducted in support of this
proposal, described in separate documents:
• Supplemental Environmental Assessment for the Reconsideration of the Effluent Guidelines and
Standards for the Steam Electric Power Generating Point Source Category (Supplemental EA; U.S.
EPA, 2019a). The Supplemental EA summarizes the environmental and human health improvements
that are expected to result from implementation of the proposed ELGs.
• Supplemental Technical Development Document for the Reconsideration of the Effluent Guidelines
and Standards for the Steam Electric Power Generating Point Source Category (Supplemental TDD;
U.S. EPA, 2019b). The Supplemental TDD provides background on the ELGs; industry description;
wastewater characterization and identification of pollutants of concern; and treatment technologies
and pollution prevention techniques. It also documents the EPA's engineering analyses to support the
regulatory options including plant-specific compliance cost estimates, pollutant loadings, and non-
water quality impact assessment.
• Regulatory Impact Analysis for Proposed Revisions to the Effluent Limitations Guidelines and
Standards for the Steam Electric Power Generating Point Source Category (RIA; U.S. EPA, 2019c).
The RIA describes the EPA's analysis of the costs and economic impacts of the regulatory options.
This analysis provides the basis for social cost estimates presented in Chapter 12 of this document.
The RIA also provides information pertinent to meeting several legislative and administrative
requirements, including the Regulatory Flexibility Act of 1980 (as amended by the Small Business
Regulatory Enforcement Fairness Act [SBREFA] of 1996), the Unfunded Mandates Reform Act of
1995, Executive Order 13211 on Actions Concerning Regulations That Significantly Affect Energy
Supply, Distribution, or Use, and others.
The rest of this chapter discusses aspects of the regulatory options that are salient to EPA's analysis of the
social benefits and social costs of the proposal and summarizes key analytic assumptions used throughout this
document.
The analyses of the regulatory options are based on data generated or obtained in accordance with the EPA's
Quality Policy and Information Quality Guidelines. The EPA's quality assurance (QA) and quality control
(QC) activities for this rulemaking include the development, approval and implementation of Quality
Assurance Project Plans for the use of environmental data generated or collected from all sampling and
analyses, existing databases and literature searches, and for the development of any models which used
environmental data. Unless otherwise stated within this document, the data used and associated data analyses
were evaluated as described in these quality assurance documents to ensure they are of known and
EPA-821-R-19-011
1-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
documented quality, meet the EPA's requirements for objectivity, integrity and utility, and are appropriate for
the intended use.
1.1 Steam Electric Power Plants
The ELGs for the Steam Electric Power Generating Point Source Category apply to a subset of the electric
power industry, namely those plants "with discharges resulting from the operation of a generating unit by an
establishment whose generation of electricity is the predominant source of revenue or principal reason for
operation, and whose generation of electricity results primarily from a process utilizing fossil-type fuel (coal,
oil, or gas), fuel derived from fossil fuel (e.g., petroleum coke, synthesis gas), or nuclear fuel in conjunction
with athermal cycle employing the steam water system as the thermodynamic medium" (40 CFR423.10).
Based on data the EPA obtained from the U.S. Department of Energy (DOE) (U.S. DOE, 2017), the 2010
Questionnaire for the Steam Electric Power Generating Effluent Guidelines (industry survey; U.S. EPA,
2010c) and other sources (see Supplemental TDD, U.S. EPA, 2019b), as well as adjustments to the 2015 rule
universe to account for actual or announced unit and plant retirements or conversions, the EPA estimates that
there are 951 plants in the steam electric power generating industry. Of these, only a subset may incur
compliance costs under the regulatory options: coal fired power plants that discharge bottom ash transport
water or FGD wastewater. See Supplemental TDD and RIA for details (U.S. EPA, 2019b; 2019c).
1.2 Baseline and Regulatory Options Analyzed
The EPA presents four regulatory options (see Table 1-1). These options differ in the stringency of controls
and applicability of these controls to units or plants based on generation capacity, net power generation, and
scrubber purge flow (see Supplemental TDD for a detailed discussion of the options and the associated
treatment technology bases). Additionally, under Options 1, 2 and 3, steam electric power plants may elect to
participate in the Voluntary Incentive Program (VIP) which requires them to meet more stringent limits for
FGD wastewater in exchange for additional time to comply with those limits.
The baseline for this analysis reflects applicable requirements (in absence of any new final EPA action).2 The
agency estimated and presents in this report the water quality and other environmental effects of bottom ash
transport water and FDG wastewater discharges under both this 2015 rule baseline and each of the four
regulatory options presented in Table 1-1. The Agency calculated the difference between the baseline and the
regulatory options to determine the net effect of any regulatory options. The changes attributable to the
regulatory options are the difference between each option and the baseline.
This includes the 2015 rule as well as the September, 2017 postponement rule which delayed the earliest compliance date for the
ELGs applicable to FGD wastewater and bottom ash transport water.
EPA-821-R-19-011 1-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
Table 1-1: Regulatory Options
Wastestream
Subcategory
Technology Basis for BAT/PSES Regulatory Options3
2015 Rule (Baseline)
Option 1
Option 2
Option 3
Option 4
FGD
Wastewater
NAb
Chemical Precipitation
+ HRTR Biological
Treatment
Chemical Precipitation
Chemical Precipitation
+ LRTR Biological
Treatment
Chemical Precipitation
+ LRTR Biological
Treatment
Membrane Filtration
High FGD Flow Facilities:
Plant-level scrubber
purge flow >4 MGD
NS
NS
Chemical Precipitation
Chemical Precipitation
Chemical Precipitation
Low Utilization Boilers:
All units have net
generation < 876,000
MWh
NS
NS
Chemical Precipitation
NS
NS
Boilers retiring by 2028
NS
Surface Impoundment
Surface Impoundment
Surface Impoundment
Surface Impoundment
FGD Wastewater Voluntary Incentives
Program (Direct Dischargers Only)
Chemical Precipitation
+ Evaporation
Membrane Filtration
Membrane Filtration
Membrane Filtration
NA
Bottom Ash
Transport
Water
NAb
Dry Handling / Closed
loop
Dry Handling or High
Recycle Rate Systems
Dry Handling or High
Recycle Rate Systems
Dry Handling or High
Recycle Rate Systems
Dry Handling or High
Recycle Rate Systems
Low Utilization Boilers:
All units have net
generation < 876,000
MWh
NS
NS
Surface Impoundment+
BMP Plan
NS
NS
Boilers retiring by 2028
NS
Surface Impoundment
Surface Impoundment
Surface Impoundment
Surface Impoundment
Abbreviations: BMP = Best Management Practice; HRTR = High Residence Time Reduction; LRTR = Low Residence Time Reduction; NS = Not subcategorized; NA = Not applicable
a. See Supplemental TDD for a description of these technologies.
b. The 2015 rule subcategorized units with nameplate capacity 50 MW or less and the EPA is not revising requirements for these units in this proposal.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
1-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
1.3 Analytic Framework
The analytic framework of this benefit-cost analysis (BCA) includes basic components used consistently
throughout the analysis of social benefits and social costs3 of the regulatory options:
1. All values are presented in 2018 dollars;
2. Future benefits and costs are discounted using rates of 3 percent and 7 percent back to 2020, which is
the projected promulgation year for a final rule;
3. Benefits and costs are analyzed over a 27-year period (2021 to 2047);
4. Benefits and costs are annualized;
5. Positive values represent improvements in environmental conditions, whereas negative values
represent forgone benefits of the regulatory options compared to the baseline; and
6. Future values account for annual U.S. population and income growth, unless noted otherwise.
These components are discussed in the sections below.
The EPA's analysis of the regulatory options generally follows the methodology the Agency used previously
to analyze the 2015 rule (see Benefit and Cost Analysis for the Final Effluent Limitations Guidelines and
Standards for the Steam Electric Power Generating Point Source Category (U.S. EPA, 2015a). In analyzing
the regulatory options, however, the EPA made several changes relative to the analysis of the 2015 rule:
• The EPA used revised inputs that reflect the costs and loads estimated for the regulatory options (see
TDD and RIA for details).
• The EPA updated the baseline information to incorporate changes in the universe and operational
characteristics of steam electric power plants such as electricicty generating unit retirements and fuel
conversions since the analysis of the 2015 rule. The EPA also incorporated updated information on
the technologies and other controls that plants employ. See the Supplemental TDD for details on the
changes (U.S. EPA, 2019b). The current analysis focuses only on the two wastestreams addressed in
this proposal: bottom ash transport water and FGD wastewater..
• Given the changes in the universe of steam electric power plants since the 2015 rule was promulgated
and advances in treatment technologies, and that this proposal is specific to a subset of wastestreams
from the 2015 rule, the EPA first modeled the compliance response, pollutant loadings, costs, and
benefit estimates for the baseline requirements (see Supplemental TDD for a detailed discussion of
the baseline). The EPA then modeled the same for each regulatory option.
• The EPA revised assumptions to use more recent data (e.g., analysis year, compliance period, dollar
year adjustments).
Unless otherwise noted, costs represented in this document are social costs.
EPA-821-R-19-011 1-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
• Finally, the EPA made certain changes to the methodologies to address environmental stressors not
quantified as part of the 2015 rule analysis, be consistent with approaches used by the Agency for
other rules, and/or incorporate recent advances in the health risk and resource valuation research.
These changes are described in the relevant sections of this document, and summarized in Appendix A.
1.3.1 Constant Prices
This BCA applies a year 2018 constant price level to all future monetary values of costs and benefits. Some
monetary values of benefits and costs are based on actual past market price data for goods or services, while
others are based on other measures of values, such as household willingness-to-pay (WTP) surveys used to
monetize ecological changes resulting from surface water quality changes. This BCA updates market and
non-market prices using the Consumer Price Index (CPI), Gross Domestic Product (GDP) implicit price
deflator, or Construction Cost Index (CCI).4
1.3.2 Discount Rate and Year
This BCA estimates the annualized value of future benefits using two discount rates: 3 percent and 7 percent.
The 3 percent discount rate reflects society's valuation of differences in the timing of consumption; the
7 percent discount rate reflects the opportunity cost of capital to society. In Circular A-4, the Office of
Management and Budget (OMB) recommends that 3 percent be used when a regulation affects private
consumption, and 7 percent in evaluating a regulation that would mainly displace or alter the use of capital in
the private sector (OMB, 2003; updated 2009). The same discount rates are used for both benefits and costs.
All future cost and benefit values are discounted back to 2020, which is the anticipated rule promulgation
year.5
1.3.3 Period of Analysis
Benefits are expected to begin accruing when each plant implements the control technologies needed to
comply with any applicable BAT effluent limits or pretreatment standards. As discussed in the RIA (in
Chapter 3: Compliance Costs), for the purpose of the economic impact and benefit analysis, the EPA
generally assumes that plants would implement for bottom ash transport water control technologies to meet
the applicable rule limitations and standards as their permits are renewed over the period of 2021 through
2023. However, some regulatory options provide a longer period to meet FGD effluent limits. Under Options
1, 2 and 3 plants may implement FGD controls as late as 20286 and under Option 4, plants have until 2028 to
meet FGD wastewater controls based on the membrane technology.7 This schedule reflects differing levels of
4 To update the value of a Statistical Life (VSL), the EPA used the GDP deflator and the elasticity of VSL with respect to income
of 0.4, as recommended in EPA's Guidelines for preparing Economic Analysis (U.S. EPA 2010a). The EPA used the GDP
deflator to update the value of an IQ point, CPI to update the WTP for surface water quality improvements, cost of illness (COI)
estimates, and the price of water purchase, and the CCI to update the cost of dredging navigational waterways and reservoirs.
5 In its analysis of the 2015 rule, the EPA presented benefits in 2013 dollars and discounted these benefits costs to 2015 (see U.S.
EPA, 2015a).
6 The VIP program under Options 1, 2 and 3 allows facilities to implement FGD controls as late as 2028. Plants that are not
participating in the VIP program may implement FGD controls as late as 2023 under Option 1 and as late as 2025 under Options
2 and 3.
7 Other dates may apply to subcategories of facilities as described in Section 3.2.1.
EPA-821-R-19-011
1-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
controls that may be needed to meet limits under different options as compared to the baseline and recognizes
that control technology implementation is likely to be staggered over time across the universe of steam
electric power plants.
The different compliance years between options, wastestreams, and plants means that environmental changes
may occur in a staggered fashion over the analysis period as plants implement control technologies to meet
applicable limits under each option. To analyze environmental changes from the baseline and resulting
benefits, the EPA used the annual average of loadings or other environmental changes (e.g., air emissions,
water withdrawals) projected over the analysis period (2021-2047) and assumed that any resulting benefits
would begin in 2021.
The period of analysis extends to 2047 to capture the estimated life of the compliance technology at any
steam electric power plant (20 or more years), starting from the year of technology implementation, which
can be as late as 2028.
1.3.4 Annualization of future costs and benefits
Consistent with the analysis of the costs, the EPA assumes that plants implement necessary technologies to
meet the specified limits at the start of each year. The EPA used the following equation to annualize the future
stream of costs and benefits:
Equation 1-1.
r(PV)
AV = —
(1 + r)[l — (1 + r)_n]
Where A V is the annualized value, PVis the present value, r is the discount rate (3 percent or 7 percent), and n
is the number of years (27 years).
1.3.5 Direction of Environmental Changes and Benefits
The technology bases or subcategorizations shown in Table 1-1 for some regulatory options yield effluent
limits that may be less stringent than the baseline. This is true, for example, for options that base FGD
effluent limits on chemical precipitation only, or for subcategory options under which some plants can use
best management practices (BMP) to control bottom ash wastewater discharges. Additionally, the delayed
effective deadline for FGD limits under some options, such as the 2028 deadline for meeting FGD limits
based on membrane technology under Option 4, prolong the period when plants would continue to operate
their existing systems and discharge at current levels. The combination of these factors means that some
options can be expected to provide negative benefits (disbenefits) when compared to the baseline. This
document uses the generic term "benefits" whether the changes are truly beneficial or are detrimental to
society (reduce social welfare). The sign, positive or negative, communicates the direction of the effects.
Under this convention, positive benefit values indicate improvements in social welfare under the option as
compared to the baseline. This effect is typically in the opposite direction as the change in environmental
effects. For example, lower effluent pollutant concentrations (negative changes) reduce the incidence of the
health effects being quantified (negative changes) and avoid excess mortality resulting from the exposure
(positive changes).
EPA-821-R-19-011
1-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
1.3.6 Population and Income Growth
To account for future population growth or decline, the EPA used the U.S. Census Bureau population
forecasts for the United States from 2017 through 2060 (U.S. Census Bureau, 2017a). The EPA used the
growth projections for each year to adjust affected population estimates for future years (i.e.. from 2021 to
2047).
Also, since WTP is expected to increase as income increases, the EPA accounted for income growth for
estimating the value of avoided premature mortality based on the value of a statistical life (VSL) and WTP for
water quality (WQ) improvements. To develop adjustment factors for VSL, the EPA first used income growth
factors in the Environmental Benefits Mapping and Analysis Program (BenMAP) database between 1990 and
2024 to estimate a linear regression model. Using coefficient estimates from the linear regression, the EPA
extrapolated the income growth factors for years 2025-2047. The EPA applied the projected income data
along with the income elasticity for the respective models (VSL and meta-regression) to adjust the VSL and
WQ meta-analysis estimates of WTP in future years.8
1.4 Organization of the Benefit and Cost Analysis Report
This BCA report presents the EPA's analysis of the benefits of the regulatory options, assessment of the total
social costs, and comparison of the social costs and monetized benefits.
The remainder of this report is organized as follows:
• Chapter 2 provides an overview of the main benefits expected to result from the implementation of
the proposed regulatory options.
• Chapter 3 describes the EPA's estimates of the environmental changes resulting from the regulatory
options, including water quality modeling that underlays estimates of several categories of benefits.
• Chapters 4 and 5 details the methods and results of the EPA's analysis of human health benefits from
changes in pollutant exposure via the drinking water and fish ingestion pathways, respectively.
• Chapter 6 discusses the EPA's analysis of the nonmarket benefits of changes in surface water quality
resulting from the regulatory options.
• Chapter 7 discusses expected changes in benefits to threatened and endangered (T&E) species.
• Chapter 8 describes the EPA's analysis of benefits associated with changes in emissions of air
pollutants associated with energy use, transportation, and the profile of electricity generation for the
regulatory options.
• Chapter 9 discusses benefits arising from changes in groundwater withdrawals.
• Chapter 10 describes benefits from changes in maintenance dredging of navigational channels and
reservoirs.
These extrapolated income growth factors were originally developed for the EPA's COBRA tool
(http://epa.gov/statelocalclimate/resources/cobra.html). The latest public version is 3,2 released in May 2018.
EPA-821-R-19-011
1-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
1: Introduction
• Chapter 11 summarizes monetized benefits across benefit categories.
• Chapter 12 summarizes the social costs of the four regulatory options.
• Chapter 13 addresses the requirements of Executive Orders that the EPA is required to satisfy for this
proposal, notably Executive Order 12866, which requires the EPA to compare the benefits and social
costs of its actions.
• Chapter 14 details the EPA's analysis of the distribution of benefits across socioeconomic groups to
fulfill requirements under Executive Order (E.O.) 12898 on Environmental Justice.
• Chapter 15 provides references cited in the text.
Several appendices provide additional details on selected aspects of analyses described in the main text of the
report.
EPA-821-R-19-011
1-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
2 Benefits Overview
This chapter provides an overview of the welfare effects to society resulting from changes in pollutant
loadings due to implementation of the regulatory options. The EPA expects the regulatory options to change
discharge loads of various categories of pollutants when fully implemented. The categories of pollutants
include conventional (such as total suspended solids (TSS), biochemical oxygen demand (BOD), and oil and
grease), priority (such as mercury (Hg), arsenic (As), and selenium (Se)), and non-conventional pollutants
(such as total nitrogen (TN), total phosphorus (TP), chemical oxygen demand (COD) and total dissolved
solids (TDS)).
Table 2-1 presents estimated annual pollutant loads under full implementation of the effluent limits for the
baseline and the four regulatory options. The Supplemental TDD provides further detail on the loading
changes (U.S. EPA, 2019b).
Table 2-1: Estimated Annual Pollutant Loadings and Changes in Loadings for
Baseline and Regulatory Options
Estimated Total Industry Pollutant
Estimated Changes3 in Pollutant
Regulatory Option
Loadings
Loadings from Baseline
(pounds per year)
(pounds per year)
Baseline
1,670,000,000
NA
1
1,680,000,000
13,400,000
2
1,560,000,000
-104,000,000
3
1,390,000,000
-276,000,000
4
342,000,000
-1,320,000,000
NA: Not applicable to the baseline
Note: Pollutant loadings values are rounded to three significant figures. See EA for details (U.S. EPA, 2019a).
a. Negative values represent loading reductions and positive values represent loading increases, compared to
the baseline.
Source: U.S. EPA Analysis, 2019
As discussed in Section 1.3.4, some of the options may increase pollutant loads for some plants,
wastestreams, pollutants, or years, when compared to the baseline. Consequently, technology options
resulting in overall increase in pollutant loads would result in forgone benefits to society while options
resulting in load reduction would result in realized benefits. Furthermore, whether a regulatory option
increases or reduces loadings depends on the particular plant, pollutant, and timing of the comparison to
baseline conditions. Section 3.2 discusses the temporal profile of pollutant loads in further detail.
Changes estimated for proposed Option 2 and Option 3 include effects of the VIP. Because the VIP is
voluntary, the set of plants participating in the program is uncertain. For the purpose of the costs and benefits
analyses, the EPA estimated VIP participants by comparing the estimated costs of the two technologies for
each affected facility and assuming that a plant owner would select the less costly of the two. The Agency
estimated that 18 steam electric power plants may choose to participate in the VIP under Option 2 and 23
plants may choose to participate in the VIP under Option 3. The facilities which the EPA estimates VIP may
be the least-cost option range in FGD wastewater flows, nameplate capacity, capacity utilization, and
location. The EPA cost estimates for the VIP tend to be lower at facilities where no treatment has been
EPA-821-R-19-011
2-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
installed beyond surface impoundments, however even for this group of facilities biological systems are still
often least-cost.
Effects of the regulatory options in comparison to the 2015 rule also include other effects of the
implementation of control technologies or other changes in plant operations, such as changes in emissions of
air pollutants (e.g., carbon dioxide (CO2), nitrogen oxides (NOx), and sulfur dioxide (SO2)) which result in
benefits forgone to society in the form of increased mortality and CO2 impacts on environmental quality and
economic activities. Other effects include changes in water use, which provide benefits in the form of
increased availability of surface water and groundwater.
This chapter also provides a brief discussion of the effects of pollutants found in bottom ash transport water
and FGD wastewater addressed by the regulatory options on human health and ecosystem services, and a
framework for understanding the benefits expected to be achieved by these options. For a more detailed
description of steam electric pollutants, their fate, transport, and impacts on human health and environment,
see the Supplemental EA document (U.S. EPA, 2019a).
Figure 2-1 summarizes the potential effects of the regulatory options, the expected environmental changes,
and categories of social welfare effects as well as the EPA's approach to analyzing those welfare effects. The
EPA was not able to bring the same depth of analysis to all categories of social welfare effects because of
imperfect understanding of the link between discharge changes or other environmental effects of the
regulatory options and welfare effect categories, and how society values some of these effects. The EPA was
able to quantify and monetize some welfare effects, quantify but not monetize other welfare effects, and
assess still other welfare effects only qualitatively. The remainder of this chapter provides a qualitative
discussion of the social welfare effect categories applicable to this rule, including human health effects,
ecological effects, economic productivity, and changes in air pollution, solid waste generation, and water
withdrawals. Some estimates of the monetary value of social welfare changes presented in this document rely
on models with a variety of assumptions, limitations and uncertainties discussed in more detail in Chapters 3
through 10 for the relevant benefit categories.
EPA-821-R-19-011
2-2
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
2: Benefits Overview
Figure 2-1: Summary of Benefits Resulting from the Regulatory Options.
Environmental
Change
Effect of
Proposed ELGs
Fish tissue
contamination
Surface water
quality
Avoided cost of dredging
Qualitative discussion
Changes in toxic,
bioaccumulative,
and other harmful
pollutants to surface
waters
Valuation
CO
Benefit Category
Value of an IQ point
Count of human health criteria
exceedances (non-monetized)
Qualitative discussion
COI
VSL
WTP for water quality improvements
Count of aquatic life criteria
exceedances (non-monetized)
Economic Productivity
Changes in dredgjng costs for maintaining navigational
waterways and reservoir capacity
Changes in drinking water treatment costs
Changes in property values
Changes in tourism
Changes in commercial fishery yields
Human Health Effects (Fish Consumption)
Changes in IQ losses in children from mercury and lead exposure
Changes in cardiovascular disease from lead exposure
Changes in cancer cases from arsenic exposure
Changes in cases of other cancer and non-cancer health effects
Human Health Effects (Drinking Water Exposure)
Changes in incidence of bladder cancer from exposure to
halogenated disinfection byproducts in treated water
Ecological Conditions
Changes in recreational and non-use values
Changes in threatened and endangered (T&E) species protection
Change in:
• Power
requirements
• Trucking
• Electricity
generation
Conversion to dry
systems
Change in air
emissions of CO2,
NOx, and S02
Changes in water
use
N
Ability to market ash
for beneficial use
Groundwater
withdrawals
Surface water
withdrawals
Change in premature mortality, non-fatal heart attacks, hospital
admissions, emergency department visits, upper and lower
respiratory symptoms, acute bronchitis, aggravated asthma, lost
work days and acute respiratory symptoms
Changes in C02 impacts
Social cost of carbon
Qualitative discussion
Changes in disposal costs
Changes in life-cycle impacts and costs of virgin raw materials
Qualitative discussion
Changes in groundwater availability
Avoided cost of water purchase
Changes in vulnerability to drought
Changes in impingement and entrainment mortality
Qualitative discussion
COI = Cost of illness; VSL = Value of Statistical Life; WTP = Willingness to Pay
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
2-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
2.1 Human Health Impacts Associated with Changes in Surface Water Quality
Pollutants present in steam electric power plant discharges can cause a variety of adverse human health
effects. Chapter 3 describes the approach the EPA used to estimate changes in pollutant levels in waters.
More details on the fate, transport, and exposure risks of steam electric pollutants are provided in the EA
(U.S. EPA, 2015b; 2019a).
Human health effects are typically analyzed by estimating the change in the expected number of adverse
human health events in the exposed population resulting from changes in effluent discharges. While some
health effects (e.g., cancer) are relatively well understood and can be quantified in a benefits analysis, others
are less well characterized and cannot be assessed with the same rigor, or at all.
The regulatory options affect human health risk by changing exposure to pollutants in water via two principal
exposure pathways discussed below: (1) treated water sourced from surface waters affected by steam electric
power plant discharges and (2) fish and shellfish taken from waterways affected by steam electric power plant
discharges. The regulatory options also affect human health risk by changing air emissions of pollutants via
shifts in the profile of electricity generation, changes in auxiliary electricity use, and transportation; these
effects are discussed separately in Section 2.4.
2.1.1 Drinking Water
Pollutants discharged by steam electric power plants to surface waters may affect the quality of water used for
public drinking supplies. People may then be exposed to harmful constituents in treated water through oral
ingestion, as well as inhalation and dermal absorption (e.g., showering, bathing). The pollutants may not be
removed adequately during treatment at a drinking water treatment plant, or constituents found in steam
electric power plant discharges may interact with drinking water treatment processes and contribute to the
formation of disinfection byproducts (DBPs). For example, bromide and other halogens are precursors to the
formation of trihalomethanes, a group of potentially carcinogenic contaminants.
Public drinking water supplies are subject to legally enforceable maximum contaminant levels (MCLs)
established by EPA (U.S. EPA, 2018a). As the term implies, an MCL for drinking water specifies the highest
level of a contaminant that is allowed in drinking water. The MCL is based on the MCL Goal (MCLG), which
is the level of a contaminant in drinking water below which there is no known or expected risk to human
health. The EPA sets the MCL as close to the MCLG as possible, with consideration for the best available
treatment technologies and costs. Table 2-2 shows the MCL and MCLG for selected constituents or
constituent derivatives of steam electric power plant effluent.
Table 2-2: Drinking Water Maximum Contaminant Levels and Goal for Selected Pollutants in Steam
Electric FGD Wastewater or Bottom Ash Transport Water Discharges
Pollutant
MCL
MCLG
(mg/L)
(mg/L)
Antimony
0.006
0.006
Arsenic
0.01
0
Barium
2.0
2.0
Beryllium
0.004
0.004
Bromate
0.010
0
Cadmium
0.005
0.005
Chromium (total)
0.1
0.1
EPA-821-R-19-011
2-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 2: Benefits Overview
Table 2-2: Drinking Water Maximum Contaminant Levels and Goal for Selected Pollutants in Steam
Electric FGD Wastewater or Bottom Ash Transport Water Discharges
Pollutant
MCL
MCLG
(mg/L)
(mg/L)
Copper3
1.3
1.3
Cyanide (free cyanide)
0.2
0.2
Lead3
0.015
0
Mercury
0.002
0.002
Nitrate-Nitrite as N
10 (Nitrate); 1 (Nitrite)
10 (Nitrate); 1 (Nitrite)
Selenium
0.05
0.05
Thallium
0.002
0.0005
Total trihalomethanes15
0.080
Not available
bromodichloromethane
Not available
0
bromoform
Not available
0
dibromochloromethane
Not available
0.06
chloroform
Not available
0.07
a. MCL value is based on action level.
b. Bromide, a constituent found in steam electric power plant effluent, is a trihalomethane precursor.
Source: 40 CFR 141.53 as summarized in U.S. EPA (2018a): National Primary Drinking Water Regulation, EPA 816-F-09-004
Pursuant to MCLs, public drinking water supplies are tested and treated for pollutants that pose human health
risks. The EPA assumed compliance with existing MCLs. Nevertheless, for some pollutants that have an
MCL above the MCLG, there may be incremental benefits from reducing concentrations below the MCL.
Examples include arsenic, lead, and total trihalomethanes (TTHM).
As shown in Table 2-2, there are no "safe levels" for some these pollutants. Therefore, any reduction in
exposure to these pollutants is expected to yield benefits. The EPA estimated the changes in levels of
bromide, a trihalomethane precursor, downstream from steam electric power plant outfalls and estimated the
resulting changes in the incidence of bladder cancers associated with TTHM exposure. These benefits are
discussed in Section 4.4. The EPA did not evaluate potential benefits associated with other health endpoints
(e.g., reproductive effects, fetal development, and other cancers resulting from reduced TTHM exposure).
The value of health benefits is the monetary value that society is willing to pay to avoid the adverse health
effects. WTP to avoid morbidity or mortality is generally considered to be a comprehensive measure of the
costs of health care, losses in income, and pain and suffering of affected individuals and their caregivers. For
example, the value of a statistical life is based on estimates of society's WTP to avoid the risk of premature
mortality. The cost-of-illness (COI) approach is a less comprehensive measure: it allows valuation of a
particular type of non-fatal illness by placing monetary values on metrics, such as lost productivity and the
cost of health care and medications that can be monetized. The EPA used the VSL and COI to estimate the
benefits of changing excess mortality and morbidity associated with incremental bladder cancers in the
population estimated to be exposed to trihalomethanes attributable to of bottom ash transport water and/or
FGD wastewater bromide discharges. Arsenic and lead benefits were not modeled due to the very small
concentration changes in downstream reaches with drinking water intakes and, furthermore, because lead
found in supplied water is generally associated with water distribution rather than source water quality. See
Chapter 4 for details of this analysis.
EPA-821-R-19-011
2-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
2.1.2 Fish Consumption
Recreational anglers and subsistence fishers (and their household members) who consume fish caught in the
reaches downstream of steam electric power plants may be affected by changes in pollutant concentrations in
fish tissue. The EPA analyzed the following direct measures of change in risk to human health from exposure
to contaminated fish tissue:
1. Neurological effects to children ages 0 to 7 from exposure to lead;
2. Neurological effects to infants from in-utero exposure to mercury;
3. Incidence of skin cancer from exposure to arsenic9; and
4. Reduced risk of other cancer and non-cancer toxic effects.
The Agency evaluated changes in potential intellectual impairment, or intelligence quotient (IQ), resulting
from changes in childhood and in-utero exposures to lead and mercury. The EPA also translated changes in
the incidence of skin cancer into changes in the number of skin cancer cases.
For constituents with human health ambient water quality criteria, the change in the risk of other cancer and
non-cancer toxic effects from fish consumption is addressed indirectly in the EPA's assessment of changes in
exceedances of these criteria (see Section 5.7).
In the 2015 rule, the EPA used VSL to estimate the value of changes in the risk of premature mortality from
cardiovascular disease (CVD). The EPA performed a screening analysis of the regulatory options using the
same approach used in the 2015 analysis. This analysis showed very small changes in CVD mortality based
on changes in lead exposure under the options. See memorandum in the rule record for details (U.S. EPA,
2019g). The Agency is aware of more recent studies linking lead exposure and CVD, but determined that the
changes in lead exposure under the regulatory options are so small as to be unlikely to yield benefits. The
EPA used a COI approach to estimate the value of changes in the incidence of skin cancer, which are
generally non-fatal (see Section 5.5). Some health effects of changes in exposure to steam electric pollutants,
such as neurological effects to children and infants exposed to lead and mercury, are measured based on
avoided IQ losses. Changes in IQ cannot be valued based on WTP approaches since available economic
research provides little empirical data on society's WTP to avoid IQ losses. Instead, the EPA calculated
monetary values for changes in neurological and cognitive damages based on the impact of an additional IQ
point on an individual's future earnings and the cost of compensatory education for children with learning
disabilities. These estimates represent only one component of society's WTP to avoid adverse neurological
effects and therefore produce a partial measure of the monetary value from changes in exposure to lead and
mercury. Employed alone, these monetary values would underestimate society's WTP to avoid adverse
neurological effects. See Sections 5.3 and 5.4 for applications of this method to valuing health effects in
children and infants from changes in exposure to lead and mercury.
The EPA is currently revising its cancer assessment of arsenic to reflect new data on internal cancers including bladder and lung
cancers associated with arsenic exposure via oral ingestion (U.S. EPA, 2010b). Because cancer slope factors for internal organs
have not been finalized, the Agency did not consider these effects in the analysis of the proposed ELG.
EPA-821-R-19-011
2-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
The EPA expects that there could also be health impacts via the fish consumption pathway arising from
changes in discharges of other steam electric pollutants, such as cadmium, selenium, and zinc. Analyses of
these health effects are not possible due to lack of data on a quantitative relationship between ingestion rate
and potential adverse health effects.
Despite numerous studies conducted by the EPA and other researchers, dose-response functions are available
only for a handful of health endpoints associated with steam electric pollutants. In addition, the available
research does not always allow complete economic evaluation, even for quantifiable health effects. For
example, the EPA's analysis omits the following health effects: low birth weight and neonatal mortality from
in-utero exposure to lead, decreased postnatal growth in children ages one to 16, delayed puberty,
immunological effects, decreased hearing and motor function (U.S. EPA, 2009a; 2019h); effects to adults
from exposure to lead (e.g., cardiovascular diseases, decreased kidney function, reproductive effects,
immunological effects, cancer and nervous system disorders) (U.S. EPA, 2009d; 2013a; 2019h); effects to
adults from exposure to mercury, including vision defects, hand-eye coordination, hearing loss, tremors,
cerebellar changes, and others (Mergler et al., 2007; CDC, 2009); and other cancer and non-cancer effects
from exposure to other steam electric pollutants. Therefore, the total monetary value of changes in human
health effects included in this analysis represent only a subset of the potential health benefits (or forgone
benefits) that are expected to result from the regulatory options.
2.1.3 Complementary Measure of Human Health Impacts
The EPA quantified but did not monetize changes in pollutant concentrations in excess of human health-based
national recommended water quality criteria (NRWQC). This analysis provides a measure of the change in
cancer and non-cancer health risk by comparing the number of receiving reaches exceeding health-based
NRWQC for steam electric pollutants in the baseline to the number exceeding NRWQC under the regulatory
options (Section 5.7).
Because the NRWQC in this analysis are set at levels to protect human health through ingestion of water and
aquatic organisms, changes in the frequency at which human health-based NRWQC are exceeded could
translate into changes in risk to human health. This analysis should be viewed as an indirect indicator of
changes in risk to human health because it does not reflect the magnitude of human health risk changes or the
population over which those changes would occur.
2.2 Ecological Impacts Associated with Changes in Surface Water Quality
The composition of steam electric power plant wastewater depends on a variety of factors, such as fuel
composition, air pollution control technologies used, and wastewater management techniques used.
Wastewater often contains toxic pollutants such as aluminum, arsenic, boron, cadmium, chromium, copper,
iron, lead, manganese, mercury, nickel, selenium, thallium, vanadium, and zinc (U.S. EPA, 2019a).
Discharges of these pollutants to surface water can have a wide variety of environmental effects, including
fish kills, reduction in the survival and growth of aquatic organisms, behavioral and physiological effects in
wildlife, and degradation of aquatic habitat in the vicinity of steam electric power plant discharges (U.S. EPA,
2019a). The adverse effects associated with releases of steam electric pollutants depend on many factors such
as the chemical-specific properties of the effluent, the mechanism, medium, and timing of releases, and site-
specific environmental conditions.
EPA-821-R-19-011
2-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
The modeled changes in environmental impacts is quite small. Still, EPA expects the ecological impacts from
the regulatory options may include habitat changes for fresh- and saltwater plants, invertebrates, fish, and
amphibians, as well as terrestrial wildlife and birds that prey on aquatic organisms exposed to steam electric
pollutants. The change in pollutant loadings may affect ecosystem productivity in waterways and the health of
resident species, including threatened and endangered species. The proposed regulation may affect the general
health of fish and invertebrate populations, their propagation to waters, and fisheries for both commercial and
recreational purposes. Changes in water quality could also affect recreational activities such as swimming,
boating, fishing, and water skiing. Finally, the proposed regulation may affect nonuse values (e.g., option,
existence, and bequest values) of the affected water resources.
2.2.1 Changes in Surface Water Quality
The regulatory options may affect the value of ecosystem services provided by surface waters through
changes in the habitats or ecosystems (aquatic and terrestrial) that are affected by steam electric power plant
discharges. Society values changes in ecosystem services by a number of mechanisms, including increased
frequency of use and quality of the improved habitat for recreational activities (e.g., fishing, swimming, and
boating). Individuals also value the protection of habitats and species that may be adversely affected by FGD
wastewater and bottom ash transport water discharges, even when those individuals do not use or anticipate
future use of the affected waterways for recreational or other purposes, resulting in nonuse values.
The EPA quantified potential ecological impacts from the regulatory options by estimating in-waterway
concentrations of bottom ash transport water and FGD wastewater pollutants and translating water quality
estimates into a single numerical indicator (water quality index (WQI)). The EPA used the estimated change
in WQI as a quantitative estimate of ecological changes for this regulatory analysis. Section 3.4 of this report
provides detail on the parameters used in formulating the WQI and the WQI methodology and calculations. In
addition to estimating changes using the WQI, the EPA compared estimated pollutant concentrations to
freshwater chronic NRWQC for aquatic life (see Section 3.4.1.1). The Supplemental EA (U.S. EPA, 2019a)
details comparisons of the estimated concentrations in immediate receiving and downstream reaches to the
freshwater chronic NRWQC for aquatic life for individual pollutants.
A variety of primary methods exist for estimating recreational use values, including both revealed and stated
preference methods (Freeman, 2003). Where appropriate data are available or can be collected, revealed
preference methods can represent a preferred set of methods for estimating use values. Revealed preference
methods use observed behavior to infer users' values for environmental goods and services. Examples of
revealed preference methods include travel cost, hedonic pricing, and random utility (or site choice) models.
In contrast to direct use values, nonuse values are considered more difficult to estimate. Stated preference
methods, or benefit transfer based on stated preference studies, are the generally accepted techniques for
estimating these values (U.S. EPA, 2010a; U.S. OMB, 2003). Stated preference methods rely on carefully
designed surveys, which either (1) ask people about their WTP for particular ecological improvements, such
as increased protection of aquatic species or habitats with particular attributes, or (2) ask people to choose
between competing hypothetical "packages" of ecological improvements and household cost (Bateman et al.,
2006). In either case, values are estimated by statistical analysis of survey responses.
Although the use of primary research to estimate values is generally preferred because it affords the
opportunity for the valuation questions to closely match the policy scenario, the realities of the regulatory
process often dictate that benefit transfer is the only option for assessing certain types of non-market values
EPA-821-R-19-011
2-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
(Rosenberger and Johnston, 2007). Benefit transfer is described as the "practice of taking and adapting value
estimates from past research ... and using them ... to assess the value of a similar, but separate, change in a
different resource" (Smith et al. 2002, p. 134). It involves adapting research conducted for another purpose to
estimate values within a particular policy context (Bergstrom and De Civita, 1999). The EPA followed the
same methodology used in analyzing the 2015 rule (U.S. EPA, 2015a) and relied on a benefit transfer
approach based on a meta-analysis of surface water valuation studies to estimate the use and non-use benefits
of improved surface water quality resulting from the proposal. This analysis is presented in Chapter 6.
Valuation of water quality changes is an area of on-going research. The EPA will update the methods
employed to reflect the inclusion of rigorous and timely studies that will shed light on household values of
water quality changes and changes in the methods in the economics literature. Further research may also
include efforts to examine the feasibility of conducting regional water quality valuation studies that are
designed to be aggregated up to the national level where the study designs are consistent with the best
practices of the economics literature as well as the OMB Circular A-4 requirements.
2.2.2 Impacts on Threatened and Endangered Species
For threatened and endangered (T&E) species, even minor changes to reproductive rates and small levels of
mortality may represent a substantial portion of annual population growth. By changing the discharge of
steam electric pollutants to aquatic habitats, the regulatory options may affect the survivability of some T&E
species living in these habitats. These T&E species may have both use and nonuse values. However, given the
protected nature of T&E species and the fact that use activities, such as fishing or hunting, generally
constitute "take" which is illegal unless permitted, the majority of the economic value for T&E species comes
from nonuse values.10
The EPA quantified but did not monetize the potential effects of the regulatory options on T&E species. The
EPA constructed databases to determine which species are found in waters that may be affected by changes in
pollutant discharge from steam electric power plants. The EPA then queried these databases to identify
"affected areas" of those habitats where 1) receiving waters do not meet aquatic life-based NRWQC under the
baseline conditions; and 2) receiving waters exceed aquatic life-based NRWQC under regulatory options, or
vice versa. Because NRWQC are set at levels to protect aquatic organisms, reducing the frequency at which
aquatic life-based NRWQC are exceeded should translate into reduced risk to T&E species and potential
improvement in species population. Conversely, increasing the frequency of exceedances may increase risk to
T&E species and jeopardize their survival or recovery. Therefore, to estimate the benefits of the regulatory
options, the EPA identified the inhabited waterbodies that see changes in achievement of wildlife NRWQC as
a consequence of the regulatory options and used these data as a proxy for locations where T&E species
recovery could be affected. This analysis and results are presented in Chapter 7.
2.2.3 Changes in Sediment Contamination
Effluent discharges from steam electric power plants can also contaminate waterbody sediments. For
example, adsorption of arsenic, selenium, and other pollutants found in FGD wastewater and bottom ash
transport water discharges can result in accumulation of contaminated sediment on stream and lake beds
(Ruhl et al., 2012), posing a particular threat to benthic (i.e., bottom-dwelling) organisms. These pollutants
10 The Federal Endangered Species Act (ESA) defines "take" to mean "to harass, harm, pursue, hunt, shoot, wound, kill, trap, capture,
or collect, or to attempt to engage in any such conduct." 16 U.S. Code § 1532
EPA-821-R-19-011
2-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
can later be re-released into the water column and enter organisms at different trophic levels. Concentrations
of selenium and other pollutants in fish tissue of organisms of lower trophic levels can bio-magnify through
higher trophic levels, posing a threat to the food chain at large (Ruhl et al., 2012).
In waters receiving direct discharges from steam electric power plants, the EPA examined potential exposures
of ecological receptors (i.e.. sediment biota) to pollutants in contaminated sediment. Benthic organisms are
affected primarily by discharges of mercury, nickel, selenium, and cadmium (U.S. EPA, 2015b; 2019a). The
chemicals in steam electric power plant discharges may accumulate in living benthic organisms that obtain
their food from sediments and pose a threat to both the organism and humans consuming the organism. In
2015, the EPA evaluated potential risks to fish and waterfowl that feed on aquatic organisms with elevated
selenium levels and found that steam electric power plant selenium discharges elevated the risk of adverse
reproduction impacts among fish and mallards in immediate receiving waters (U.S. EPA, 2015b).
By changing discharges of pollutants to receiving reaches, the regulatory options may affect the future
contamination of waterbody sediments, thereby impacting benthic organisms and changing the probability
that pollutants could later be released into the water column and affect surface water quality and the
waterbody food chain. Due to data limitations, the EPA did not quantify or monetize the associated benefits.
2.3 Economic Productivity
The economic productivity changes estimated to result from the regulatory options may include changes in
beneficial use of coal ash and the resulting reduction in disposal costs. Other potential economic productivity
effects may stem from changes in contamination of public drinking water supplies and irrigation water;
changes in sediment deposition in reservoirs and navigational waterways; changes in tourism, commercial
fish harvests, and property values. Due to the small magnitude of changes in water quality estimated in the
Supplemental EA, the latter three categories are not monetized or discussed further in this document.
2.3.1 Marketability of Coal Ash for Beneficial Use
The regulatory options may prompt certain plants to convert from wet handling of bottom ash to dry handling.
This change could in turn allow plants to more readily market the CCR to beneficial uses. In particular,
bottom ashes can be used as substitutes for sand and gravel in fill applications. There are economic
productivity benefits from plants avoiding certain costs associated with disposing of the ashes as waste and
from society or users of the ash avoiding the cost and life-cycle effects associated with the displaced virgin
material. In the analysis of the 2015 rule, the EPA quantified the benefits from increased dry handling of fly
ash and bottom ash (see Chapter 10 in U.S. EPA, 2015a). That analysis showed that the economic value was
greatest for fly ash used in concrete production, and smallest for fly ash or bottom ash used as fill material.
Among the regulatory options considered for this proposal, Option 1 would not affect fly ash, the wastestream
responsible for the vast majority of projected benefits in this category in the 2015 rule (see U.S. EPA, 2015a),
while Options 2, 3, and 4 could affect fly ash to the extent facilities decide to encapsulate membrane filtration
brine with fly ash that is currently beneficially used. Since the EPA could not estimate which facilities might
use fly ash for encapsulation versus an alternative brine management method (e.g., deep well injection), this
potential change in fly ash beneficial use was not quantified, and represents an uncertainty in the analysis.
With respect to bottom ash, the EPA estimates that only Option 2 would affect the quantity of bottom ash
handled wet when compared to the baseline, and for that option the estimated increase in bottom ash handled
wet is small (total of 310,671 tons per year at 20 plants). See the Supplemental TDD for details (U.S. EPA,
EPA-821-R-19-011
2-10
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
2019b). Given the uncertainties surrounding changes in fly ash, the small changes in bottom ash, and the
uncertainty associated with projecting plant-specific changes in marketed bottom ash, the EPA did not
quantify this benefit category in the analysis of the proposed Option 2.
2.3.2 Water Supply and Use
The regulatory options are expected to change loading of steam electric pollutants to surface waters by small
amounts relative to the 2015 rule estimates (see U.S. EPA, 2015a), and thus may affect uses of these waters
for drinking water supply and agriculture:
• Drinking water treatment costs. The regulatory options have the potential to affect costs of drinking
water treatment (e.g., filtration and chemical treatment) by changing pollutant concentrations and
eutrophication in source waters (also see discussion of changes in bromide concentrations below).
Eutrophication is one of the main causes of taste and odor impairment in drinking water, which has a
major negative impact on public perceptions of drinking water safety. Additional treatment to address
foul tastes and odors can significantly increase the cost of public water supply. The Agency
conducted screening-level assessment to evaluate the potential for changes in costs incurred by public
drinking water systems and concluded that such changes, while they may exist, are likely to be
insignificant. The assessment involved identifying the pollutants for which treatment costs may vary
depending on source water quality, estimating changes in downstream concentrations of these
pollutants at the location of drinking water intakes, and determining whether modeled water quality
changes have the potential to affect drinking water treatment costs. Based on this analysis, the EPA
determined that there are no drinking water systems drawing water at levels that exceed an MCL for
metals and other toxics listed in Table 2-2 such as selenium and cyanide under either the baseline or
the four regulatory options, and only one drinking water intake is drawing water from a reach with
nitrate concentrations exceeding an MCL (10 mg/L) under the baseline. No changes in MCL
exceedances are expected under the regulatory options. At many drinking water treatment facilities,
treatment system operations do not generally respond to small incremental water quality changes for
one pollutant or a small subset of pollutants. Furthermore, associated operations costs are not
expected to change significantly due to small incremental changes in water quality. Accordingly, the
EPA did not conduct analysis of cost changes in publicly operated treatment systems.
• Reduction in bromide concentrations. Existing treatment technologies in the majority of public
drinking water sources are not designed to remove bromide (a constituent of FGD wastewater and
bottom ash transport waters) from raw surface waters. Bromide and other halogens found in source
water can react during routine drinking water treatment to generate harmful disinfection byproducts
(DBPs) (U.S. EPA, 2016). The EPA estimated the costs of controlling DBP levels to the MCL in
treated water as part of the Stage 2 Disinfectants and Disinfection Byproduct Rule (DBPR). These
costs include treatment technology changes as well as non-treatment costs such as routine monitoring
and operational evaluations. Public water systems (PWS) may adjust their current operations to
control DBP levels, such as changing disinfectant dosage, moving the chlorination point, or
enhancing coagulation and softening. These changes carry "negligible costs" (U.S. EPA, 2005a, page
7-19). Where those low-cost changes are not sufficient to meet the MCL, PWS may need to incur
irreversible capital costs to upgrade their treatment process to use alternative disinfection
technologies such as ozone, ultraviolet light, and chloride dioxide, switch to chloramines for residual
disinfection, or add a pre-treatment stage to remove DBP precursors (e.g., microfiltration,
EPA-821-R-19-011
2-11
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
ultrafiltration, aeration, or increased chlorine levels and contact time). Some drinking water treatment
facilities have already had to upgrade their treatment systems as a direct result of bromide discharges
from steam electric power plants (U.S. vs. Duke Energy, 2015; Rivin, 2015). In extreme cases, if
water treatment is not sufficient, an alternate water source needs to be substituted or developed
(Watson et al., 2012). Thus, increased bromide levels in raw source water could translate into
permanently higher drinking water treatment costs at some plants, in addition to posing increased
human health risk. Conversely, reducing bromide levels in source waters can reduce the health risk,
even where treatment changes have already occurred.11 In some cases, operation and maintenance
(O&M) costs may also be reduced. The EPA did not have data on drinking water treatment
technologies at potentially affected PWS or cost estimates for those technologies given changes in
bromide concentrations in source water. Since cost data were insufficeint, and treatment costs and
human health benefits overlap, the Agency estimated only the human health benefits of changes in
bromide discharges (see Section 2.1.1 for a discussion of this benefit category and Chapter 4 for the
analysis).
• Irrigation and other agricultural uses: Changes in steam electric pollutants discharges can affect
agricultural productivity by improving water quality used for irrigation and livestock watering (Clark
et al., 1985). Although elevated nutrient concentrations in irrigation water would not adversely affect
its usefulness for plants, concerns exist for potential residual effects due to steam electric pollutants,
such as arsenic, mercury, lead,cadmium, and selenium, entering the food chain. Further,
eutrophication promotes cyanobacteria blooms that can kill livestock and wildlife that drink the
contaminated surface water. TDS can impair the utility of water for both irrigation and livestock use.
The EPA did not quantify or monetize effects of quality changes in agricultural water sources arising
from the regulatory options due to data limitations and small estimated changes in water quality.
• Reservoir Capacity. Reservoirs serve many functions, including storage of drinking and irrigation
water supplies, flood control, hydropower supply, and recreation. Streams can carry sediment into
reservoirs, where it can settle and cause buildup of silt layers over time, at a recorded average rate of
1.2 billion kilograms per reservoir every year (USGS, 2009). Sedimentation reduces reservoir
capacity (Graf et al., 2010) and the useful life of reservoirs unless measures such as dredging are
taken to reclaim capacity (Clark et al., 1985). The EPA expects that by reducing TSS concentrations,
the regulatory options could provide cost savings by reducing dredging activity to reclaim capacity at
existing reservoirs. Conversely, an increase in TSS concentrations could lead to an increase in
dredging costs (see Chapter 10 for detail).
2.3.3 Sedimentation Changes in Navigational Waterways
Navigable waterways, including rivers, lakes, bays, shipping channels and harbors, are an integral part of the
United States' transportation network. Navigable channels are prone to reduced functionality due to sediment
build-up, which can reduce the navigable depth and width of the waterway (Clark et al., 1985). For many
navigable waters, periodic dredging is necessary to remove sediment and keep them passable. Dredging of
navigable waterways can be costly.
11 Regli et al (2015) estimated benefits of reducing bromide across various types of water treatment systems.
EPA-821-R-19-011
2-12
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
The EPA expects that the regulatory options would reduce sediment loadings to surface waters and reduce
dredging of navigational waterways under Option 4. The EPA quantified and monetized these benefits based
on the avoided cost for expected future dredging volumes. Small increases in sediment loads under Options 1,
2, and 3 would result in a modest increase in dredging costs in navigational waterways. Chapter 10 describes
this analysis.
2.3.4 Commercial Fisheries
Pollutants in steam electric power plant discharges can reduce fish populations by inhibiting reproduction and
survival of aquatic species. These changes may negatively affect commercial fishing industries as well as
consumers of fish, shellfish, and fish and seafood products. Estuaries are particularly important breeding and
nursery areas for commercial fish and shellfish species. In some cases, excessive pollutant loadings can lead
to the closures of shellfish beds, thereby reducing shellfish harvests. Improved water quality due to reduced
discharges of steam electric pollutants would enhance aquatic life habitat and, as a result, contribute to
reproduction and survival of commercially harvested species and larger fish and shellfish harvest, which in
turn could lead to an increase in producer and consumer surplus. Conversely, an increase in pollutant loadings
under some regulatory options could lead to negative impacts on fish and shellfish harvest.
The EPA did not quantify or monetize impacts to commercial fisheries from the regulatory options. The
EPA's Supplemental EA (U.S. EPA, 2019a) shows that eight steam electric power plants discharge bottom
ash transport water or FGD wastewater directly to the Great Lakes or to estuaries. Although estimated
increases or decreases in annual average pollutant loads under the regulatory options may have an impact on
local fish populations and commercial harvest, the overall effects to commercial fisheries arising from the
regulatory options are likely to be negligible. Most species of fish have numerous close substitutes. The
literature suggests that when there are plentiful substitute fish products, numerous fishers, and a strong ex-
vessel market, individual fishers are generally price takers. Therefore, the measure of consumers welfare
(consumer surplus) is unlikely to change as a result of small changes in fish landings, such as those the EPA
expects under the regulatory options.
2.3.5 Tourism
Discharges of pollutants may also affect the tourism industries (e.g., sales of fishing equipment) and, as a
result, local economies in the areas surrounding affected waters due to changes in recreational opportunities.
The effects of water quality on tourism are likely to be highly localized. Moreover, since substitute tourism
locations may be available, increased tourism in one location (e.g., the vicinity of steam electric power plants)
may lead to a reduction in tourism in other locations and vice versa. Due to the estimated small magnitude of
water quality changes expected from the regulatory options (see Section 3.4 for detail) and availability of
substitute sites the overall effects on tourism and, as a result, social welfare is likely to be trivial. Therefore,
the EPA did not quantify or monetize this benefit category.
2.3.6 Property Values
Discharges of pollutants may affect the aesthetic quality of land and water resources by changing pollutant
discharges and thus altering water clarity, odor, and color in the receiving and downstream reaches.
Technologies implemented by steam electric power plants to comply with the regulatory options remove
nutrients and sediments to varying degrees and these differences could have an effect on water eutrophication,
algae production, and water turbidity, among others. Several studies (Boyle et al., 1999; Poor et al., 2001;
Leggett and Bockstael, 2000; Gibbs et al., 2002; Bin and Czakowski, 2013; Walsh et al., 2011; Tuttle and
EPA-821-R-19-011
2-13
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
Heintzelman, 2014; Netusil et al., 2014; Liu et al., 2017; Klemick et al., 2018; Kung et al., 2018) suggest that
waterfront property is more desirable when located near unpolluted water. Therefore, the value of properties
located in proximity to waters contaminated with steam electric pollutants may increase or decrease due to
changes in the composition of bottom ash transport water or FGD wastewater discharges.
Due to data limitations, the EPA was not able to quantify or monetize the potential change in property values
associated with the regulatory options. The magnitude of the potential change depends on many factors,
including the number of housing units located in the vicinity of the affected waterbodies, community
characteristics (e.g., residential density) and housing stock (e.g., single family or multiple family) and the
effects of steam electric pollutants on aesthetic quality of surface water. Given the small changes in aesthetic
quality of surface waters that may result from the small changes in pollutant concentrations under the
regulatory options, the EPA expects impacts on property values to be small. In addition, there may be overlap
between shifts in property values and the estimated total WTP for surface water quality changes summarized
in Section 2.2.1.
2.4 Changes in Air Pollution
The regulatory options are expected to affect air pollution through three main mechanisms: 1) changes in
energy use by steam electric power plants to operate wastewater treatment, ash handling, and other systems
needed to comply with the regulatory options; 2) changes in transportation-related emissions due to changes
in trucking of CCR and other waste to on-site or off-site landfills; and 3) the change in the profile of
electricity generation due to relatively higher cost to generate electricity at plants incurring compliance costs
for the regulatory options (or conversely, lower generation costs for plants incurring cost savings under the
rule options). The different profile of generation can result in lower or higher air pollutant emissions due to
differences in emission factors for coal or natural gas combustion, or nuclear or hydroelectric power
generation.
Of the three mechanisms above, the change in the emissions profile of electricity generation at the market
level is the only one that increases under Option 2. As described in Chapter 5 of the RIA, the EPA used the
Integrated Planning Model (IPM®), a comprehensive electricity market optimization model that can evaluate
impacts of the proposed ELG options within the context of regional and national electricity markets. The EPA
analyzed proposed Option 2 and Option 4 using IPM.
Electricity market analyses using IPM indicate that in 2030 under Option 2, coal fired electric power
generation may increase by 0.6 percent and under Option 4 may increase by 0.2 percent, when compared to
the baseline without ACE (see RIA; U.S. EPA, 2019c). These small changes in generation result in air
emsission increases that are also relatively small. Changes in coal-based electricity generation as a result of
the regulatory options are compensated by changes in generation using other fuels or energy sources —
natural gas, nuclear power, solar, and wind power. The changes in air emissions reflect the differences in
emissions factors for these other fuels, as compared to coal-fueled generation. Overall for the three
mechanisms (auxiliary services, transportation, and market-level generation), the EPA estimates a net increase
in CO2 and SO2, and NOx emissions under all regulatory options as compared to the baseline.
Following the promulgation of the ACE rule finalized in June 19, 2019, the EPA also conducted a sensitivity
analysis of the impacts of proposed Option 2 relative to a baseline that includes the Affordable Clean Energy
(ACE) final rule (see U.S. EPA, 2019f). Appendix C in the RIA details this sensitivity analysis.
EPA-821-R-19-011
2-14
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
CO2 is the most prevalent of the greenhouse gases, which are air pollutants that the EPA has determined
endangers public health and welfare through their contribution to climate change. The EPA used estimates of
the domestic social cost of carbon (SC-CO2) to monetize the benefits of changes in CO2 emissions as a result
of this proposal The SC-C02 is a metric that estimates the monetary value of projected impacts associated
with marginal changes in C02 emissions in a given year. It includes a wide range of anticipated climate
impacts, such as net changes in agricultural productivity and human health, property damage from increased
flood risk, and changes in energy system costs, such as reduced costs for heating and increased costs for air
conditioning. Chapter 8 details this analysis.
The air-related benefit estimates presented in Chapter 8 rely on IPM projections from a sensitivity analysis of
proposed Option 2, which includes the effects of the ACE rule in the baseline, and from an analysis of
proposed Option 4 completed before the ACE rule was finalized (and which therefore does not include the
ACE rule).
NOx, and SO2 are known precursors to PM2 5, a criteria air pollutant that has been associated with a variety of
adverse health effects, including premature mortality and hospitalization for cardiovascular and respiratory
diseases (e.g., asthma, chronic obstructive pulmonary disease (COPD), and shortness of breath). The EPA
quantified changes in emissions of PM2 5 precursors NOx and SO2 but did not monetize the estimated changes
in secondary PM exposure that would result from changes in NOx, and SO2 emissions at this time. To map
those emission changes to air quality changes across the country, full scale air quality modeling is needed.
Prior to this proposal, the EPA's modeling capacity was fully allocated to supporting other regulatory and
policy efforts and as a result we did not do an air quality impact assessment and quantify the air disbenefits of
this proposal, were it to become a final regulation. For the final rule, the EPA intends to conducting full scale
air quality modeling to provide spatially explicit estimates of concentration changes, which is required for
characterizing uncertainty in mortality risk from changes in PM2 5 concentrations at different levels of
baseline PM25 exposure.
The Agency did not estimate the number or economic value of forgone benefits from increased exposure to
PM2 5 associated with increased SO2 and NOx emissions using a benefit per-ton (BPT) approach. Over the last
year and a half, the EPA systematically compared the changes in benefits, and concentrations where available,
from its BPT technique and other reduced-form techniques to the changes in benefits and concentrations
derived from full-form photochemical model representation of a few different specific emissions scenarios.
Reduced form tools are less complex than the full air quality modeling, requiring less agency resources and
time. That work, in which we also explore other reduced form models is referred to as the "Reduced Form
Tool Evaluation Project" (Project), began in 2017, and the initial results were available at the end of 2018.
The Agency's goal was to better understand the suitability of alternative reduced-form air quality modeling
techniques for estimating the health impacts of criteria pollutant emissions changes in EPA's benefit-cost
analysis. This research suggests that, for purposes of estimating the impacts of current emissions changes in
the EGU sector, the 2012 BPT approach (which was based off a 2005 inventory) may yield estimates of PM2 5
benefits that are as much as 25 percent greater than those estimated when using full air quality modeling. EPA
continues to work to develop refined reduced-form approaches for estimating PM2 5 benefits. The scenario-
EPA-821-R-19-011
2-15
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
specific emission inputs developed for this project are currently available online.12 The study design and
methodology will be thoroughly described in the final report summarizing the results of the project, which is
planned to be completed by the end of 2019. The agency intends to monetize the changes in PM2.5 and their
precursor emissions by conducting the full air quality modeling in the final rule.
2.5 Reduced Water Withdrawals
The regulatory options may change water withdrawals associated with wet bottom ash transport and wet FGD
scrubbers. In comparison to the baseline, these changes are estimated to be small. The regulatory options are
expected to increase water withdrawals from aquifers under Option 2 and from surface waterbodies under
Options 1, 2, and 3. The estimated increase in water withdrawal ranges from 0.22 billion gallons per year
(0.61 million gallons per day) under Option 3 to 7.7 billion gallons per year (21 million gallons per day)
under Option 2 (see Supplemental TDD for details). The EPA estimates that power plants would reduce water
withdrawals by 3.4 billion gallons per year (9.4 million gallons per day) under Option 4.
Increased water use from groundwater sources by steam electric power plants under the regulatory options
could reduce availability of groundwater supplies for alternative uses. One power plant affected by this
proposal relies on groundwater sources. The EPA's analysis of potential costs associated with an increase in
groundwater withdrawal are presented in Chapter 9.
A change in surface water intake would affect impingement and entrainment mortality. An increase in surface
water withdrawal under Options 1, 2, and 3 would increase impingement and entrainment mortality. Although
the overall increase in water withdrawal is modest, the significance of local ecological impacts is uncertain
and will depend on the overall health of the affected species population as well as species vulnerability to
impingement and entrainment (e.g., if water intakes affect a nursery habitat). A reduction in water withdrawal
under Option 4 may benefit fish species affected by impingement and entrainment mortality. Due to data
limitations and uncertainty, the EPA did not quantify and monetize these benefits as part of this analysis.
2.6 Summary of Benefits Categories
Table 2-3 summarizes the potential social welfare effects of the regulatory options and the level of analysis
applied to each category. As indicated in the table, only a subset of potential effects can be quantified and
monetized (in which case the table identifies the section of the report that discusses the analysis). The
monetized welfare effects include changes in some human health risks, use and non-use values from changes
in surface water quality, changes in costs for dredging navigational waterways, increased air pollution, and
changes in water withdrawals. Other welfare effect categories, including expected changes of pollutant
concentrations in excess of human health-based NRWQC limits, can be quantified but not monetized. Finally,
the EPA was not able to quantify or monetize other welfare effects, including impacts to commercial fisheries
or changes in the marketability of coal ash for beneficial use; the EPA evaluated these effects qualitatively as
discussed above in Sections 2.1 through 2.5.
12 The scenario-specific emission inputs developed for this project are currently available online at: https://github.com/epa-
kpc/RFMEVAL. Upon completion and publication of the final report, the final report and all associated documentation will be
online and available at this URL.
EPA-821-R-19-011
2-16
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
Table 2-3: Estimated Welfare Effects of Changes in Pollutant Discharges from Steam Electric Power
Plants
Category
Effect of Regulatory Options
Benefits Analysis
Quantified
Monetized
Methods (Report
Chapter or Section
where Analysis is
Detailed)
Human Health Benefits from Surface Water Quality Improvements
Incidence of bladder
Changes in exposure to TTHM in
V
V
VSL and COI (Section
cancer
drinking water
2.1.1)
IQ losses to children ages
Changes in childhood exposure to lead
V
V
IQ point valuation
Oto 7
from fish consumption
(.Section 5.3)
Need for specialized
Changes in childhood exposure to lead
V
V
Avoided cost (Section
education
from fish consumption
5.3)
Incidence of
Changes in exposure to lead from fish
Qualitative discussion
cardiovascular disease
consumption
IQ losses in infants
Changes in-utero mercury exposure
V
V
IQ point valuation
from maternal fish consumption
(.Section 5.4)
Incidence of cancer
Changes in exposure to arsenic from fish
V
V
COI (Section 5.5)
consumption
Other adverse health
Changes in exposure to other pollutants
Human health criteria
effects (cancer and non-
(arsenic, lead, etc.) via fish consumption
V
exceedances (Section
cancer)
or drinking water
5.7)
Reduced adverse health
Changes in exposure to pollutants from
Qualitative discussion
effects
recreational water uses
Ecological Conditions and Effects on Recreational Use from Surface Water Quality Changes
Aquatic and wildlife
habitat3
Changes in ambient water quality in
receiving reaches
V
V
Benefit transfer
(Chapter 6)
Water-based recreation3
Changes in swimming, fishing, boating,
and near-water activities from water
quality changes
Aesthetics3
Changes in aesthetics from shifts in
water clarity, color, odor, including
nearby site amenities (residing, working,
traveling)
Non-use values3
Changes in existence, option, and
bequest values from improved
ecosystem health
Aquatic and wildlife3
Changes in risks to aquatic life from
exposure to steam electric pollutants
Protection of T&E
species
Changes in T&E habitat and thus
potential effects on T&E population
V
Qualitative discussion
(Chapter 7)
Sediment contamination
Changes in deposition of toxic pollutants
to sediment
Qualitative discussion
Market and Productivity Benefits
Dredging costs
Changes in costs for maintaining
navigational waterways and reservoir
capacity
V
V
Cost of dredging
(Chapter 10)
Beneficial use of ash
Changes in disposal costs and avoided
lifecycle impacts from displaced virgin
material
Qualitative discussion
EPA-821-R-19-011
2-17
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
2: Benefits Overview
Table 2-3: Estimated Welfare Effects of Changes in Pollutant Discharges from Steam Electric Power
Plants
Category
Effect of Regulatory Options
Benefits Analysis
Quantified
Monetized
Methods (Report
Chapter or Section
where Analysis is
Detailed)
Water treatment costs
for drinking water and
irrigation water
Changes in quality of source water used
for drinking and irrigation
Qualitative discussion
Commercial fisheries
Changes in fisheries yield and harvest
quality due to aquatic habitat changes
Qualitative discussion
Tourism industries
Changes in participation in water-based
recreation
Qualitative discussion
Property values
Increased property values from water
quality improvements
Qualitative discussion
Air-Related Effects
Air emissions of NOx and
S02
Changes in mortality and morbidity from
exposure to particulate matter (PM2.5)
linked to changes in NOxand S02
emissions
V
Changes in tons of NOx
and S02 emitted
(Chapter 8)
Air emissions of C02
Climate change impacts
V
V
Domestic social cost of
carbon (SC-C02)
(Chapter 8)
Air emissions of other
pollutants
Changes in human health and other
effects from pollutants emissions
Qualitative discussion
Changes in Water Withdrawal
Groundwater
withdrawals
Decreased availability of groundwater
resources
V
V
Cost per gallon of
water withdrawn
(Chapter 9)
Surface water
withdrawals
Changes in vulnerability to drought and
impingement and entrainment mortality
Qualitative discussion
a. These values are implicit in the total WTP for water quality improvements.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
2-18
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
3: Environmental Effects
3 Water Quality Effects of Regulatory Options
Changes in the quality of surface waters, aquatic habitats and ecological functions due to the regulatory
options depend on a number of factors, including the operational characteristics of steam electric power
plants, treatment technologies implemented to control pollutant levels, the timing required for plants to
comply with the regulatory options, and the hydrography of reaches receiving steam electric pollutant
discharges, among others. This chapter describes the surface water quality changes projected under the
regulatory options. The EPA modeled water quality based on loadings estimated for the baseline and for
each of the four regulatory options (Options 1-4). The differences in predicted concentrations between the
baseline and option scenarios represent the changes attributable to the regulatory options. These changes
inform the analysis of several of the benefits described in Chapter 2.
The analyses use pollutant loading estimates detailed in in the Supplemental TDD (U.S. EPA, 2019b) and
expand upon the analysis of immediate receiving waters described in the Supplemental. EA document (U.S
EPA, 2019a) by estimating changes in both receiving and downstream reaches. The Supplemental EA
provides additional information on the effects of steam electric power plant discharges on surface waters
and how they may change under the regulatory options.
3.1 Waters Affected by Steam Electric Power Plant Discharges
The regulatory options affect pollutant discharges to receiving waters downstream of 116 steam electric
power plants. The EPA used the United States Geological Survey (USGS) medium-resolution National
Hydrography Dataset (NHD) (USGS, 2018) to represent and identify waters affected by steam electric
power plant discharges, and used additional attributes provided in version 2 of the NHDPlus dataset (U.S.
EPA, 2018c) to characterize these waters.
Of the 116 plants modeled, 112 had non-zero pollutant discharges under the baseline or the regulatory
options.13 In the aggregate, these plants discharge bottom ash transport water or FGD wastewater to 112
waterbodies (as categorized in NHDPlus), including lakes, rivers, and estuaries. NHDPlus also provides
the Strahler Stream Order14 for each reach, where the order increases as one moves from headwaters
(order 1) to downstream segments (orders 2-9). Table 3-1 summarizes Strahler Stream Order for the 112
receiving reaches. Stream order is one of the factors considered in evaluating potential uses of reaches
(e.g., whether the reach is likely to be fishable), when estimating benefits of water quality changes.
Table 3-1: Strahler Stream Order Designation for Reaches Receiving
Steam Electric Power Plant Discharges
Stream Order
Number of Reaches
1
15
2
9
3
6
4
9
13 Two plants have multiple receiving waters to which different waste streams are discharged — one receiving inputs from
FGD discharges and the other from BA discharges. There are also two reaches that receive discharges from two separate
plants.
14 Strahler Stream Order is a numerical measure of stream branching complexity. First order streams are the origin or
headwaters of a flowline. The confluence of two first order streams forms a second order stream, the confluence of two
second order streams forms a third order stream, and so on.
EPA-821-R-19-011
3-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
3: Environmental Effects
Table 3-1: Strahler Stream Order Designation for Reaches Receiving
Steam Electric Power Plant Discharges
Stream Order
Number of Reaches
5
9
6
18
7
17
8
20
9
3
Not classified
6
Receiving reaches that lack NHD classification for both waterbody area type and stream order generally
correspond to reaches that do not have valid flow paths15 for analysis of the fate and transport of steam
electric power plant discharges (Section 3.3). While eight steam electric power plants discharge bottom
ash transport water and/or FGD wastewater to tidal reaches or the Great Lakes,16 the EPA did not assess
pollutant loadings and water quality changes associated with these waterbodies because of the lack of a
defined flow path in NHDPlus, the complexity of flow patterns, and the relatively small changes in
concentrations expected.17 The EPA did not quantify the water quality changes and resulting benefits (or
forgone benefits) to these systems. Thus, the total number of plants for which the EPA estimated
downstream water quality changes is 104 (112 plants with nonzero pollutant discharges minus the eight
plants discharging to the Great Lakes or tidal waterbodies).
3.2 Changes in Pollutant Loadings
The EPA estimated post-compliance pollutant loadings for each plant under the baseline and the four
regulatory options. The TDD details the methodology (U.S. EPA, 2019b). The sections below discuss the
approach the EPA used to develop a profile of loading changes over time and summarize the results.
3.2.1 Timing of ELG Implementation
Benefits analyses account for the temporal profile of environmental changes as the public values changes
occurring in the future less than those that are more immediate (OMB, 2003). As described in the
proposal, the regulatory options incorporate varying compliance deadlines for meeting the revised limits
depending on the wastestream and technology basis, including providing more time to plants that
participate in the VIP to meet more stringent FGD wastewater effluent limits.
Table 3-2 summarizes the expected implementation schedules for the baseline and the four regulatory
options. This implementation schedule means that plants may be installing wastewater treatment
technologies in different years across the industry and potentially even within a given plant (e.g.,
complying with bottom ash transport water requirements in 2021 and FGD wastewater requirements in
2028). This in turn can translate into variations in pollutant loads to waters over time.
15 In NHDPlus, the flow path represents the distance traveled as one moves downstream from the reach to the terminus of the
stream network. An invalid flow path suggests that a reach is disconnected from the stream network.
10 Six reaches, one of which receives discharges from two steam electric power plants, are located in the Great Lakes (four
reaches along or near Lake Michigan, one reach along Lake Erie, and one reach on Saginaw Bay near Lake Huron). One
additional reach is located in Hillsborough Bay and is influenced by tidal processes.
17 The EPA looked at the changes in pollutant loadings and impacts to these systems in selected case studies as part of the
analysis of the 2015 rule (see EA document for details; U.S. EPA, 2015a).
EPA-821-R-19-011
3-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
3: Environmental Effects
To support estimating the benefits of the regulatory options, the EPA estimated the annual average
loadings discharged by each plant during the period of analysis (2021-2047), accounting for when each
plant would implement technologies to comply with the regulatory options. Using average annual values
instead of a year-by-year profile masks potential transitional effects of the regulatory options, including
temporary increases in loadings relative to the 2015 final rule baseline due to an extended status quo from
delayed implementation of new requirements. However, because the categories of benefits that the EPA is
analyzing generally result from changes in long-term processes (e.g., bladder cancer from chronic
exposure to trihalomethanes), annual average pollutant levels are likely an appropriate measure of
changes in environmental stressors under the regulatory options.
As discussed in the RIA (U.S. EPA, 2019c), there is uncertainty in the exact timing when individual steam
electric power plants would be implementing technologies to meet the ELGs. This benefits analysis uses
the same plant-specific technology installation years used in the cost and economic impact analysis. To
the extent that technologies are implemented earlier or later, the annualized loading values presented in
this section may under or overstate the annual loads during the analysis period.
EPA-821-R-19-011
3-3
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
3: Environmental Effects
Table 3-2: Implementation Schedule by Wastestream and Regulatory Option
Bottom Ash Transport Water
FGD Wastewater
Year(s)
Baseline
Option 1
Option 2
Option 3
Option 4
Baseline
Option 1
Option 2
Option 3
Option 4
2020
Current
Current
Current
Current
Current
Current
Current
Current
Current
Current
2021
Transition
Transition
Transition
Transition
Transition
Transition
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Current
2022
Transition
Transition
Transition
Transition
Transition
Transition
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Current
2023
Transition
Transition
Transition
Transition
Transition
Transition
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Current
2024
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Transition
2025
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Transition
(non-VIP
plants)
Transition
2026
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Interim Loads
Interim Loads
Interim Loads
Transition
2027
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Interim Loads
Interim Loads
Interim Loads
Transition
2028
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Transition
(VIP plants)
Transition
(VIP plants)
Transition
(VIP plants)
Transition
2029-
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
Revised ELG
2048
Current = Current loadings
Transition = Some plants meet the revised limits, based on permitting schedule (see Section 3.1.3 in the RIA (U.S. EPA, 2019c) for details on the modeled plant-specific compliance schedule).
Aggregate loadings are lower than under current conditions but greater than under the revised ELG.
Interim loads = Non-VIP plants have reached the steady-state post-compliance loadings, but loadings for VIP plants are still at the current level.
Revised ELG = All plants meet revised limits. Loadings are at their minimum steady-state post-compliance level.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
3-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
3.2.2 Results
Differences in the stringency of effluent limits and pretreatment standards and the timing of their applicability
to steam electric power plants (and the resulting treatment technology implementation) means that changes in
pollutant loads between the regulatory options and the baseline vary over the period of analysis. Table 3-3
summarizes the average annual changes in FGD wastewater, bottom ash transport water, and total loads for
selected pollutants that inform the EPA's analysis of the benefits discussed in Chapters 4 through 7. Negative
values in the table indicate reductions in pollutant loadings under an option as compared to the baseline. As
shown in the table, total aggregate annual average pollutant loads increase under Option 1 across all
pollutants. Options 2 and 3 show a decline in total bromide loads, with Option 3 also reducing total
phosphorus and thallium loads. Option 4 reduces total loadings of additional pollutants, but still shows
increases in total nitrogen, selenium, and zinc, among others. While this is not apparent from the total values,
the direction of the changes for a particular pollutant is not necessarily uniform across all plants under a given
option. For example, plants that participate in the VIP program under Options 2 and 3 may see reduced
pollutant loadings in their FGD wastewater when compared to the baseline, whereas pollutant loads may
increase for non-VIP plants implementing chemical precipitation with LRTR biological treatment control
technologies. Additionally, while Option 4 reduces total bromide loads, plants with bottom ash wastestreams
only may discharge greater quantities of bromide under Option 4 than under the baseline. These differences
are expected to have varying impacts on benefit estimates depending on the location of the plants and their
proximity to sensitive populations or environmental receptors.
EPA-821-R-19-011
3-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
4: Human Health Benefits via Drinking Water
Table 3-3: Annual Average Changes in Total Pollutant Loading in 2021-2047 for Selected Pollutants in Steam Electric Power Plant Discharges,
Relative to Baseline (lb/year)
Option la
Option 2a
Option 3a
Option 4a
Pollutant
FGD
Bottom
Ash
Total
FGD
Bottom
Ash
Total
FGD
Bottom
Ash
Total
FGD
Bottom
Ash
Total
Arsenic
0
86.9
86.9
-20.9
579
558
-31.3
86.9
55.6
-204
86.9
-117
Bromide
0
50,000
50,000
-10,200,000
338,000
-9,890,000
-11,100,000
50,000
-11,100,000
-23,900,000
50,000
-23,900,000
Cadmium
0
6.73
6.73
287
44.7
332
409
6.73
416
337
6.73
344
Chromium
0
47.4
47.4
3.46
315
319
3.33
47.4
50.8
-182
47.4
-135
Copper
0
36.8
36.8
33.9
245
279
47.5
36.8
84.4
-55.7
36.8
-18.8
Lead
0
97.2
97.2
-11.5
647
635
-17.2
97.2
80.0
-117
97.2
-20.3
Mercury
5.14
0.954
6.09
23.3
6.33
29.6
32.1
0.954
33.1
35.0
0.954
35.9
Nickel
164
163
327
2,490
1,090
3,570
3,510
163
3,670
3,770
163
3,930
Nitrogen,
Total
5,550,000
24,600
5,570,000
2,040,000
164,000
2,210,000
1,710,000
24,600
1,740,000
2,150,000
24,600
2,180,000
Phosphorus,
Total
0
2,070
2,070
-1,420
13,900
12,500
-2,100
2,070
-35.9
-11,700
2,070
-9,660
Selenium
54,000
114
54,100
21,100
766
21,900
18,000
114
18,100
24,400
114
24,500
Thallium
0
10.6
10.6
-32.4
70.7
38.2
-48.8
10.6
-38.2
-340
10.6
-330
TSS
0
125,000
125,000
16,400
840,000
856,000
21,200
125,000
146,000
-228,000
125,000
-103,000
Zinc
0
316
316
3,780
2,100
5,880
5,400
316
5,710
5,480
316
5,800
a. Negative values represent a reduction in pollutant loadings as compared to the baseline.
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
3-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
3.3 Water Quality Downstream from Steam Electric Power Plants
The EPA used the estimated annual average changes in total pollutant loadings to estimate concentrations
downstream from each plant. The methodology uses two main models to estimate downstream
concentrations:
• A dilution model to estimate pollutant concentrations downstream from the plants. The approach,
which for the purpose of this analysis is referred to as the D-FATE model (Downstream Fate And
Transport Equations), involves calculating concentrations in each downstream medium-resolution
NHD reach assuming conservation of mass and annual average Enhanced Runoff Method (EROM)
flows from NHDPlus v2. The calculations are similar to the methodology the EPA used in 2015 ELG
rule analysis (U.S. EPA, 2015a), but use updated data (e.g., flow). Appendix A summarizes
differences between the 2015 rule analysis and the present analysis.
• USGS's SPAtially Referenced Regressions On Watershed attributes (SPARROW) to estimate flow-
weighted nutrient and sediment concentrations. The SPARROW models provided baseline and post-
compliance concentrations of total nitrogen, total phosphorus, and total suspended solids. These
calibrated national models are the same models used by the EPA in the 2015 ELG rule analysis. Refer
to the BCA document for the 2015 rule for more details on this analysis (U.S. EPA, 2015a).
The models include only discharges to rivers and streams, which represent the vast majority of plants affected
by the regulatory options (104 plants out of 116 plants affected by the regulatory options). As discussed in
Section 3.1, the EPA omitted steam electric power plants that discharge to the Great Lakes or to estuaries
from this analysis.
In the D-FATE model, the EPA used stream routing and flow information from the medium-resolution
NHDPlus v2 to track masses of pollutants from steam electric power plant discharges and other pollutant
sources as they travel through the hydrographic network. For each point source discharger, the D-FATE
model estimates pollutant concentrations for the receiving reach and all downstream reaches based on NHD
mean annual flows. The model assumes that the discharges do not affect in-stream flows. The EPA notes that
steam electric power plant discharges frequently constitute a return of flow withdrawn for plant use from the
same surface water. In addition, FGD and BA wastewater discharges generally comprise a very small fraction
of annual mean flows in the NHDPlus v2 dataset.18
Following the approach used in the analysis of the 2015 rule (U.S. EPA, 2015a) to estimate pollutant
concentrations, the EPA included loadings from major dischargers (in addition to the steam electric power
plants) that reported to the 2016 Toxics Release Inventory (TRI). TRI data were available for a subset of
toxics: arsenic, barium, chromium, copper, lead, manganese, mercury, nickel, selenium, thallium, and zinc.
The EPA summed reach-specific background concentrations from TRI dischargers and concentration
estimates resulting from steam electric power plant loadings to represent water quality impacts from multiple
sources. The pollutant concentrations calculated in the D-FATE model are used to analyze nonmarket benefits
18 Steam electric power plant FGD discharge rates are typically about 1 million gallons per day (MGD), whereas the annual mean
stream flows in receiving waters average approximately 15,000 MGD.
EPA-821-R-19-011
3-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
of water quality improvements (see Chapter 6) and to derive fish tissue concentrations used to analyze human
health effects from consuming self-caught fish (See Chapter J).
3.4 Overall Water Quality Changes
Overall water quality changes modeled as a result of all evaluated options is relatively small compared to the
2015 rule analysis. Following the approach used in the analysis of the 2015 ELG (U.S. EPA, 2015a), the EPA
used a WQI to link water quality changes from reduced metal, nutrient and sediment discharges to effects on
human uses and support for aquatic and terrestrial species habitat. The WQI translates water quality
measurements, gathered for multiple parameters (e.g., dissolved oxygen (DO) concentrations) that are
indicative of various aspects of water quality, into a single numerical indicator. The WQI value, which is
measured on a scale from 0 to 100, reflects varying water quality, with 0 for poor quality and 100 for
excellent.
As detailed in U.S. EPA (2015a), the WQI includes seven parameters: DO, BOD, fecal coliform (FC), TN,
TP, TSS, and one aggregate subindex for toxics. The pollutants considered in the aggregate subindex for
toxics are those that are discharged by modeled steam electric power plants or 2016 TRI dischargers and that
have chronic aquatic life-based NRWQC. Pollutants that meet these qualifications include arsenic, chromium,
copper, lead, manganese, mercury, nickel, selenium, and zinc.19 The only update from the suite of pollutants
used for the 2015 rule analysis is the addition of copper exceedances in the toxics subindex, meaning the
subindex reflects nine toxics instead of eight. As a result, the subindex curve for toxics assigns the lowest
WQI value of 0 to waters where exceedances are observed for the nine toxics analyzed, and a maximum WQI
value of 100 to waters where there are no exceedances. Intermediate values are distributed between 100 and 0
in proportion to the number of exceedances.
3.4.1 WQI Data Sources
To calculate the WQI, the EPA used modeled NRWQC exceedances for toxics (using concentrations from D-
FATE) and modeled concentrations for TN, TP, and TSS (from the respective SPARROW models). The
USGS National Water Information System (NWIS) provided concentration data from 2007-2017 for three
parameters that are assumed to remain constant between the baseline and options: 1) fecal coliform, 2)
dissolved oxygen, and 3) biochemical oxygen demand (see Section 3.4.1.2).20
3.4.1.1 Exceedances of Water Quality Standards and Criteria
For each regulatory option, the EPA identified reaches that do not meet national recommended chronic water
quality criteria for aquatic life.21 There are 18 reaches with NRWQC exceedances in the baseline; five of
19 Barium and thallium are included in the 2016 TRI dataset but do not have chronic NRWQC in EPA's Aquatic Life Criteria Table
and thus are excluded from the aggregate toxics subindex.
20 USGS's NWIS dataset provides information on the occurrence, quantity, quality, distribution, and movement of surface and
underground waters based on data collected at approximately 1.5 million sites in all 50 States, the District of Columbia, and U.S.
territories. More information on NWIS can be found at http://waterdata.usgs.gov/nwis/
21 Aquatic life criteria are the highest concentration of pollutants in water that are not expected to pose a significant risk to the
majority of species in a given environment. For most pollutants, aquatic NRWQC are more stringent than human health NRWQC
and thus provide a more conservative estimate of potential water quality impairment. Chronic criteria are derived using longer
term (7-day to greater than 28-day) toxicity tests if available, or an acute-to-chronic ratio procedure where the acute criteria is
EPA-821-R-19-011
3-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
these reaches show improved water quality for at least one pollutant under the regulatory options. There are
22 reaches with no NRWQC exceedance in the baseline that have exceedances under one or more of the
regulatory options; 12 of these reaches have NRWQC exceedances for at least one pollutant under all four of
the regulatory options. Refer to the Supplemental EA for additional discussion of comparisons of receiving
and downstream water pollutant concentrations to acute and chronic aquatic NRWQCs (U.S. EPA, 2019a).
3.4.1.2 Sources for Ambient Water Quality
The EPA used average monitoring values for fecal coliform, dissolved oxygen, and biochemical oxygen
demand for 2007-2017 where available. Where more recent data were not available, the EPA used the same
averages as for the 2015 rule analysis. The EPA used a successive average approach to assign average values
for the three WQI parameters not explicitly modeled (i.e., DO, BOD, fecal coliform). The approach, which
adapts a common sequential averaging imputation technique, involves assigning the average of ambient
concentrations for a given parameter within a hydrologic unit to reaches within the same hydrologic unit with
missing data, and progressively expanding the geographical scope of the hydrologic unit (Hydrologic unit
code (HUC)8, HUC6, HUC4, and HUC2) to fill in all missing data.22 This approach assumes that reaches
located in the same watershed generally share similar characteristics. Using this estimation approach, the EPA
compiled ambient water quality data and/or estimates for all analyzed NHD reaches. As discussed below, the
values of the three WQI parameters not explicitly modeled are kept constant for the baseline and regulatory
policy scenarios. This approach has not been peer reviewed, but it has been used by EPA for several other
rules and previously subject to public comment.
3.4.1.3 Spatial Reference for Water Quality Index Calculations
The EPA used two different reach classification frameworks to assess in-stream water quality under the
baseline and each of the regulatory options: the medium-resolution NHD network23 and the USGS's
Enhanced River File 1 (E2RF1). Pollutant concentrations and exceedances were estimated for reaches
indexed to the NHD network, the SPARROW data are available for reaches indexed to the E2RF1 network,
and NWIS and STOrage and RETrieval Data Warehouse (STORET) data (U.S. EPA, 2008a) were averaged
to USGS's HUC watersheds. The WQI and benefits are ultimately calculated at the resolution of NHD
reaches, but with adjustments made to data available only at the E2RF1 level to reflect differences in spatial
scale. Thus, to reconcile the two levels of resolution, the EPA mapped all modeled reaches from the E2RF1 to
the NHD network using GIS.
derived using short term (48-hour to 96-hour) toxicity tests (U.S. EPA, 2017b). More information on aquatic NRWQC can be
found at https://www.epa.gov/wac/national-recommended-water-aualitv-criteria-aauatic-life-criteria-table.
22 Hydrologic Unit Codes (HUCs) are cataloguing numbers that uniquely identify hydrologic features such as surface drainage
basins. The HUCs consist of 8 to 14 digits, with each set of 2 digits giving more specific information about the hydrologic
feature. The first pair of values designate the region (of which there are 21), the next pair the subregion (total of 222), the third
pair the basin or cataloguing unit (total of 352), and the fourth pair the subbasin, or accounting unit (total of 2,262) (USGS,
2007). Digits after the first eight offer more detailed information, but are not always available for all waters. In this discussion, a
HUC level refers to a set of waters that have that number of HUC digits in common. For example, the HUC6 level includes all
reaches for which the first six digits of their HUC are the same.
23 The Watershed Boundary Dataset (WBD) is a companion dataset to the NHD and, therefore, was not considered a separate
hydrologic unit classification framework.
EPA-821-R-19-011
3-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
The water quality analysis included a total of 10,315 medium-resolution NHD reaches that are potentially
affected by steam electric power generating plants under the baseline. Of these 10,315 NHD reaches, the EPA
estimated concentrations for 10,284 reaches, selected based on their Strahler stream order and mean annual
flow rates. Table 3-4 summarizes the data sources used to estimate baseline and post-compliance values by
water quality parameters.
Table 3-4: Water Quality Data used in Calculating WQI for the Baseline and Regulatory
Options
Parameter
Baseline
Regulatory Option
TN
Concentrations calculated using
SPARROW (baseline run) at the E2RF1
level and indexed to NHD reaches
Concentrations calculated using
SPARROW (regulatory option run) at the
E2RF1 level and indexed to NHD reaches
TP
Concentrations calculated using
SPARROW (baseline run) at the E2RF1
level and indexed to NHD reaches
Concentrations calculated using
SPARROW (regulatory option run) at the
E2RF1 level and indexed to NHD reaches
TSS
Concentrations calculated using
SPARROW (baseline run) at the E2RF1
level and indexed to NHD reaches
Concentrations calculated using
SPARROW (regulatory option run) at the
E2RF1 level and indexed to NHD reaches
DO
Observed values averaged at the WBD
watershed level3
No change. Regulatory option value set
equal to baseline value
BOD
Observed values averaged at the WBD
watershed level3
No change. Regulatory option value set
equal to baseline value
Fecal Coliform
Observed values averaged at the WBD
watershed level3
No change. Regulatory option value set
equal to baseline value
Toxics
Baseline exceedances calculated using
D-FATE model at the NHD level
Regulatory option exceedances
calculated using D-FATE model at the
NHD level
WBD = Watershed Boundary Dataset
a. Values based on STORET and NWIS data, averaged for progressively larger geographical units (HUC8, HUC6,
HUC4, and HUC2), as needed to fill in all missing data.
Source: U.S. EPA Analysis, 2019.
3.4.2 WQI Calculation
The EPA used the approach described in the BCA document for the 2015 rule (U.S. EPA, 2015a) to estimate
WQI values for each reach under the baseline and each option. Implementing the WQI methodology involves
three key steps: 1) obtaining water quality levels for each of seven parameters included in the WQI; 2)
transforming parameter levels to subindex values expressed on a common scale; and 3) aggregating the
individual parameter subindices to obtain an overall WQI value that reflects waterbody conditions across the
seven parameters. These steps are repeated to calculate the WQI value for the baseline (i.e., the 2015 rule),
and for each analyzed regulatory option. See details of the calculations in Appendix B, including the subindex
curves used to transform levels of individual parameters.
3.4.3 Baseline WQI
Based on the estimated WQI value under the baseline scenario (WQI-BL), the EPA categorized each of these
10,284 NHD reaches using five WQI ranges (WQI < 25, 25
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
recreational use with the lowest required WQI), whereas WQI values greater than 70 indicate that waters are
swimmable (the recreational use with the highest required WQI).24
Table 3-5: Estimated Percentage of Potentially Affected Inland Reach Miles by WQI Classification:
Baseline Scenario
Water Quality
Classification
Baseline WQ
Number of
Reaches
Percent of
Affected
Reaches
Number of
Reach Miles
Percent of
Affected Reach
Miles
Unusable
WQK25
5
0.0%
2
0.0%
Suitable for
Boating
25
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 3-6: Ranges of Estimated Water Quality Changes for Regulatory Options
Options
Minimum AWQIa
Maximum AWQI
Median AWQI
AWQI Interquartile
Range
Option 1
-5.29
0.00
-1.02x10-3
0.01
Option 2
-2.95
1.30
-4.69x10-4
1.68x10-3
Option 3
-2.95
1.30
-2.29x10-4
7.79x10-4
Option 4
-2.62
1.31
-2.30x10-5
1.25x10-3
a. Negative changes in WQI values indicate degrading water quality.
Source: U.S. EPA Analysis, 2019
3.5 Limitations and Uncertainty
The methodologies and data used in the estimation of environmental effects of regulatory options involve
limitations and uncertainties. Table 3-7 summarizes the limitations and uncertainties and indicates the
direction of the potential bias. Uncertainties associated with some of the input data are covered in greater
detail in other documents. Regarding the uncertainties associated with use of the NHDPlus data, see U.S EPA
(2018c). Regarding the uncertainties associated with estimated loads, see the TDD (U.S. EPA, 2019b).
Table 3-7: Limitations and Uncertainties in Estimating Environmental Effects of Regulatory Options
Uncertainty/Assumption
Effect on
Environmental
Effects Estimation
Notes
Limited data are available to validate
water quality concentrations
estimated in D-FATE
Uncertain
The modeled concentrations reflect only a subset of
pollutant sources {e.g., steam electric power plant
discharges and TRI releases) whereas measured data
also reflect other sources such as bottom sediments,
air deposition, and other point and non-point sources
of pollution. EPA comparisons of D-FATE estimates to
monitoring data available for selected locations and
parameters {e.g., bromide concentrations downstream
of steam electric power plant discharges) confirmed
that D-FATE provides reasonable values. Also refer to
the EA for the 2015 rule for discussion of model
validation for selected case studies (U.S. EPA, 2015a)
In-stream concentrations assume
that stream flows are unaffected by
steam electric power plant
discharges
Overestimate
The degree of overestimation, if any, would be small
given that steam electric power plant discharge flows
tend to be very small as compared to stream flows in
modeled receiving and downstream reaches.
In-stream toxics concentrations are
based only on loadings from steam
electric power plants and other TRI
discharges.
Underestimate
Concentration estimates do not account for
background concentrations of these pollutants from
other sources, such as legacy pollution in sediments,
non-point sources, point sources that are not required
to report to TRI, air deposition, etc.
Annual loadings are estimated based
on estimated plant-specific
technology implementation years
Uncertain
To the extent that technologies are implemented
earlier or later, the annualized loading values
presented in this section may under or overstate the
annual loads during the analysis period. The effect of
this uncertainty is limited to the early years of the
analysis since loads reach a steady-state level by the
compliance deadlines applicable to the regulatory
options {e.g., by 2028)
EPA-821-R-19-011
3-12
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 3-7: Limitations and Uncertainties in Estimating Environmental Effects of Regulatory Options
Uncertainty/Assumption
Effect on
Environmental
Effects Estimation
Notes
The EPA used constant values for
fecal coliform, dissolved oxygen, and
biochemical oxygen demand.
Uncertain
The use of constant values for these parameters omits
the potential impacts of changes in stream electric
plant discharges under the regulatory options on these
water quality indicators, most notably dissolved
oxygen.
The EPA used regional averages of
monitoring data from 2007-2017 for
fecal coliform, dissolved oxygen, and
biochemical oxygen demand, when
location-specific data were not
available. In cases where more
recent data were not available, the
EPA used the same averages as used
in the 2015 rule analysis (U.S. EPA,
2015a).
Uncertain
The monitoring values were averaged over
progressively larger hydrologic units to fill in any
missing data. As a result, WQI values may not reflect
certain constituent fluctuations resulting from the
various regulatory options and/or may be limited in
their temporal and spatial relevance. Note that the
analysis keeps these parameters constant under both
the baseline and regulatory options. Modeled changes
due to the regulatory options are not affected by this
uncertainty.
Use of nonlinear subindex curves
Uncertain
The methodology used to translate in-stream sediment
and nutrient concentrations into subindex scores (see
Section 3.4.2 and Appendix B) employs nonlinear
transformation curves. Water quality changes that fall
outside of the sensitive part of the transformation
curve {i.e., above/below the upper/lower bounds,
respectively) yield no change in the analysis and no
benefit in the analysis described later in Chapter 6.
EPA-821-R-19-011
3-13
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
4 Human Health Benefits from Changes in Pollutant Exposure via
Drinking Water Pathways
The EPA expects that the small changes in pollutant loadings from the regulatory options relative to the 2015
analysis (U.S. EPA, 2015a) could affect several aspects of human health by changing bromide and other
pollutant discharges to surface waters and, as a result, pollutant concentrations in the reaches that serve as
sources of drinking water. The Supplemental EA (U.S. EPA, 2019a) provides details on the health effects of
steam electric pollutants.
As described in Section 2.1, human health benefits deriving from changes in pollutant loadings to receiving
waters include those associated with changes in exposure to pollutants via treated drinking water and fish
ingestion. This chapter addresses the first exposure pathway: drinking water. Chapter 5 addresses the fish
consumption pathway.
Section 4.1 presents background information regarding the potential impacts of bromide discharges on
drinking water quality and human health. Sections 4.2 through 4.4 present the EPA's analysis of human health
effects from changes in bromide discharges. Section 4.5 summarizes potential impacts on source waters from
changes in other pollutant discharges. Section 4.6 discusses uncertainty and limitations associated with the
analysis presented in this chapter.
In general, the estimated effects of the proposed regulatory option, Option 2, on pollutant exposure via
drinking water pathways are small compared to those estimated in 2015 (U.S. EPA, 2015a).
4.1 Background
Bottom ash transport water and FGD wastewater discharges contain variable quantities of bromide due to the
natural presence of bromide in coal feedstock and from additions of halogens, including bromide-containing
salts, and use of brominated activated carbon products to enhance air emissions control (Kolker et al, 2012).
Wastewater treatment technologies employed at steam electric power plants vary widely in their ability to
remove bromide. A number of studies have documented elevated bromide levels in surface water due to steam
electric power plant discharges (e.g., Cornwell et al., 2018; Good and VanBriesen, 2016, 2017; McTigue et
al., 2014; Ruhl et al., 2012; States et al., 2013; U.S. EPA, 2017a, 2019d) and have attributed measured
changes in bromide levels to the installation of wet FGD devices at an increasing number of steam electric
power plants. FGD wastewaters have been shown to contain relatively high levels of bromide relative to other
industrial wastewaters. Modeling studies have sought to quantify the potential for drinking water sources to
be affected by FGD wastewater discharges (Good and VanBriesen, 2019).
Bromide does not undergo significant physical (e.g., sorption, volatilization), chemical or biological
transformation in freshwater environments and is commonly used as a tracer in solute transport and mixing
field studies. Surface waters transport bromide discharges to downstream drinking water treatment facility
intakes where they are drawn into the treatment systems.
Although the bromide ion has a low degree of toxicity (WHO, 2009), it can contribute to the formation of
brominated DBPs during drinking water disinfection processes, including chlorination, chloramination, and
ozonation. Bromate, a regulated DBP under the Safe Drinking Water Act (SDWA), forms when bromine
reacts directly with ozone. Chlorine reacts with bromide to produce hypobromite (BrO), which reacts with
EPA-821-R-19-011
4-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
organic matter to form brominated and mixed chloro-bromo DBPs, including three of the four regulated
trihalomethanes25 (THM4, also referred to as total trihalomethanes (TTHM) in this discussion) and two of the
five regulated haloacetic acids26 (HAA5). Additional unregulated brominated DBPs have been cited as an
emerging class of water supply contaminants that can potentially pose health risks to humans (Richardson et
al., 2007; NTP, 2018; U.S. EPA, 2016).
There is a substantial body of literature on trihalomethane precursor occurrence, trihalomethane formation
mechanisms in drinking water treatment plants, and relationships between source water bromide levels and
TTHM levels in treated drinking water. The formation of TTHM in a particular drinking water treatment plant
is a function of several factors including chlorine, bromide, organic material, temperature, and pH levels as
well as system residence times. There is also substantial evidence linking TTHM exposure to bladder cancer
incidence (see U.S. EPA, 2016 for a review of recent studies). Bromodichloromethane and bromoform are
likely to be carcinogenic to humans by all exposure routes and there is evidence suggestive of
dibromochloromethane's carcinogenicity (National Toxicology Program, 2018; U.S. EPA, 2016). The
relationships between exposure to DBPs, specifically TTHMs and other halogenated compounds resulting
from water chlorination, and bladder cancer are further discussed in Section 4.3.3.2 and U.S. EPA (2019a).
4.2 Overview of the Analysis
Figure 4-1 illustrates the EPA's approach for quantifying and valuing the human health effects of altering
bromide discharges from steam electric power plants. The analysis entails estimating in-stream changes in
bromide levels between conditions under the baseline and each of the four regulatory options (Step 1);
estimating the change in source water bromide levels and corresponding changes in TTHM concentrations in
treated water supplies (Step 2); relating these changes to changes in the incidence of bladder cancers in the
exposed population (Step 3); and estimating the associated monetary value of benefits (Step 4).
The approach in Step 3 builds on the approach the Agency previously used to analyze the effects of the Stage
2 Disinfectants and Disinfection Byproduct Rule (DBPR) (U.S. EPA, 2005a) and incorporates studies, data,
and methodological advances that have become available following the promulgation of the DBPR.
Specifically, this analysis includes findings from a peer-reviewed paper by Regli et al (2015) that built on the
approach taken in the DBPR to derive a slope factor to relate changes in lifetime bladder cancer risk to
changes in TTHM exposure. The paper was published after promulgation of the DBPR and includes many of
the methodological components that supported the DBPR, such as the pooled analysis of Villanueva et al.
(2004). The approach used for this analysis also incorporates more recent National Cancer Institute's
Surveillance, Epidemiology, and End Results (SEER) program data to model incidence of bladder cancers by
age and sex, cancer stage, changes in lifetime cancer risk attributable to the proposed rule options, and
survival outcomes. The life table modeling approach used by the EPA to estimate changes in health outcomes
is a widely used method in public health, insurance, medical research, and other studies and was used by the
EPA in the analysis of lead-associated health effects in the 2015 Rule (U.S. EPA, 2015a) and of PM2 5-related
health effects in revisions to the National Ambient Air Quality Standards for ground-level ozone (U.S. EPA,
20087; 2008b). Other examples include the Occupational Safety and Health Administration (OSHA)'s use of
25 The four regulated trihalomethanes are bromodichloromethane, bromoform, chloroform, and dibromochloromethane.
26 The five regulated haloacetic acids are dibromoacetic acid, dichloroacetic acid, monobromoacetic acid, monochloroacetic acid,
and trichloroacetic acid.
EPA-821-R-19-011
4-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
a life table approach to estimate lifetime excess lung cancer, NMRD mortality, and silicosis risks from
exposure to respirable crystalline silica (81 FR 16285, March 25, 2016; OSHA, undated). The main advantage
of the life table approach is that it explicitly accounts for age and cancer stage-specific patterns in cancer
outcomes, as well as for other causes of mortality in the affected population.
The TTHM MCL is set higher than the health-based trihalomethane MCLGs in order to balance protection
from human health risks from DBP exposure with the need for adequate disinfection to control human health
risks from microbial pathogens. Actions that reduce TTHM levels below the MCL can therefore further
reduce human health risk. The EPA's analysis quantifies the human health effects associated with incremental
changes between the MCL and the MCLG. Recent TTHM compliance monitoring data indicate that the
drinking water treatment facilities contributing most significantly to total estimated benefits for the proposal
have TTHM levels below the MCL but in excess of the MCLGs for trihalomethanes.
This qualitative relationship between bladder cancer and bromide demonstrates the relative size of the benefit
to other benefits associated with this proposal. Should this analysis be used to justify an economically
significant rulemaking, EPA intends to peer review the analysis consistent with OMB's Information Quality
Bulletin for Peer Review. That review would include robust examination of the strengths and limitations of
the methods and an exploration of the sensitivity of the results to the assumptions made. If the analysis is
designated a highly influential scientific assessment (HISA), one way the EPA may seek such a review is via
the EPA's Science Advisory Board (SAB), which is particularly well suited to provide a peer review of
HISAs. The EPA's SAB is a statutorily established committee with a broad mandate to provide advice and
recommendations to the Agency on scientific and technical matters.
EPA-821-R-19-011
4-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Figure 4-1: Overview of Analysis of Human Health Benefits of Altering Bromide Discharges.
Transport and
ABr -> ATTHM
conversion factor
Dose-response
function
Center for
Disease Control
health statistics
Cost of illness and
value of statistical
... life
Public drinking
water treatment
system
Health benefits
Br loadings from
steam electric
plants
Population
exposure
Change in in-
stream Br
concentrations
Estimate for
baseline and
policy options
Change in
incidence of
bladder cancers
SDWIS surface
intake locations
SDWIS
population
served
Demographic
profile for
county
National Cancer
Institute SEER
program data
lost of i
value of s
life
Legend:
Analysis
component
Data/Inputs
Analysis step
Valuation
endpoint
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
4-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
4.3 Analysis Steps
4.3.1 Step 1: Modeling Bromide Concentrations in Surface Water
As described in the Supplemental TDD (U.S. EPA, 2019b), the EPA estimated steam electric power plant-
level bromide loadings associated with bottom ash transport water and FGD wastewater for the baseline and
four regulatory options. Total plant loadings are calculated as the sum of bottom ash transport water and FGD
wastewater loadings under each scenario. This chapter presents benefits estimated using the EPA's best
estimate of changes in bromide loadings under each of the four regulatory options. Appendix C includes
results of a sensitivity analysis using alternative loading estimates.
The EPA used the D-FATE model described in Section 3.3 to estimate in-stream bromide concentrations
downstream from 104 steam electric power plants with estimated non-zero bromide loads under the analyzed
scenarios. The EPA first estimated the annual average bromide load over the period of analysis. The EPA then
estimated concentrations in the receiving reach and each downstream reach, assuming conservation of mass,
until the load reaches the network terminus (e.g., Great Lake, estuary).27 The EPA summed individual
contributions from all plants to estimate total in-stream concentrations under the baseline and the four
regulatory options. Finally, the EPA estimated the change in bromide concentrations in each reach as the
difference between each regulatory option and the baseline. This change is not dependent on bromide
contributions from other sources (i.e.. receiving waterbody background levels).
4.3.2 Step 2: Modeling Changes in Trihalomethanes in Treated Water Supplies
4.3.2.1 Affected Public Water Systems
The population potentially exposed to trihalomethanes deriving from bromide discharges from steam electric
power plants includes individuals served by PWS whose source waters receive steam electric power plant
discharges.
The EPA's Safe Drinking Water Information System (SDWIS) database28 provides the latitude and longitude
of surface water facilities29, including source water intakes for public drinking water treatment systems. To
identify potentially affected PWS, the Agency georeferenced each permanent surface water facility associated
with non-transient community water systems to the NHD medium-resolution stream network used in D-
FATE.30 Appendix E describes the methodology the EPA used to determine the NHD water feature for each
facility. The SDWIS database also includes information on PWS primary sources (e.g., whether a PWS relies
primarily on groundwater or surface water for their source water), operational status, and population served,
27 As discussed in Section 3.1, the EPA did not estimate concentration changes in the Great Lakes or estuaries.
28 The EPA used intake locations as of January 2018 and PWS data as of June 2018, which reflects the second quarter report for
2018. Intake location data are protected from disclosure due to security concerns. SDWIS public data records are available from
the Federal Reporting Services system at https://ofmpub.epa.gov/apex/sfdw/.
29 Surface water facilities include any part of a public water system that aids in obtaining, treating, and distributing drinking water.
Facilities in the SDWIS database may include groundwater wells, consecutive connections between buyer and seller PWS, pump
stations, reservoirs, and intakes, among others.
30 This analysis does not include intakes that draw from the Great Lakes or other water bodies not analyzed in the D-FATE model.
EPA-821-R-19-011
4-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
among other attributes. For this analysis, the EPA used the subset of facilities that identify surface water as
their primary water source (specifically surface water intakes and reservoirs) and were categorized as "active"
and "permanent" in SDWIS. This subset of facilities corresponds to PWS that are more likely to be affected
by upstream bromide releases on an ongoing basis, as compared to other systems that may use surface water
sources only sporadically. This approach identifies populations most likely to experience changes in long-
term TTHM exposures and associated health effects due to the regulatory options.
PWS can be either directly or indirectly affected by steam electric power plant discharges. Directly affected
PWS are systems with surface water intakes drawing directly from reaches downstream from steam electric
power plants discharging bromide.31 Other PWS are indirectly affected because they purchase their source
water from another PWS via a "consecutive connection" instead of withdrawing directly from a surface water
or groundwater source. For these systems, SDWIS provides information on the PWS that supplies the
purchased water. The EPA used SDWIS data to identify PWS that may be indirectly affected by steam
electric power plant discharges because they purchase water from a directly affected PWS. The total
potentially exposed population consists of the people served by both directly and indirectly affected systems.
Table 4-1 summarizes the intakes, PWS, and populations potentially affected by steam electric power plant
discharges. Fourteen PWS are both directly and indirectly affected in that they both have intakes downstream
from steam electric power plants and purchase water from another directly affected PWS. In this analysis, the
average distance from the steam electric discharge point to the drinking water treatment plant intake is
approximately 40 miles.
Table 4-1: Estimated Reaches, Surface Water Intakes, Public Water Systems, and Populations
Potentially Affected by Bromide Discharges from Steam Electric Power Plants
Impact category
Number of reaches
with drinking water
intakes
Number of intakes
downstream of
steam electric power
plants
Number of PWS
Total population
served (million
people)
Direct3
278
348
294
20.2
Indirect
Not applicable
Not applicable
721
11.2
Total
278
348
1,015
31.4
a. Includes 14 systems that are both directly and indirectly affected by steam electric power plant discharges.
Source: U.S. EPA analysis, 2019
4.3.2.2 System-Level Changes in Bromide Concentrations in Source Water
The EPA estimated the changes in TTHM concentrations to which populations served by affected PWS are
exposed by first estimating the change in bromide concentrations in the source water for each public water
system that would result from the regulatory options, and then estimating the resulting changes in TTHM
concentration in the treated water. In this discussion, the term "system" refers to public water systems and
31 To identify potentially affected PWS, the EPA looked at all downstream reaches starting from the immediate reach receiving the
steam electric power plant discharge to the reach identified as the terminus of the stream network.
EPA-821-R-19-011
4-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
their associated drinking water treatment operations, whereas the term "facility" refers to the intake that is
drawing untreated water from a source reach for treatment at the PWS level.
To estimate changes in bromide concentrations at the PWS level, the EPA obtained the number of active
permanent surface water sources used by each PWS based on SDWIS data. SDWIS does not provide any
information on respective source flow contributions from surface water and groundwater facilities for a given
PWS. For drinking water treatment systems that have both surface water and groundwater facilities, the EPA
assessed changes from surface water sources only. This approach is reasonable given that the analysis is
limited to the PWS for which SDWIS identifies surface water as primary source.
For intakes located on reaches modeled in D-FATE, the EPA calculated the reach-level change in bromide
concentration as the difference between the regulatory option and the baseline conditions. Some PWS rely on
a single intake facility for their source water supply. If the source water reach associated with this single
intake is affected by steam electric power plant bromide discharges, the system-level changes in bromide
concentration at the PWS would equal the estimated change in bromide concentration of the source water
reach. Other PWS rely on multiple intake facilities that may be located along different source water reaches.
System-level changes in bromide concentrations at these PWS are an average of the estimated changes in
bromide concentrations associated with each source water reach. For any additional intakes not located on the
modeled reaches and for intakes relying on groundwater sources, the EPA assumed zero change in bromide
concentration. Because SDWIS does not provide information on source flows contributed by intake facilities
used by a given PWS, the EPA calculated the system-level change in bromide concentration assuming each
active permanent source facility contributes equally to the total volume of water treated by the PWS. For
example, the PWS-level change in bromide concentration for a PWS with three intakes, of which one intake is
directly affected by steam electric power plant discharges, is estimated as one third of the modeled reach
concentration change ([ABr + 0 + 0]/3).
The EPA addressed water purchases similarly, but with the change in bromide concentration associated with
the consecutive connection set equal to the PWS-level change estimated for the seller PWS instead of a reach-
level change. For facilities affected only indirectly by steam electric power plant discharges, the EPA
assumed zero change in bromide concentrations for any other unaffected source facility associated with the
buyer. The EPA also assumed that each permanent source facility contributes an equal share of the total
volume of water distributed by the buyer. For the 14 intakes classified as both directly and indirectly affected
by steam electric power plant bromide discharges, the EPA assessed the total change in bromide
concentration as a blended average of the change in concentration from both directly-drawn and purchased
water.
Table 4-2 summarizes the distribution of changes in bromide concentrations under the four regulatory
options. The direction of the changes depends on the option, source water reach, and PWS. Overall, Option 1
would result in an increase in bromide concentrations. Options 2, 3, and 4 would result in both increases and a
decreases in bromide concentrations. Option 4 has a higher frequency and magnitude of reduction in bromide
concentrations than the other regulatory options. All modeled changes to PWS bromide concentrations are
small. Refer to Table 3-3 for a summary of changes in bromide loadings associated with FGD and bottom ash
transport wastewaters under each regulatory option.
EPA-821-R-19-011
4-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 4-2: Estimated Distribution of Changes in Source Water and PWS-level Bromide
Concentrations by Regulatory Option
Number of source water reaches
Mumber of PWSa
ABr range (ng/L)
Positive13 ABr
Negative13
ABr
No ABr
(ABr = 0)
Positive13 ABr
Negative13
ABr
No ABr
(ABr = 0)
Option 1
Oto 10
212
0
66
699
0
316
10 to 30
0
0
0
0
0
0
30 to 50
0
0
0
0
0
0
50 to 75
0
0
0
0
0
0
>75
0
0
0
0
0
0
Option 2
Oto 10
154
33
38
502
168
125
10 to 30
0
47
0
0
193
0
30 to 50
0
3
0
0
8
0
50 to 75
0
0
0
0
0
0
>75
0
3
0
0
19
0
Option 3
Oto 10
110
72
39
374
274
143
10 to 30
0
50
0
0
196
0
30 to 50
0
3
0
0
8
0
50 to 75
0
1
0
0
1
0
>75
0
3
0
0
19
0
Option 4
Oto 10
66
94
9
243
383
46
10 to 30
0
89
0
0
280
0
30 to 50
0
10
0
0
24
0
50 to 75
0
2
0
0
10
0
>75
0
8
0
0
29
0
a- Includes systems potentially directly and/or indirectly affected by steam electric power plant discharges.
b. Positive values indicate higher estimated bromide concentrations under the regulatory option as compared to the baseline,
whereas negatives values indicate lower bromide concentrations under the regulatory option.
Source: U.S. EPA Analysis, 2019.
4.3.2.3 Changes in TTHM Concentration in Treated Water Supplies
The prior step provides the estimated PWS-level change in bromide concentration in the blend of source
waters used by a given system. The step described in this section provides the estimated PWS-level change in
TTHM concentration associated with this change in bromide concentration.
Reglietal. (2015) applied the Surface Water Analytical Tool (SWAT) version 1.1, which models TTHM
concentrations in drinking water treatment plants as a function of precursor levels, source water quality (e.g.,
bromide and organic material levels), water temperature, treatment processes (e.g., pH, residence time), and
disinfectant dose (e.g., chlorine levels) to predict the distribution of changes in TTHM concentrations in
finished water associated with defined increments of changes in bromide concentration in source waters. That
study estimated the distribution of increments of change in TTHM concentration for a subset of the
population of PWS characterized in the 1997-1998 Information Collection Rule (ICR) dataset. Table 4-3
summarizes the results from the Regli et al. (2015) analysis.
EPA-821-R-19-011
4-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 4-3: Estimated Increments of Change in TTHM Levels (ng/L) as a Function of Change in
Bromide Levels (ng/L)
Change in bromide
Change in TTHM concentration (iig/L)
concentration
Minimum
5th
Median
Mean
95th
Maximum
(Hg/L)
Percentile
Percentile
10
0.0
0.1
1.1
1.3
3.4
10.1
30
0.0
0.3
2.6
3.2
8.3
23.7
50
0.0
0.5
3.7
4.6
11.6
33.2
75
0.0
0.6
4.9
6.0
14.8
42.1
100
0.0
0.8
5.8
7.1
17.5
49.3
Source: Regli et al (2015), Table 2.
For this analysis, the EPA used the results from Regli et al. (2015) to predict TTHM concentration changes
for each water treatment plant with changes in bromide concentrations in their source water due to the
regulatory options. Figure 4-2 shows the relationship (dashed line) between the change in bromide
concentration and the change in TTHM concentration based on fitting a polynomial curve through the median
estimates from Table 4-3 (circular markers). The EPA used the equation of the best-fit curve32 to estimate
changes in TTHM concentration as a function of changes in bromide concentration within the bromide
concentration range presented in Regli, et al (2015) (0 to 100 (ig/L). For changes in bromide greater than
100 (ig/L. the EPA extrapolated values by continuing the slope of the best-fit curve for a 100 (ig/L change in
bromide concentration (equivalent to 0.022 (ig/L ATTHM per 1 (ig/L ABr). Estimates of TTHM concentration
changes presented in the remainder of this section reflect median changes from Regli et al. (2015).33 The EPA
developed similar relationships for the 5th and 95th percentile estimates in Table 4-3 to evaluate the sensitivity
of the benefits estimates to the relationship between changes in bromide and changes in TTHM. Appendix C
summarizes the results of this sensitivity analysis.
32 The polynomial curve fits observations in Table 4-3 with residuals of zero over the range of observations.
33 While Regli et al. (2015) show similar mean and median changes in TTHM concentrations across the range of changes in
bromide concentrations, the EPA used the median to minimize potential influence of outlier values or skew in the distribution.
EPA-821-R-19-011
4-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Figure 4-2: Modeled Relationship between Changes in Bromide Concentration and Changes in TTHM
Concentrations based on Median Values in Regli et al. (2015).
If ABr > 100 ^g/L;
ATTHM.m - 5,80 + 0,022 (ABr -100)
y
If ABr i 100 ng/L:
ATTHMm - -8 30x10" ABr * 196x10 * ABr' -1.81x10'ABR' + 1.26x10'ABr
*
0 20 40 60 80 100 120 140 160
ABr (ur/L)
Source: U.S. EPA Analysis, 2019, based on Regli et al. (2015).
Table 4-4 shows the distribution of modeled absolute changes in TTHM concentrations and the potentially
exposed populations under each of the regulatory options. As shown in the table, the magnitude of estimated
bromide concentration changes is generally less than 10 |ag/L, corresponding to estimated changes in TTHM
concentrations of less than 1.1 |ag/L. Compared to the baseline, Option 1 is estimated to increase TTHM
concentrations in treated water. Options 2, 3, and 4 are estimated to increase TTHM concentrations at some
PWS and decrease them at the majority of PWS.
Table 4-4: Distribution of Estimated Changes in TTHM Concentration by the Number of PWS
and Population Served
Absolute ABr range3
(Hg/L)
Absolute ATTHM
range3 (ng/L)
Number of PWSb
Total population served
(million people)c
Option 1
>0 to 10
0.000103 to 0.0844
699
25.27
10 to 30
-
-
-
30 to 50
-
-
-
50 to 75
--
--
--
>75
--
--
--
EPA-821-R-19-011
4-10
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 4-4: Distribution of Estimated Changes in TTHM Concentration by the Number of PWS
and Population Served
Absolute ABr range3
(Hg/L)
Absolute ATTHM
range3 (ng/L)
Number of PWSb
Total population served
(million people)c
Option 2
>0 to 10
0.000114 to 1.07
670
24.76
10 to 30
1.10 to 2.25
193
1.46
30 to 50
2.86 to 3.43
8
0.02
50 to 75
No data
No data
No data
>75
5.87 to 9.88
19
0.67
Option 3
>0 to 10
0.000114 to 1.07
648
24.62
10 to 30
1.11 to 2.46
196
1.54
30 to 50
2.87 to 3.43
8
0.02
50 to 75
3.96 to 3.96
1
0.01
>75
5.87 to 9.88
19
0.67
Option 4
>0 to 10
0.000114 to 1.09
626
24.53
10 to 30
1.11 to 2.60
280
5.08
30 to 50
2.61 to 3.61
24
0.15
50 to 75
3.78 to 4.55
10
0.60
>75
4.93 to 10.2
29
0.93
a. Shows only non-zero absolute changesAA. Modeled PWS-level changes under individual options may be zero, positive,
or negative.
b. Includes systems potentially directly and/or indirectly affected by steam electric power plant discharges.
c. Approximately 0.3 percent to 20 percent (depending on the regulatory option) of the total population served by PWS
potentially affected by bromide discharges from steam electric power plants are served by PWS with no change in source
water bromide concentrations.
Source: U.S. EPA analysis, 2019.
4.3.3 Step 3: Quantifying Population Exposure and Health Effects
The EPA used the following steps to quantify changes in human health resulting from changes in TTHM
levels in drinking water supplies:
• Characterize the exposed populations;
• Estimate changes in individual health risk; and
• Quantify the changes in adverse health outcomes.
4.3.3.1 Exposed Populations
SDWIS provides the total population served by each PWS and identifies the counties constituting the PWS
service area. For this analysis, the EPA assumed that all individuals served by a given PWS are exposed to the
same modeled changes in TTHM levels for the PWS, i.e., there are no differences in TTHM concentrations in
different parts of the water distribution system.
EPA-821-R-19-011
4-11
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
The EPA used county-level data from the 2017 American Community Survey (ACS, U.S. Census Bureau,
2018) to distribute the total exposed population for each PWS by age group to model health effects as
described in Section 4.3.3.3.34
4.3.3.2 Health Impact Function
The relationship between exposure to DBPs, specifically trihalomethanes and other halogenated compounds
resulting from water chlorination, and bladder cancer has been the subject of multiple epidemiological studies
(Cantor et al., 2010; U.S. EPA, 2016; NTP, 2018), a meta-analysis (Villanueva et al., 2003; Costet et al.,
2011), and pooled analysis (Villanueva et al., 2004). The relationship between trihalomethane levels and
bladder cancer in the Villanueva et al. (2004) study was used to support the benefits analysis for the EPA's
Stage 2 DBP Rule35 which specifically aimed to reduce the potential health risks from DBPs (U.S. EPA,
2005a).
Regli et al. (2015) conducted an analysis of potential bladder cancer risks associated with increased bromide
levels in surface source water. To estimate risks associated with modeled TTHM levels, they built on the
approach taken in EPA's Stage 2 DBP Rule, i.e., deriving a slope factor from the pooled analysis of
Villanueva et al. (2004). They showed that, while the original analysis deviated from linearity, particularly at
low doses, the overall pooled exposure-response relationship for TTHM could be well-approximated by a
linear slope factor that predicted an incremental lifetime cancer risk of 1 in ten thousand exposed individuals
(10 4) per 1 (ig/L increase in TTHM. The linear model proposed by Regli et al. (2015) provides a basis for
estimating the dose-response relationship associated with changes in TTHM levels estimated for the
regulatory options. The linear slope factor enables estimates of the total number of cancer cases associated
with lifetime exposures to different TTHM levels.
The EPA used the relationship estimated by Regli et al. (2015) to model the impact of changes in TTHM
concentration in treated water on the lifetime bladder cancer risk:
Equation 4-1. O(x) = 0(0) ¦ exp (0.00427 * x),
where 0 (x) are the odds of lifetime bladder cancer incidence for an individual exposed to a lifetime average
TTHM concentration in residential water supply of x (ig/L and 0 (0) are the odds of lifetime bladder cancer in
the absence of exposure to TTHM in residential water supply. The log-linear relationship (Equation 4-1) has
the advantage of being independent from the baseline TTHM exposure level, which is highly uncertain for
most affected individuals due to lack of historical data.
The EPA used 2012 to 2016 Census county-level data to distribute the exposed population by racial/ethnic group and poverty
status to support analysis of environmental justice (EJ) considerations in baseline exposure to pollutants in steam electric power
plant discharges and to evaluate how regulatory options may mitigate EJ concerns (see Chapter 14 for details).
See DBP Rule documentation at https://www.epa.gov/dwreginfo/stage-l-and-stage-2-disinfectants-and-disinfection-bvproducts-
rules
EPA-821-R-19-011
4-12
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
4.3.3.3 Health Risk Model and Data Sources
The EPA estimated changes in lifetime bladder cancer cases due to estimated changes in lifetime TTHM
exposure using a dynamic microsimulation model that estimates affected population life tables under different
exposure conditions. Life table approaches are standard among practitioners in demography and risk sciences
and provide a flexible method for estimating the probability of health impacts during a defined period (Miller
and Hurley, 2003; Rockett, 2010).36 In this application, the life table approach estimates age-specific changes
in bladder cancer probability and models subsequent bladder cancer mortality, which is highly dependent on
the age at the time of diagnosis. This age-specific cancer probability addresses variability in age-specific life
expectancy across the population alive at the time the change occurs. This model allows for quantification of
relatively complex policy scenarios, including those that involve variable contaminant level changes over
time.
For this analysis, the EPA assumed that the population affected by estimated changes in bromide discharges
from steam electric power plants is exposed to baseline TTHM levels prior to implementation of the
regulatory options - i.e., prior to 2021 - and to alternative TTHM levels from 2021 through 2047. As
described in Section 1.3.3, the period of analysis is based on the approximate life span of the longest-lived
compliance technology for any steam electric power plant (20 or more years) and the final year of
implementation (2028). The change in TTHM exposure affects the risk of developing bladder cancer beyond
this period, however, because the majority of cancer cases manifest during the latter half of the average
individual life span (Hrudey el al., 2015). To capture these effects while being consistent with the cost-benefit
framework of the regulatory options, the EPA modeled changes in health outcomes resulting from changes in
exposure in 2021-2047. To capture long term benefits of reduced exposure to TTHM from 2021 to 2047, EPA
modeled associated changes in cancer incidence through 2121.
Lifetime health risk model data sources, detailed in Table 4-5 (next page), include EPA SDWIS, ACS 2017
(U.S. Census Bureau, 2018), the Surveillance, Epidemiology, and End Results (SEER) program database
(National Cancer Institute), and the Center for Disease Control (CDC) National Center for Health Statistics.
36 The EPA has used life table approaches to estimate health risks associated with radon in homes, formaldehyde exposure, and
Superfund and RCRA site chemicals exposure, among others (Pawel and Puskin, 2004; Munns and Mitro, 2006; National
Research Council, 2011).
EPA-821-R-19-011
4-13
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
4: Human Health Benefits via Drinking Water
Table 4-5: Summary of Data Sources Used in Lifetime Health Risk Model
Data element
Modeled variability
Data source
Notes
Number of persons in the
affected population in 2021
Age: 1-year groups (ages 0 to
100)
Sex: males, females
Location: county for PWS service
area from SDWIS3
2017 American Community Survey
(ACS) (data on age- and sex-specific
county-level population [U.S. Census
Bureau, 2017b]).
Location-specific number of exposed
persons as described in Appendix C.
ACS data were in 5-year age groups. The EPA assumed
uniform distribution within each age interval to
represent data as 1-year age groups. The EPA then
computed relevant age- and sex- population shares
and used them to distribute location-specific affected
population within each county.
Bladder cancer incidence
rate (IR) per 100,000
persons
Age at diagnosis: 1-year groups
(ages 0 to 100)
Sex: males, females
Surveillance, Epidemiology, and End
Results (SEER)b 18 bladder cancer
incidence rates by age and sex at
diagnosis
Distinct SEER 18 IR data were available for ages 0,1-4,
5-9,10-14,15-19, 20-24, 25-29, 30-34, 35-39, 40-44,
45-49, 50-54, 55-59, 60-64, 65-69, 70-74, 75-79, 80-
84, 85+. The EPA assumed that the same IR applies to
all ages within each age group.
General population
mortality rate
Age: 1-year groups (ages 0 to
100)
Sex: males, females
Center for Disease Control
(CDC)/National Center for Health
Statistics (NCHS) United States Life
Tables, 2014
The EPA extracted age- and sex-specific probabilities
of dying within the integer age intervals.
Share of bladder cancer
incidence at specific cancer
stage
Age at diagnosis: 1-year groups
(ages 0 to 100)
Sex: males, females
Cancer stage: localized, regional,
distant, unstaged
SEER 18 distribution of bladder cancer
incidence over stages by age and sex at
diagnosis
Distinct SEER 18 data were available for ages 0-44, 45-
54, 55-64, 65-74, 75+. The EPA assumed that the
same cancer incidence shares by stage apply to all
ages within each age group.
Relative bladder cancer
survival by cancer stage
Age at diagnosis: 1-year groups
(ages 0 to 100)
Sex: males, females
Duration: 1-year groups
(durations 0 to 100 years)
Cancer stage: localized, regional,
distant, unstaged
SEER 18 relative bladder cancer
survival by age at diagnosis, sex, cancer
stage and duration with diagnosis
For males, distinct SEER 18 data were available for
ages at diagnosis 0-44, 45-54, 55-64, 65-74, 75+. For
females, data were available for ages at diagnosis 0-
49, 50-54, 55-64, 65-74, 75+. The EPA assumed that
the same cancer relative survival patterns apply to all
ages within each age group.
SEER 18 contained data on relative survival among
persons that had bladder cancer for 0,1,2,3,4,5 years.
For disease durations >5 years the EPA applied 5-year
relative survival rates.
a. EPA's Safe Drinking Water Information System SDWIS: https://www3.epa.gov/enviro/facts/sdwis/search.html
b. SEER program, National Cancer Institute, National Institute of Health
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
4-14
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 4-6 summarizes sex- and age group-specific general population mortality rates and bladder cancer
incidence rates used in the model simulations, as well as the sex-specific share of the affected population for
each age group. Appendix C summarize sex- and age group-specific distribution of bladder cancer cases over
four analyzed stages as well as the age of onset-specific relative survival probability for each stage.
Using available data on cancer incidence and mortality, the EPA then calculated changes in bladder cancer
cases resulting from the regulatory options using the relationship between the change in TTHM
concentrations and the change in lifetime bladder cancer risk estimated by Regli et al. (2015) (see Section
4.3.3.2). The analysis accounts for the gradual changes in lifetime exposures to TTHM following small
estimated changes in annual average bromide discharges and associated TTHM exposure under the regulatory
options compared to the baseline.
Table 4-6: Summary of Sex- and Age-specific Mortality and Bladder Cancer Incidence Rates
General population
General population
bladder cancer incidence
Age
Sex-specific share of the
mortality rate
rate
Sex
group
affected population3
(per 100,000)b
(per 100,000)bc
Male
Os
0.1296
80.5565
0.0225
Male
10s
0.1305
39.8300
0.0615
Male
20s
0.1425
131.3614
0.4104
Male
30s
0.1399
171.6144
1.5856
Male
40s
0.1266
314.4912
7.6596
Male
50s
0.1356
762.5491
31.6385
Male
60s
0.1095
1522.7546
96.5131
Male
70s
0.0577
3355.4487
215.9884
Male
80s
0.0188
8252.4234
333.1737
Male
90s
0.0094
31453.2483
366.5350
Female
0s
0.1177
66.5815
0.0000
Female
10s
0.1190
18.9204
0.0290
Female
20s
0.1350
51.6802
0.1986
Female
30s
0.1355
91.9849
0.6450
Female
40s
0.1248
202.5000
2.5795
Female
50s
0.1366
468.1672
9.2859
Female
60s
0.1170
951.2290
25.2908
Female
70s
0.0681
2357.4268
50.1267
Female
80s
0.0285
6420.9784
74.3698
Female
90s
0.0177
27743.8548
78.1720
a. Shares calculated for the total population served by potentially affected PWS, based on county-level data.
b. Based on the general population of the United States.
c. Single age-specific rates were aggregated up to the age groups reported in the table using the individual age-specific
number of affected persons as weights.
Source: U.S. EPA analysis (2019) of 2017 ACS Data.
4.3.3.4 Model Implementation
The EPA analyzed effects of the regulatory options using the dynamic microsimulation model and data
sources described in Section 4.3.3.3. As described above, the EPA models TTHM changes (ATTHM) due to
the regulatory options as being in effect for the years 2021 through 2047. After 2047, the EPA does not
EPA-821-R-19-011
4-15
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
attribute costs or changes in bromide loadings to the rule, and therefore does not model incremental changes
in exposures to TTHM.37
To estimate changes in bladder cancer incidence, the EPA defined and quantified a set of 102,414 unique
combinations38 of the following parameters:
• Location and TTHM changes: 507 PWS groups;39
• Age : age of the population at the start of the evaluation period (2021), ranging from 0 to 100;
• Sex: population sex (male or female).
4.3.4 Step 4: Quantifying the Monetary Value of Benefits
The EPA estimated total monetized benefits from avoided morbidity and mortality (also referred to as avoided
cancer cases and avoided cancer deaths, respectively, in this discussion) from estimated changes in bromide
discharges, and estimated changes in TTHM exposure and the resulting estimated bladder cancer incidence
rate using 3 percent and 7 percent discount rates for each of the four regulatory options.40
• Morbidity: To value changes in the economic burden associated with cancer morbidity the EPA used
estimates of annual medical expenses for bladder cancer treatment from Greco et al. (2018) and the
estimated life years with cancer morbidity (differentiating between first and subsequent years after
cancer diagnosis). For invasive cancer, the medical treatment costs are $42,750 and $2,850 per case
for the first and subsequent years respectively. For non-invasive cancer, medical treatment costs are
$15,618 and $1,026 per case for the first and subsequent years, respectively.
• Mortality: To value changes in excess mortality from bladder cancer the EPA used a default central
tendency VSL estimate of $11,021 million per death (U.S. EPA 2010a). The product ofVSL and the
estimated aggregate reduction in risk of death in a given year represents the affected population's
aggregate WTP to reduce its probability of death in one year.
4.4 Results of Analysis of Human Health Benefits from Estimated Changes in Bromide
Discharges Analysis
Using the data the EPA assembled on cancer incidence and mortality, the Agency estimated changes in
bladder cancer cases for the regulatory options using the relationship between TTHM concentrations and the
lifetime bladder cancer risk estimated by Regli et al. (2015). Figure 4-3 and Figure 4-4 show the estimated
37 In other words, costs after 2047 = $0 and Abromide after 2047 is zero (hence ATTHM after 2047 is zero).
38 The set of 102,414 combinations was determined by multiplying the number of PWS groups by the number of ages and sexes
considered (507 x 101 x 2).
39 The PWS groups represent unique combinations of location (county) and ATTHM values and typically consist of a directly
affected PWS and other PWSs serving populations located in the same county and purchasing water from the directly affected
PWS. The number of PWS in each PWS group ranges from 1 to 41.
40 In some cases, benefits are derived from a delay in cancer morbidity and mortality.
EPA-821-R-19-011 4-16
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
number of bladder cancer cases and premature deaths avoided, respectively, under the four regulatory options
by decade.
Consistent with the small increase in bromide loadings for Option 1 in Table 3-3, this option would result in a
small increase in cancer incidence as compared to the baseline. Options 2, 3, and 4 generally show decreases
in cancer incidence over the period of analysis. More than 50 percent of the modeled avoided bladder cancer
incidence associated with Options 2, 3, and 4 occurs between 2021 and 2050. This pattern is consistent with
existing cancer cessation lag models (e.g., Hrubec and McLaughlin 1997, Hartge et al. 1987, and Chen and
Gibbs 2003) that show between 61 and 94 percent reduction in cancer risk in the first 25 years after exposure
cessation (see Appendix C for detail). After 2050, the benefits attributable to exposures incurred under the
regulatory options in 2021-2047 decline due to comparably fewer people surviving to mature ages.41 In the
years after 2080, the avoided cases decline considerably and in the last decade considered in the analysis, the
cancer incidences increase relative to baseline incidences.42
Figure 4-3: Estimated Number of Bladder Cancer Cases Avoided under the Regulatory Options.
170
50
30
-10
2021-2030 2031-2040 2041-20S0 2051-2060 2061-2070 2071-2080 2081-2090 2091-2100 2101-2110 2111-2120
¦ Option 1 ¦ Option 2 ¦ Option 3 ¦ Option 4
Source: U.S. EPA Analysis, 2019.
41 In the period between 2051 and 2080, the estimated avoided cases decline slowly as the living people exposed to the estimated
changes in TTHM levels reach 70 years (the age at which the highest annual incidence of bladder cancer is observed). According
to American Cancer Society, about 9 out of 10 people diagnosed with bladder cancer are over the age of 55. The average age at
the time of diagnosis is 73 (ACS, 2019).
4- The increase in cancer cases in the last decade is due to the connection between survival and cancer incidence. Lower estimated
TTHM exposure due to reductions in bromide loadings under certain regulatory options reduces the estimated number of people
developing bladder cancer during the earlier years of the analysis and increases overall survival rates. Higher estimated rates of
survival lead to longer life spans and more people developing cancer later in life. This effect becomes more apparent closer to the
end of the evaluation period, at which point there are fewer people estimated to be alive in the baseline population compared to
the estimated number of people alive under certain regulatory option scenarios.
EPA-821-R-19-011
4-17
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Figure 4-4: Estimated Number of Cancer Deaths Avoided under the Regulatory Options.
70
60
50
40
2021-2030 2031-2040 2041-2050 2051-2060 2061-2070 2071-2080 2081-2090 2091-2100 2101-2110 2111-2120
I Option 1 ¦ Option 2 ¦ Option 3 ¦ Option 4
Source: U.S. EPA Analysis, 2019.
Table 4-7 summarizes the estimated changes in the incidence of bladder cancer from exposure to TTHM due
to the regulatory options and the value of benefits from avoided cancer cases, including avoided mortality and
morbidity.
Table 4-7: Estimated Bromide-related Bladder Cancer Mortality and Morbidity Monetized Benefits
Regulatory
Changes in cancer cases from
changes in TTHM exposure
2021-20473
Benefits (million 2018$, discounted to 2020)
Option
Total bladder
cancer cases
avoided
Total cancer
deaths
avoided
Annualizedb
benefits from
avoided mortality
Annualizedb
benefits from
morbidity avoided
Total annualized11
benefits
3%
7%
3%
7%
3%
7%
1
-3
-1
-$0.36
-$0.23
$0.00
$0.00
-$0.36
-$0.23
2
343
139
$37,42
$24.08
$0.19
$0.12
$37.61
$24.21
3
387
157
$42.36
$27.34
$0.21
$0.14
$42.57
$27.48
4
769
311
$83.90
$54.03
$0.42
$0.28
$84.32
$54.30
a. The analysis accounts for the persisting health effects (up until 2121) from changes in TTHM exposure during the period of
analysis (2021-2047).
b. Benefits are annualized over 27 years.
Source: U.S. EPA Analysis, 2019
These estimated total benefits are not uniformly distributed across plants that discharge bromide. For
example, out of the 104 steam electric power plants included in this analysis, under Option 2 more than
EPA-821-R-19-011
4-18
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
85 percent of total benefits are attributable to discharge changes at only five steam electric power plants.
Similarly, approximately 78 percent of the benefits of Option 4 come from changes at ten steam electric
power plants. Figure 4-5 illustrates the plant-level contributions to total annualized benefits for each of the
four regulatory options. Orange and blue bars show negative and positive benefits, respectively.
EPA-821-R-19-011
4-19
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Figure 4-5: Contributions of Individual Steam Electric Power Plants to Total Annualized Benefits of
Changes in Bromide Discharges under the Regulatory Options (3 Percent Discount Rate)
Option 1
$0.0 |
$0.2
-$0.4
C.
O
-$0.8
| &
CO (J
±± oo
% Q
is si°
a S 512
~ 00
ro t—l
13 O
c r\i
c
<
-$1.4
-$1.6
-$1.8
-$2.0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
Cumulative Number of Steam Electric Plants
Option 2
$90
I „ $8°
= £ $70
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Figure 4-5: Contributions of Individual Steam Electric Power Plants to Total Annualized Benefits of
Changes in Bromide Discharges under the Regulatory Options (3 Percent Discount Rate)
Option 3
$90
c $80
O
:> i
— O
CO (J
00
$70
$60
:> §
— O
CO U
—
^ ^ $50
(L)
co j4o
"O 03
jj $30
03
I s $2°
< $10
$0
I
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49
Cumulative Number of Steam Electric Plants
Option 4
$90
£= $80
0
$70
$60
"5 O $50
a)
co $40
~D 03
| # $30 |-
OJ
1 S 520
< $10
I
¦
I
$0
I
I
3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51
Cumulative Number of Steam Electric Plants
4.5 Additional Measures of Human Health Effects from Exposure to Steam Electric Pollutants
via Drinking Water Pathway
The regulatory options may result in small changes to source water quality for additional parameters that can
adversely affect human health (see Section 2.1.1). Many pollutants in steam electric power plant discharges
have MCLs that set allowable levels in treated water. For some pollutants that have an MCL above the
MCLG, there may be incremental benefits from reducing concentrations below the MCL. In addition to
EPA-821-R-19-011
4-21
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
certain brominated DBPs discussed in the previous sections, there are no "safe levels" for lead and arsenic
and therefore any reduction in exposure to these pollutants is expected to yield benefits.43
To assess potential additional drinking water-related health benefits of the regulatory options for pollutants
found in steam electric power plant discharges, the EPA estimated the expected changes in the number of
receiving reaches with drinking water intakes that have modeled pollutant concentrations in excess of MCLs.
The EPA did this analysis for all of the pollutants listed in Table 2-2, except bromate and TTHM.44 This
analysis showed no changes in the number of MCL exceedances under the regulatory options, when
compared to the baseline. Furthermore, the EPA found no reaches with drinking water intakes that had
modeled lead or arsenic concentrations in excess of MCLs under either the baseline or the regulatory
options.45 The Agency concluded, based on these screening analyses, that any additional benefits from
changes in exposure to other pollutants via the drinking water pathway would be minimal.
4.6 Limitations and Uncertainties
Table 4-8 summarizes principal limitations and sources of uncertainties associated with the estimated changes
in incidences of bladder cancer cases from exposure to TTHM in drinking water affected by steam electric
power plant discharges. Additional limitations and uncertainties are associated with the estimation of bromide
discharges (see U.S. EPA, 2019a) and derivation of other analysis inputs such as cancer incidence and
mortality rates. Note that the effect on benefits estimates indicated in the second column of the table refers to
the magnitude of the benefits rather than the direction (/'. e., a source of uncertainty that tends to underestimate
benefits indicates expectation for larger forgone benefits).
Table 4-8: Limitations and Uncertainties in the Analysis of Human Health Benefits from Changes in
Bromide Discharges
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
Characterizing the exposed population
Analysis does not account for
births and migration within
the exposed population.
Underestimate
The analysis does not account for people born after 2021, nor
does it account for people leaving or moving into the service
area. The analysis does account for mortality. To the extent
that population growth exceeds migration out of the area,
omitting those additional individuals understates the affected
population and benefits.
43 Even in cases where the MCLG is equal to the MCL, there may be incremental health-related benefits associated with changes in
concentrations arising from the regulatory options since detection of the pollutants is subject to imperfect monitoring and
treatment may not remove all contaminants from the drinking water supplies, as evidenced by reported MCL violations for
inorganic and other contaminants at community water systems (U.S. EPA, 2013c).
44 Only reaches designated as fishable (i.e., Straliler Stream Order larger than 1) were included in the human health ambient water
quality criteria exceedances analysis.
45 The EPA also found that there are no reaches with drinking water intakes that have pollutant concentrations in excess of human
health ambient water quality criteria for either the consumption of water and organism or the consumption of organism only.
EPA-821-R-19-011
4-22
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 4-8: Limitations and Uncertainties in the Analysis of Human Health Benefits from Changes in
Bromide Discharges
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
Bladder cancer risks are
estimated for populations for
which changes in TTHM
exposures relative to baseline
exposures start at different
ages, including children.
Uncertain
The relative cancer potency of TTHM in children is unknown,
which may bias benefits estimates either upward or
downward. Past reviews found no clear evidence that children
are at greater risk of adverse effects from bromoform or
dibromochloromethane exposure (U.S. EPA, 2005c) although
certain modes of action and health effects may be associated
with exposure to TTHM during childhood (U.S. EPA, 2016).
Because bladder cancer incidence in children is very small, the
EPA assesses any bias to be negligible.
Modeling changes in TTHM in publicly supplied water
The analysis does not
consider bromide sources
beyond those associated with
steam electric power plants.
Uncertain
The approach to modeling bromide concentrations in source
water excludes other bromide sources such as oil and gas
production, active and abandoned coal mines, and certain
types of chemical manufacturing. To the degree that the
relationship between changes in bromide levels and changes
in TTHM formation is non-linear and depends on absolute
bromide concentrations in source waters, this analysis uses a
linear model and therefore may overstate or understate the
impacts of changes in bromide levels.
For PWS with multiple
sources of water, the analysis
assumes equal contributions
from each source.
Uncertain
Data on the flow rates of individual source facilities are not
available and the EPA therefore assumed that all permanent
active sources contribute equally to a PWS's total supply.
Effects of the regulatory option may be greater or smaller
than estimated, depending on actual supply shares.
Changes in bromide
concentrations are analyzed
for active permanent surface
water intakes and reservoirs
only.
Underestimate
The analysis includes only permanent active surface water
facilities associated with non-transient PWS classified as
"community water systems" that use surface water as primary
source. To the extent that PWS using surface waters as
secondary source or other non-permanent surface water
facilities are affected, this assumption understates the effects
of the regulatory options.
Changes in TTHM formation
depends only on changes in
bromide levels.
Uncertain
The regulatory options are expected to affect bromide levels
in source water. Other factors such as disinfection method,
pH, temperature, and organic content affect TTHM formation.
The EPA assumes that PWS and source waters affected by
steam electric power plant discharges have similar
characteristics as those modeled in Regli et al (2015).
Use of a national relationship
from Regli et al (2015) to
relate changes in bromide
concentration to changes in
TTHM concentration.
Uncertain
The EPA did not collect site-specific information on factors
affecting TTHM formation at each potentially affected drinking
water treatment plant, but instead used the median from a
sample population of approximately 200 drinking water
treatment systems. Actual changes in TTHM concentrations
for a given change in bromide concentrations at any specific
drinking water treatment system could be higher or lower
than that estimated using the national relationship.
EPA-821-R-19-011
4-23
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 4-8: Limitations and Uncertainties in the Analysis of Human Health Benefits from Changes in
Bromide Discharges
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
Modeling changes in health risks
Change in risk is based on
changes in exposure to
TTHMs rather than to
brominated trihalomethanes
specifically.
Underestimate
As noted in Section 4.3.3.2, brominated species play a
prominent role in the overall toxicity of DBP exposure. Given
that the regulatory options predominantly affect the
formation of brominated DBPs, the changes in risk could be
greater than that which the EPA estimated in this analysis. See
U.S. EPA (2016) for additional information about health effects
of DBPs.
The analysis relies on public-
access SEER 18 5-year relative
bladder cancer survival data
to model mortality patterns
in the bladder cancer
population.
Uncertain
Reliance on these data generates both a downward and an
upward bias. The downward bias is due to the short, 5-year
excess mortality follow-up window. Survival rates beyond 5
years following the initial diagnosis are likely to be lower. The
upward bias comes from the inability to determine how many
of the excess deaths were deaths from bladder cancer.
The dose-response function
used to estimate risk assumes
causality of bladder cancer
from exposure to disinfected
drinking water
Overestimate
While the evidence supporting causality has increased since
EPA's Stage 2 DBP Rule, the weight of evidence is still not
definitive (see Regli et al., 2015).
The relationship from Regli et
al. (2015) is a linear
approximation of the odds
ratios reported in Villanueva
et al. (2004).
Uncertain
Given the uncertainty about the historical, location-specific
TTHM baselines, Regli et al. (2015) provides a reasonable
approximation of the risk. However, depending on the
baseline TTHM exposure level, the impact computed based on
Regli et al. (2015) may be larger or smaller than the impact
computed using the Villanueva et al. (2004)-reported odds
ratios directly.
The analysis does not account
for the relationship between
TTHM exposure and bladder
cancer within certain
subpopulations.
Overestimate
There is literature suggesting that TTHM effects could be
possibly greatest for the smoker population, whose members
are already at higher risk for bladder cancer. Smoking
prevalence has declined in the United States and relationships
estimated with data from the 1980s and 1990s may
overestimate future bladder cancer impact. Robust synthesis
estimates of the relationship between TTHM and bladder
cancer in the smoker population are lacking, limiting the EPA's
ability to account for smoking when modeling health effects.
The change in risk for a given
change in TTHM is uncertain
for changes in TTHM
concentrations that are less
than 1 ng/L.
Uncertain
While there is greater uncertainty for smaller changes in
TTHM concentrations, the EPA assesses that it is appropriate
to include these predicted changes when estimating benefits
rather than assuming zero benefits by omitting the results.
The EPA notes that the majority of the regulatory options
benefits are associated with PWS for which predicted changes
in TTHM concentration are greater than 1 ng/L.
EPA-821-R-19-011
4-24
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 4: Human Health Benefits via Drinking Water
Table 4-8: Limitations and Uncertainties in the Analysis of Human Health Benefits from Changes in
Bromide Discharges
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
Potential health effects other
than bladder cancer are not
quantified in this analysis.
Uncertain
U.S. EPA (2016) discusses potential linkages between DBP
exposures and other health endpoints, e.g., developmental
effects (with a short-term exposure) and cancers other than
bladder cancers (with a long-term exposure), but there is
insufficient to fully evaluate these endpoints.
EPA-821-R-19-011
4-25
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
5 Human Health Effects from Changes in Pollutant Exposure via Fish
Ingestion Pathway
The EPA expects the regulatory options to affect human health risk by changing effluent discharges to surface
waters and, as a result, ambient pollutant concentrations in the receiving reaches. The EPA's Supplemental
EA (U.S. EPA, 2019a) provides details on the health effects of steam electric pollutants. Recreational anglers
and subsistence fishers (and their household members) who consume fish caught in the reaches receiving
steam electric power plant discharges could benefit from reduced pollutant concentrations in fish tissue. This
chapter presents the EPA's analysis of human health effects resulting from changes in exposure to pollutants
in bottom ash transport water and FGD wastewater via the fish consumption pathway. The analyzed health
effects include:
• Changes in exposure to lead: This includes changes in neurological and cognitive damages in children
(ages 0-7) based on the impact of an additional IQ point on an individual's future earnings and the
cost of compensatory education for children with learning delays.
• Changes in exposure to mercury: Changes in neurological and cognitive damages in infants from
exposure to mercury in-utero.
• Changes in exposure to arsenic: Changes in incidence of cancer cases.
The total quantified human health effects included in this analysis represent only a subset of the potential
health benefits expected to result from the regulatory options. While additional adverse health effects are also
associated with pollutants in bottom ash transport water and FGD wastewater (such as kidney damage from
cadmium or selenium exposure, gastrointestinal problems from zinc, thallium, or boron exposure, and others),
the lack of data on dose-response relationships46 between ingestion rates and these effects precluded EPA
from quantifying the associated health effects.
The EPA's analysis of the monetary value of human health effects utilizes data and methodologies described
in Chapter 3 and in the Supplemental EA (U.S. EPA, 2019a). The relevant data include COMIDs47 for
receiving waters, estimated baseline and regulatory options annual plant-level loadings of each discharged
pollutant, estimated ambient pollutant concentrations in receiving reaches, and estimated fish consumption
rates among different age and ethnic cohorts for affected recreational anglers and subsistence fishers.
Section 5.1 describes how the EPA identified the population potentially exposed to pollutants from steam
electric power plant discharges via fish consumption. Section 5.2 describes the methods for estimating fish
tissue pollutant concentrations and potential exposure via fish consumption in the affected population.
Sections J.J to 5.5 describe EPA's analysis of various human health endpoints potentially affected by the
regulatory options. Section 5.7 provides additional measures of human health benefits. Section 5.8 describes
these assumptions, limitations, and uncertainties.
46 A dose response relationship is an increase in incidences of an adverse health outcome per unit increase in exposure to a toxin.
47 A COMID is a unique numeric identifier for a given waterbody, assigned by a joint effort of the United States Geological Survey,
EPA, and Horizon Systems, Inc.
EPA-821-R-19-011
5-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
In general, the estimated human health effects of the proposed regulatory option, Option 2, are small
compared to those estimated in 2015 (see U.S. EPA, 2015a).
5.1 Affected Population
The affected population (/. e., individuals potentially exposed to steam electric pollutants via consumption of
contaminated fish tissue) includes recreational anglers and subsistence fishers who fish reaches affected by
steam electric power plant discharges (including receiving and downstream reaches), as well as their
household members. The EPA estimated the number of people who are likely to fish affected reaches based
on typical travel distances to a fishing site and presence of substitute fishing locations. The EPA notes that the
universe of sites potentially visited by recreational anglers includes reaches subject to fish consumption
advisories (FCA).48 Angler's response to FCA's presence is assumed to be reflected in their catch and release
practice, as discussed below. Since fish consumption rates vary across different age, racial and ethnic groups,
and fishing mode (recreational versus subsistence fishing), the EPA estimated potential health effects
separately for a number of age-, ethnicity-, and mode-specific cohorts.
First, for each Census Block Group (CBG) within 50 miles of an affected reach, the EPA assembled 2016
American Community Survey data on the number of people in 7 age categories (0 to 1, 2, 3 to 5, 6 to 10, 11 to
15, 16 to 21, and 21 or higher), and then subdivided each group according to 7 racial/ethnic categories:49 1)
White non-Hispanic; 2) African-American non-Hispanic; 3) Tribal/Native Alaskan non-Hispanic;
4) Asian/Pacific Islander non-Hispanic; 5) Other non-Hispanic (including multiple races); 6) Mexican
Hispanic; and 7) Other Hispanic50. Within each racial/ethnic group, the EPA further subdivided the
population according to recreational and subsistence groups. The Agency assumed that the 95th percentile of
the general population consumption rate is representative of the subsistence fisher consumption rate.
Accordingly, the Agency assumed that 5 percent of the angler population practices subsistence fishing.51
Finally, the EPA also subdivided the affected population by income into poverty and non-poverty groups,
based on the share of people below the federal poverty line.52 After subdividing population groups by age,
race, fishing mode, and the poverty indicator, each CBG has 196 unique population cohorts (7 age groups x 7
ethnic/racial groups x 2 fishing modes [recreational vs. subsistence fishing] x 2 poverty status designations).
48 Based on the EPA's review of studies documenting anglers' awareness of FCA and their behavioral responses to FCAs, 57.0
percent to 61.2 percent of anglers are aware of FCAs, and 71.6 percent to 76.1 percent of those who are aware ignore FCAs
(Burger, 2004, Jakus et al., 1997; Jakus et al., 2002; Williams et al., 2000). Therefore, only 17.4 percent of anglers may adjust
their behavior in response to FCA (U.S. EPA 2015a). As noted above, we assumed the angler's response to FCA is reflected in
their catch and release practice.
49 The racial/ethnic categories are based on available fish consumption data as well as the breakout of ethnic/racial populations in
Census data, which distinguishes racial groups within Flispanic and non-Hispanic categories.
50 The Mexican Flispanic and Flispanic block group populations were calculated by applying the Census tract percent Mexican
Elispanic and Flispanic to the underlying block-group populations, since these data were not available at the block-group level.
51 Data are not available on the share of the fishing population that practices subsistence fishing. The EPA assumed that 5 percent of
people who fish practice subsistence fishing, based on the assumed 95th percentile fish consumption rate for this population in the
EPA's Exposure Factors Elandbook (see U.S. EPA, 2011).
52 Poverty status is based on data from the Census Bureau's American Community Survey which determines poverty status by
comparing annual income to a set of dollar values called poverty thresholds that vary by family size, number of children, and the
age of the householder.
EPA-821-R-19-011
5-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
The EPA distinguished the exposed population by racial/ethnic group and poverty status to support analysis
of potential EJ considerations in baseline exposure to pollutants in steam electric power plant discharges, and
to allow evaluation of the effects of the regulatory options on mitigating any EJ concerns. See Chapter 14 for
details of the EJ analysis. As noted below, distinguishing the exposed population in this manner also allows
the Agency to account for differences in exposure among demographic groups, where supported by available
data.
Equation 5-1 shows how the EPA estimated the affected population, ExPop(i)(s)(c), for CBG i in state 5 for
cohort c.
Equation 5-1. ExPop(l)(s)(c) = Pop(l)(c)x %Flsh(s) x CaR(c)
Where:
Pop(i)(c) = Total CBG population in cohort c. Age and racial/ethnicity-specific populations in each
CBG are based on data from the 2016 American Community Survey, which provides
population numbers for each CBG broken out by age and racial/ethnic group separately.
To estimate the population in each age- and ethnicity/race-specific group, the EPA
calculated the share of the population in each racial/ethnic group and applied those
percentages to the population in each age group.
%Fish(s) = Fraction of people who live in households with anglers. To determine what percentage of the
total population participates in fishing, the EPA used region-specific U.S. Fish and
Wildlife Service (U.S. FWS, 2016) estimates of the population 16 and older who fish.53
The EPA assumed that the share of households that includes anglers is equal to the
fraction of people over 16 who are anglers.
CaR(c) = Adjustment for catch-and-re lease practices. According to U.S. FWS (2006) data,
approximately 23.3 percent of anglers release all the fish they catch ("catch-and-release"
anglers). Anglers practicing "catch-and-release" would not be exposed to steam electric
pollutants via consumption of contaminated fish. For all recreational anglers, the EPA
reduced the affected population by 23.3 percent. The EPA assumed that subsistence
fishers do not practice "catch-and-release" fishing.
Table 5-1 summarizes the population living within 50 miles of reaches affected by steam electric power plant
discharges (see Section 5.2.1 for a discussion of this distance buffer) and the EPA's estimate of the population
potentially exposed to the pollutants via consumption of subsistence- and recreationally-caught fish (based on
2016 population data and not adjusted for population growth during the analysis period). Of the total
population, 16.0 percent live within 50 miles of an affected reach and participate in recreational and/or
subsistence fishing, and 12.4 percent are potentially exposed to fish contaminated by steam electric pollutants
in bottom ash transport water and/or FGD wastewater discharges.
53 The share of the population who fishes ranges from 8 percent in the Pacific region to 20 percent in the East South Central region.
Other regions include the Middle Atlantic (10 percent), New England (11 percent), South Atlantic (15 percent), Mountain (15
percent), West South Central (17 percent), East North Central (17 percent), and West North Central (18 percent).
EPA-821-R-19-011
5-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
Table 5-1: Summary of Potentially Affected Population Living within 50 Miles of Affected Reaches
(baseline, as of 2016)
Total population
123,829,132
Total angler population3
19,772,063
Population potentially exposed to contaminated fishb c
15,395,517
a. Total population living within 50 miles of an affected reach multiplied by the state-specific share of the population who fishes
based on U.S. FWS (2016; between 8 percent and 20 percent, depending on the state).
b. Total angler population adjusted to reflect lower consumption rates from catch-and-release practices.
c. Analysis accounts for projected population growth so that the average affected population over the period of 2021 through 2047
is 12 percent higher than the population in 2016 presented in the table, or 17.2 million people. The analysis further assumes that
the fraction of the U.S. population engaged in recreational and subsistence fishing remains constant from 2021 through 2047.
Source: U.S. EPA Analysis, 2019
5.2 Pollutant Exposure from Fish Consumption
The EPA calculated an average fish tissue concentration for each pollutant for each CBG based on a length-
weighted average concentration for all reaches within 50 miles. For each combination of pollutant, cohort and
CBG, the EPA calculated the average daily dose (ADD) and lifetime average daily dose (LADD) consumed
via the fish consumption pathway.
5.2.1 Fish Tissue Pollutant Concentrations
The set of reaches that may represent a source of contaminated fish for recreational anglers and subsistence
fishers in each CBG depends on the typical travel distance anglers travel to fish. The EPA assumed that
anglers typically travel up to 50 miles to fish54, using this distance to estimate the relevant fishing sites for the
population of anglers in each CBG.
Anglers may have several fishable sites to choose from within 50 miles of travel. To account for the effect of
substitute sites, the EPA assumed that anglers are uniformly distributed among all the available fishing sites
within 50 miles from the CBG (travel zone) and alternate their travels across all the sites. For each CBG, the
EPA identified all fishable COMIDs within 50 miles (where distance was determined based on the Euclidian
distance between the centroid of the CBG and the midpoint of the COMID) and the COMID length in miles.
The EPA then calculated, for each CBG within the 50-mile buffer, the fish tissue concentration of As, Hg, and
lead (Pb). Appendix D describes the approach used to calculate fish tissue concentrations of steam electric
pollutants in the baseline and under each of the regulatory options.
For each CBG, the EPA then calculated the reach length (Lengtht) weighted fish fillet concentration (C
Fish Fmet (CBG)) based on all fishable COMIDS within the 50 mile radius according to Equation 5-2:
Equation 5-2. C^^CBG) =
5.2.2 Average Daily Dose
Exposure to steam electric pollutants via fish consumption depends on the cohort-specific fish consumption
rates. Table 5-2 summarizes the average fish consumption rates, expressed in daily grams per kilogram of
54 Studies of angler behavior and practices have made similar assumptions (e.g., Sohngen et al., 2015 and Sea Grant, Illinois-
Indiana, 2018).
EPA-821-R-19-011
5-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
body weight (BW), according to the race/ethnicity and fishing mode. The rates reflect recommended values
for consumer-only intake of finfish in the general population from all sources, based on the EPA's Exposure
Factors Handbook (U.S. EPA, 2011). For more details on these fish consumption rates, see U.S. EPA (2019a)
and the uncertainty discussion in Section 5.8.
Table 5-2: Summary of Group-specific Consumption Rates for Fish Tissue Consumption Risk
Analysis
Race/ Ethnicity3
EA Cohort"
Consumption Rate (g/kg BW/day)
Recreational
Subsistence
White (non-Hispanic)
Non-Hispanic White
0.67
1.9
African American (non-Hispanic)
Non-Hispanic Black
0.77
2.1
Asian/Pacific Islander (non-Hispanic)
Other, including Multiple Races
0.96
3.6
Tribal/Native Alaskan (non-Hispanic)
Other, including Multiple Races
0.96
3.6
Other non-Hispanic
Other, including Multiple Races
0.96
3.6
Mexican Hispanic
Mexican Hispanic
0.93
2.8
Other Hispanic
Other Hispanic
0.82
2.7
a. Each group is also subdivided into seven age groups (0-1, 2, 3-5, 6-10,11-15,16-20, Adult (21 or higher) and two income groups
(above and below the poverty threshold)).
b. U.S. EPA (2019a).
Source: U.S. EPA Analysis, 2019
Equation 5-3 and Equation 5-4 show the cohort- and CBG-specific ADD and LADD calculations based on
fish tissue concentrations, consumption rates, and exposure duration and averaging periods from U.S. EPA
(2019a), as shown below.
_ .. _ - cFish Fillet (') XCRpish(c)XFpish
Equation 5-3. ADD(c)(i) = —
Where:
ADD(c)(i) = average daily dose of pollutant from fish consumption for cohort c in CBG i
(milligrams[mg] per kilogram [kg] body weight [BW] per day)
Cfishjmetij) = average fish fillet pollutant concentration consumed by humans for CBG /' (mg per kg)
CRfishic) = consumption rate of fish for cohort c (grams per kg BW per day); see Table 5-2.
Ffsh = fraction of fish from reaches within the analyzed distance from the CBG (percent; assumed value
of 100%)
.. _ . W ADD(c)(i) XED(c) XEF
Equation 5-4. LADD(c)(i) = —
^ J ATX365
Where:
LADD (c)(i) = lifetime average daily dose (mg per kg BW per day) for cohort c in CBG /'
ADD (c)(j) = average daily dose (mg per kg BW per day) for cohort c in CBG /'
ED(c) = exposure duration (years) for cohort c
EF= exposure frequency (days; assumed value of 350)
AT= averaging time (years; assumed value of 70)
EPA-821-R-19-011
5-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
The EPA used the doses of steam electric pollutants as calculated above from fish caught through recreational
and subsistence fishing in its analysis of benefits associated with the various human health endpoints
described below.
5.3 Health Effects in Children from Changes in Lead Exposure
The EPA estimated changes in lead exposure as a result of the regulatory options are small compared to those
estimated in the 2015 analysis (see U.S. EPA, 2015a).
Lead is a highly toxic pollutant that can cause a variety of adverse health effects in children of all ages. In
particular, elevated lead exposure may induce a number of adverse neurological effects in children, including
decline in cognitive function, conduct disorders, attentional difficulties, internalizing behavior55, and motor
skill deficits (see National Toxicology Program 2012, U.S. EPA 2013b, U.S. EPA, 2019a, and U.S. EPA
2019h). Elevated blood lead (PbB) concentrations in children may also result in slowed postnatal growth in
children ages one to 16, delayed puberty in 8- to 17-year-olds, decreased hearing and motor function
(National Toxicology Program 2012, U.S. EPA 2019h). Lead exposure is also associated with adverse health
outcomes related to the immune system, including atopic and inflammatory responses (e.g., allergy and
asthma) and reduced resistance to bacterial infections. Studies have also found a relationship between lead
exposure in expectant mothers and lower birth weight in newborns (National Toxicology Program 2012; U.S.
EPA 2019h; Zhu et al., 2010). Because of data limitations, the EPA estimated only the effects of changes in
neurological and cognitive damages to pre-school (ages 0 to 7) children using the dose-response relationship
for IQ decrements (Crump et al. 2013).
The EPA estimated health effects from changes in exposure to lead to preschool children using PbB as a
biomarker of lead exposure. The EPA first modeled PbB under the baseline and post-compliance scenarios,
and then used a concentration-response relationship between PbB and IQ loss to estimate avoided IQ losses in
the affected population of children and changes in incidences of extremely low IQ scores (less than 70, or two
standard deviations below the mean). The EPA calculated the monetary value of changes in children's health
effects based on the impact of an additional IQ point on an individual's future earnings and the cost of
compensatory education for children with learning disabilities (including children with IQ less than 70 and
PbB levels above 20 |a,g/dL).
The EPA used the methodology described in Section 5.1 to estimate the population of children from birth to
age seven who live in recreational angler and subsistence fisher households and are potentially exposed to
lead via consumption of contaminated fish tissue. The EPA notes that fish tissue is not the only route of
exposure to lead among children. Other routes of exposure may include drinking water, dust, and other food.
The EPA used reference exposure values for these other routes of lead exposures and held these values
constant for the baseline and regulatory options scenarios. Since this health effect applies to children up to the
seventh birthday only, the EPA restricted the analysis to the relevant age cohorts of angler household
members.
55 Behavioral difficulties in children may include both externalizing behavior (e.g., such as inattention, impulsivity, conduct
disorders), and internalizing behaviors (e.g., withdrawn behaviors, symptoms of depression, fearfulness, and anxiety).
EPA-821-R-19-011
5-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
5.3.1 Methods
This analysis considers children who are born after implementation of the regulatory options and live in
recreational angler and subsistence fisher households. It relies on the EPA's Integrated Exposure, Uptake, and
Biokinetics (IEUBK) Model for Lead in Children (U.S. EPA, 2009c), which uses lead concentrations in a
variety of media - including soil, dust, air, water, and diet - to estimate total exposure to lead for children in
seven one-year age cohorts from birth through the seventh birthday. Based on this total exposure, the model
generates a predicted geometric mean PbB for a population of children exposed to similar lead levels (See the
2013 BCA report (U.S. EPA, 2013a) for more detail).
For each CBG, the EPA used the cohort-specific ADD based on Equation 5-3. The EPA then multiplied the
cohort-specific ADD by the average body weight for each age group56 to calculate the "alternative source"
input for the IEUBK model. Lead bioavailability and uptake after consumption varies for different chemical
forms. Many factors complicate the estimation of bioavailability, including nutritional status and timing of
meals relative to lead intake. For this analysis, the EPA used the default media-specific bioavailability factor
for the "alternative source" provided in the IEUBK model, which is 50 percent for oral ingestion.
The EPA used the IEUBK model to generate the geometric mean PbB for each cohort in each CBG under the
baseline and post-compliance scenarios. Note the IEUBK model processes daily intake to two decimal places
(fig/day). For this analysis, this means that some of the change between the baseline and regulatory options is
missed by using the model (i.e.. it does not capture very small changes), since the estimated changes in health
effects are driven by very small changes across large populations. This aspect of the model contributes to
potential underestimation of the actual monetary value of lead-related health effects in children arising from
the regulatory options.
5.3.1.1 Estimating Changes in IQ Point Losses
The EPA used the Crump et al. (2013) dose-response function to estimate changes in IQ losses between the
baseline and post-compliance scenarios. Comparing the baseline and post-compliance results provides the
changes in IQ loss per child. Crump et al. (2013) concluded that there was statistical evidence that the
exposure-response is non-linear over the full range of PbB. Equation 5-5 shows an exposure-response
function that represents this non-linearity:
Equation 5-5. AIQ = p1 x ln(PbB + 1)
Where:
Pi = -3.315 (log-linear regression coefficient on the lifetime blood lead level57)
Multiplying the result by the number of affected pre-school children yields the total change in the number of
IQ points for the affected population of children for the baseline and each regulatory option.
56 The average body weight values are 11.4 kg for ages 0 to 2, 13.8 kg for ages 2 to <3, 18.6 kg for ages 3 to <6, and 31.8 kg for
ages 6 to 7.
57 The lifetime blood lead level in children ages 0 to 7 is defined as a mean from six months of age to present (Crump et al. 2013).
EPA-821-R-19-011
5-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
The IEUBK model estimates the mean of the PbB distribution in children, assuming a continuous exposure
pattern for children from birth through the seventh birthday. The 2016 American Community Survey
indicates that children ages 0 to 7 are approximately evenly distributed by age. To get an annual estimate of
the number of children that would benefit from implementation of the regulatory options, the EPA divided the
estimated number of affected pre-school children by 7. This division adjusts the equation to apply only to
children age 0 to 1. The estimated changes in IQ loss is thus an annual value (i.e.. it would apply to the cohort
of children born each year after implementation).58 Equation 5-6 shows this calculation for the annual
increase in total IQ points.
Equation 5-6. A/Q(i)(c) = (in(AGM(l)(c)) * CRF *
Where:
AIQ(i)(c) = the difference in total IQ points between the baseline and regulatory option scenarios for
cohort c in CBG i
Ln(AGM(i)(c)) = the log-linear change in the average PbB in affected population of children (|_ig/dL) in
cohort c in CBG i
CRF= -3.315 (log-linear regression coefficient from Crump et al. (2013))
ExCh(i) = the number of affected children aged 0 to 7 for CBG i
The available economic literature provides little empirical data on society's overall WTP to avoid a decrease
in children's IQ. To determine the value of avoided IQ losses, the EPA used estimates of the changes in a
child's future expected lifetime earnings per one IQ point reduction and the cost of compensatory education
for children with learning disabilities.
The EPA monetized the value of an IQ point based on the methodology from Salkever (1995). The EPA
estimated the value of an IQ point using the methodology presented in Salkever's (1995) analysis but with
more recent data from the 1997 National Longitudinal Survey of Youth (U.S. EPA, 2019e). Updated results
based on Salkever (1995) indicate that a one-point IQ reduction reduces expected lifetime earnings by
2.63 percent. Table 5-3 summarizes the estimated values of an IQ point based on the updated Salkever (1995)
analysis using 3 percent and 7 percent discount rates. These values are discounted to the third year of life to
represent the midpoint of the exposed children population. The EPA also used an alternative value of an IQ
point from Lin et al. (2018) in a sensitivity analysis (Appendix G).
58 Dividing by seven undercounts overall benefits. Children from ages 1 to 7 (i.e., born prior to the base year of the analysis) are not
accounted for in the analysis, although they are also affected by lead exposure.
EPA-821-R-19-011
5-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
Table 5-3: Value of an IQ Point (2018$) based on Expected
Reductions in Lifetime Earnings
Discount Rate
Value of an IQ Pointa b (2018$)
3 percent
$20,832
7 percent
$4,358
a. Values are adjusted for the cost of education.
b. The EPA adjusted the value of an IQ point to 2018 dollars using the GDP
deflator.
Source: U.S. EPA (2019e) re-analysis of data from Salkever (1995)
5.3.1.2 Reduced Expenditures on Compensatory Education
Children whose PbB exceeds 20 |a,g/dL are more likely to have IQs less than 70, which means that they would
require compensatory education tailored to their specific needs. Costs of compensatory education and special
education are not reflected in the IQ point dollar value. Reducing exposure to lead at an early age is expected
to reduce the incidence of children requiring compensatory and/or special education, which would in turn
lower associated costs. Though these costs are not a substantial component of the overall benefits, they do
represent a potential benefit of reducing lead exposure. While the EPA quantitatively assessed this benefit
category using the methodology from the 2015 BCA (U.S. EPA, 2015a), the estimated cost savings from the
expected changes in the need of compensatory education are negligible and are not included in the total
monetized benefits.
5.3.2 Results
Table 5-4 shows the social welfare effects associated with changes in IQ losses from lead exposure via fish
consumption. The EPA estimated that regulatory options 1 and 2 lead to slight increases in lead exposure and,
as a result, forgone benefits, whereas Options 3 and 4 result in slight reductions. The total net change in IQ
points over the entire population of children with changes in lead exposure ranges from -11.1 points to
0.9 points. Annualized monetary values of increased IQ losses range from -$9,140 (Option 2) to $740 (Option
4) using a 3 percent discount, and -2,070 (Option 2) to $170 (Option 4) using a 7 percent discount.
Table 5-4: Estimated Monetary Value of Changes in IQ Points for Children Exposed to Lead
Regulatory
Option
Average Annual
Number of Affected
Children 0 to 7C
Total Change in IQ
Points, 2021 to 2047
in All Affected
Children Oto 7
Annualized Value of Changes in IQ Pointsa b
(Thousands 2018$)
3% Discount Rate
7% Discount Rate
Option 1
1,521,036
-3.58
-$2.96
-$0.67
Option 2
1,521,036
-11.07
-$9.14
-$2.07
Option 3
1,521,036
0.35
$0.29
$0.07
Option 4
1,521,036
0.90
$0.74
$0.17
a. Assumes that the loss of one IQ point results in the loss of 2.63 percent of lifetime earnings (following updated Salkever (1995)
values from U.S. EPA (2019e)).
b. Negative values represent forgone benefits.
c. The number of affected children is based on reaches analyzed across the four options. Some of the children included in this
count see no changes in exposure under some options.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
5-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
5.4 Heath Effects in Children from Changes in Mercury Exposure
The EPA estimated small changes in mercury exposure as a result of the regulatory options, compared to
those estimated in the 2015 analysis (see U.S. EPA, 2015a).
Mercury can have a variety of adverse health effects on adults and children (see U.S. EPA, 2019a). The
regulatory options may change the discharge of mercury to surface waters by steam electric power plants and
therefore affect a range of human health effects. Due to data limitations, however, the EPA estimated only the
monetary value of the changes in IQ losses among children exposed to mercury in-utero as a result of
maternal consumption of contaminated fish.
The EPA identified the population of children exposed in-utero starting from the CBG-specific affected
population described in Section 5.1. Because this analysis focuses only on infants born after implementation
of the regulatory options, the EPA further limited the affected population by estimating the number of women
between the ages of 15 and 44 potentially exposed to contaminated fish caught in the affected waterbodies,
and multiplying the result by ethnicity-specific average fertility rates.59 This yields the cohort-specific annual
number of births for each CBG.
The U.S. Department of Health and Human Services provides fertility rates by race for 2015 in the National
Vital Statistics Report (Martin et al., 2017). The fertility rate measures the number of births occurring per
1,000 women between the ages of 15 and 44 in a particular year. Fertility rates were highest for Hispanic
women at 71.7, followed by African Americans at 64.1, Caucasians at 59.3, Asian or Pacific Islanders at 58.5,
and Tribal/Other at 43.9.
5.4.1 Methods
The EPA used the same ethnicity- and mode-specific consumption rates shown in Table 5-2 and calculated
the CBG- and cohort-specific mercury ADD based on Equation 5-3. In this analysis, the EPA used a linear
dose-response relationship between maternal mercury hair content and subsequent childhood IQ loss from
Axelrad et al. (2007). Axelrad et al. (2007) developed a dose-response function based on data from three
epidemiological studies in the Faroe Islands, New Zealand, and Seychelle Islands. According to their results,
there is a 0.18 point IQ loss for each 1 part-per-million (ppm) increase in maternal hair mercury.
To estimate maternal hair mercury concentrations based on the daily intake (see Section 5.2.2), the EPA used
the median conversion factor derived by Swartout and Rice (2000), who estimated that a 0.08 j^ig/kg body
weight increase in daily mercury dose is associated with a 1 ppm increase in hair concentration. Equation 5-7
shows the EPA's calculation of the total annual IQ changes for a given receiving reach.
Equation 5-7. IQL(i)(c) = InExPop(i) * MADD(i)(c) * * DRF
Where:
IQL(i) = IQ changes associated with in-utero exposure to mercury from maternal consumption of fish
contaminated with mercury for cohort c in CBG i
InExPop(i) = affected population of infants in CBG /' (the number of births)
59 The EPA acknowledges that fertility rates vary by age. However, the use of a single average fertility rate for all ages is not
expected to bias results because the average fertility rate reflects the underlying distribution of fertility rates by age.
EPA-821-R-19-011
5-10
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
MADD(i)(c) = maternal ADD for cohort c in CBG i (^g/kg BW/day)
Conv = conversion factor for hair mercury concentration based on maternal mercury exposure
(0.08 (ig/kg BW/day per 1 ppm increase in hair mercury)
DRF= dose response function for IQ decrement based on marginal increase in maternal hair mercury
(0.18 point IQ decrement per 1 ppm increase in hair mercury)
Summing estimated IQ changes across all analyzed CBGs yields the total changes in the number of IQ points
due to in-utero mercury exposure from maternal fish consumption under each analyzed regulatory option. The
benefits of the regulatory options are calculated as the change in IQ points lost between the baseline and
modeled post-compliance conditions under each of the regulatory options.
The available economic literature provides little empirical data on society's overall WTP to avoid a decrease
in children's IQ. To determine the value of avoided IQ losses, the EPA used estimates of the changes in a
child's future expected lifetime earnings per one IQ point reduction and the cost of additional education. The
values of an IQ point presented in Section 5.3.1 are discounted to the third year of life to represent the
midpoint of the exposed children population. EPA further discounted the present value of lifetime income
differentials three additional years to reflect the value of an IQ point at birth and better align the benefits of
reducing exposure to mercury with in-utero exposure (U.S. EPA, 2019i). The IQ values discounted to birth
range from $3,704 to $19,064. The EPA also used an alternative value of an IQ point from Lin et al. (2018) in
a sensitivity analysis (Appendix G).
5.4.2 Results
Table 5-5 shows the estimated changes in IQ point losses for infants exposed to mercury in-utero and the
corresponding monetary values, using a 3 percent and 7 percent discount rates. All regulatory options result in
a small net increase in IQ losses and, as a result, in forgone benefits to society. Using a 3 percent discount
rate, monetary values of an increased IQ losses range from -$2.85 million (Option 3) to -$0.31 million
(Option 1). Using a 7 percent discount rate, estimates range from -$0.58 million (Option 3) to -$0.06 million
(Option 1).
Table 5-5: Estimated Monetary Values from Changes in IQ Points for Infants from Mercury Exposure
Regulatory Option
Number of
Affected Infants
per Yearc
Total Change in IQ
Points, 2021 to 2047
in All Affected Infants
Annualized Value of Changes in IQ Pointsa b (Millions
2018$)
3% Discount Rate
7% Discount Rate
Option 1
225,272
-411
-$0.31
-$0.06
Option 2
225,272
-3,785
-$2.84
-$0.57
Option 3
225,272
-3,777
-$2.85
-$0.58
Option 4
225,272
-2,021
-$1.49
-$0.30
a. Assumes that the loss of one IQ point results in the loss of 2.63 percent of lifetime earnings discounted to birth (following
updated Salkever (1995) values from U.S. EPA (2019i)).
b. Negative values represent forgone benefits.
c. The number of affected infants is based on reaches analyzed across the four options. Some of the children included in this count
see no changes in exposure under some options.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
5-11
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
5.5 Estimated Changes in Cancer Cases from Arsenic Exposure
Among steam electric pollutants that can contaminate fish tissue and are analyzed in the Supplemental EA,
arsenic is the only confirmed carcinogen with a published dose response function (see U.S. EPA, 2010b).6"
The EPA used the methodology presented in Section 3.6 of the 2015 BCA document (U.S. EPA 2015a) to
estimate the number of annual cancer cases associated with consumption of fish contaminated with arsenic
from steam electric power plant discharges under the baseline and the change corresponding to each
regulatory option and the associated monetary values. Based on the EPA's analysis, no changes in cancer
cases from exposure to arsenic via fish consumption are expected under the regulatory options. Accordingly,
the expected social welfare effects are zero under all regulatory options.
5.6 Total Monetary Values of Estimated Changes in Human Health Effects
Table 5-6 presents the estimated monetary value of changes in adverse human health outcomes under the
regulatory options. Using a 3 percent discount rate, the estimated monetary values range from -$2.85 million
to -$0.31 million. Using a 7 percent discount rate, the estimated monetary values range from -$0.58 million to
-$0.06 million. Negative values reflect forgone benefits. Changes in mercury exposure for children account
for the majority of total monetary values from increases in adverse health outcomes.
Table 5-6: Total Monetary Values of Changes in Human Health Outcomes Associated with Fish
Consumption for Regulatory Options (millions of 2018$)
Discount Rate
Regulatory
Option
Reduced Lead
Exposure for
Children8'"''
Reduced Mercury
Exposure for
Children3"
Reduced Cancer
Cases from
Arsenic
Total3"
1
<$0.00
-$0.31
$0.00
-$0.31
3%
2
-$0.01
-$2.84
$0.00
-$2.85
3
<$0.00
-$2.85
$0.00
-$2.85
4
<$0.00
-$1.49
$0.00
-$1.49
1
<$0.00
-$0.06
$0.00
-$0.06
7%
2
<$0.00
-$0.57
$0.00
-$0.57
3
<$0.00
-$0.58
$0.00
-$0.58
4
<$0.00
-$0.30
$0.00
-$0.30
a. Negative values represent forgone benefits and positive values represent realized benefits.
b. Assumes that the loss of one IQ point results in the loss of 2.63 percent of lifetime earnings (following updated Salkever (1995)
values from U.S. EPA (2019e)).
c. "<$0.00" indicates that monetary values are greater than -$0.01 million but less than $0.00 million. Benefits to children from
changes in exposure to lead range from -$9.1 to $0.7 thousands per year, using a 3 percent discount rate, and from -$2.1 to $0.2
thousands, using a 7 percent discount rate.
Source: U.S. EPA Analysis, 2019
5.7 Additional Measures of Potential Changes in Human Health Effects
As noted in the introduction to this chapter, untreated pollutants in steam electric power plant discharges have
been linked to additional adverse human health effects. The EPA compared immediate receiving water
concentrations to human health-based NRWQC in U.S. EPA (2019a). To provide an additional measure of the
o0 Although other pollutants, such as cadmium, are also likely to be carcinogenic (see U.S. Department of Health and Human
Services (U.S. DHHS), 2012), the EPA did not identify dose-response functions to quantify the effects of changes in these other
pollutants.
EPA-821-R-19-011
5-12
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
potential health effects of the regulatory options, the EPA also estimated the expected changes in the number
of receiving and downstream reaches with pollutant concentrations in excess of human health-based
NRWQC. This analysis and its findings are not additive to the preceding analyses in this chapter, but instead
represent another way of characterizing potential health effects resulting from changes in exposure to steam
electric pollutants. This analysis compares in-stream pollutant concentrations estimated for the baseline and
each analyzed regulatory option in receiving reaches and downstream reaches to criteria established by the
EPA for protection of human health. The EPA compared estimated in-water concentrations of antimony,
arsenic, barium, cadmium, chromium, cyanide, copper, lead, manganese, mercury, nitrate-nitrite as N, nickel,
selenium, thallium, and zinc to EPA's national recommended water quality criteria protective of human health
used by states and tribes (U.S. EPA, 2018b).61 Estimated pollutant concentrations in excess of these values
indicate potential risks to human health.
Table 5-7 shows the results of this analysis.62 The EPA estimates that with baseline steam electric pollutant
discharges, in-stream concentrations of steam electric pollutants exceed human health criteria for at least one
pollutant in 141 reaches based on the "consumption of water and organism" criteria, and 37 reaches based on
the "consumption of organism only" criteria nationwide. The EPA estimates that the total number of reaches
with exceedances to increase under Options 1, 2, and 3 and decrease under Option 4. Table 5-7 presents the
estimated number of stream reaches that may change for each regulatory option.
Table 5-7: Estimated Number of Reaches Exceeding Human Health Criteria for Steam Electric
Pollutants
Number of Reaches with
Ambient Concentrations
Number of Reaches with Higher
Number of Reaches with Lower
Regulatory
Option
Exceeding Human Health
Number of Exceedances,
Number of Exceedances,
Criteria for at Least One
Pollutant3
Relative to Baseline
Relative to Baseline
Consumption
Consumption
Consumption
Consumption
Consumption
Consumption
of Water +
of Organism
of Water +
of Organism
of Water +
of Organism
Organism
Only
Organism
Only
Organism
Only
Baseline
141
37
-
-
-
-
Option 1
165
37
30
3
0
0
Option 2
222
71
85
37
12
5
Option 3
171
38
34
4
12
5
Option 4
110
27
16
4
66
21
a. Pollutants for which there was at least one exceedance include antimony, arsenic, cyanide, lead, manganese, nitrate-nitrite as N,
selenium, and thallium.
Source: U.S. EPA Analysis, 2019
5.8 Limitations and Uncertainties
The analysis presented in this chapter does not include all possible human health effects associated with post-
compliance changes in pollutant discharges due to lack of data on a dose-response relationship between
ingestion rates and potential adverse health effects. Therefore, the total quantified human health effects
01 For pollutants that do not have national recommended water quality criteria protective of human health, EPA used MCLs. These
pollutants include cadmium, chromium, lead, and mercury.
02 Only reaches designated as fishable {i.e., Strahler Stream Order larger than 1) were included in the human health ambient water
quality criteria exceedances analysis.
EPA-821-R-19-011
5-13
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
included in this analysis represent only a subset of the potential health effects expected to result from the
regulatory options.
Additionally, the methodologies and data used in the analysis of health effects associated with changes in
incidences of adverse health outcomes due to consumption of fish contaminated with steam electric pollutants
involve limitations and uncertainties. Table 5-8 summarizes the limitations and uncertainties and indicates the
direction of the potential bias. Note that the effect on benefits estimates indicated in the second column of the
table refers to the magnitude of the benefits rather than the direction (i.e.. a source of uncertainty that tends to
underestimate benefits indicates expectation for larger forgone benefits). Additional limitations and
uncertainties associated with the EA analysis and data are discussed in the Supplemental EA (see U.S. EPA,
2019a).
Table 5-8: Limitations and Uncertainties in the Analysis of Human Health Effects
U nee rta i nty/Assu m ptio n
Effect on Benefits
Estimate
Notes
The EPA estimated the annual
average loadings during the
period of analysis and
estimated annual average
concentrations to which
individuals or environmental
receptors are exposed over
the period of analysis.
Uncertain
The timing of changes in pollutant levels is an important
factor for analyzing the benefits of the regulatory options.
However, the approach for estimating the benefits of changes
in pollutant concentration cannot readily incorporate a
complex temporal profile of pollutant loadings, nor would the
analysis necessarily gain by doing so for benefits depend on
long-term processes such as adverse health effects from
lifetime exposures to toxic pollutants.
The EPA's analysis uses annual
average values for stream
flows.
Uncertain
The EPA recognizes that low-flow periods may coincide with
higher pollutant loadings and result in higher pollutant
concentrations, and vice versa. There may be human health
effects from short-duration exposure to higher steam electric
pollutant levels. The Agency's analysis focused on long-term
exposure only given that concentrations are not likely to
reach levels of concern for acute exposure, and adverse
health effects for non-acute short-duration exposures are
generally not well understood.
Anglers are assumed to be
distributed evenly (over the
reach miles) over all available
fishing sites within the 50-mile
travel distance.
Uncertain
The EPA assumed that all anglers travel up to 50 miles and
distribute their visits over all fishable sites within the area. In
fact, recreational anglers may have preferred sites {e.g., a site
located closer to their home) that they visit more frequently.
The characteristics of these sites, notably ambient water
concentrations and fishing advisories, affects exposure to
pollutants, but the EPA does not have data to support a more
detailed analysis of fishing visits. The impact of the
assumption on monetary estimates is uncertain since
fewer/more anglers may be exposed to higher/lower fish
tissue concentrations than assumed by the EPA in the
analysis.
The exposed population is
estimated based on
households in proximity to
affected reaches and the
fraction of the general
population who fish.
Uncertain
The EPA assumed that the share of households that includes
anglers is equal to the fraction of people over 16 who are
anglers. On the one hand, this may double-count households
with two anglers over 16. On the other hand, the exposed
population may also include non-household members who
also consume the catch.
EPA-821-R-19-011
5-14
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 5: Human Health Benefits from Fish Ingestion
Table 5-8: Limitations and Uncertainties in the Analysis of Human Health Effects
U nee rta i nty/Assu m ptio n
Effect on Benefits
Estimate
Notes
Fish intake rates used in
estimating exposure are based
on recommended values for
the entire consuming
population and include all fish
sources.
Overestimate
The fish consumption rates used in the analysis account for all
fish sources, i.e., store-bought or recreationally-caught fish.
This assumption may overestimate exposure from
recreationally-caught fish. The degree of the overestimate is
unknown as the fish consumption rates for the general
consuming population are within the range of freshwater
recreational fish intake rates reported in EPA's Exposure
Factors Handbook (U.S. EPA, 2011).
The number of subsistence
fishers was assumed to equal
5 percent of the total number
of anglers fishing the affected
reaches.
Uncertain
The magnitude of subsistence fishing in the United States or
individual states is not known. Assuming 5 percent may
understate or overstate the number of potentially affected
subsistence fishers (and their households) overall, and ignores
potential variability in subsistence rates across racial/ethnic
groups.
There is a linear 0.18 point IQ
loss for each 1 ppm increase
in maternal hair mercury.
Uncertain
This dose-response function may over- or underestimate IQ
impacts arising from mercury exposure if a linear function is
not the best representation of the relationship between
maternal body burden and IQ losses.
For the mercury- and lead-
related health impact
analyses, the EPA assumed
that IQ losses are an
appropriate endpoint for
quantifying adverse cognitive
and neurological effects
resulting from childhood or in-
utero exposures to lead and
mercury (respectively).
Underestimate
IQ may not be the most sensitive endpoint. Additionally, there
are deficits in cognitive abilities that are not reflected in IQ
scores, including acquisition and retention of information
presented verbally and many motor skills (U.S. EPA, 2005b).
To the extent that these impacts create disadvantages for
children exposed to mercury at current exposure levels or
result in the absence of (or independent from) measurable IQ
losses, this analysis may underestimate the social welfare
effects of the regulatory options of increased lead and
mercury exposure.
The IEUBK model processes
daily intake from "alternative
sources" to 2 decimal places
(Hg/day).
Underestimate
Since the fish-associated pollutant intakes are small, some
variation is missed by using this model {i.e., it does not
capture very small changes).
The EPA did not quantify the
health effects associated with
changes in adult exposure to
mercury.
Underestimate
The scientific literature suggests that exposure to mercury
may have significant adverse health effects for adults; if
measurable effects are occurring at current exposure levels,
excluding the effects of increased adult exposure results in an
underestimate of forgone benefits.
EPA-821-R-19-011
5-15
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
6: Non-Market Benefits
6 Nonmarket Benefits from Water Quality Changes
As discussed in the Supplemental EA (U.S. EPA, 2019a), heavy metals, nutrients, and other pollutants
discharged by steam electric power plants can have a wide range of effects on water resources located in the
vicinity and downstream from the plants. These environmental changes affect environmental goods and
services valued by humans, including recreation; commercial fishing; public and private property ownership;
navigation; water supply and use; and existence services such as aquatic life, wildlife, and habitat designated
uses. Some environmental goods and services (e.g., commercially caught fish) are traded in markets, and thus
their value can be directly observed. Other environmental goods and services (e.g., recreation and support of
aquatic life) cannot be bought or sold directly and thus do not have observable market values. These second
types of environmental goods and services are classified as "nonmarket". The expected changes in the
nonmarket values of the water resources affected by the regulatory options (hereafter nonmarket benefits) are
additive to the market benefits (e.g., avoided costs of producing various market goods and services).
The analysis of the nonmarket value of water quality changes resulting from the regulatory options follows
the same approach the EPA used in the analysis of the 2015 rule (U.S. EPA, 2015a). This approach, which is
briefly summarized below, involves:
• characterizing the change in water quality for the regulatory options relative to the baseline using a
WQI and linking these changes to ecosystem services or potential uses that are valued by society (see
Section 6.1),
• monetizing changes in the nonmarket value of affected water resources attributable to the regulatory
options using a meta-analysis of surface water valuation studies that provide data on the public's
WTP for water quality changes (see Section 6.2).
The analysis accounts for changes in water quality resulting from changes in nutrient, sediment, and toxics
concentrations in reaches potentially affected by bottom ash transport water and FGD wastewater discharges.
In general, the analysis shows that the estimated effects of the proposed regulatory option, Option 2, on the
nonmarket value of water quality are small compared to those estimated in 2015 (see U.S. EPA, 2015a).
6.1 Linking Changes in Water Quality to Valuation
Once an overall WQI value is calculated (see Section 3.4 for detail), it can be related to suitability for
potential uses. Vaughan (1986) developed a water quality ladder (WQL) that can be used to indicate whether
water quality is suitable for various human uses (i.e., boating, rough fishing, game fishing, swimming, and
drinking without treatment). Vaughan identified "minimally acceptable parameter concentration levels" for
each of the five potential uses. Vaughan used a scale of zero to 10 instead of the WQI scale of zero to 100 to
classify water quality based on its suitability for potential uses. Therefore, the WQI value corresponding to a
given water quality use classification equals the WQL value multiplied by 10. See Table 3-5 (in Chapter 3)
for the correspondence between WQI scores and use classifications.
6.2 Total WTP for Water Quality Changes
The EPA estimated economic values of water quality changes at the CBG level using results of a meta-
analysis of 140 estimates of total WTP (including both use and nonuse values) for water quality
EPA-821-R-19-011
6-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
6: Non-Market Benefits
improvements, provided by51 original studies conducted between 1981 and 2011.63 The estimated
econometric model allows calculation of total WTP for changes in a variety of environmental services
affected by water quality and valued by humans, including changes in recreational fishing opportunities, other
water-based recreation, and existence services such as aquatic life, wildlife, and habitat designated uses. The
model also allows EPA to adjust WTP values based on the core geospatial factors predicted by theory to
influence WTP, including: scale (the size of affected resources or areas), market extent (the size of the market
area over which WTP is estimated), and the availability of substitutes. The meta-analysis regression is based
on two models: Model 1 provides the EPA's central estimate of non-market benefits and Model 2 develops a
range of estimates that account for uncertainty in the WTP estimates. Appendix H provides details on how the
EPA used the meta-analysis to predict household WTP for each CBG and year as well as the estimated
regression equation intercept, variable coefficients for the two models used in this analysis, and the
corresponding independent variable names and assigned values.
Based on the meta-analysis results, the EPA multiplied the coefficient estimates for each variable (see Model
1 and Model 2 in Table H-l) by the variable levels calculated for each CBG or fixed at the levels indicated in
"Assigned Value" in Table H-l. The sum of these products represents the predicted natural log of marginal
household WTP (In MWTP) for a representative household in each CBG. Equation 6-1 provides the discount
formula used to calculate household benefits for each CBG.
Equation 6-1. HWTPYB = MWTPY B x AWQIB
where:
HWTPy,b = Annual household WTP in 2018$ in year 7 for households located in
the CBG (5),
MWTPy,b = Marginal WTP for water quality for a given year (7) and the CBG (B)
estimated by the meta-analysis function and evaluated at the midpoint
of the range over which water quality is changed,
AWQIb = Estimated annual average water quality change for the CBG (B).
As summarized in Table 6-1, average annual household WTP estimates for the regulatory options range from
-$0.11 under Option 1 to $ 1.04 under Option 4, for the four regulatory options the EPA analyzed.
63 Although the potential limitations and challenges of benefit transfer are well established (Desvousges et al., 1998), benefit
transfers are a nearly universal component of benefit cost analyses conducted by and for government agencies. As noted by Smith
et al. (2002; p. 134), "nearly all benefit cost analyses rely on benefit transfers, whether they acknowledge it or not."
EPA-821-R-19-011
6-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
6: Non-Market Benefits
Table 6-1: Estimated Household Willingness-to-Pay for Water Quality Changes
Regulatory
Option
Number of Affected
Households (Millions)
Average Annual WTP Per Household (2018$)a,b
Low
Central
High
Option 1
85.24
-$0.11
-$0.14
-$0.62
Option 2
86.86
$0.10
$0.14
$0.56
Option 3
84.64
$0.16
$0.22
$0.87
Option 4
86.51
$0.19
$0.26
$1.04
a. Negative values represent forgone benefits and positive values represent realized benefits
b. Model 2 provides low and high estimates for each option, while Model 1 provides central estimates. We note that the central
estimate does not fall at the midpoint of the range, but instead represents the value from Model 1 which falls between the low
and high bound estimates provided by Model 2.
Source: U.S. EPA Analysis, 2019
To estimate total WTP (TWTP) for water quality changes for each CBG, the EPA multiplied the per-
household WTP values for the estimated water quality change by the number of households within each block
group in a given year. The EPA then calculated annualized total WTP values for each CBG with both a
3 percent and 7 percent discount rate as shown below in Equation 6-2. As discussed in Chapter 1, monetary
values of water quality changes are estimated for all years between 2021 and 2047.
Equation 6-2.
where:
2047
TWTPb = f ^
HWTPyb X HHy b
m =2021
(1 + 0
,'^-2018
X
i x (1 + i)n
(1 + i)n+1 - 1
TWTPb = Total household WTP in 2018$ for households located in the CBG
CB),
HWTPy,b = Annual household WTP in 2018$ for households located in the CBG
(B) in year (Y),
HHy,b = the number of households residing in the CBG (B) in year (Y).
T = Year when benefits are realized
i = Discount rate (3 or 7 percent)
n = Duration of the analysis (27 years)64
The EPA generated annual household counts for each CBG through the period of analysis based on projected
population growth following the method described in Section 1.3.6. Table 6-2 presents the results for the
3 percent and 7 percent discount rates.
04 See Section 1.3.3 for detail on the period of analysis.
EPA-821-R-19-011 6-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 6: Non-Market Benefits
Table 6-2: Estimated Total Annualized Willingness-to-Pay for Water Quality Changes Compared to
Baseline (Millions 2018$)
Regulatory
Number of Affected
3% Discount Ratea
7% Discount Rate3
Option
Households (Millions)
Low
Central
High
Low
Central
High
Option 1
85.2
-$10.0
-$12.5
-$55.5
-$8.6
-$10.9
-$48.1
Option 2
86.9
$11.8
$16.7
$65.6
$10.1
$14.3
$56.1
Option 3
84.6
$16.3
$22.5
$90.7
$14.0
$19.4
$77.8
Option 4
86.5
$19.8
$27.3
$110.2
$17.0
$23.6
$94.6
a. Negative values represent forgone benefits and positive benefits represent realized benefits.
Source: U.S. EPA Analysis, 2019
The total annualized benefits of water quality changes resulting from reduced toxics, nutrient and sediment
discharges in these reaches range from -$55.5 million under Option 1 (3 percent discount rate) to $110.2
million under Option 4 (3 percent discount rate). The negative values under Option 1 represent forgone
benefits, while the positive values for Option 2, 3, and 4 represent realized benefits. Appendix H provides a
detailed description of the results in Table 6-2.
6.3 Limitations and Uncertainties
Table 6-3 summarizes the limitations and uncertainties in the analysis of benefits associated with changes in
surface water quality and indicates the direction of any potential bias. Note that the effect on benefits
estimates indicated in the second column of the table refers to the magnitude of the benefits rather than the
direction (/'. e., a source of uncertainty that tends to underestimate benefits indicates expectation for larger
forgone benefits).
Table 6-3: Limitations and Uncertainties
in the Analysis of Nonmarket Water Quality Benefits
Issue
Effect on Benefits
Estimate
Notes
Limitations inherent to the meta-analysis model and benefit transfer
Use of 100-mile buffer
for calculating water
quality benefits for each
CBG
Underestimate
The distance between the surveyed households and the affected
waterbodies is not well measured by any of the explanatory variables
in the meta-regression model. The EPA would expect values for water
quality changes to diminish with distance (all else equal) between the
home and affected waterbody. The choice of 100 miles is based on
typical driving distance to recreational sites {i.e., 2 hours or 100 miles).
Therefore, the EPA used 100 miles to approximate the distance decay
effect on WTP values. The analysis effectively assumes that people
living farther than 100 miles place no value on water quality
improvements for these waterbodies despite literature that shows
that while WTP tends to decline with distance from the waterbody,
people place value on the quality of waters outside their region.
Selection of the WQI
parameter value for
estimating low and high
WTP values
Uncertain
The EPA set AWQI to 5 and 50 units to estimate high and low benefit
values based on Model 2. These values were based on the lowest and
highest water quality changes included in the meta-data. To the
extent that AWQI = 50 is significantly larger than the change in water
quality expected from the regulatory options, it is likely to significantly
understate the estimated WTP value. AWQI = 5 is more consistent
with the magnitude of water quality changes resulting from the
regulatory options.
EPA-821-R-19-011 6-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
6: Non-Market Benefits
Table 6-3: Limitations and Uncertainties in the Analysis of Nonmarket Water Quality Benefits
Issue
Effect on Benefits
Estimate
Notes
Whether potential
hypothetical bias is
present in underlying
stated preference results
Uncertain
Following standard benefit transfer approaches, this analysis proceeds
under the assumption that each source study provides a valid,
unbiased estimate of the welfare measure under consideration (cf.
Moeltner et al. 2007; Rosenberger and Phipps 2007). To minimize
potential hypothetical bias underlying stated preference studies
included in meta-data, the EPA set independent variable values to
reflect best benefit transfer practices.
Use of different water
quality measures in the
underlying meta-data
Uncertain
The estimation of WTP may be sensitive to differences in the
environmental water quality measures across studies in the meta
data. Studies that did not use the WQI were mapped to the WQI so a
comparison could be made across studies. In preliminary model runs,
the EPA tested a dummy variable (WQI) that captures the effect of a
study using (WQI=1) or not using (WQI=0) the WQI. The variable
coefficient was not statistically different from zero, indicating no
evidence of systematic bias in the mapping of studies that did not use
the WQI.
Transfer error
Uncertain
Transfer error may occur when benefit estimates from a study site are
adopted to forecast the benefits of a policy site. Rosenberger and
Stanley (2006) define transfer error as the difference between the
transferred and actual, generally unknown, value. While meta-analysis
is fairly accurate when estimating benefit function, transfer error may
be a problem in cases where the sample size is small. Meta-analyses
have been shown to outperform other function-based transfer
methods in many cases, but this result is not universal (Shrestha et al.
2007). This notwithstanding, results reviewed by Rosenberger and
Phipps (2007) are "very promising" for the performance of meta-
analytic benefit transfers relative to alternative transfer methods.
Use of the WQI to link water quality changes to human uses and support for aquatic and terrestrial species
Omission of Great Lakes
and estuaries from
analysis of benefits from
water quality changes
Underestimate
Eight out of 112 (7 percent) steam electric power generating plants
discharge to the Great Lakes or estuaries. Due to limitations of the
water quality models used in the analysis of the regulatory options,
these waterbodies were excluded from the analysis. This omission is
likely to underestimate benefits of water quality changes from the
regulatory options.
Changes in WQI reflect
only reductions in toxics,
nutrient, and total
suspended sediment
concentrations
Uncertain
The estimated changes in WQI reflect only water quality changes
resulting directly from changes in toxics, nutrient and sediment
concentrations. They do not include changes in other water quality
parameters {e.g., BOD, dissolved oxygen) that are part of the WQI. If
the omitted water quality parameters also change, then the analysis
underestimates the expected water quality changes.
EPA-821-R-19-011
6-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
6: Non-Market Benefits
Table 6-3: Limitations and Uncertainties in the Analysis of Nonmarket Water Quality Benefits
Issue
Effect on Benefits
Estimate
Notes
In-stream toxics
concentrations are
based only on loadings
from steam electric
power generating plants
and other TRI
dischargers
Uncertain
In-stream concentrations for toxics were estimated based on loadings
from steam electric power plant and other TRI dischargers only and, as
a result, do not account for background concentrations of these
pollutants from other sources, such as contaminated sediments, non-
point sources, point sources that are not required to report to TRI, air
deposition, etc. Not including other contributors to background toxics
concentrations in the analysis is likely to result in understatement of
baseline concentrations of these pollutants and therefore of NRWQC
exceedances. The overall impact of this limitation on the estimated
WTP for water quality changes is uncertain but is expected to be small
since the WTP function used in this analysis is most sensitive to the
change in water quality.
EPA-821-R-19-011
6-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 7: Threatened & Endangered Species Benefits
7 Impacts and Benefits to Threatened and Endangered Species
7.1 Introduction
Threatened and endangered (T&E) species are species vulnerable to future extinction or at risk of extinction
in the near future, respectively. These designations reflect low or rapidly declining population levels, loss of
essential habitat, or life history stages that are particularly vulnerable to environmental alteration. In many
cases, T&E species are given special protection due to inherent vulnerabilities to habitat modification,
disturbance, or other human impacts. This chapter examines the change in environmental impacts of steam
electric power plant discharges on T&E species and the benefits associated with changes resulting from the
regulatory options.
As described in the Supplemental EA (U.S. EPA, 2019a), the untreated chemical constituents of steam electric
power plant waste streams can pose serious threats to ecological health due to the bioaccumulative nature of
many pollutants, high concentrations, and high loadings. Pollutants such as selenium, arsenic and mercury
have been associated with fish kills, disruption of growth and reproductive cycles and behavioral and
psychological alterations in aquatic organisms (U.S. EPA, 2015a; Appendix I). Additionally, high nutrient
loads can lead to the eutrophication of waterbodies. Eutrophication can lead to increases in the occurrence and
intensity of water column phytoplankton, including harmful algal blooms (e.g., nuisance and/or toxic species),
which have been found to cause fatal poisoning in other animals, fish, and birds (Williams et al., 2001).
Eutrophication may also result in the loss of critical submerged rooted aquatic plants (or macrophytes), and
reduced DO, levels, leading to anoxic or hypoxic waters.
For species vulnerable to future extinction, even minor changes to growth and reproductive rates and small
levels of mortality may represent a substantial portion of annual population growth. To quantify the effects of
the regulatory options compared to baseline, the EPA identified the inhabited waterbodies that see changes in
achievement of wildlife NRWQC, relative to the baseline, as a consequence of the regulatory options and
used these data to estimate the number of the geographic locations where the options are likely to affect T&E
species recovery. Because NRWQC are set at levels to protect aquatic organisms, reducing the frequency at
which aquatic life-based NRWQC are exceeded is likely to translate into reduced risk to T&E species and
potential improvement in species population. Conversely, increasing the frequency of exceedances may
increase risk to T&E species and jeopardize their survival or recovery.
In this chapter, the EPA explores the current conservation status of major freshwater taxa and identifies the
extent to which the regulatory options can be expected to benefit species protected by the Endangered Species
Act (ESA).
In general, the analysis shows the estimated effects of the proposed regulatory option, Option 2, on T&E
species to be small compared to those estimated in 2015 for the baseline.
7.2 Baseline Status of Freshwater Fish Species
Reviews of aquatic species' conservation status over the past three decades have documented the effect of
cumulative stressors on freshwater aquatic ecosystems, resulting in a significant decline in the biodiversity
and condition of indigenous communities (Deacon et al., 1979; Williams et al., 1989; Williams et al., 1993;
Taylor et al., 1996; Taylor et al., 2007; Jelks et al., 2008). Overall, aquatic species are disproportionately
imperiled relative to terrestrial species. For example, while 39 percent of freshwater and diadromous fish
EPA-821-R-19-011
7-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 7: Threatened & Endangered Species Benefits
species (Jelks et al., 2008) are classified as T&E, a similar status review found that only 7 percent of North
American bird and mammal species are currently imperiled (Wilcove and Master, 2005).
Approximately 39 percent of described fish species in North America are imperiled, with 700 fish taxa
classified as vulnerable (230), threatened (190), or endangered (280) in addition to 61 taxa presumed extinct
or functionally extirpated from nature (Jelks et al., 2008). These data show that the number of T&E species
have increased by 98 percent and 179 percent when compared to similar reviews conducted by the American
Fisheries Society in 1989 (Williams et al., 1989) and 1979 (Deacon et al., 1979), respectively. Despite recent
conservation efforts, including the listing of several species under the ESA, only 6 percent of the fish taxa
assessed in 2008 had improved in status since the 1989 inventory (Jelks et al., 2008).
Several families of fish have strikingly high proportions of T&E species. Approximately 46 percent and
44 percent of species within families Cyprinidae (carps and true minnows) and Percidae (darters and perches)
are imperiled, respectively. Some families with few, wide-ranging species have even higher rates of
imperilment, including the Acipenseridae (sturgeons; 88 percent) and Polyodontidae (paddlefish;
100 percent). Families with species important to sport and commercial fisheries ranged from a low of
22 percent for Centrarchidae (sunfishes) to a high of 61 percent for Salmonidae (salmon) (Jelks et al., 2008).
7.3 T&E Species Affected by the Regulatory Options
To assess the potential effects of the regulatory options on T&E species, the EPA constructed databases to
determine which species are found in waters expected to improve or degrade due to changes in pollutant
discharge from steam electric power plants. Notably, these databases exclude all species considered
threatened or endangered by scientific organizations (e.g., the American Fisheries Society [Williams et al.,
1993; Taylor et al., 2007; Jelks et al., 2008]) but not protected by the ESA.65 These databases allowed EPA to
estimate the changes in potential impacts of steam electric power plant discharges on surface waters
overlapping critical habitat of T&E species, a quantitative, but unmonetized proxy of the benefits associated
with the regulatory options.
7.3.1 Identifying T&E Species Potentially Affected by the Regulatory Options
To estimate the effects of the regulatory options on surface waters overlapping with critical habitat of T&E
species, all affected species must first be identified. The EPA identified all species currently listed or in
consideration for listing under the ESA using the U.S. FWS Environmental Conservation Online System
(U.S. FWS, 2014a). Whenever possible, the EPA obtained the geographical distribution of T&E species in
geographic information system (GIS) format as polygon (shape) files, line files (for inhabitants of small
creeks and rivers) and as a subset of geodatabase files. Data sources include U.S. FWS (2014b), the National
Oceanic and Atmospheric Administration's (NOAA's) Office of Response and Restoration (NOAA, 2010),
NatureServe (NatureServe, 2014), and NOAA National Marine Fisheries Service (NMFS, 2014a; NMFS,
2014b; NMFS, 2014c). For several freshwater species, geographic ranges were available only as 8-digit
HUCs (NatureServe, 2014; U.S. FWS, 2014b). For these species, the EPA compared 8-digit HUCs for T&E
species to 8-digit HUCs associated with affected reaches.
65 The EPA chose to limit its analysis to the species protected by the ESA due to limitations of the data provided by other scientific
organizations as well as time and resource constraints.
EPA-821-R-19-011
7-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 7: Threatened & Endangered Species Benefits
To determine the probability that an individual T&E species has critical habitat overlapping with surface
waters which could benefit from the regulatory options, the EPA compiled data on locations of steam electric
power plants and receiving waterbodies. See Supplemental EA for details on approach used to determine
outfall locations (U.S. EPA, 2019a). The result of this analysis consists of the NHDPlus reaches that receive
bottom ash transport water or FGD wastewater discharges from steam electric power plants and indicators of
water quality under the baseline and each analyzed regulatory option based on comparison of modeled
concentrations to aquatic life criteria (see Section 3.4.1.1). The EPA queried these data to identify "affected
areas" as those habitats where 1) receiving waters do not meet water quality benchmark values for pollutants
recognized to cause harm in organisms under baseline conditions but meet the benchmarks under one or more
of the regulatory options; and 2) receiving waters meet the benchmarks under baseline conditions but do not
meet the benchmarks under one or more of the regulatory options. The EPA used these data in ArcGIS to
determine the T&E species with habitat extents overlapping the affected areas.
The EPA identified T&E species living in aquatic habitats for several life history stages and/or species that
obtain a majority of their food from aquatic sources. Life history data used to classify species were obtained
from a wide variety of sources (Froese and Pauly, 2009; NatureServe, 2014; Alaska Fisheries Science Center
(AFSC), 2010; Atlantic States Marine Fisheries Commission (ASMFC), 2010; Northeast Fisheries Science
Center (NEFSC), 2010; Pacific Islands Fisheries Science Center (PIFSC), 2010a; PIFSC, 2010b; Southeast
Fisheries Science Center (SEFSC), 2010; Southwest Fisheries Science Center (SWFSC), 2010; U.S. FWS,
2010). For these species, the EPA conducted further analyses to remove from further consideration:
• Species presumed to be extinct, including those not collected for a minimum of 30 years.
• Endemic species living in waterbodies (e.g., isolated headwaters, natural springs) unlikely to be
affected by steam electric power plant discharges.
• Species protected by the ESA whose recovery plans i) do not include pollution or water quality issues
as factors preventing recovery, and ii) identify habitat destruction (due to damming, stream
channelization, water impoundments, wetland drainage, etc.) as a primary factor preventing recovery.
• Listings due to non-native species introductions and/or hybridization with native or non-native
congeners .
• Listings where water quality issues are identified as the primary issue preventing recovery, but where
a specific industry or entity not within the scope of the regulatory options is identified as the culprit..
• Species about which very little is known, including geographic distribution.
After eliminating the T&E species meeting these criteria, the EPA identified a total of 24 species, listed in
Table 7-1, whose known critical habitat overlaps with surface waters which may be affected by the regulatory
options.
EPA-821-R-19-011
7-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 7: Threatened & Endangered Species Benefits
Table 7-1: T&E Species with High Vulnerability Habitat Occurring within Waterbodies Affected by
Steam Electric Power Plants
Species Group
Species Count
Species
Common Name
Amphibians
1
Cryptobranchus alleganiensis
Hellbender salamander
Birds
1
Sterna antillarum
Least tern
Clams and Mussels
17
Cyprogenia stegaria
Fanshell
Dromus dromas
Dromedary pearlymussel
Epioblasma obliquata obliquata
Purple cat's paw
Epioblasma obliquata perobliqua
White cat's paw
Fusconaia cor
Shiny pigtoe
Fusconaia cuneolus
Finerayed pigtoe
Hemistena lata
Cracking pearlymussel
Lampsilis abrupta
Pink mucket
Lampsilis virescens
Alabama lampmussel
Lemiox rimosus
Birdwing pearlymussel
Obovaria retusa
Ring pink
Plethobasus cicatricosus
White wartyback
Plethobasus cooperianus
Orangefoot pimpleback
Plethobasus cyphyus
Sheepnose mussel
Pleurobema clava
Clubshell
Pleurobema plenum
Rough pigtoe
Quadrula fragosa
Winged mapleleaf
Fishes
3
Acipenser brevirostrum
Shortnose sturgeon
Acipenser oxyrinchus oxyrinchus
Atlantic sturgeon
Etheostoma trisella
Trispot darter
Reptiles
1
Clemmys muhlenbergii
Bog turtle
Snails
1
Athearnia anthonyi
Anthony's riversnail
Total
24
Source: U.S. EPA Analysis, 2019.
7.3.2 Estimating Effects of the Proposed Rule on T&E Species
The regulatory options, if implemented, have the potential to positively affect surface waters overlapping
known critical habitat for six T&E species. To assess effects of the regulatory options on these surface waters,
the EPA compared the estimated pollutant concentrations under the baseline and each regulatory option to
NRWQC for wildlife. For each of the six species considered in this analysis, the EPA estimated the
magnitude of potential benefits by identifying inhabited waterbodies likely to meet or fail to meet NRWQC
for aquatic life as a consequence of the regulatory options and comparing these areas to the overall area of
habitat occupied by T&E species.
First, for each T&E species affected by steam electric power plant discharges, the EPA estimated water
quality in each of the waterbodies inhabited by each T&E species under baseline conditions, and under
regulatory options conditions. Then, the EPA identified waterbodies that 1) do not meet NRWQC for wildlife
under baseline conditions, but have no wildlife NRWQC exceedances following implementation of the
regulatory options and 2) do meet NRWQC for wildlife under baseline conditions, but have wildlife NRWQC
exceedances following implementation of the regulatory options.
EPA-821-R-19-011
7-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 7: Threatened & Endangered Species Benefits
As shown in Table 7-2, six T&E species under Option 4 and two species (Atlantic sturgeon and Trispot
darter) under Options 2 and 3 have known critical habitat which overlaps with surface waters that may benefit
from habitat improvements.
Table 7-2: T&E Species Whose Habitat May Benefit from the Regulatory Options
Number of Reaches with NRWQC Exceedances for at Least
Species Common Namea
State(s)
One Pollutant
Baseline
Option 1
Option 2
Option 3
Option 4
Atlantic sturgeon
GA
1
1
0
0
0
Clubshell
PA
1
1
1
1
0
Hellbender salamander
PA
1
1
1
1
0
Least tern
KS
1
1
1
1
0
Trispot darter
GA
1
1
0
0
0
Winged mapleleaf
KS
1
1
1
1
0
Total number of reaches with NRWQC
exceedances
a. Species Latin names are listed in Table 7-1
Source: U.S. EPA Analysis, 2019.
The EPA's analysis also shows that 23 T&E species have known critical habitat which overlaps with surface
waters which may be adversely affected by water degradation under one or more of the regulatory options
(see Table 7-3 for detail). Note that there are five species listed in both Table 7-2 and Table 7-3 (Atlantic
sturgeon, Clubshell, Hellbender salamander, Least tern, and Trispot darter); two of these species (Atlantic
sturgeon and Trispot darter) inhabit the same reach that is expected to experience both improvement and
degradation, depending on the pollutants considered. Specifically, this reach meets NRWQC in the baseline
for selenium and shows degradation under all four regulatory options. However, the reach does not meet
NRWQC in the baseline for cadmium and shows improvements under the regulatory options for Options 2, 3,
and 4. The remaining three species have known critical habitat that overlaps surface waters which may be
affected by the options differently based on the state in which they reside (i.e.. the Clubshell critical habitat
overlaps surface waters which would experience water quality improvement under Option 4 in Pennsylvania
but would experience degradation under Option 4 in Ohio). The actual effect of the regulatory options on
these surface waters would depend on the effects of improvements in ambient concentrations of some
pollutants outweigh the effects of water quality degradation associated with other pollutants.
Table 7-3: T&E Species Whose Habitat May be Adversely Affected by the Regulatory Options
Number of Reaches with Ambient Water Quality Criteria
Species Common Name
State(s)
Exceedances for at Least One Pollutant
Baseline
Option 1
Option 2
Option 3
Option 4
Alabama lampmussel
TN
0
0
0
1
0
Anthony's riversnail
TN
0
0
0
1
0
Atlantic sturgeon
GA
0
3
3
3
3
Birdwing pearlymussel
TN
0
0
0
1
0
Bog turtle
NY
0
2
2
0
0
EPA-821-R-19-011
7-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 7: Threatened & Endangered Species Benefits
Table 7-3: T&E Species Whose Habitat May be Adversely Affected by the Regulatory Options
Number of Reaches with Ambient Water Quality Criteria
Species Common Name
State(s)
Exceedances for at Least One Pollutant
Baseline
Option 1
Option 2
Option 3
Option 4
Clubshell
OH
0
1
0
0
0
Cracking pearlymussel
KY, OH, TN, WV
0
3
0
1
0
Dromedary pearlymussel
TN
0
0
0
1
0
Fanshell
KY, OH, TN, WV
0
7
4
5
4
Finerayed pigtoe
TN
0
0
0
1
0
Hellbender salamander
KY
0
1
1
1
1
Least tern
KY
0
1
1
1
1
Orangefoot pimpleback
IL, KY, OH, TN, WV
0
3
1
2
1
Pink mucket
OH, TN
0
1
0
1
0
Purple cat's paw
KY
0
6
5
5
5
Ring pink
IL, KY, OH, TN, WV
0
4
1
2
1
Rough pigtoe
KY, OH, TN, WV
0
7
4
5
4
Sheepnose mussel
TN
0
0
0
1
0
Shiny pigtoe
TN
0
0
0
1
0
Shortnose sturgeon
SC
0
1
1
1
1
Trispot darter
GA, TN
0
3
3
4
3
White cat's paw
KY
0
1
1
1
1
White wartyback
OH, TN
0
1
0
1
0
Total number of reaches with NRWQC exceedances
0
45
27
40
25
a. Species Latin names are listed in Table 7-1.
Source: U.S. EPA Analysis, 2019.
7.4 Limitations and Uncertainties
The main limitation of the EPA's analysis of the regulatory options" impacts on T&E species habitat is the
lack of data necessary to quantitively estimate population changes of T&E species and to monetize these
effects. First, data required to estimate the response of T&E populations to improved habitats are rarely
available. Second, the contribution of T&E species to ecosystem stability, ecosystem function, and life history
remains relatively unknown. Third, there is a paucity of economic data focused on the benefits of preserving
habitat for T&E species because nonuse values comprise the principal source of benefit estimates for most
T&E species. Additional caveats, omissions, biases, and uncertainties known to affect the EPA's assessment
of ELG's impacts on T&E species are summarized in Table 7-4.
Table 7-4: Limitations and Uncertainties in the Analysis of T&E Species Benefits
Issue
Effect on Benefits
Estimate
Notes
Change in T&E populations
due to the effect of revised
ELGs is uncertain
Uncertain
Data necessary to quantitatively estimate population changes
are unavailable. Therefore, the EPA used the methodology
described in Section 7.3.1 to assess whether the regulatory
options is likely to contribute to recovery of T&E populations.
Only those T&E species listed
as threatened or endangered
on the Endangered Species
Act are included in the
analysis
Underestimate
The databases used to estimate benefits to T&E species
exclude all species considered threatened or endangered by
scientific organizations but not protected by the ESA. The
magnitude of the underestimate is likely to be significant, since
the proportion of imperiled fish and mussel species is high
(e.g., Jelks et al 2008, Taylor et al 2007).
EPA-821-R-19-011
7-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 7: Threatened & Endangered Species Benefits
Table 7-4: Limitations and Uncertainties in the Analysis of T&E Species Benefits
Issue
Effect on Benefits
Estimate
Notes
Lack of available and/or
precise spatial data for T&E
habitats
Uncertain
For several freshwater T&E species, geographic ranges were
available only as 8-digit HUCs. Because of this, the exact
location of T&E habitats was estimated based on
correspondence with the geographic range 8-digit HUC.
EPA-821-R-19-011
7-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
8 Air-Related Benefits
The regulatory options may affect air quality through three main mechanisms: 1) changes in energy use by
steam electric power plants to operate wastewater treatment, ash handling, and other systems needed to
comply with the regulatory options; 2) transportation-related emissions due to the changes in trucking of coal
combustion residuals and other waste to on-site or off-site landfills; and 3) electricity generation profile
changes due to changes in the cost to generate electricity at plants incurring compliance costs for the
regulatory options. The different profile of generation can result in lower or higher air pollutant emissions due
to differences in emission factors. Thus, small increases in coal-based electricity generation as a result of the
regulatory options are compensated by reductions in generation using other fuels or energy sources - natural
gas, nuclear, solar, wind, hydro, and biomass. For example, as detailed in Chapter 5 of the RIA (U.S. EPA,
2019c), the Integrated Planning Model (IPM) projects a 0.7 percent increase in electricity generation from
coal (6,278 GWh), under Option 2; this increase is partially offset by a 0.2 percent decline in natural gas
generation (3,171 GWh) and additional declines in electricity generation from nuclear, wind, and hydro
sources.66 The changes in air emissions reflect the differences in emissions factors for these other fuels or
sources of energy, as compared to coal.
In this analysis, which follows the same general methodology the EPA used in the analysis of the 2015 rule
(U.S. EPA, 2015a), the EPA estimated the human health and other benefits resulting from net changes in
emissions of three pollutants: NOx, SO2, and CO2.
NOx and SOx (which include SO2 emissions quantified in this analysis) are known precursors to fine particles
(PM2 5) air pollution, a criteria air pollutant that has been associated with a variety of adverse health effects —
most notably, premature mortality.67 In addition, in the presence of sunlight, NOx and volatile organic
compounds (VOCs) can undergo a chemical reaction in the atmosphere to form ozone. Depending on
localized concentrations of VOCs, reducing NOx emissions would also reduce human exposure to ozone and
the incidence of ozone-related health effects. Reducing emissions of SO2 and NOx would also reduce ambient
exposure to SO2 and NO2, respectively. For the purpose of this analysis, the EPA quantified only those health
effects and associated benefits from associated reductions PM2.5.68
CO2 is the most prevalent of the greenhouse gases, which are air pollutants that the EPA has determined
endangers public health and welfare through their contribution to climate change. The EPA used estimates of
the domestic social cost of carbon (SC-CO2) to monetize the benefits of changes in CO2 emissions as a result
of this proposal The SC-CO2 is a metric that estimates the monetary value of projected impacts associated
with marginal changes in CO2 emissions in a given year. It includes a wide range of anticipated climate
66 These projections are based on the IPM sensitivity analysis scenario that includes the Affordable Clean Energy (ACE) rule in the
baseline.
67 Sulfur oxides (SOx) include sulfur monoxide (SO), sulfur dioxide (SO2), sulfur trioxide (SO3) and other sulfur oxides. In this
analysis, the EPA analyzed changes in emissions of SO2 only.
68 The Integrated Science Assessment for Particulate Matter (PM ISA) (U.S. EPA, 2009b) identified the human health effects
associated with ambient PM2.5 exposure, which include premature morality and a variety of morbidity effects associated with
acute and chronic exposures. Similarly, the Integrated Science Assessment for Ozone and Related Photochemical Oxidants
(Ozone ISA) (U.S. EPA, 2013b) identified the human health effects associated with ambient ozone exposure, which include
premature morality and a variety of morbidity effects associated with acute and chronic exposures.
EPA-821-R-19-011
8-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
impacts, such as net changes in agricultural productivity and human health, property damage from increased
flood risk, and changes in energy system costs, such as reduced costs for heating and increased costs for air
conditioning.
8.1 Data and Methodology
8.1.1 Changes in Air Emissions
As discussed in the RIA (Chapter 5: Electricity Market Analyses), the EPA used IPM to estimate the
electricity market-level effects of two of the four regulatory options (Options 2 and 4; see Chapter 5 in RIA
(U.S. EPA, 2019c)). IPM outputs include NOx, SO2, and CO2 emissions to air from electricity generating
units (EGU). Comparing these emissions to those projected for the baseline scenario provides an assessment
of the changes in air emissions resulting from changes in the profile of electricity generation under the
regulatory options. The EPA used seven run years, shown in Table 8-1, to represent the 2021-2047 period of
analysis (for a more detailed discussion of the IPM analysis, refer to Chapter 5 in RIA).
Table 8-1: IPM Run Years
Run Year
Years Represented
2021
2021
2023
2022-2023
2025
2024-2027
2030
2028-2032
2035
2033-2037
2040
2038-2042
2045
2043-2047
Source: U.S. EPA, 2018f
The EPA used the IPM sensitivity scenario that includes the ACE rule in the baseline (IPM-ACE) as the basis
for estimating changes in emissions for proposed Option 2 and the IPM analysis scenario that does not
include the ACE rule in the baseline as the basis for estimating changes in emissions for proposed Option 4.
As part of its analysis of non-water quality environmental impacts, the EPA developed separate estimates of
changes in energy requirements for operating wastewater treatment systems and ash handling systems, and
changes in transportation needed to landfill solid waste and combustion residuals (see Supplemental TDD for
details; U.S. EPA, 2019b). The EPA estimated NOx, SO2, and CO2 emissions associated with changes in
energy requirements to power wastewater treatment systems by multiplying plant-specific changes in
electricity consumption by plant- or North American Electric Reliability Corporation (NERC)-specific
emission factors obtained from IPM for each analysis year. The EPA estimated air emissions associated with
changes in transportation by multiplying the number of miles by average emission factors.
Table 8-2 and Table 8-3 summarize the estimated changes in emissions for the three mechanisms, the three
pollutants, and the two regulatory options covered in this analysis. As shown in the tables, the EPA estimates
that changes in power requirements and transportation (Table 8-2) would result in a decrease in emissions
under Option 2 and an increase in emissions under Option 4. These values reflect full compliance with the
EPA-821-R-19-011
8-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
regulatory options, which is projected to occur no later than by the end of 2028. For the purpose of this
analysis, however, the EPA used these same values for the years 2021-2028.69
As shown in Table 8-3, projected changes in the profile of electricity generation generally lead to increased
CO2, SO2 and NOx emissions, with the exception of a decline in CO2 and NOx emissions during 2033-2042
under Option 4. Table 8-4 presents the net emissions changes across the three mechanisms.
The largest effect on projected air emissions is due to the change in the emissions profile of electricity
generation at the market level. As presented in the RIA (U.S. EPA, 2019c; see Section 5.2), IPM projects
increases in electricity generation coming from coal as a result of the either of the two regulatory options
analyzed (about 0.6 percent for Option 2; 0.2 percent for Option 4), while decreases are projected for
generation from other fuels or energy sources - natural gas and renewables, including biomass, wind, and
solar. The changes in air emissions reflect the differences in emissions factors for these other fuels, as
compared to coal.
Table 8-2: Estimated Changes in Air Pollutant Emissions due to Increase in Power Requirements and
Trucking at Steam Electric Power Plants 2021-2047, Relative to Baseline
Source
C02 (Metric
Tonnes/Year)3
NOx (Tons/Year)3
S02 (Tons/Year)3
Option 2
Power requirements'5
-44,084.2
-32.2
-54.3
Trucking
-490.0
-0.5
0.0
Option 4
Power requirements0
59,320.0
31.3
20.4
Trucking
1,440.0
1.4
0.0
a. Negative values indicate a reduction in emissions; positive values indicate increased emissions
b. Estimates are based on the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE).
c. Estimates are based on IPM analysis scenario that does not include the ACE rule in the baseline.
Source: U.S. EPA Analysis, 2019
Table 8-3: Estimated Changes in Annual Air Pollutant Emissions due to Changes in Electricity
Generation Profile, Relative to Baseline
Regulatory
Option
Year
C02 (Metric
Tonnes/Year)3
NOx (Tons/Year)3
S02 (Tons/Year)3
Option 2b
2021
502,249
2,932
1,701
2022-2023
2,724,221
5,081
3,659
2024-2027
3,516,021
6,083
6,654
2028-2032
5,655,615
4,654
4,928
2033-2037
4,530,351
3,214
3,846
2038-2042
3,739,662
2,190
3,417
2043-2047
3,256,567
2,172
2,063
09 The EPA estimated transportation emissions for the aggregate industry to avoid the disclosure of Confidential Business
Information, preventing precise allocation of these emissions to individual years within the period of 2021 to 2028. These
emissions are small when compared to emissions from changes in the electricity generation profile. Assuming that the changes
occur in the first year of the analysis (2021) does not materially affect benefit estimates.
EPA-821-R-19-011
8-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
Table 8-3: Estimated Changes in Annual Air Pollutant Emissions due to Changes in Electricity
Generation Profile, Relative to Baseline
Regulatory
Option
Year
C02 (Metric
Tonnes/Year)3
NOx (Tons/Year)3
S02 (Tons/Year)3
Option 4C
2021
773,369
3,055
5,424
2022-2023
2,184,188
4,241
4,699
2024-2027
296,007
2,928
1,748
2028-2032
1,183,400
999
1,870
2033-2037
-685,296
-276
1,204
2038-2042
-353,497
-143
3,679
2043-2047
1,354,002
1,807
3,290
a. Negative values indicate a reduction in emissions; positive values increase increased emissions.
b. Estimates are based on the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE).
c. Estimates are based on IPM analysis scenario that does not include the ACE rule in the baseline.
Source: U.S. EPA Analysis, 2019; See Chapter 5 in RIAfor details on IPM (U.S. EPA, 2019c).
Table 8-4: Estimated Net Changes in Air Pollutant Emissions due to Changes in Power
Requirements, Trucking, and Electricity Generation Profile, Relative to Baseline
Regulatory
Option
Year
C02 (Metric Tonnes/Year)3
NOx (Tons/Year)3
S02 (Tons/Year)3
Option 2b
2021
457,675
2,899
1,646
2022-2023
2,679,647
5,048
3,605
2024-2027
3,471,447
6,050
6,599
2028-2032
5,611,041
4,621
4,874
2033-2037
4,485,777
3,181
3,791
2038-2042
3,695,087
2,157
3,362
2043-2047
3,211,993
2,140
2,009
Option 4C
2021
834,033
3,088
5,445
2022-2023
2,244,852
4,274
4,720
2024-2027
356,671
2,961
1,768
2028-2032
1,244,064
1,032
1,890
2033-2037
-624,632
-244
1,224
2038-2042
-292,833
-110
3,699
2043-2047
1,414,666
1,840
3,310
a. Negative values indicate a reduction in emissions; positive values increase increased emissions.
b. Estimates are based on the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE).
c. Estimates are based on IPM analysis scenario that does not include the ACE rule in the baseline.
Source: U.S. EPA Analysis, 2019
8.1.2 N0xandS02
NOx and SO2 are known precursors to PM2.5. Several adverse health effects have been associated with PM2.5,
including premature mortality, non-fatal heart attacks, hospital admissions, emergency department visits,
upper and lower respiratory symptoms, acute bronchitis, aggravated asthma, lost work days and acute
respiratory symptoms. For the analysis of the 2015 rule, the EPA relied on estimates of national monetized
benefits per ton of emissions avoided, which represented the total monetized human health benefits from
changes in the adverse outcomes mentioned above (U.S. EPA, 2015a)). Table 8-5 presents EPA's estimates of
benefits perton for NOx and SO2 forthe years 2016, 2020, 2025, and 2030. The estimates vary based on the
EPA-821-R-19-011
8-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
epidemiology study used as the basis for premature mortality estimates (Krewski et al. (2009) or Lepeule et
al. (2012)), the discount rate (3 percent or 7 percent), and the emissions source (EGUs or on-road mobile
sources).7"
Table 8-5: National Benefits per Ton Estimates for NOx and SO2 Emissions (2018$/ton) from the
Benefits per Ton Analysis Reported by U.S. EPA (2018d)
Discount
Rate
Year
EGUa
Mobile Source (On-road)a
Krewski et al. (2009)
Lepeule et al. (2012)
Krewski et al. (2009)
Lepeule et al. (2012)
S02
NOx
S02
NOx
S02
NOx
S02
NOx
3 percent
2016
$41,718
$6,258
$95,950
$14,601
$21,902
$8,656
$50,061
$19,816
2020
$43,803
$6,466
$100,122
$14,601
$23,988
$9,074
$54,233
$20,859
2025
$47,975
$6,988
$104,294
$15,644
$26,073
$9,804
$59,448
$21,902
2030
$51,104
$7,509
$114,723
$16,687
$29,202
$10,429
$66,748
$23,988
7 percent
2016
$37,546
$5,632
$86,564
$12,515
$19,816
$7,822
$44,846
$17,730
2020
$39,632
$5,840
$89,693
$13,558
$21,902
$8,135
$49,018
$18,773
2025
$42,761
$6,258
$96,993
$14,601
$23,988
$8,865
$54,233
$19,816
2030
$46,932
$6,779
$104,294
$15,644
$26,073
$9,595
$59,448
$21,902
a. Estimation of benefits per ton for 2016, 2020, 2025, and 2030 were based on year 2016 emissions modeling. Values were
updated from 2015 dollars to 2018 dollars using the GDP deflator (1.043).
Source: U.S. EPA Analysis, 2019 based on U.S. EPA (2018d)
For this proposed rule, the Agency quantified, but did not monetize, changes in emissions of PM25 precursors
NOx and SO2. To map those emission changes to air quality changes across the country, full scale air quality
modeling is needed. Prior to this proposal, the EPA's modeling capacity was fully allocated to supporting
other regulatory and policy efforts and as a result we did not do an air quality impact assessment and quantify
the air disbenefits of this proposal, were it to become a final regulation. Full scale air quality modeling
provides spatially explicit estimates of concentration changes, which is required for characterizing uncertainty
in mortality risk from changes in PM25 concentrations at different levels of baseline PM25 exposure. If the
EPA estimates PM2 5 concentration changes and monetizes these effects for the final rule, it will do so
consistent with methods current at that time.
8.1.3 C02
The EPA estimated the monetary value of CO2 emission changes using a measure of the domestic social cost
of carbon (SC-CO2). The SC-CO2 is a metric that estimates the monetary value of projected impacts
associated with marginal changes in CO2 emissions in a given year. It includes a wide range of anticipated
climate impacts, such as net changes in agricultural productivity and human health, property damage from
increased flood risk, and changes in energy system costs, such as reduced costs for heating and increased
costs for air conditioning. It is used to assess the avoided damages as a result of regulatory actions {i.e.,
benefits of rulemakings that lead to an incremental reduction in cumulative global CO2 emissions). The SC-
70 While all of the health effects enumerated in the paragraph above are included in the estimation of benefits per ton values in U.S.
EPA (2018d, a very large percentage, 98 percent, of the total monetized benefits of changes in PM2.5 concentrations are
attributable to premature mortality. U.S. EPA (2018d) presents two benefit per ton estimates for valuing changes in premature
mortality based on the change in the incidence of premature mortality associated with a given change in exposure to PM2.5:
Krewski et al. (2009) and Lepeule et al. (2012).
EPA-821-R-19-011
8-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
CO2 estimates used in this analysis focus on the projected impacts of climate change that are anticipated to
directly occur within U.S. borders.
The SC-CO2 estimates used in this analysis are interim values developed under E.O. 13783 for use in
regulatory analyses until an improved estimate of the impacts of climate change to the U.S. can be developed
based on the best available science and economics. E.O. 13783 directed agencies to ensure that estimates of
the social cost of greenhouse gases used in regulatory analyses "are based on the best available science and
economics" and are consistent with the guidance contained in OMB Circular A-4, "including with respect to
the consideration of domestic versus international impacts and the consideration of appropriate discount rates"
(E.O. 13783, Section 5(c)). In addition, E.O. 13783 withdrew the technical support documents (TSDs) used in
the benefits analysis of the 2015 ELG for describing the global social cost of greenhouse gas estimates
developed under the prior Administration as no longer representative of government policy.
Regarding the two analytical considerations highlighted in E.O. 13783 - how best to consider domestic versus
international impacts and appropriate discount rates - current guidance in OMB Circular A-4 is as follows.
Circular A-4 states that analysis of economically significant proposed and final regulations "should focus on
benefits and costs that accrue to citizens and residents of the United States." (OMB, 2003) We follow this
guidance by adopting a domestic perspective in our central analysis. Regarding discount rates, Circular A-4
states that regulatory analyses "should provide estimates of net benefits using both 3 percent and 7 percent."
(OMB, 2003) The 7 percent rate is intended to represent the average before-tax rate of return to private capital
in the U.S. economy. The 3 percent rate is intended to reflect the rate at which society discounts future
consumption, which is particularly relevant if a regulation is expected to affect private consumption directly.
The EPA follows this guidance below by presenting estimates based on both 3 and 7 percent discount rates in
the main analysis. See Appendix I for a discussion the modeling steps involved in estimating the domestic SC-
CO2 estimates based on these discount rates. These SC-CO2 estimates developed under E.O. 13783 presented
below will be used in regulatory analysis until more comprehensive domestic estimates can be developed,
which would take into consideration recent recommendations from the National Academies of Sciences,
Engineering, and Medicine (2017) to further update the current methodology to ensure that the SC-CO2
estimates reflect the best available science.71
Table 8-6 presents the average domestic SC-CO2 estimate across all of the integrated assessment model runs
used to estimate the SC-CO2 for each discount rate for the years 2015 to 2050.72 As with the global SC-CO2
estimates, the domestic SC-CO2 increases overtime because future emissions are expected to produce larger
incremental damages as economies grow and physical and economic systems become more stressed in
response to greater climate change.
The EPA estimates the dollar value of the C02-related effects for each analysis year between 2021 and 2047
by applying the SC-CO2 estimates, shown in Table 8-6, to the estimated changes in CO2 emissions in the
71 See National Academies of Sciences, Engineering, and Medicine, Valuing Climate Damages: Updating Estimation of the Social
Cost of Carbon Dioxide, Washington, D.C., January 2017. http://www.nap.edu/catalog/24651/valuingclimate-changes-updating-
estimation-of-the-social-cost-of
72 The SC-CO2 estimates rely on an ensemble of three integrated assessment models (IAMs): Dynamic Integrated Climate and
Economy (DICE) 2010; Climate Framework for Uncertainty, Negotiation, and Distribution (FUND) 3.8; and Policy Analysis of
the Greenhouse Gas Effect (PAGE) 2009.
EPA-821-R-19-011
8-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
corresponding year under the regulatory options. The EPA then calculates the present value and annualized
benefits from the perspective of 2020 by discounting each year-specific values to the year 2020 using the
same 3 percent and 7 percent discount rates.
Table 8-6: Interim Domestic Social Cost of Carbon Values (2018$/metric tonne CO2)
Year
3% Discount Rate, Average
7% Discount Rate, Average
2015
$6
$1
2020
$7
$1
2025
$7
$1
2030
$8
$1
2035
$9
$2
2040
$10
$2
2045
$10
$2
2050
$11
$2
Note: These SC-C02 values are stated in $/metric ton C02 and rounded to the nearest dollar. The estimates vary
depending on the year of C02 emissions and are defined in real terms, i.e., adjusted for inflation using the GDP
implicit price deflator. Values updated from 2016 dollars to 2018 dollars using GDP deflator (1.030). The EPA
interpolated annual values for intermediate years.
Source: U.S. EPA Analysis, 2019 based on U.S. EPA (2019f)
The limitations and uncertainties associated with the SC-CO2 analysis, which were discussed in the 2015
BCA (U.S. EPA, 2015a), likewise apply to the domestic SC-CO2 estimates presented in this Chapter. Some
uncertainties are captured within the analysis, as discussed in detail in Appendix I, while other areas of
uncertainty have not yet been quantified in a way that can be modeled. For example, limitations include the
incomplete way in which the integrated assessment models capture catastrophic and non-catastrophic impacts,
their incomplete treatment of adaptation and technological change, the incomplete way in which inter-
regional and intersectoral linkages are modeled, uncertainty in the extrapolation of damages to high
temperatures, and inadequate representation of the relationship between the discount rate and uncertainty in
economic growth over long time horizons. The science incorporated into these models understandably lags
behind the most recent research, and the limited amount of research linking climate impacts to economic
damages makes this comprehensive global modeling exercise even more difficult. These individual
limitations and uncertainties do not all work in the same direction in terms of their influence on the SC-CO2
estimates. In accordance with guidance in OMB Circular A-4 on the treatment of uncertainty, Appendix I
provides a detailed discussion of the ways in which the modeling underlying the development of the SC-CO2
estimates used in this RIA addressed quantified sources of uncertainty and presents a sensitivity analysis to
show consideration of the uncertainty surrounding discount rates over long time horizons.
Recognizing the limitations and uncertainties associated with estimating the SC-CO2, the research community
has continued to explore opportunities to improve SC-CO2 estimates. Notably, the National Academies of
Sciences, Engineering, and Medicine conducted a multidiscipline, multi-year assessment to examine potential
approaches, along with their relative merits and challenges, for a comprehensive update to the current
methodology. The task was to ensure that the SC-CO2 estimates that are used in Federal analyses reflect the
best available science, focusing on issues related to the choice of models and damage functions, climate
science modeling assumptions, socioeconomic and emissions scenarios, presentation of uncertainty, and
discounting. In January 2017, the Academies released their final report, "Assessing Approaches to Updating
the Social Cost of Carbon," and recommended specific criteria for future updates to the SC-CO2 estimates, a
EPA-821-R-19-011
8-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
modeling framework to satisfy the specified criteria, and both near-term updates and longer-term research
needs pertaining to various components of the estimation process (National Academies, 2017).
The Academies" 2017 report also discussed the challenges in developing domestic SC-CO2 estimates, noting
that current integrated assessment models do not model all relevant regional interactions - i.e., how climate
change impacts in other regions of the world could affect the United States, through pathways such as global
migration, economic destabilization, and political destabilization. The Academies concluded that it "is
important to consider what constitutes a domestic impact in the case of a global pollutant that could have
international implications that impact the United States. More thoroughly estimating a domestic SC-CO2
would therefore need to consider the potential implications of climate impacts on, and actions by, other
countries, which also have impacts on the United States." (National Academies, 2017, pg. 12-13). In addition
to requiring reporting of impacts at a domestic level, Circular A-4 states that when an agency ""cvaluatc|s| a
regulation that is likely to have effects beyond the borders of the United States, these effects should be
reported separately" (OMB, 2003; page 15). This guidance is relevant to the valuation of damages from CO2
and other greenhouse gases (GHGs), given that GHGs contribute to damages around the world independent of
the country in which they are emitted. Therefore, in accordance with this guidance in OMB Circular A-4,
Appendix I presents the global climate benefits from this proposed rulemaking using global SC-CO2 estimates
based on both 3 and 7 percent discount rates. The EPA did not quantitatively project the full impact of the
ELG on international trade and the location of production, so it is not possible to present analogous estimates
of international costs resulting from the regulatory options. However, to the extent that the electricity market
analysis endogenously models international electricity and natural gas trade (see Chapter 5 in RIA: U.S. EPA,
2019c), and to the extent that affected firms have some foreign ownership, some of the costs accruing to
entities outside U.S. borders is captured in the compliance costs presented in the RIA (U.S. EPA, 2019c).
8.2 Results
Table 8-7 shows the estimated monetary value of the estimated changes in CO2 emissions in each of several
selected years for the two regulatory options the EPA analyzed. Negative values indicate forgone benefits of
the proposed regulatory option as compared to the baseline.
Table 8-7: Estimated Domestic Climate Benefits from Changes in CO2 Emissions for Selected Years
(millions; 2018$)
Regulatory Option
Year
3% Discount Rate
7% Discount Rate
Option 2a
2021
-$3.0
-$0.4
2025
-$22.2
-$3.0
2030
-$33.7
-$4.0
2035
-$25.3
-$2.7
2040
-$19.4
-$1.8
2045
-$15.7
-$1.3
2047
-$15.2
-$1.2
EPA-821-R-19-011
8-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
Table 8-7: Estimated Domestic Climate Benefits from Changes in CO2 Emissions for Selected Years
(millions; 2018$)
Regulatory Option
Year
3% Discount Rate
7% Discount Rate
Option 4b
2021
-$5.6
-$0.8
2025
-$2.3
-$0.3
2030
-$7.5
-$0.9
2035
$3.5
$0.4
2040
$1.5
$0.1
2045
-$6.9
-$0.6
2047
-$6.7
1A
O
In
a. Estimates are based on the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE).
b. Estimates are based on IPM analysis scenario that does not include the ACE rule in the baseline.
Source: U.S. EPA Analysis, 2019
Table 8-8 shows the total annualized monetary values associated with changes in CO2 emissions for the two
regulatory options the EPA analyzed and by category of emissions. The EPA annualized monetary value
estimates to enable consistent reporting across benefit categories (e.g., benefits from improvement in water
quality). All monetary values are negative, indicating that the regulatory options result in forgone benefits
when compared to the baseline. The annualized values for Options 2 are -$31.6 million and -$5.2 million,
using discount rates of 3 and 7 percent, respectively. For Option 4, the estimated benefits are -$4.8 million
and -$0.9 million, using discount rates of 3 and 7 percent, respectively. The vast majority of the forgone
benefits arise from changes in the profile of electricity generation.
Table 8-8: Estimated Total Annualized Domestic Climate Benefits from Changes in CO2 Emissions
(Millions; 2018$)
Regulatory Option
Category of Air Emissions
3% Discount Rate
7% Discount Rate
Electricity Generation
-$32.0
-$5.2
Option 2a
Trucking
$0.0
$0.0
Energy use
$0.4
$0.1
Total
-$31.6
-$5.2
Electricity Generation
-$4.3
00
O
vv
Option 4b
Trucking
$0.0
$0.0
Energy use
1A
O
In
-$0.1
Total
-$4.8
-$0.9
a. Estimates are based on the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE).
b. Estimates are based on IPM analysis scenario that does not include the ACE rule in the baseline.
Source: U.S. EPA Analysis, 2019
The EPA did not analyze domestic climate benefits for Options 1 and 3 but extrapolated values to enable
comparison of total benefits of the four options. To estimate domestic climate benefits for Options 1 and 3,
the EPA scaled Option 2 benefits in proportion to the social costs of the respective options (see Section 12.2)
since changes in the profile of electricity generation accounts for the majority of changes in air emissions and
this generation profile is affected most directly by the incremental compliance costs. Specifically, the EPA
EPA-821-R-19-011
8-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
calculated the ratio of the domestic climate benefits to total social costs for Option 2,73 then multiplied total
social costs for Options 1 and 3 by this ratio. Table 8-9 presents extrapolated annualized benefits for changes
in air emissions for Options 1 and 3. Extrapolated domestic climate benefits are -$30.3 million
to -$4.8 million for Option 1 and from -$20.9 million to -$3.7 million for Option 3, depending on the discount
rate.
Table 8-9: Extrapolated Annualized Domestic Climate Benefits from Changes in CO2 Emissions
(Millions; 2018$)
Regulatory Option3'b
3% Discount Rate
7% Discount Rate
Option 1
-$30.3
00
^t"
vv
Option 3
-$20.9
-$3.7
a. The EPA estimated air-related benefits for Options 1 and 3 by multiplying the total social costs for each option (see Table 12-1)
by the ratio of [air-related benefits / total social costs] for Option 2.
b. Results are based on the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE).
Source: U.S. EPA Analysis, 2019
As discussed above, the EPA used different baselines for estimating the air-related benefits of proposed
Options 2 and 4. Estimates for proposed Option 2 are based on an IPM sensitivity scenario that includes the
ACE rule in the baseline (IPM-ACE), whereas proposed Option 4 estimates are based on an IPM scenario that
does not include the ACE rule in the baseline. The EPA did not extrapolate air-related benefits estimated for
Option 2 using the IPM-ACE scenario outputs to corresponding values for Option 4 due to anticipated
inaccuracies such extrapolation would introduce in benefit estimates for Option 4 (e.g., suggesting positive
benefits where they may be negative). As discussed in the preamble for the proposed rule, the EPA solicits
comments on the significance of using two different IPM baselines for estimating benefits of Options 1
through 3 (with the ACE rule) and Option 4 (without the ACE rule) and intends to include the ACE rule in the
baseline for IPM analyses for the final rulemaking.
8.3 Limitations and Uncertainties
Table 8-10 summarizes the limitations and uncertainties associated with the analysis of the air-related
benefits. The effect on benefits estimates indicated in the second column of the table refers to the magnitude
of the benefits rather than the direction (i.e.. a source of uncertainty that tends to underestimate benefits
indicates expectation for larger forgone benefits). The analysis also inherits uncertainties associated with IPM
modeling, which are discussed in Chapter 5 in the RIA (U.S. EPA, 2019c).
73 The ratios are 0.23 for the 3 percent discount rate estimates, and 0.03 for the 7 percent discount rate estimates.
EPA-821-R-19-011
8-10
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
8: Air-Related Benefits
Table 8-10: Limitations and Uncertainties in Analysis of Air-related Benefits
Issue
Effect on Benefits
Estimate
Notes
Domestic SC-C02 estimates
do not capture the full range
of impacts from climate
change
Underestimate
Current integrated assessment models (lAMs) used in
developing the SC-C02 do not model all relevant regional
interactions - i.e., how climate change impacts in other
regions of the world could affect the United States, through
pathways such as global migration, economic
destabilization, and political destabilization.
The EPA did not monetize
air-related benefits of
changes in NOx, S02, and
other pollutants emitted by
electricity generating units
Underestimate
NOx and S02are precursors to PM25, which causes a variety
of adverse health effects including premature death, non-
fatal heart attacks, hospital admissions, emergency
department visits, upper and lower respiratory symptoms,
acute bronchitis, aggravated asthma, lost work days, and
acute respiratory symptoms.
There are additional direct benefits from changes in levels of
NOx, S02 and other air pollutants emitted by electricity
generating units. As described in U.S. EPA (2019f), these
include health benefits from changes in ambient N02 and
S02 exposure, health benefits from changes in mercury
deposition, ecosystem benefits associated with changes in
emissions of NOx, S02, PM, and mercury, and visibility
impairment.
The EPA used predicted
changes in air emissions
from an IPM run that does
not include the ACE rule in
the baseline to estimate air-
related benefits for
proposed Option 4
Uncertain
Effects of the ACE rule on predicted changes in emissions for
proposed Option 4 are unknown. The EPA assessed that the
air-related benefits for Option 4 estimated using the IPM
scenario without the ACE rule provide an approximate
representation of the benefits of this option.
EPA-821-R-19-011
8-11
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
9: Water Withdrawals
9 Changes in Water Withdrawals
Steam electric power plants use water for ash transport and for operating wet FGD scrubbers. The regulatory
options are estimated to change water withdrawal from surface waterbodies and aquifers by affecting sluicing
operations or incentives to recycle water within the plants. The change in water use depends on the regulatory
option, but are small compared to those estimated in 2015 (see U.S. EPA, 2015a).
Table 9-1 shows estimated changes in water withdrawals for each evaluated regulatory option.
Table 9-1: Industry-level Total Changes in Water Withdrawals
(Surface Water and Aquifers)
Regulatory Option
Change in water withdrawals
(billion gallons per year)
Option 1
1.2
Option 2
7.7
Option 3
0.2
Option 4
-3.4
Source: U.S. EPA Analysis, 2019
The sections below discuss the benefits resulting specifically from estimated changes in groundwater
withdrawals. Benefits associated with surface water withdrawals are discussed qualitatively in Chapter 2.
9.1 Methods
The analysis follows the same general methodology the EPA used in the analysis of the 2015 rule (U.S. EPA,
2015a). Changes in water withdrawal from groundwater sources by steam electric power plants may affect
availability of groundwater for local municipalities that rely on groundwater aquifers for drinking water
supplies. These municipalities may incur incremental costs for supplementing drinking water supplies through
alternative means, such as bulk drinking water purchases as water withdrawals by steam electric power plants
change. The EPA estimated the monetary value of changes in groundwater withdrawals based on costs of
purchasing drinking water during periods of shortages in groundwater supply.
9.2 Results
The EPA's analysis of the regulatory options indicates that one plant would increase the volume of
groundwater withdrawn under Option 2. No changes in groundwater withdrawal are estimated under Options
1, 3, and 4. See details in the Supplemental TDD (U.S. EPA, 2019b).
The EPA estimated that the plant would increase withdrawals by a total of 21,971 gallons per day (8 million
gallons per year) under Option 2. To estimate the value of reduced groundwater supply, the EPA used state-
specific prices of bulk drinking water supplies, based on the assumption that municipalities may need to
purchase supplementary supplies in response to any change in groundwater availability arising from
additional withdrawals. While this is an approximate assumption, the analysis provides screening-level
indication of the potential forgone benefits.
To estimate the monetary value of the specific reduced groundwater withdrawal due to the one facility's
increase in groundwater use, the EPA relied on the current state-specific drinking water prices of $1,192 per
EPA-821-R-19-011
9-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
9: Water Withdrawals
acre/foot for the affected location. The EPA multiplied the increase in groundwater withdrawal (in gallons per
year) by the estimated price of drinking water per gallon.74
Table 9-2 shows estimated annual forgone benefits from increased groundwater withdrawals under Option 2.
The annual forgone benefits from Option 2 are $0.02 million using a 3 percent discount rate ($0.02 million
using a 7 percent discount rate). As described above, there are no changes in groundwater withdrawals
associated with any of the other regulatory options and, therefore, no change in monetary benefits under those
options.
Table 9-2: Estimated Annualized Benefits from Increased Groundwater Withdrawals
(Millions; 2018$)
Regulatory Option
Increase in Groundwater
Intakes (million gallons
per year)
3% Discount Rate
7% Discount Rate
Option 1
0.0
$0.00
$0.00
Option 2
8.0
-$0.02
-$0.02
Option 3
0.0
$0.00
$0.00
Option 4
0.0
$0.00
$0.00
a Reflects changes after full compliance with requirements for Option 2 in 2023.
Source: U.S. EPA Analysis, 2019
9.3 Limitations and Uncertainties
Table 9-3 summarizes the limitations and uncertainties in the analysis of benefits associated with changes in
groundwater withdrawals.
Table 9-3: Limitations and Uncertainties in Analysis of Changes in Groundwater Withdrawals
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
The EPA assumed that municipalities
would need to replace lost
groundwater supplies with bulk
drinking water purchases.
Uncertain
See below.
Municipalities may not need to replace groundwater
withdrawn by steam electric power plants (in which
case the benefits of the ELG may be overstated), or
they may choose to replace the groundwater through
other means.
The EPA assumed a direct
relationship between groundwater
withdrawals in water-stressed states
and groundwater shortages, i.e.,
that reducing demand for limited
groundwater supplies would result
in avoided costs for purchased
water.
Overestimate
The EPA assumed that demand for additional water
supply exists in the affected areas due to potential
drought. However, the extent of this demand is
uncertain.
74 The EPA used a conversion factor of 325,851 to convert acre foot to gallons.
EPA-821-R-19-011
9-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
9: Water Withdrawals
Table 9-3: Limitations and Uncertainties in Analysis of Changes in Groundwater Withdrawals
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
The EPA estimated cost of bulk
water purchases based on state-
wide averages
Uncertain
Costs of water may vary within a state and assuming
the average value may result in under- or overstating
of the cost for any given location. This uncertainty is
more significant in cases where there are few affected
locations, as is the case for this analysis which shows
only one plant with changes in groundwater
withdrawals.
Data on the characteristics of
affected aquifers are not available
Uncertain
If the affected aquifers are used for private wells only,
the estimated benefits of improved groundwater
recharge could be under- or overstated, depending on
households WTP for protecting groundwater quantity.
EPA-821-R-19-011
9-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
10: Dredging
10 Estimated Changes in Dredging Costs
As summarized in Table 3-3, the regulatory options could result in small changes to total suspended solid
(TSS) discharged by steam electric power plants, which could have an impact on the rate of sediment
deposition to affected waterbodies, including navigable waterways and reservoirs that require dredging for
maintenance.
Navigable waterways, including rivers, lakes, bays, shipping channels and harbors, are an integral part of the
United States' transportation network. They are prone to reduced functionality due to sediment build-up,
which can reduce the navigable depth and width of the waterway (Clark et al., 1985). In many cases, costly
periodic dredging is necessary to keep them passable. The regulatory options could increase or reduce costs
for government and private entities responsible for maintenance of navigable waterways by changing the need
for dredging.
Reservoirs serve many functions, including storage of drinking and irrigation water supplies, flood control,
hydropower supply, and recreation. Streams and rivers carry sediment into reservoirs, where it can settle and
cause buildup of silt layers at a recorded average rate of 1.2 billion kilograms per reservoir every year (USGS,
2009. Sedimentation reduces reservoir capacity (Graf et al., 2010) and the useful life of reservoirs unless
measures such as dredging are taken to reclaim capacity (Clark et al., 1985).
The EPA estimated that the proposed regulatory option, Option 2, would have a small effect on dredging
costs when compared to those estimated in 2015 (see U.S. EPA, 2015a).
10.1 Methods
In this analysis, the EPA followed the same general methodology for estimating changes in costs associated
with changes in sediment depositions in navigational waterways and reservoirs that the EPA used in the 2015
rule (U.S. EPA 2015a; see Appendix K). The methodology utilizes information on historic dredging locations,
frequency of dredging, the amount of sediment removed, and dredging costs in conjunction with the estimated
changes in sediment deposition and removal in dredged waterways and reservoirs under the regulatory
options. Benefits are equal to avoided costs, calculated as the difference in total annualized dredging costs at
baseline and under each regulatory option. Negative values represent cost increases (i.e.. forgone benefits to
society).
10.1.1 Estimated Changes in Navigational Dredging Costs
The EPA identified 22 unique dredging jobs and 91 dredging occurrences75 within the affected reaches
equivalent to 0.8 percent of the dredging occurrences with coordinates reported in the Dredging Information
System (USACE, 2013). The recurrence interval for dredging jobs ranged from 1 to 15 years across all
affected reaches and averaged 9.6 years. Costs vary considerably across affected reaches, from approximately
$ 1.72 per cubic yard at Establishment Bar in North Carolina to $30.30 per cubic yard at Bonum Creek in
Virginia. The average unit cost of dredging for the entire conterminous United States is $6.00 per cubic yard.
75 Dredging jobs refer to unique sites/locations defined by USACE where dredging was conducted, whereas dredging occurrences are
unique instances when dredging was conducted and may include successive dredging at the same location.
EPA-821-R-19-011
10-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
10: Dredging
Table 10-1 presents estimates of baseline sediment dredging in navigational waterways that may be affected
by bottom ash transport water and FGD wastewater discharges from 2021 to 2047 and low, mean, and high
cost estimates. The EPA generated low, medium, and high estimates for navigational dredging by varying
assumptions for projected future dredging occurrence, including dredging frequency and job start as well as
cost of dredging for locations that did not report location specific costs (see U.S. EPA 2015a, Appendix K for
detail). Estimated total baseline navigational dredging costs range from $49.4 to $58.5 million per year, using
a 3 percent discount rate, and from $43.4 to $55.6 million using a 7 percent discount rate.
Table 10-1: Estimated Annualized Dredging Costs at Affected Reaches under the Baseline (Millions
of 2018$)
Total Sediment Dredged
(millions cubic yards)
Costs at 3% discount rate
(millions of 2018$ per year)
Costs at 7% discount rate
(millions of 2018$ per year)
Low
Mean
High
Low
Mean
High
Low
Mean
High
118.5
118.7
129.1
$49.4
$49.6
$58.5
$43.4
$43.5
$55.6
Source: U.S. EPA analysis, 2019.
The difference between the estimated dredging costs under the baseline and a particular regulatory option
represents the avoided costs (or forgone benefits) of that regulatory option. Table 10-2 presents estimated cost
changes for navigational dredging for the four regulatory options.
Table 10-2: Estimated Annualized Changes in Navigational Dredging Costs (Thousands of 2018$)
Regulatory
Option
Total Reduction in Sediment
3% discount rate
7% discount rate
Dredged
millions cubic yards)
(millions of 2018$ per year)a
(millions of 2018$ per year)a
Low
Mean
High
Low
Mean
High
Low
Mean
High
Option 1
-0.1
-0.1
-0.1
-$0.04
-$0.04
-$0.05
-$0.03
-$0.03
-$0.05
Option 2
-0.1
-0.1
-0.1
-$0.04
-$0.04
-$0.07
-$0.03
-$0.03
-$0.07
Option 3
-0.1
-0.1
-0.1
-$0.04
-$0.04
-$0.05
-$0.03
-$0.03
-$0.05
Option 4
1.1
1.1
1.3
$0.49
$0.49
$0.62
$0.42
$0.42
$0.60
a. Positive values represent cost savings; negative values represent cost increases.
Source: U.S. EPA analysis, 2019.
10.1.2 Estimated Changes in Reservoir Dredging Costs
The EPA identified 217 reservoirs within the affected reaches with changes in sediment loads under at least
one of the regulatory options, equivalent to 0.3 percent of the reservoirs included in the E2RF1 file (USGS,
2002). Table 10-3 presents the total amount of sediment that is estimated to be dredged in 2021 to 2047 from
these reservoirs, and the estimated annualized cost of dredging under the baseline scenario, including low,
mean, and high estimates. Estimated dredging costs for the reservoirs range between $460.4 million and
$624.8 million with a 3 percent discount rate and $385.4 million and $574.2 million with a 7 percent discount
rate under the baseline scenario.
Table 10-3: Estimated Annualized Reservoir Dredging Costs under Baseline (Millions 2018$)
Total Sediment Dredged
(millions cubic yards)
Costs at 3% Discount Rate (millions of
2018$ per year)
Costs at 7% Discount Rate (millions of
2018$ per year)
Low
Mean
High
Low
Mean
High
Low
Mean
High
2,090.2
2,508.2
2,717.2
$460.4
$557.2
$624.8
$385.4
$483.5
$574.2
Source: U.S. EPA analysis, 2019.
EPA-821-R-19-011
10-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
10: Dredging
The difference between the estimated dredging costs under the baseline and a particular regulatory option
represents the avoided costs of that regulatory option. Table 10-4 presents estimated cost changes for
reservoir dredging under the four regulatory options, including low, mean, and high estimates.
Table 10-4: Estimated Total Annualized Changes in Reservoir Dredging Costs (2018$)
Total Reduction in Sediment
Costs at 3% Discount Rate3
Costs at 7% Discount Rate3
Regulatory
Dredged (millions cubic yards)
(millions of 2018$ per year)
(millions of 2018$ per year)
Option
Low
Mean
High
Low
Mean
High
Low
Mean
High
Option 1
-0.1
-0.2
-0.2
-$0.03
-$0.04
-$0.04
-$0.02
-$0.03
-$0.04
Option 2
-0.5
-0.6
-0.6
-$0.10
-$0.13
-$0.14
-$0.09
-$0.11
-$0.13
Option 3
-0.1
-0.1
-0.1
-$0.02
-$0.02
-$0.02
-$0.01
-$0.02
-$0.02
Option 4
0.3
0.4
0.4
$0.08
$0.09
$0.10
$0.06
$0.08
$0.10
a. Positive values represent cost savings; negative values represent cost increases.
Source: U.S. EPA analysis, 2019.
10.2 Limitation and Uncertainty
Key uncertainties and limitations in the analysis of sediment dredging benefits are summarized in Table 10-5.
Detailed description is provided in Appendix K of the 2015 BCA document (U.S. EPA, 2015a). Note that the
effect on benefits estimates indicated in the second column of the table refers to the magnitude of the benefits
rather than the direction (i.e.. a source of uncertainty that tends to underestimate benefits indicates expectation
for larger forgone benefits). Uncertainties and limitations associated with SPARROW model estimates of
sediment deposition are discussed in U.S. EPA (2009a).
Table 10-5: Limitations and Uncertainties in Analysis of Changes in Dredging Costs
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
The analysis of navigational
waterways is restricted to jobs
reported in USACE Database for
1998 to 2012 (USACE, 2013).
Underestimate
Because some dredging jobs included in the USACE
Database lack latitude and longitude and the database
does not use standardized job names the EPA was only
able to map about 71 percent of all dredging
occurrences with records in the data. This may lead to
potential underestimation of baseline and changes in
dredging costs under the regulatory options.
The EPA's analysis for modeled
watersheds explicitly omits any
reservoirs that are not located on
the E2RF1 network.
Underestimate
The omission of other reservoirs would understate the
magnitude of estimated baseline and changes in
reservoir dredging benefits in cases where there are
additional reservoirs located downstream from steam
electric power plants that discharge bottom ash
transport water or FGD wastewater.
EPA-821-R-19-011
10-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
11: Total Monetized Benefits
11 Summary of Estimated Total Monetized Benefits
Table 11-1 and Table 11-2, on the next two pages, summarize the total annualized monetary value of social
welfare changes using 3 percent and 7 percent discount rates, respectively.
The monetary value of social welfare changes does not account for all effects of the regulatory options,
including changes in certain non-cancer health risk (e.g., effects of cadmium on kidney functions and bone
density), impacts of pollutant load changes on threatened and endangered species habitat, and ash marketing
changes. See Chapter 2 for a discussion of categories of social welfare effects the EPA did not monetize.
Chapter 4 through Chapter 10 provide more detail on the estimation methodologies for each benefit category.
EPA-821-R-19-011
11-1
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
11: Total Monetized Benefits
Table 11-1: Summary of Estimated Total Annualized Benefits at 3 Percent (Millions; 2018$)
Benefit Category
Option la
Option T
Option 3a
Option 4a
Low
Mid
High
Low
Mid
High
Low
Mid
High
Low
Mid
High
Human Health
-$0.7
$34.8
$39.7
$82.8
Changes in IQ losses in children from exposure to leadb
<$0.0
<$0.0
<$0.0
<$0.0
Changes in IQ losses in children from exposure to mercury
-$0.3
-$2.8
-$2.9
-$1.5
Changes in cancer risk from DBPs in drinking water
-$0.4
$37.6
$42.6
$84.3
Ecological Conditions and Recreational Uses Changes
-$10.0
-$12.5
-$55.5
$11.8
$16.7
$65.6
$16.3
$22.5
$90.7
$19.8
$27.3
$110.2
Use and nonuse values for water quality changes
-$10.0
-$12.5
-$55.5
$11.8
$16.7
$65.6
$16.3
$22.5
$90.7
$19.8
$27.3
$110.2
Market and Productivity
-$0.1
-$0.1
-$0.1
-$0.2
-$0.2
-$0.2
-$0.1
-$0.1
-$0.1
$0.6
$0.6
$0.7
Changes in dredging costs
-$0.1
-$0.1
-$0.1
-$0.1
-$0.2
-$0.2
-$0.1
-$0.1
-$0.1
$0.6
$0.6
$0.7
Reduced water withdrawals'5
$0.0
<$0.0
$0.0
$0.0
Air-related effects
-$30.3
-$31.6
-$20.9
-$4.8
Changes in C02 air emissions0
-$30.3
-$31.6
-$20.9
-$4.8
Totald
-$41.0
-$43.6
-$86.6
$14.8
$19.6
$68.5
$35.1
$41.3
$109.4
$98.4
$105.9
$188.9
a. Negative values represent forgone benefits and positive values represent realized benefits.
b. "<$0.0" indicates that monetary values are greater than -$0.1 million but less than $0.00 million.
c. The EPA estimated the air-related benefits for Option 2 using the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE). EPA extrapolated estimates
for Options 1 and 3 air-related benefits from the estimate for Option 2 that is based on IPM-ACE outputs. The values for Option 4 air-related benefits were estimated using the IPM
analysis scenario that does not include the ACE rule in the baseline. See Chapter 8 for details.
d. Values for individual benefit categories may not sum to the total due to independent rounding.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
11-2
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
11: Total Monetized Benefits
Table 11-2: Summary of Estimated Total Annualized Benefits at 7 Percent (Millions; 2018$)
Benefit Category
Option V
Option 2a
Option 3a
Option 4a
Low
Mid
High
Low
Mid
High
Low
Mid
High
Low
Mid
High
Human Health
-$0.3
$23.6
$26.9
$54.0
Changes in IQ losses in children from exposure to leadb
<$0.0
<$0.0
<$0.0
<$0.0
Changes in IQ losses in children from exposure to
mercuryb
-$0.1
-$0.6
-$0.6
-$0.3
Changes in cancer risk from DBPs in drinking water
-$0.2
$24.2
$27.5
$54.3
Ecological Conditions and Recreational Uses Changes
-$8.6
-$10.9
-$48.1
$10.1
$14.3
$56.1
$14.0
$19.4
$77.8
$17.0
$23.6
$94.6
Use and nonuse values for water quality Changes
-$8.6
-$10.9
-$48.1
$10.1
$14.3
$56.1
$14.0
$19.4
$77.8
$17.0
$23.6
$94.6
Market and Productivity
-$0.1
-$0.1
-$0.1
-$0.1
-$0.2
-$0.2
$0.0
-$0.1
-$0.1
$0.5
$0.5
$0.7
Changes in dredging costs
-$0.1
-$0.1
-$0.1
-$0.1
-$0.1
-$0.2
$0.0
-$0.1
-$0.1
$0.5
$0.5
$0.7
Reduced water withdrawals'5
$0.0
<$0.0
$0.0
$0.0
Air-related Effects
-$4.8
-$5.2
-$3.7
-$0.9
Changes in C02 air emissions0
-$4.8
-$5.2
-$3.7
-$0.9
Totald
-$13.7
-$16.0
-$53.3
$28.4
$32.6
$74.4
$37.1
$42.5
$100.9
$70.6
$77.2
$148.4
a. Negative values represent forgone benefits and positive values represent realized benefits.
b. "<$0.0" indicates that monetary values are greater than -$0.1 million but less than $0.00 million.
c. The EPA estimated the air-related benefits for Option 2 using the IPM sensitivity analysis scenario that includes the ACE rule in the baseline (IPM-ACE). EPA extrapolated estimates for
Options 1 and 3 air-related benefits from the estimate for Option 2 that is based on IPM-ACE outputs. The values for Option 4 air-related benefits were estimated using the IPM analysis
scenario that does not include the ACE rule in the baseline. See Chapter 8 for details.
d. Values for individual benefit categories may not sum to the total due to independent rounding.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
11-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
12: Total Social Costs
12 Summary of Total Social Costs
This chapter discusses the EPA's estimates of the costs to society under the regulatory options. Social costs
include costs incurred by both private entities and the government (e.g., in implementing the regulation). As
described further in Chapter 10 of the RIA (U.S. EPA, 2019c), the EPA did not evaluate incremental cost to
state governments to evaluate and incorporate best professional judgment into National Pollutant Discharge
Elimination System (NPDES) permits. Consequently, the only category of costs used to calculate social costs
are estimated compliance costs for steam electric power plants. As discussed below, these costs may be
positive or negative, with the latter occurring when a regulatory option provides savings as compared to the
baseline.
12.1 Overview of Costs Analysis Framework
RIA Chapter 3: Compliance Costs presents the EPA's development of costs to the 951 steam electric power
plants subject to the regulatory options (U.S. EPA, 2019c). These costs (pre-tax) are used as the basis of the
social cost analysis.
As described in Chapter 1, the EPA assumed that steam electric power plants, in the aggregate, would
implement control technologies between 2021 and 2028, with the compliance schedule varying across
wastestreams and regulatory options. For the analysis of social costs, the EPA estimated a plant- and year-
explicit schedule of compliance cost outlays over the period of 2021 through 2047.76 After creating a cost-
incurrence schedule for each cost component, the EPA summed the costs expected to be incurred in each year
for each plant, then aggregated these costs to estimate the total costs for each year in the analysis period.
Following the approach used for the 2015 ELG analysis (U.S. EPA, 2015a), after compliance costs were
assigned to the year of occurrence, the Agency adjusted these costs for change between their stated year and
the year(s) of their incurrence as follows:
• All technology costs, except planning, were adjusted to their incurrence year(s) using the
Construction Cost Index (CCI) from McGraw Hill Construction and the Gross Domestic Product
(GDP) deflator index published by the U.S. Bureau of Economic Analysis (BEA);
• Planning costs were adjusted to their incurrence year(s) using the Employment Cost Index (ECI)
Bureau of Labor Statistics (BLS) and GDP deflator.
The CCI and ECI adjustment factors were developed only through the year 2027; after these years, the EPA
assumed that the real change in prices is zero - that is, costs are expected to change in line with general
inflation. The EPA judges this to be a reasonable assumption, given the fact that capital expenditures would
occur by 2028 and uncertainty of long-term future price projections.
After developing the year-explicit schedule of total costs and adjusting them for predicted real change to the
year of their incurrence, the EPA calculated the present value of these cost outlays as of the rule promulgation
year by discounting the cost in each year back to 2020, using both 3 percent and 7 percent discount rates.
These discount rate values reflect guidance from the OMB regulatory analysis guidance document, Circular
76 The period of analysis extends to 2047 to capture a substantive portion of the life of the compliance technology at any steam
electric power plant (20 or more years), and the last year of technology implementation (2028).
EPA-821-R-19-011
12-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
12: Total Social Costs
A-4 (U.S. OMB, 2003). The EPA calculated the constant annual equivalent value (annualized value), again
using the two values of the discount rate, 3 percent and 7 percent, over a 27-year social cost analysis period.
The EPA assumed no re-installation of compliance technology during the period covered by the social cost
analysis.
To assess the economic costs of the regulatory options to society, the EPA relied first on the estimated costs
to steam electric power plants for the labor, equipment, material, and other economic resources needed to
comply with the regulatory options (see U.S. EPA (2019b) for detail). In this analysis, the market prices for
labor, equipment, material, and other compliance resources represent the opportunity costs to society for use
of those resources in regulatory compliance. The EPA assumed in its social cost analysis that the regulatory
options do not affect the aggregate quantity of electricity that would be sold to consumers and, thus, that the
rule's social cost would include no changes in consumer and producer surplus from changes in electricity
sales by the electricity industry in aggregate. Given the small impact of the regulatory options on electricity
production cost for the total industry, this assumption is reasonable for the social cost analysis (for more
details on the impacts of the regulatory options on electricity production cost, see RIA Chapter 5: Electricity
Market Analyses). The social cost analysis considers costs on an as-incurred, year-by-year basis - that is, this
analysis associates each cost component to the year(s) in which they are assumed to occur relative to the
assumed promulgation and technology implementation years.77
Finally, as discussed in Chapter 10 of the RIA document (U.S. EPA, 2019c; see Section 10.7: Paperwork
Reduction Act of 1995), the regulatory options would not result in additional administrative costs for plants to
implement, and state and federal National Pollutant Discharge Elimination System (NPDES) permitting
authorities to administer, the revised ELGs, once promulgated. As a result, the social cost analysis focuses on
the resource cost of compliance as the only direct cost incurred by society as a result of the regulatory options.
12.2 Key Findings for Regulatory Options
Table 12-1 presents annualized costs for the baseline and each of the four regulatory options. The table also
provides the incremental costs attributable to the regulatory options, calculated as the difference between each
option and the baseline. As shown in the table, the regulatory options generally result in cost savings across
the four options and discount rates, with the exception of Option 4 which results in incremental costs at
3 percent discount rate. Thus, incremental costs range from -$136.3 million to $11.9 million at a 3 percent
discount rate, and from -$166.2 million to -$27.3 million at a 7 percent discount rate.
Table 12-1: Summary of Estimated Annualized Costs (Millions; $2018)
Annualized Costs
Incremental Costs
Regulatory Option
3% Discount Rate
7% Discount Rate
3% Discount Rate
7% Discount Rate
Baseline
$364.9
$417.0
Option 1
$234.3
$263.0
-$130.6
-$154.0
Option 2
$228.6
$250.8
-$136.3
-$166.2
Option 3
$274.8
$297.5
-$90.1
-$119.5
Option 4
$376.8
$389.7
$11.9
-$27.3
Source: U.S. EPA Analysis, 2019.
11 The specific assumptions of when each cost component is incurred can be found in Chapter 3: Compliance Costs of the RL4.
EPA-821-R-19-011
12-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
12: Total Social Costs
Table 12-2 provides additional detail on the social cost calculations. The table compiles, for the baseline and
each of the four regulatory options, the time profiles of compliance costs incurred. The table also reports the
estimated annualized values of costs at 3 percent and 7 percent discount rates. The maximum compliance
outlays differ across the options but are incurred over the years 2021 through 2028, i.e., during the estimated
window when steam electric power plants are expected to implement compliance technologies.
Table 12-2: Time Profile of Costs to Society (Millions; $2018)
Compliance Costs
Incremental Costs
Year
Baseline
Option 1
Option 2
Option 3
Option 4
Option 1
Option 2
Option 3
Option 4
2020
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
$0.0
2021
$1,211.9
$673.5
$635.9
$683.3
$554.9
-$538.4
-$576.0
-$528.6
-$657.0
2022
$746.5
$487.0
$375.9
$475.2
$405.3
-$259.5
-$370.6
-$271.3
-$341.1
2023
$2,070.8
$1,231.1
$907.4
$996.0
$870.9
-$839.7
-$1,163.4
-$1,074.8
-$1,200.0
2024
$192.8
$135.2
$178.8
$203.1
$321.3
-$57.6
-$14.0
$10.3
$128.5
2025
$195.1
$136.6
$181.8
$222.8
$460.0
-$58.4
-$13.3
$27.7
$265.0
2026
$190.7
$131.4
$109.4
$134.7
$647.0
-$59.3
-$81.3
-$56.1
$456.2
2027
$201.5
$141.2
$115.1
$141.0
$456.4
-$60.3
-$86.3
-$60.4
$255.0
2028
$189.6
$129.4
$445.2
$575.0
$657.8
-$60.3
$255.5
$385.3
$468.2
2029
$204.8
$144.6
$146.5
$185.2
$287.0
-$60.3
-$58.4
-$19.6
$82.2
2030
$201.8
$141.6
$144.7
$183.7
$285.1
-$60.3
-$57.1
-$18.1
$83.3
2031
$205.1
$144.9
$149.8
$190.1
$288.4
-$60.3
-$55.3
-$15.0
$83.3
2032
$207.0
$146.8
$149.9
$191.2
$294.1
-$60.3
-$57.1
-$15.8
$87.1
2033
$214.7
$154.4
$148.1
$188.8
$299.5
-$60.3
-$66.6
-$25.9
$84.8
2034
$201.7
$141.4
$146.7
$185.6
$289.3
-$60.3
-$55.0
-$16.1
$87.6
2035
$204.8
$144.6
$147.6
$186.3
$289.7
-$60.3
-$57.3
-$18.5
$84.9
2036
$193.9
$133.6
$142.8
$181.8
$286.9
-$60.3
-$51.1
-$12.1
$93.0
2037
$199.5
$139.2
$146.2
$185.0
$288.7
-$60.3
-$53.3
-$14.5
$89.2
2038
$189.5
$129.3
$140.8
$182.9
$287.5
-$60.3
-$48.7
-$6.6
$97.9
2039
$203.1
$142.8
$146.1
$185.5
$287.6
-$60.3
-$57.0
-$17.6
$84.5
2040
$201.7
$141.4
$145.0
$183.5
$285.4
-$60.3
-$56.7
-$18.2
$83.7
2041
$208.8
$148.6
$150.4
$190.3
$290.3
-$60.3
-$58.4
-$18.5
$81.5
2042
$207.3
$147.0
$150.4
$191.9
$294.8
-$60.3
-$56.9
-$15.4
$87.5
2043
$212.7
$152.4
$151.4
$192.9
$299.7
-$60.3
-$61.3
-$19.8
$87.1
2044
$201.6
$141.3
$146.4
$185.8
$289.7
-$60.3
-$55.2
-$15.8
$88.1
2045
$202.5
$142.2
$146.7
$186.2
$290.0
-$60.3
-$55.7
-$16.2
$87.6
2046
$200.5
$140.3
$146.3
$185.3
$289.5
-$60.3
-$54.3
-$15.3
$88.9
2047
$202.2
$141.9
$146.9
$186.0
$289.9
-$60.3
-$55.3
-$16.2
$87.7
Annualized
$364.9
$234.3
$228.6
$274.8
$376.8
-$130.6
-$136.3
-$90.1
$11.9
Costs, 3%
Annualized
$417.0
$263.0
$250.8
$297.5
$389.7
-$154.0
-$166.2
-$119.5
-$27.3
Costs, 7%
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
12-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
13: Benefits and Social Costs
13 Benefits and Social Costs
This chapter compares total monetized benefits and costs for the four regulatory options analyzed. Benefits
and costs are compared on two bases: (1) incrementally for each of the options analyzed as compared to the
baseline and (2) incrementally across options. The comparison of benefits and costs also satisfies the
requirements of Executive Order 12866: Regulatory Planning and Review and Executive Order 13563:
Improving Regulation and Regulatory Review (see Chapter 9: Other Administrative Requirements of the RIA:
U.S. EPA, 2019c).
13.1 Comparison of Benefits and Costs by Option
Chapter 11 and Chapter 12 present estimates of the benefits and costs, respectively, for the regulatory options
as compared to the baseline.
Table 13-1 presents the EPA's estimates of benefits and costs of the regulatory options, at 3 percent and
7 percent discount rates, and annualized over 27 years. These values are all in 2018 dollars and are based on
the discounting of costs and benefits to 2020, the rule promulgation year.
Table 13-1: Total Estimated Annualized Benefits and Costs by Regulatory Option and Discount Rate
(Millions; 2018$)
Regulatory Option
Total Monetized Benefits3
Total Costs
Low
Mid
High
3% Discount Rate
Option 1
-$41.0
-$43.6
-$86.6
-$130.6
Option 2
$14.8
$19.6
$68.5
-$136.3
Option 3
$35.1
$41.3
$109.4
-$90.1
Option 4
$98.4
$105.9
$188.9
$11.9
7% Discount Rate
Option 1
-$13.7
-$16.0
-$53.3
-$154.0
Option 2
$28.4
$32.6
$74.4
-$166.2
Option 3
$37.1
$42.5
$100.9
-$119.5
Option 4
$70.6
$77.2
$148.4
-$27.3
a. The EPA estimated the air-related benefits for Option 2 using the IPM sensitivity analysis scenario that includes the ACE rule in
the baseline (IPM-ACE). EPA extrapolated estimates for Options 1 and 3 air-related benefits from the estimate for Option 2 that is
based on IPM-ACE outputs. The values for Option 4 air-related benefits were estimated using the IPM analysis scenario that does
not include the ACE rule in the baseline. See Chapter 8 for details.
Source: U.S. EPA Analysis, 2019.
13.2 Analysis of Incremental Benefits and Costs
In addition to comparing estimated benefits and costs for each regulatory option relative to the baseline, as
presented in the preceding section, the EPA also estimated the benefits and costs of the options on an
incremental basis. The comparison in the preceding section addresses the simple quantitative relationship
between estimated benefits and costs for each option and determines whether costs or benefits are greater for
a given option and by how much. In contrast, incremental analysis looks at the differential relationship of
benefits and costs across options and poses a different question: as increasingly more costly options are
considered, by what amount do benefits, costs, and net benefits (/'. e., benefits minus costs) change from option
EPA-821-R-19-011
13-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
13: Benefits and Social Costs
to option? Incremental net benefit analysis provides insight into the net gain to society from imposing
increasingly more costly requirements.
The EPA conducted the incremental net benefit analysis by calculating, for the four regulatory options, the
change in net benefits, from option to option, in moving from the least stringent option to successively more
stringent options, where stringency is determined based on total pollutant loads. As described in Chapter 1,
the regulatory options differ in the technology basis for different wastestreams. Thus, the difference in
benefits and costs across the options derives from the characteristics of the wastestreams controlled by an
option, the relative effectiveness of the control technology in reducing pollutant loads, the timing of control
technology implementation, and the distribution and characteristics of steam electric power plants that would
implement the technologies and of the receiving waterbodies.
As reported in Table 13-2, the EPA estimated that cost savings exceed forgone monetized benefits under
Option 1, with mid-range net annual monetized benefits of $87.0 million using a 3 percent discount rate.
Options 2 and 3 have positive benefits and cost savings, with mid-range net annual monetized benefits
ranging from $131.4 million under Option 3 to $155.9 million under Option 2 (3 percent discount rate).
Option 4 has both positive benefits and costs, with mid-range net annual monetized benefits of $94.0 million
using a 3 percent discount rate. Among the regulatory options, the proposed option (Option 2) results in the
highest net annual monetized benefits.
Using a 3 percent discount rate, the incremental net annual monetized benefits of moving from Option 1 to
Option 2 is $68.9 million. The positive value indicates that net annual monetized benefits are higher for
Option 2 than for Option 1. Moving from Option 2 to Option 3, the change is negative, at -$24.6 million,
which indicates that the increase in costs is greater than the increase in benefits. The change of moving from
Option 3 to Option 4 is also negative, at -$37.3 million, again indicating that the increase in costs is larger
than the increase in benefits.
Table 13-2: Estimated Incremental Net Benefit Analysis (Millions; 2018$)
Regulatory
Net Annual Monetized Benefits313
Incremental Net Annual Monetized Benefitsc
Option
Low
Mid
High
Low
Mid
High
3% Discount Rate
Option 1
$89.6
$87.0
$44.0
NA
NA
NA
Option 2
$151.1
$155.9
$204.8
$61.5
$68.9
$160.8
Option 3
$125.2
$131.4
$199.5
-$25.9
-$24.6
-$5.3
Option 4
$86.5
$94.0
$177.0
-$38.7
-$37.3
-$22.5
7% Discount Rate
Option 1
$140.3
$138.0
$100.7
NA
NA
NA
Option 2
$194.6
$198.8
$240.6
$54.4
$60.9
$139.8
Option 3
$156.6
$162.0
$220.4
-$38.0
-$36.8
-$20.1
Option 4
$97.9
$104.5
$175.7
-$58.8
-$57.5
-$44.7
EPA-821-R-19-011
13-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
13: Benefits and Social Costs
Table 13-2: Estimated Incremental Net Benefit Analysis (Millions; 2018$)
Regulatory
Net Annual Monetized Benefitsa b
Incremental Net Annual Monetized Benefitsc
Option
Low
Mid
High
Low
Mid
High
NA: Not applicable for Option 1
a. Net benefits are calculated by subtracting total annualized costs from total annual monetized benefits, where both costs and
benefits are measured relative to the baseline.
b. The EPA estimated the air-related benefits for Option 2 using the IPM sensitivity analysis scenario that includes the ACE rule in
the baseline (IPM-ACE). EPA extrapolated estimates for Options 1 and 3 air-related benefits from the estimate for Option 2 that is
based on IPM-ACE outputs. The values for Option 4 air-related benefits were estimated using the IPM analysis scenario that does
not include the ACE rule in the baseline. See Chapter 8 for details.
c. Incremental net benefits are equal to the difference between net benefits of an option and net benefits of the previous, less
stringent option.
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
13-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
14: Environmental Justice
14 Environmental Justice
Executive Order (E.O.) 12898 (59 FR 7629, February 11, 1994) requires that, to the greatest extent
practicable and permitted by law, each Federal agency must make the achievement of EJ part of its mission.
E.O. 12898 provides that each Federal agency must conduct its programs, policies, and activities that
substantially affect human health or the environment in a manner that ensures such programs, policies, and
activities do not have the effect of (1) excluding persons (including populations) from participation in, or (2)
denying persons (including populations) the benefits of, or (3) subjecting persons (including populations) to
discrimination under such programs, policies, and activities because of their race, color, or national origin.
To meet the objectives of E.O. 12898, the EPA examined whether the change in benefits from the regulatory
options may be differentially distributed among population subgroups in the affected areas. The EPA
considered the following factors in this analysis: population characteristics, proximity to affected waters,
exposure pathways, cumulative risk exposure, and susceptibility to environmental risk. For example,
subsistence fishers rely on self-caught fish for a larger share of their food intake than do recreational
fishermen, and as such may incur a larger share of effects arising from the regulatory options.
As described in the following sections, the EPA conducted two types of analyses to evaluate the EJ
implications of the regulatory options: (1) summarizing the demographic characteristics of the households
living in proximity to steam electric power plant discharges; (2) analyzing the distribution of human health
impacts among minority and/or low-income populations from changes in exposure to pollutants via the
consumption of self-caught fish and drinking water. The first analysis provides insight on the distribution of
regulatory options effects (e.g., changes in air emissions and effects of water quality changes) on communities
in close proximity to steam electric power plants. The second analysis seeks to provide more specific insight
on the distribution of changes in adverse health effects and benefits and to assess whether minority and/or
low-income populations incur disproportionally high environmental impacts and/or are disproportionally
excluded from realizing the benefits of this regulatory options.
The following two sections describe (1) a comparison of the socio-economic characteristics of the populations
that live in proximity to steam electric power plants to state and national averages, and (2) the evaluation of
human health effects and benefits that accrue to populations in different socio-economic cohorts.
14.1 Socio-economic Characteristics of Populations Residing in Proximity to Steam Electric
Power Plants
For the first analysis, the EPA assessed the demographic characteristics of the populations within specified
distances of steam electric power plants. The analysis is analogous to the profile the EPA developed to
support the 2015 rule (U.S. EPA, 2015a).
EPA-821-R-19-011
14-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs 14: Environmental Justice
The EPA collected population-specific the U.S. Census Bureau's ACS data on:
• the percent of the population below the poverty threshold,78 and
• the population categorized in various racial/ethnic groups, from which EPA calculated the percent of
the total population that belongs to a minority racial/ethnic group.79
The EPA compiled these data for CBGs located within specified distances (e.g., one mile, three miles,
15 miles, 30 miles, and 50 miles) of steam electric power plants. The EPA compared demographic metrics to
state and national averages to identify communities where EJ concerns may exist. EJ concerns may exist in
areas where the percent of the population living below the poverty threshold or that is minority is higher than
the respective state or national averages.
This first analysis considers the spatial distribution of low-income and minority groups to determine whether
these groups are more or less represented in the populations in proximity to steam electric power plants that
discharge bottom ash transport water or FGD wastewater. The specified distance buffers from the reaches are
denoted below as the "benefit region." Populations within the regions included in the analysis may be affected
by steam electric power plant discharges and other environmental impacts in the baseline and would be
affected by environmental changes resulting from the regulatory options, whether those changes are beneficial
or detrimental. If the population within a given region has a larger proportion of minority or low-income
families than the state average, it may indicate that the regulatory options may affect communities that have
been historically exposed to a disproportionate share of environmental impacts and the proposal may thus
contribute to redressing or exacerbating existing EJ concerns, depending on the direction of the changes.
The EPA used the U.S. Census Bureau's ACS data for 2012 to 2016 to identify poverty and minority status at
the state and CBG levels. The EPA overlaid the data with GIS data of buffer zones of specified distances from
steam electric power plants to characterize the communities living in proximity to the affected reaches. Table
14-1 summarizes the socio-economic characteristics of the regions defined using radial distances of one,
three, 10, 15, 30 and 50 miles from the steam electric power plants.
Poverty status is based on data from the Census Bureau's American Community Survey which determines poverty status by
comparing annual income to a set of dollar values called poverty thresholds that vary by family size, number of children, and the
age of the householder.
The racial/ethnic categories are based on available fish consumption data as well as the breakout of ethnic/racial populations in
Census data, which distinguishes racial groups within Hispanic and non-Hispanic categories. Minority groups include: African
American (non-Hispanic); Asian (non-Hispanic); Native Hawaiian/Pacific Islander (non-Hispanic); American Indian/Alaska
Native (non-Hispanic); Other non-Hispanic; Hispanic/Latino.
EPA-821-R-19-011
14-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
14: Environmental Justice
Table 14-1: Socio-economic Characteristics of Communities Living in Proximity to Steam Electric
Power Plants, Compared to National Average
Distance from
receiving reach
Total population
(millions)
Percent minority
Percent below
poverty level
Demographic lndexa
1 mile
0.49
16.7%
12.9%
14.8%
3 miles
1.56
19.9%
14.0%
17.0%
15 miles
18.49
31.2%
14.9%
23.1%
30 miles
56.56
33.0%
14.4%
23.7%
50 miles
107.84
33.2%
14.6%
23.9%
United States
318.6
26.7%
15.1%
20.9%
a. The demographic index is an average of the two demographic indicators explicitly named in EO 12898: low-income and
minority.
b. Communities are based on Census Block Groups within the specified distance of one or more steam electric power plants.
Source: U.S. EPA analysis, 2019
As shown in Table 14-1 approximately 490,000 people live within one mile of steam electric power plants
currently discharging bottom ash transport water or FGD wastewater to surface waters, over 1.5 million live
within three miles, and nearly 56.6 million people live within 30 miles. The statistics also show that a greater
fraction of the communities living between 15 to 50 miles from steam electric power plants is minority, when
compared to the national average. Approximately 31 to 33 percent of households in communities within 15 to
50 miles from steam electric power plants belong to minority racial or ethnic groups as compared to a national
average of 27 percent. Communities between one to three miles from steam electric power plants have a
smaller fraction of their population belonging to minority groups compared to the national average. A smaller
fraction of the population within all analyzed radial distances from the plants have income below the poverty
level compared to the national average (15 percent), but the difference is generally small. As one moves
farther away from the steam electric power plants, the fraction of the community that is below the poverty
threshold fluctuates below the national average while the percent minority increases, so that the overall
demographic index approaches and then exceeds that of the U.S. population overall.
The simple comparison to the national average may not account for important differences, however, between
states, particularly given the non-uniform geographical distribution of steam electric power plants across the
country. The EPA therefore also compared the demographic profile of affected communities within the state
where plants are located. Table 14-2 summarizes the results of this comparison. For this analysis, the
demographic profile of each affected community (defined at the CBG level) located within a given distance of
a steam electric power plant is compared to the average profile within the relevant state. Although the results
in Table 14-1 show that poverty and some minority percentages within the various radial distances from
steam electric facilities are below the national average, the comparison to state averages show affected
communities within the various distance buffers with greater poverty or minority percentages than the state
average. This pattern derives, in part, from variances of state average poverty and minority levels from
national levels. For example, of the 346 communities within one mile of steam electric power plants, 120 (35
percent) have a higher percentage of households living below the poverty threshold than their state average,
62 (18 percent) have a higher percent of the population that is minority, and 38(11 percent) have a higher
proportion of households that are both living below the poverty level and minority. Details of this analysis are
included in the docket for this proposed rule (DCN SE07640). These results highlight the potential for
localized differences indicative of potential EJ concern, but the overall comparison shows no indication that
EPA-821-R-19-011
14-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
14: Environmental Justice
any communities with EJ concern would be precluded from the benefits of the regulatory options, or
conversely, would be disproportionally affected by the resulting environmental changes.
Table 14-2: Socio-economic Characteristics of Affected Communities, Compared to State Average
Distance from
plant
Number of
Affected
Communities3
Number of Communities that...
are Poorer
have a Higher
Proportion of Minority
Population
are Poorer and have a
Higher Proportion of
Minority Population
... than the State Average
1 mile
346
120
62
38
3 miles
1,105
432
266
172
15 miles
13,032
5,604
5,001
3,345
30 miles
38,811
15,584
14,872
9,277
50 miles
73,628
29,896
27,975
17,585
a. "Affected communities" are Census Block Groups within the specified distance of one or more steam electric power plants.
Source: U.S. EPA analysis, 2019
14.2 Distribution of Human Health Impacts and Benefits
The second type of analysis looks at the distribution of environmental effects and benefits to further inform
understanding of the potential EJ concerns and the extent to which the regulatory options may mitigate or
exacerbate them.
A significant share of the benefits of the regulatory options comes from the small estimated changes in the
discharges of harmful pollutants to surface waters and associated changes in fish tissue contamination and
drinking water quality. The sections below discuss the distribution of health effects for these two pathways.
This analysis allows the Agency to report the distribution of benefits or forgone benefits across population
subgroups, including subgroups who may have been historically exposed to a disproportionate share of
environmental impacts.
The EPA did not analyze the potential EJ concerns associated with changes in air emissions since the
approach used to estimate air-related benefits from changes in EGU emissions does not provide explicit
airsheds to overlay with population data, nor does it break out effects for sensitive subgroups. See Chapter 8
for details.
14.2.1 Socio-economic Characteristics of Populations Impacted by Changes in Exposure to Pollutants via
Drinking Water Pathway
The EPA quantified the human health benefits resulting from the small estimated changes in exposure to
TTHM in drinking water in individuals served by PWS either directly or indirectly affected by steam electric
power plants" bromide discharges. The analysis relied on county-level and tribal area data to estimate the
number and characteristics of individuals exposed to steam electric pollutants through the consumption of
drinking water, and race and ethnicity-specific assumptions to estimate exposure. The EPA did not quantify
or monetize health effects associated with exposure to other pollutants in drinking water (see Chapter 2 for a
qualitative discussion).
This section presents estimates for populations affected by changes in exposure to TTHM associated with
bromide in drinking water in two types of geographic areas: tribal areas and counties. Chapter 4 discusses the
EPA-821-R-19-011
14-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
14: Environmental Justice
approach used to identify the affected population, estimate exposure levels, quantify health effects, and
monetize benefits.
Table 14-3 summarizes the estimated affected population exposed to TTHM through consumption of drinking
water in the general population and in population subgroups that may be indicative of EJ concerns. The
analysis is conducted at the county level and compares the demographic profile of the affected counties to that
of the state where they are located. Over 43 million people, across more than 300 counties and 29 states,
would be affected by changes resulting from the regulatory options. Of the 29 states affected, the majority of
states (24) have affected counties that are poorer than the state average, 23 have affected counties that have a
higher proportion of minority population, and 21 have affected counties that are both poorer and have a higher
proportion of minority population. Details of this analysis are included in the docket for this proposed rule
(DCN SE07640).
Table 14-3: Socio-economic Characteristics of Affected Counties, Compared to State Average
Number of States where Affected Counties...
Total Affected
Population
(millions)
have a Higher
are Poorer and have
Number of Affected
Counties
Number of
Affected States
are Poorer
Proportion of
Minority
Population
a Higher Proportion
of Minority
Population
... than the State Average
303
43.92
29
24
23
21
Source: U.S. EPA analysis, 2019
Table 14-4 summarizes the estimated tribal area population and population subgroups indicative of EJ
concerns that are potentially exposed to trihalomethanes in drinking water as a result of steam electric power
plant discharges. The analysis is conducted at the tribal area level and compares the demographic profile of
the affected tribal areas to that of the state where they are located. Based on the population affected by stream
electric plant discharges, an average of 40 percent of tribal area population is expected to be affected by the
regulatory options (see Table 14-4).
Affected tribal areas consistently have a higher minority population than the state average; half of the tribal
areas have minority population percentages greater than 90 percent. Nearly all of the affected tribal areas have
a higher low-income population than the state average. The Otoe-Missouria Oklahoma Tribal Statistical Area
(OTSA) has a lower low-income population percentage (12.6 percent) compared to the state average
(16.5 percent). This difference, however, is not significant enough to influence the demographic index for the
Otoe-Missouria OTSA. Therefore, all affected tribal areas have higher demographic indices compared to the
state average.
EPA-821-R-19-011
14-5
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
Table 14-4: Socio-economic Characteristics of Affected Tribal Areas, Compared to State Average
Affected Tribal Areas
States
Population
Percent Minority
Percent Below Poverty
Level
Demographic Index
Affected
Population3
Total
Population of
Tribal Area
Affected Tribal
Area Population
Percentage
Tribal Area
State
Tribal Area
State
Tribal Area
State
Crow Creek
Reservation
SD
1,357
2,190
62.0%
92.8%
17.1%
38.4%
13.9%
65.6%
15.5%
Fort Berthold
Reservation
ND
5,846
7,435
79.0%
74.8%
13.6%
22.5%
11.0%
48.6%
12.3%
Lake Traverse
Reservation
ND; SD
230
11,269
2.0%
47.2%
15.5%
21.3%
12.6%
34.2%
14.0%
Lower Brule
Reservation
SD
2,116b
1,531
80.0%c
94.3%
17.1%
43.4%
13.9%
68.8%
15.5%
Navajo Nation
AZ; NM; UT
1,198
174,692
0.7%
98.2%
41.1%
41.4%
16.1%
69.8%
28.6%
Otoe-Missouria OTSA
OK
250
921
27.1%
51.0%
33.1%
12.6%
16.5%
31.8%
24.8%
Pine Ridge Reservation
NE; SD
8b
19,698
0.0%d
90.0%
18.9%
50.8%
12.6%
70.4%
15.7%
Rosebud Indian
Reservation
SD
9b
11,324
0.1%
91.8%
17.1%
49.6%
13.9%
70.7%
15.5%
Standing Rock
Reservation
ND; SD
6,839
8,612
79.4%
78.9%
15.5%
42.2%
12.6%
60.5%
14.0%
Yankton Reservation
SD
1,064
6,700
15.9%
50.3%
17.1%
27.6%
13.9%
39.0%
15.5%
a. Affected population is defined as the population served as reported in the EPA SDWIS database for PWS affected by steam electric power plant discharges associated with each tribal
area.
b. PWS ID 84690026 serves several reservations and counties. Therefore, SDWIS reported population served was equally distributed over the three reservations served: Lower Brule
Reservation, Pine Ridge Reservation, and Rosebud Indian Reservation.
c. PWS ID 84690441 serves the Lower Brule Reservation and surrounding South Dakota counties. As a result, the SDWIS reported population served exceeds the Census reported total
population of the reservation. The affected percentage of tribal area was adjusted to 80 percent to reflect that the majority of the reservation is likely served by the affected PWS.
d. Value less than 0.1% but greater than 0.0%.
Source: U.S. EPA analysis, 2019
EPA-821-R-19-011
14-6
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
14.2.2 Socio-economic Characteristics of Populations Impacted by Changes in Exposure to Pollutants via
Fish Ingestion Pathway
The EPA quantified the human health effects resulting from the small estimated changes in exposure to
pollutants in fish tissue in individuals who consume fish caught in reaches immediately receiving or
downstream from steam electric power plant discharges. The analysis relied on CBG-level data to estimate
the number and characteristics of individuals exposed to steam electric pollutants through the consumption of
self-caught fish, and race and ethnicity-specific data to estimate exposure.
This section presents results for the two types of anglers analyzed: recreational anglers and subsistence
fishers. Chapter 5 provides more details on the approach used to identify the affected population, estimate
exposure, quantify health effects, and monetize benefits.
The EPA limited its analysis of the distribution of health effects and potential benefits to two pollutants (lead
and mercury) since the regulatory options did not generate any changes in arsenic-related health effects.
Further, for recreational anglers, the EPA focused on health effects in infants and children.
Table 14-5 summarizes the estimated number of individuals exposed to steam electric pollutants through
consumption of self-caught fish in the general population and in population subgroups that may be indicative
of EJ concerns. The population values included in Table 14-5 are based on the population potentially exposed
to lead, which is larger than the population potentially exposed to mercury. As shown in the table, of the
approximately 1.5 million people potentially exposed to steam electric pollutants through fish tissue
consumption, 13.9 percent live below the poverty level, 63.5 percent are minority, and 11.5 percent both live
below the poverty level and are minority. Overall, 68.5 percent of potentially exposed individuals are
categorized in at least one or more EJ subgroup based on their poverty level or race/ethnicity, while
31.5 percent are neither minority nor live below the poverty level.
Table 14-5: Characteristics of Population Potentially Exposed to Lead from Steam Electric Power
Plants via Consumption of Self-caught Fish
Subgroup
Minority
Non-Minority
Total
Below Poverty
Level
175,198
11.5%
77,077
5.1%
252,275
13.9%
Above Poverty
Level
790,136
51.9%
478,694
31.5%
1,268,829
86.1%
Total
965,334
63.5%
555,770
36.5%
1,521,104
100.0%
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
14-7
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
The distribution of adverse health effects is a function of the characteristics of the affected population (Table
14-5), including age and sex,8" ethnicity-specific exposure factors,81 and reach water quality. Table 14-6
shows the distribution of selected adverse health effects in the baseline. Table 14-7 shows the distribution of
changes in adverse health effects under each of the four regulatory options. Note that monetary values follow
the same distribution as changes in adverse health effects since each case is valued equally, irrespective of the
socio-economic subgroup.
The two tables show results for three selected subgroups:
• Below the poverty level and minority (PAM) (11.5 percent of the exposed population),
• Below the poverty level or minority (POM) (i.e.. but not both; 57.0 percent of the exposed
population), and
• All others (i.e., above the poverty level and white; 31.5 percent of the exposed population).
The first two subgroups are the primary interest of this analysis as potentially indicative of EJ concerns.
The distribution health effects under baseline and regulatory options can be compared to the relative share of
the population exposed to steam electric pollutants (from Table 14-5) to assess the degree to which the
regulatory options contribute to mitigating or increasing any EJ concerns that may be present in the baseline.
Table 14-6 and Table 14-7 both summarize the percent of estimated changes that are incurred by the specific
population subgroups (in the table body) and contrast this distribution to the share of the affected population
represented by each subgroup (in the column headings).
Table 14-6: Estimated Distribution of Baseline IQ Point Decrements by Pollutant (2021 to 2047)
Pollutant
Below the poverty
level and Minority
(PAM)
(11.5% of Population)
Below the poverty level
or Minority
(POM) (57.0% of
Population)
All Others
(31.5% of Population)
Total
Lead
2,791,727
11.6%
13,784,647
57.1%
7,567,405
31.3%
24,143,779
100%
Mercury
88,042
12.7%
442,798
63.8%
162,782
23.5%
693,622
100%
Source: U.S. EPA Analysis, 2019
80 Some adverse health effects are analyzed only for individuals in certain age groups. For example, IQ point decrements from
exposure to lead are calculated for children 0 to 7 years old and the baseline exposure therefore depends on the number of
children within this age group in the affected population in each socio-economic subgroup. IQ point decrements from exposure to
mercury are calculated for infants bom within the analysis period and baseline exposure depends on the number of women of
childbearing age (and fertility rates) in the affected population.
81 Ethnicity-specific factors that determine exposure to pollutants in fish tissue include the assumed fish consumption rates and
average fertility rate. For example, Asian/Pacific Islander anglers have daily consumption rates that are 1.4 times and 1.9 times
those of While (non-Hispanic) anglers for recreational and subsistence fishing modes, respectively.
EPA-821-R-19-011
14-8
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
Table 14-7: Distribution of Changes in IQ Point Relative to the Baseline, by Pollutant (2021 to 2047)
Pollutant
and
Population
Below the poverty level and
Minority (PAM)
(11.5% of Population)
Below the poverty level or
Minority (POM) (57.0% of
Population)
All Others
(31.5% of Population)
Total
Regulatory
Option
Positive IQ
Change
(percent of
exposed
population)
Negative IQ
Change
(percent of
exposed
population)
Positive IQ
Change
(percent of
exposed
population)
Negative IQ
Change (percent
of exposed
population)
Positive IQ
Change
(percent of
exposed
population)
Negative IQ
Change
(percent of
exposed
population)
Positive IQ
Change
(percent of
exposed
population)
Negative IQ
Change (percent
of exposed
population)
Children
Exposed to
Lead3
Option 1
0.06
0.03%
-0.38
4.87%
0.24
0.10%
-1.87
29.9%
-
-
-1.63
31.3%
0.30
0.13%
-3.89
66.1%
Option 2
0.33
0.04%
-0.44
4.86%
1.50
0.20%
-3.16
29.8%
-
-
-9.30
31.3%
1.83
0.25%
-12.90
66.0%
Option 3
0.52
0.04%
-0.37
4.86%
2.41
0.20%
-1.66
29.8%
-
-
-0.54
31.3%
2.92
0.25%
-2.57
66.0%
Option 4
0.54
0.04%
-0.33
4.86%
2.54
0.20%
-1.43
29.8%
-
-
-0.42
31.3%
3.08
0.25%
-2.18
66.0%
Infants
Exposed to
Mercury3
Option 1
-
--
-54
12.7%
-
-
-264
63.8%
-
-
-93
23.5%
-
-
-411
100%
Option 2
-
--
-497
12.7%
-
-
-2,489
63.8%
-
-
-799
23.5%
-
-
-3,785
100%
Option 3
-
--
-503
12.7%
-
-
-2,483
63.8%
-
-
-791
23.5%
-
-
-3,777
100%
Option 4
-
--
-274
12.7%
-
-
-1,320
63.8%
-
-
-427
23.5%
-
-
-2,021
100%
a. Negative values represent forgone benefits and positive values represent realized benefits.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
14-9
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
As shown in Table 14-6, the PAM subgroup represents 11.5 percent of the potentially affected population, but
accounts for 11.6 percent and 12.7 percent of the baseline estimated IQ point changes from lead and mercury
exposure, respectively, in the exposed population. As shown in Table 14-8, a smaller percentage of children
(4.9 percent of the exposed population) in the PAM subgroup experiences forgone benefits from an increase
in exposure to lead than its share in the affected population (11.5 percent), while an even smaller percentage
of children (less than 0.1 percent) experience realized benefits. However, a larger share of children
(12.7 percent) in this subgroup experience forgone benefits from an increase in mercury exposure.
The POM group represents 57.0 percent of the potentially affected population, but accounts for 57.1 percent
and 63.8 percent of the baseline estimated IQ point decrements from lead and mercury exposure, respectively,
in the exposed population. Similar to the PAM subgroup, the POM subgroup experiences a smaller share of
children lead exposure forgone benefits (29.8-29.9 percent, depending on the option) than its population but a
larger share of mercury exposure forgone benefits (63.8 percent).
In the analysis of health benefits for the fish ingestion pathway (see Chapter 5), the EPA assumed that
5 percent of the exposed population are subsistence fishers, and that the remaining 95 percent are recreational
anglers. This is based on the assumed 95th percentile fish consumption rate for subsistence fishers.
Subsistence fishers consume more self-caught fish than recreational anglers and can therefore be expected to
experience higher health risks associated with steam electric pollutants in fish tissue.
The results of the human health analysis suggest that subsistence fishers may be disproportionally exposed to
pollutants in steam electric power plant discharges via fish consumption and may disproportionally incur
adverse health effects from this exposure. As shown in Table 14-8, subsistence fishers incur 7 percent to 17
percent of the baseline IQ decrements, even though they represent only 5 percent of the overall population. As
shown in Error! Reference source not found., 6 percent to 17 percent of the total exposed population are
subsistence fishers who experience health changes and forgone benefits of the regulatory options.
Table 14-8: Estimated Distribution of Baseline IQ Point Decrements by Pollutant and Fishing Mode
(2021 to 2047)
Pollutant and Exposed
Population
Subsistence Fishers
(5 percent of population)
Recreational Fishers
(95 percent of population)
Total
Children Exposed to Lead
1,605,246
6.6%
22,538,533
93.4%
24,143,779
100%
Infants Exposed to Mercury
119,747
17.3%
573,875
82.7%
693,622
100%
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
14-10
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
Table 14-9: Estimated Distribution of Changes in IQ Point Decrements Relative to the Baseline by Fishing Mode, and Pollutant (2021 to
2047)
Pollutant and
Exposed
Population
Regulatory
Option
Subsistence Fishers
(5 percent of population)
Recreational Fishers
(95 percent of population)
Total
Positive IQ Change
(percent of exposed
population)
Negative IQ Change
(percent of exposed
population)
Positive IQ
Change
(percent of
exposed
population)
Negative IQ
Change (percent of
exposed
population)
Positive IQ
Change
(percent of
exposed
population)
Negative IQ Change
(percent of exposed
population)
Children Exposed
to Lead3
Opt
on 1
0.30
0.13%
-2.68
6.52%
-
-
-1.21
59.6%
0.30
0.13%
-3.89
66.1%
Opt
on 2
1.83
0.25%
-11.7
6.40%
-
-
-1.21
59.6%
1.83
0.25%
-12.9
66.0%
Opt
on 3
2.92
0.25%
-1.36
6.40%
-
-
-1.21
59.6%
2.92
0.25%
-2.57
66.0%
Opt
on 4
3.08
0.25%
-0.97
6.40%
-
-
-1.21
59.6%
3.08
0.25%
-2.18
66.0%
Infants Exposed
to Mercury3
Opt
on 1
--
--
-71
17.3%
-
-
-340
82.7%
-
-
-411
100%
Opt
on 2
--
--
-652
17.3%
-
-
-3,134
82.7%
-
-
-3,785
100%
Opt
on 3
--
--
-650
17.3%
-
-
-3,127
82.7%
-
-
-3,777
100%
Opt
on 4
--
--
-347
17.3%
-
-
-1,673
82.7%
-
-
-2,021
100%
a. Negative values represent forgone benefits and positive values represent realized benefits.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
14-11
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
14.3 EJ Analysis Findings
Based on the EJ analyses discussed above, the EPA determined that the regulatory options would not exclude
communities from the benefits of environmental improvements expected to result from the 2015 rule
requirements, but the regulatory options may disproportionally affect communities in cases where the small
changes in water quality increase pollutant exposure compared to the baseline.
Communities in close proximity to a steam electric power plant (between one and fifteen miles) tend to be
less poor (i.e., fewer people living below the poverty level) and minority than the national average. However,
when compared to state averages, as shown in Table 14-2, a greater share of affected communities are poorer
and/or are more minority than the state average. The communities in close proximity to steam electric power
plants may be more likely to be affected by changes in pollutant discharges from these plants.
The majority of county populations potentially exposed to TTHM in drinking water as a result of steam
electric power plant discharges are poorer and more minority than the state average. In addition, all affected
tribal areas have lower demographic indices compared to the state average. Options 2, 3 and 4 would benefit
the EJ communities served by affected PWS by reducing exposure to TTHM via drinking water. Conversely,
an increase in exposure to TTHM has the potential to harm the same communities of concern under Option 1.
Recreational anglers and members of their household, including children, are estimated to experience small
forgone benefits from an increase in pollutant concentrations in fish tissue compared to baseline. A large
portion of forgone benefits to children (IQ decrements) from increased mercury exposure are estimated to
occur within the PAM group and POM group. Increased lead exposure, however, is estimated to impact a
smaller proportion of the PAM and POM. Close to 50 percent of greater IQ decrements are expected to occur
within the non-minority, above the poverty level population.
Because communities at the census block, county, and tribal area levels are poorer and more minority than
state averages, the regulatory options could benefit or harm populations with EJ concerns depending on the
direction of changes in pollutant loadings for the regulatory options and the resulting change in potential
exposure.
14.4 Limitations and Uncertainties
This EJ analysis inherits the limitations and uncertainties of the human health effects analysis (see Chapter 4,
Chapter 5, and Chapter 8) regarding pollutant exposure, incidence of adverse health outcomes, and valuation.
In addition, the EJ analysis embeds uncertainty derived from the application of uniform assumptions across
the estimated population exposed to pollutant discharges when factors may instead vary across socioeconomic
characteristics (see Table 14-10 for detail). In summary, use of average values across the entire population of
the United States (or within a state or SDWIS-identified population served) instead of assumptions that reflect
specific socioeconomic factors may over- or understate inequities present in the baseline and the differential
impacts or benefits to populations living below the poverty level or minority populations from changes due to
the regulatory options.
EPA-821-R-19-011
14-12
-------
BCAfor Proposed Revisions to Steam Electric Power Generating ELGs
14: Environmental Justice
Table 14-10: Limitations and Uncertainties in EJ Analysis
Uncertainty/Assumption
Effect on EJ Analysis
Notes
The EPA does not have data to
delineate airsheds affected by
changes in air quality resulting from
IPM-projected changes in emissions
from EGUs, limiting EPA's ability to
analyze the distribution of forgone
benefits from increased air emissions.
Underestimate
Some population subgroups may be more susceptible to
changes in air quality. While EPA estimated changes in
the amount of air pollutants emitted by EGUs, EPA did
not model the associated changes in the distribution of
air pollutant levels to which populations may be
exposed. The EJ analysis therefore does not account for
EJ concerns that may arise from this exposure pathway.
The EPA assumed that all fishers
travel up to 50 miles
Uncertain
Some anglers stay closer to home and certain EJ or
sensitive subpopulations may tend to stay closer to
home {e.g., people living below the poverty level and
subsistence fishers). To the extent that these people fish
predominantly from waters receiving discharges from
steam electric power plants, they may be exposed to
relatively higher concentrations of pollutants.
Conversely, people who live farther from steam electric
power plants may predominantly fish from waters not
affected by pollutants in steam electric power plant
discharges and be exposed to relatively lower
concentrations of pollutants.
The EPA assumed that subsistence
fishers are 5 percent of all anglers,
with this assumption applied
uniformly across all socioeconomic
groups.
Underestimate
A relatively higher share of EJ groups may be subsistence
fishers. This would tend to increase the inequities
already in the baseline and affect the extent to which
the regulatory options may address or further these
inequities.
The EPA applied uniform fishing
participation rates, FCAs, and catch
and release practices across the entire
population.
Uncertain
Differences in behavior across socioeconomic groups
may result in different distribution of baseline and
regulatory option impacts.
The EPA assumed that the counties
served by PWS, as reported in the
SDWIS database, are representative of
the population affected by changes in
TTHM levels due to steam electric
power plant discharges.
Uncertain
Counties and tribal areas can be served by multiple PWS
and some PWS serve people across multiple counties,
such that the affected population may have different
socioeconomic characteristics.
The EPA used the SDWIS database to
identify counties served by affected
PWS. For any PWS IDs without any
associated county information, the
EPA used the PWS Name and the PWS
latitude and longitude to identify
associated tribal areas.
Uncertain
There may be some PWSs that serve counties and tribal
areas. However, if only the county was listed in SDWIS,
the EJ analysis does not account for the associated tribal
area.
EPA-821-R-19-011
14-13
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
15 Cited References
Alaska Fisheries Science Center (AFSC) of NOAA National Marine Fisheries Service. 2010. Alaska Fisheries
Science Center Publications Database.
American Cancer Society. 2019. Key Statistics for Bladder Cancer. Webpage.
https://www.cancer.org/cancer/bladder-cancer/about/key-statistics.html. Accessed 2019.
Anthoff, D., and R. J. Tol. 2010. On international equity weights and national decision making on climate
change. Journal of Environmental Economics and Management, 60(1): 14-20.
Anthoff, D. and Tol, R.S.J. 2013. "The uncertainty about the social cost of carbon: a decomposition analysis
using FUND." Climatic Change, 117: 515-530.
Arrow, K., M. Cropper, C. Gollier, B. Groom, G. Heal, R. Newell, W. Nordhaus, R. Pindyck, W. Pizer, P.
Portney, T. Sterner, R.S.J. Tol, and M. Weitzman. 2013. "Determining Benefits and Costs for Future
Generations." Science, 341: 349-350.
Atlantic States Marine Fisheries Commission (ASMFC). 2010. "Managed Species." from
http ://www .asmfc. org/managedspecie s .htm.
Axelrad, D.A., D.C. Bellinger, L.M. Ryan, and T.J. Woodruff. 2007. Dose-Response Relationship of Prenatal
Mercury Exposure and IQ: An Integrative Analysis of Epidemiological Data. Environmental Health
Perspectives 115 (4): 609-615.
Bateman, I. J., B.H. Day, S. Georgiou and I. Lake. 2006. "The Aggregation of Environmental Benefit Values:
Welfare Measures, Distance Decay and Total WTP." Ecological Economics 60(2): 450-460.
Bergstrom, J.C. and L.O. Taylor. 2006. "Using Meta-Analysis for Benefits Transfer: Theory and Practice."
Ecological Economics 60(2):351-360.
Bergstrom, J.C. and P. De Civita. 1999. "Status of Benefits Transfer in the United States and Canada: A
Review." Canadian Journal of Agricultural Economics 47(1): 79-87.
Bin, O., and J. Czajkowski. 2013. "The Impact of Technical and Non-technical Measures of Water Quality on
Coastal Waterfront Property Values in South Florida." Marine Resource Economics 28(1): 43-63.
Boyle, K.J., P.J. Poor, and L.O. Taylor. 1999. Estimating the Demand for Protecting Freshwater Lakes from
Eutrophication. American Journal of Agricultural Economics 81:1118-1122.
Burger, J. 2004. Fish Consumption Advisories: Knowledge, Compliance and Why People Fish in an Urban
Estuary. Journal of Risk Research, 7(5): 463-479.
Canadian Council of Ministers of the Environment (CCME). 2001. Canadian water quality guidelines for the
protection of aquatic life: CCME Water Quality Index 1.0, Technical Report. In: Canadian
environmental quality guidelines, 1999, Canadian Council of Ministers of the Environment,
Winnipeg.
Cantor, K.P., C.M. Villanueva, C.T. Silverman, J.D. Figueroa, F. X. Real, M. Garcia-Closas, N. Malats, S.
Chanock, M. Yeager, A. Tardon, R. Garcia-Closas, C. Serra., A. Carrato, G. Castano-Vinyals, C.
Samanic, N. Rothman, and M. Kogevinas. 2010. Polymorphisms in GSTT1, GSTZ1, and CYP2E1,
EPA-821-R-19-011
15-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
disinfection by-products, and risk of bladder cancer in Spain. Environmental Health Perspectives,
118(11): 1545-1550.
Carruthers, T. and C. Wazniak. 2003. Development of a Water Quality Index for the Maryland Coastal Bays.
Maryland's Coastal Bays: Ecosystem Health Assessment, Chapter 4.4. 2003.
Centers for Disease Control and Prevention (CDC), Department of Health and Human Services. 2005.
Preventing Lead Poisoning in Young Children. A Statement by the Centers for Disease Control and
Prevention. U.S. Department of Health and Human Services, Public Health Service.
Center for Disease Control and Prevention (CDC), Department of Health and Human Services. 2009. Fourth
National Report on Human Exposure to Environmental Chemicals.
Chen C.W. and H. Gibb. 2003. Procedures for calculating cessation lag. Regulatory Toxicology and
Pharmacology. 38(2): 157-165.
Clark, E., J.A. Haverkamp, and W. Chapman. 1985. "Eroding Soils: The Off-Farm Impacts." Washington,
DC: The Conservation Foundation.
Cornwell, D. A., B. K. Sidhu, R. Brown, and N. E. McTigue. 2018. Modeling Bromide River Transport and
Bromide Impacts on Disinfection Byproducts. Journal of the American Water Works Association,
110: E1-E23.
Costet, N., Villanueva, C. M., Jaakkola, J. J. K., Kogevinas, M., Cantor, K. P., King, W. D., Lynch, C. F.,
Nieuwenhuijsen, M. J., Cordier, S. Water disinfection by-products and bladder cancer: is there a
European specificity? A pooled and meta-analysis of European case-control studies. Occup. Environ.
Med. 2011, 68(5): 379-385.
Crump, K.S., C. Van Landingham, T.S. Bowers, D. Cahoy, and J.K. Chandalia. 2013. "A statistical
reevaluation of the data used in the Lanphear et al. (2005) pooled-analysis that related low levels of
blood lead to intellectual deficits in children." Critical Reviews in Toxicology, 43(9): 785-799.
Cude, C. G. 2001. Oregon Water Quality Index: "A Tool for Evaluating Water Quality Management
Effectiveness." Journal of the American Water Resources Association 37(1): 125-137.
Deacon, J. E., J. D. Kobetich, J.D. Williams, S. Contreras, et al. 1979. "Fishes of North America endangered,
threatened, or of special concern: 1979." Fisheries 4(2): 1-69.
Desvousges, W.H., F.R. Johnson, and H.S. Banzhaf. 1998. Environmental Policy Analysis with Limited
Information: Principles and Applications of the Transfer Method. Northampton, MA: Edward Elgar
Publishers.
Dunnette, D. A. 1979. A Geographically Variable Water Quality Index Used in Oregon. As cited in Cude
2001.
Fann N, K.R. Baker, and C. Fulcher. 2012. Characterizing the PM2.5-related health benefits of emission
reductions for 17 industrial, area and mobile emission sectors across the U.S. Environment
International 49: 141-151.
Fraas, A., R. Lutter, S. Dudley, T. Gayer, J. Graham, J.F. Shogren, and W.K. Viscusi. 2016. Social Cost of
Carbon: Domestic Duty. Science, 351(6273): 569.
EPA-821-R-19-011
15-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
Freeman III, A.M. 2003. The Measurement of Environmental and Resource Values: Theory and Methods.
Washington, DC: Resources for the Future.
Froese, R. and D. Pauly. 2009. "Fishbase (version 07/2009)." http://www.fishbase.org.
Gayer, T., and K. Viscusi. 2016. Determining the Proper Scope of Climate Change Policy Benefits in U.S.
Regulatory Analyses: Domestic versus Global Approaches. Review of Environmental Economics and
Policy, 10(2): 245-63.
Gibbs, J. P., J. M. Halstead, K. J. Boyle, and J. Huang. 2002. "An Hedonic Analysis of the Effects of Lake
Water Clarity on New Hampshire Lakefront Properties." Agricultural and Resource Economics
Review 31(1): 39-46.
Good, K.D., and J.M. VanBriesen. 2016. Current and potential future bromide loads from coal-fired power
plants in the Allegheny River Basin and their effects on downstream concentrations. Environmental
Science and Technology 50:9078-908.
Good, K.D., and J.M. VanBriesen. 2017. Power plant bromide discharges and downstream drinking water
systems in Pennsylvania. Environmental Science and Technology 51:11829-11838.
Good, K.D., and J.M. VanBriesen. 2019. Coal-Fired Power Plant Wet Flue Gas Desulfurization Bromide
Discharges to U.S. Watersheds and Their Contributions to Drinking Water Sources. Environmental
Science and Technology 53:213-223.
Graf, W. L., E. Wohl, T. Sinha, and J.L. Sabo. 2010. "Sediment and sustainability of western American
reservoirs." Water Resources Research 46(W12535).
Greco S.L., A. Belova, J. Haskell, L. Backer. 2019. Estimated Burden of Disease from Arsenic in Drinking
Water Supplied by Domestic Wells in the United States. Journal of Water and Health.
https://iwaponline.com/iwh/article-pdf/doi/10.2166/wh.2019.216/582364/wh2019216.pdf
Hartge, P., D. Silberman, R. Hoover, C. Schairer, R. Altman, D. Austin, K. Cantor, M. Child, C. Key, and
L.D. Marrett. 1987. Changing Cigarette Habits and Bladder Cancer Risk: A Case-Control Study.
Journal of the National Cancer Institute. 78(6): 1119-1125.
Hope, Chris. 2013. "Critical issues for the calculation of the social cost of C02: why the estimates from
PAGE09 are higher than those from PAGE2002." Climatic Change, 117: 531-543.
Hrubec, Z. and McLaughlin, J.K. 1997. Former cigarette smoking and mortality among U.S. veterans: a 26-
year follow-up. In: Changes in cigarette related disease risks and their implication for prevention
and control. D.M. Burns, L. Garfinkel, J.M. Samet (eds.). NIH Monograph No. 8, National Institutes
of Health. 501-530. Washington, DC: National Cancer Institute.
Hrudey, Steve E., Lorraine C. Backer, Andrew R. Humpage, Stuart W. Krasner, Dominique S. Michaud, Lee
E. Moore, Philip C. Singer, and Benjamin D. Stanford. 2015. Evaluating Evidence for Association of
Human Bladder Cancer with Drinking-Water Chlorination Disinfection By-Products. J Toxicol
Environ Health B Crit Rev. 2015 Jul 4; 18(5): 213-241.
Institute of Medicine. 2013. Environmental Decisions in the Face of Uncertainty. Washington, DC: The
National Academies Press, https://doi.org/10.17226/12568.
EPA-821-R-19-011
15-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
Interagency Working Group on Social Cost of Carbon (IWGSCC), United States Government. 2010.
Technical Support Document: Social Cost of Carbon for Regulatory Impact Analysis under Executive
Order 12866. February 2010.
Interagency Working Group on Social Cost of Carbon (IWGSCC), United States Government. 2013a.
Technical Update of the Social Cost for Regulatory Impact Analysis under Executive Order 12866.
Revised November 2013.
Interagency Working Group on Social Cost of Carbon (IWGSCC), United States Government. 2013b.
Technical Update of the Social Cost for Regulatory Impact Analysis under Executive Order 12866.
Revised June 2015.
Interagency Working Group on Social Cost of Carbon (IWGSCC), United States Government. 2015a.
Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact
Analysis Under Executive Order 12866. Revised July 2015.
Jakus, P., M. Downing, M.S. Bevelhimer, J.M. Fly. 1997. Do Sportfish Consumption Advisories Affect
Reservoir Anglers' Site Choice? Agricultural and Resource Economics Review. Vol. 26, 2, Oct.
1997.
Jakus, P., M. McGuinness, and A. Krupnick. 2002. The Benefits and Costs of Fish Consumption Advisories
for Mercury. Resources for the Future Discussion Paper 02-55.
Jelks, H. L., S. J. Walsh, et al. 2008. "Conservation Status of Imperiled North American Freshwater and
Diadromous Fishes." Fisheries 33(8): 372-407.
Johnston, R.J., E.Y. Besedin, R. Iovanna, C.J. Miller, R.F. Wardwell and M.H. Ranson. 2005. "Systematic
Variation in Willingness to Pay for Aquatic Resource Improvements and Implications for Benefit
Transfer: A Meta-Analysis." Canadian Journal of Agricultural Economics 53(2-3): 221-248.
Klemick, H., C. Griffiths, D. Guignet and P. Walsh. 2018. "Improving Water Quality in an Iconic Estuary: An
Internal Meta-analysis of Property Value Impacts Around the Chesapeake Bay." Environmental and
Resource Economics 69: 265-292.
Kolker, A., J.C. Quick, C.L. Senior, and H.E. Belkin. 2012. Mercury and halogens in coal - their role in
determining mercury emissions from coal combustion. USGS. Reston, VA.
Kopp, R.J., A.J. Krupnick, and M. Toman. 1997. Cost-Benefit Analysis and Regulatory Reform: An
Assessment of the Science and the Art. Report to the Commission on Risk Assessment and Risk
Management.
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.
Research report. Health Effects Institute; 2009. p. 5-114. (discussion 115-36).
Kung, M., D. Guignet, and P. Walsh 2018. Comparing Pollution Where You Live and Play: A Hedonic
Analysis of Enterococcus in the Long Island Sound. Manuscript under review.
EPA-821-R-19-011
15-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
Lanphear, B.P., R. Hornung, J. Khoury, et al. 2005. Low-level Environmental Lead Exposure and children's
Intellectual Function: An International Pooled Analysis. Environmental Health Perspectives 113(7):
894-899. As cited in ATSDR (2007).
Leggett, C.G. and N.E. Bockstael. 2000. Evidence of the Effects of Water Quality on Residential Land Prices.
Journal of Environmental Economics and Management 39: 121-144.
Lepeule, J., F. Laden, D. Dockery, and J. Schwartz. 2012. Chronic exposure to fine particles and mortality: an
extended follow-up of the Harvard Six Cities study from 1974 to 2009. Environmental Health
Perspectives 120(7), 965-70.
Lin, D., R. Lutter, and C.J. Ruhm. 2018. Cognitive performance and labour market outcomes. Labour
Economics 51, 121-135.
Liu, T., J. J. Opaluch and E. Uchida. 2017. "The impact of water quality in Narragansett Bay on housing
prices." Water Resources Research 53(8), 6456-6471.
Martin, J.A., B.E. Hamilton, M.J.K. Osterman, A.K. Driscoll, and T.J. Matthews. 2017. National Vital
Statistics Reports: Volume 66, Number 1. Births: Final Data for 2015.
McTigue, N., K. Graf, & R. Brown. 2014. Occurrence and consequences of increased bromide in drinking
water sources. Journal of the American Water Works Association, 106(11), E492-E508.
Menke, A., Muntner, P., Batuman, V., Silbergeld, E. K., & Guallar, E. 2006. Blood lead below 0.48
micromol/L (10 microg/dL) and mortality among US adults. Circulation, 774(13), 1388-1394. doi:
10.1161 /circulationaha. 106.628321
Mergler, D., H.A. Anderson, L.H.M. Chan, K.R. Mahaffey, M. Murray, M. Sakamoto, and A.H. Stern. 2007.
Methylmercury Exposure and Health Effects in Humans: A Worldwide Concern. AMBIO: A Journal
of the Human Environment 36: 3-11.
Miller, B. G., and J. F. Hurley. 2003. "Life table methods for quantitative impact assessments in chronic
mortality." Journal of Epidemiology & Community Health 57.3: 200-206.
Mitchell, R.C. and R.T. Carson. 1989. Using Surveys to Value Public Goods: The Contingent Valuation
Method. Resources for the Future, Washington, DC.
Moeltner, K., K.J. Boyle, and R.W. Paterson. 2007. "Meta-Analysis and Benefit Transfer for Resource
Valuation—Addressing Classical Challenges with Bayesian Modeling." Journal of Environmental
Economics and Management 53(2): 250-269.
Munns and Mitro. 2006. Assessing Risks to Populations at Superfund and RCRA Sites Characterizing Effects
on Populations. U.S. Environmental Protection Agency Office of Research and Development.
National Health and Environmental Effects Research Laboratory. Atlantic Ecology Division.
Narragansett, Rhode Island. EPA 600-R-06-038.
National Academies of Sciences, Engineering, and Medicine. 2017. Valuing Climate Damages: Updating
Estimation of the Social Cost of Carbon Dioxide. National Academy Press: Washington, D.C.,
January 2017. http://www.nap.edu/catalog/24651/valuingclimate-changes-updating-estimation-of-the-
social-cost-of
EPA-821-R-19-011
15-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
National Marine Fisheries Service (NMFS) of the National Oceanic and Atmospheric Administration
(NOAA). 2014a. Fish Stock Sustainability Index: 2013 Quarter 4 Update Through December 31,
2013. Silver Spring, Maryland, U.S. Department of Commerce, NOAA, National Marine Fisheries
Service. Available at:
http://www.nmfs.noaa.gov/sfa/fisheries eco/status of fisheries/status updates.html
National Marine Fisheries Service (NMFS) of the National Oceanic and Atmospheric Administration
(NOAA). 2014b. NOAA Fisheries Geographical Information Systems, Fisheries Data: Critical
Habitat. Available at: http://www.nmfs.noaa.gov/gis/data/listed.htm
National Marine Fisheries Service (NMFS) of the National Oceanic and Atmospheric Administration
(NOAA). 2014c. "NOAA National Marine Fisheries Service Southwest Regional Office, GIS Data."
from http://swr.nmfs.noaa.gov/salmon/layers/finalgis.htm.
National Oceanic and Atmospheric Administration (NOAA). 2010. Environmental Sensitivity Index (ESI)
Maps, Office of Response and Restoration.
National Research Council. 2011. Review of the environmental protection agency's draft IRIS assessment of
formaldehyde. National Academies Press, 2011.
National Toxicology Program. 2012. Prepublication Copy: NTP Monograph on Health Effects of Low-Level
Lead.
National Toxicology Program. 2018. Report on Carcinogens: Monograph on Haloacetic Acids Found as
Water Disinfection By-Products. Office of the Report on Carcinogens, Division of the National
Toxicology Program, National Institute of Environmental Health Sciences, US Department of Health
and Human Services. Research Triangle Park, NC. ISSN# 2331-267X, 200 pp.
NatureServe. 2014. "NatureServe Explorer: An Online Encyclopedia of Life." from
http: //www .nature serve. org/cxplorcr/.
Netusil, N. R., M. Kincaid and H. Chang. 2014. "Valuing water quality in urban watersheds: A comparative
analysis of Johnson Creek, Oregon, and Burnt Bridge Creek, Washington." Water Resources
Research 50(5): 4254-4268.
Nordhaus, W. 2014. "Estimates of the Social Cost of Carbon: Concepts and Results from the DICE-2013R
Model and Alternative Approaches." Journal of the Association of Environmental and Resource
Economists, 1(1/2): 273-312.
Nordhaus, William D. 2017. "Revisiting the social cost of carbon." Proceedings of the National Academy of
Sciences of the United States, 114 (7): 1518-1523.
Northeast Fisheries Science Center (NEFSC) of NOAA National Marine Fisheries Service. 2010. "Northeast
Fisheries Science Center Publications." from http://www.nefsc.noaa.gov/publications/.
Occupational Safety and Health Administration (OSHA). Undated. Occupational Exposure to Respirable
Crystalline Silica ~ Review of Health Effects Literature and Preliminary Quantitative Risk
Assessment
EPA-821-R-19-011
15-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory Analysis. September 2003.
Available at https://www.whitehouse.gov/sites/whitehouse.gov/files/omb/circulars/A4/a-4.pdf
Office of Management and Budget (OMB). 2004. Revised Information Quality Bulletin for Peer Review.
April 15, 2004. Available at
https://www.whitehouse.gov/sites/whitehouse.gov/files/omb/assets/OMB/inforeg/peer review041404
.pdf
Pacific Islands Fisheries Science Center (PIFSC) of the NOAA National Marine Fisheries Service. 2010a.
"Fishery Biology and Stock Assessment Division." from http://www.pifsc.noaa.gov/fbsad/index.php.
Pacific Islands Fisheries Science Center (PIFSC) of the NOAA National Marine Fisheries Service. 2010b.
Pacific Islands Fisheries Science Center Staff Publications Database.
Pawel, D. J., and J. S. Puskin. 2004. "The US Environmental Protection Agency's assessment of risks from
indoor radon." U.S. Environmental Protection Agency Office of Radiation and Indoor Air. Health
physics 87.1 (2004): 68-74.
Poor P.J., K.J. Boyle, L.O. Taylor, Bouchard R. 2001. Water clarity in hedonic property value models. Land
Economics 77:482-493.
Regli, S., J. Chen, M. Messner, M.S. Elovitz, F.J. Letkiewicz, R.A. Pegram, T.J. Pepping, S.D. Richardson,
J.M. Wright. 2015. Estimating potential increased bladder cancer risk due to increased bromide
concentrations in sources of disinfected drinking waters. Environmental Science & Technology,
49(22), 13094-13102.
Revesz R.L., J.A. Schwartz., P.H. Howard Peter H., K. Arrow, M.A. Livermore, M. Oppenheimer, and T.
Sterner Thomas. 2017. The social cost of carbon: A global imperative. Review of Environmental
Economics and Policy, 11(1): 172-173.
Richardson, S. D., Plewa, M. J., Wagner, E. D., Schoeny, R., and Demarini, D. M. 2007. Occurrence,
genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in drinking
water: a review and roadmap for research. Mutation Research, 636(1-3), 178-242.
Richardson, L. and J. Loomis. 2009. "The total economic value of threatened, endangered and rare species:
An updated meta-analysis." Ecological Economics 68(5): 1535-1548.
Rivin, Gabe. 2015. Duke Energy Required to Pay Towns, Cities for Degraded Water. North Carolina Health
News, https://www.northcarolinahealthnews.org/2015/06/19/duke-energv-required-to-pav-towns-
cities-for-degraded-water/. Accessed 2019.
Rockett, Ian RH. 2010. "Eliminating injury: an international life table analysis.". Centers for disease control
and prevention. Available at: www.cdc.gov/nchs/data/ice/ice95v2/cl4.pdf.
Roe, G., and M. Baker. 2007. "Why is climate sensitivity so unpredictable?" Science, 318:629- 632.
EPA-821-R-19-011
15-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
Rosenberger, R.S. and R.J. Johnston. 2007. Selection Effects in Meta-Valuation Function Transfers. Meta-
Analysis in Economics Conference, September 27-30, 2007, Sandbjerg Manor, Sonderborg,
Denmark.
Rosenberger, R.S. and T.D. Stanley. 2006. "Measurement, Generalization, and Publication: Sources of Error
in Benefit Transfers and Their Management." Ecological Economics 60(2): 372-378.
Rosenberger, R.S. and T.T. Phipps. 2007. "Correspondence and Convergence in Benefit Transfer Accuracy:
Meta-Analytic Review of the Literature." In: Navrud, S. and R. Ready (eds.), Environmental Value
Transfer: Issues and Methods. Dordrecht, The Netherlands: Springer, pp.23-43.
Ruhl, L., A. Vengosh, G.S. Dwyer, H. Hsu-Kim, G. Schwartz, A. Romanski, S.D. Smith. 2012. The Impact of
Coal Combustion Residue Effluent on Water Resources: A North Carolina Example. Environmental
Science and Technology 2012 Oct 15.
Salkever, D.S. 1995. Updated Estimates of Earnings Benefits from Reduced Exposure of Children to
Environmental Lead. Environmental Research 70: 1-6.
Schwartz, J. 1994. Low-Level Lead Exposure and Children's IQ: A Meta-analysis and Search for a
Threshold. Environmental Research 65: 42-55.
Sea Grant, Illinois-Indiana. 2018. Lake Michigan anglers boost local Illinois and Indiana economies.
https://iiseagrant.org/lake-michigan-anglers-boost-illinois-and-indiana-local-economies/. Accessed
2019.
Shrestha, R., R. Rosenberger and J. Loomis. 2007. "Benefit Transfer Using Meta-Analysis in Recreation
Economic Valuation, in Benefit Transfer Accuracy: Meta-Analytic Review of the Literature." In:
Navrud, S. and R. Ready (eds.). Environmental Value Transfer: Issues and Methods. Dordrecht, The
Netherlands: Springer, pp.161-177.
Smith, V.K., G. Van Houtven, and S.K. Pattanayak. 2002. "Benefit Transfer via Preference Calibration:
'Prudential Algebra' for Policy." Land Economics 78(1): 132-152.
Sohngen, B., W. Zhang, J. Bruskotter, and B. Sheldon. 2015. "Results from a 2014 survey of Lake Erie
anglers." Columbus, OH: The Ohio State University, Department of Agricultural, Environmental and
Development Economics and School of Environment & Natural Resources.
Southeast Fisheries Science Center (SEFSC) of NOAA National Marine Fisheries Service. 2010. SouthEast
Data, Assessment and Review (SEDAR).
Southwest Fisheries Science Center (SWFSC) of NOAA National Marine Fisheries Service. 2010.
"Southwest Fisheries Science Center Fisheries Biology Publications Database." from
http://swfsc.noaa.gov//publications/fedbin/qrvpublications.asp?ParentMenuId=54
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.
Taylor, C. A., G. A. Schuster, et al. 2007. "A Reassessment of the Conservation Status of Crayfishes of the
United States and Canada after 10+ Years of Increased Awareness." Fisheries 32(8): 372-389.
EPA-821-R-19-011
15-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
Taylor, C. A., M. L. Warren, et al. 1996. "Conservation Status of Crayfishes of the United States and
Canada." Fisheries 21(4): 25-38.
Tuttle, C.M., and M.D. Heintzelman. 2014. "A Loon on Every Lake: A Hedonic Analysis of Lake Water
Quality in the Adirondacks." Resource and Energy Economics 39: 1-15.
United States of America vs. Duke Energy Business Services, LLC, Duke Energy Carolinas, LLC, Duke
Energy Progress, Inc (U.S. vs. Duke Energy). 2015. United States District Court for the Eastern
District of North Carolina Western Division. No. 5:15-CR-62-H. No. 5:15-CR-67-H. No. 5:15-CR-
68-H. Joint Factual Statement, https://www,iustice.gov/usao-ednc/file/771581/download. Accessed
2019.
U.S. Army Corps of Engineers (USACE). (2013). Dredging Information System.
United States Census Bureau. 2016. American Community Survey Demographic and Housing Estimates.
DP05: 2012-2016 American Community Survey 5-Year Estimates. Retrieved from
http ://factfinder.census. gov/
United States Census Bureau. 2017a. National Population Projections Tables: 2017 to 2060. Retrieved from
https://www.census.gov/data/tables/2017/demo/popproi/2017-summarv-tables.html
United States Census Bureau. 2017b. American Community Survey Demographic and Housing Estimates.
DP05: 2013-2017 American Community Survey 5-Year Estimates. Retrieved from
http ://factfinder.census. gov/
U.S. Census Bureau. 2018. Annual Estimates of the Resident Population for Selected Age Groups by Sex for
the United States, States, Counties and Puerto Rico Commonwealth and Municipios: April 1, 2010 to
July 1, 2017. Retrieved from
https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=PEP_2017_PEPA
GESEX&prodType=table.
United States Department of Energy (U.S. DOE). 2017. Form EIA-860 Detailed Data: Final 2016 Data.
Released November 9, 2017.
United States Department of Health and Human Services (U.S. DHHS). 2012. Toxicological Profile for
Cadmium.
United States Environmental Protection Agency (U.S. EPA). 1990. Review of the National Ambient Air
Quality Standards for Lead: Assessment of Scientific and Technical Information. OAQPS Staff
Paper. EPA-450/2-89-022.
United States Environmental Protection Agency (U.S. EPA). 2003. Economic and Environmental Benefits
Analysis Document for the Final Effluent Limitations Guidelines and Standards for the Metal
Products and Machinery Point Source Category. EPA-821-B-03-002.
United States Environmental Protection Agency (U.S. EPA). 2005a. Economic Analysis for the Final Stage 2
Disinfectants and Disinfection Byproducts Rule. EPA 815-R-05-010.
United States Environmental Protection Agency (U.S. EPA). 2005b. Regulatory Impact Analysis of the Clean
Air Mercury Rule. EPA-452/R-05-003.
EPA-821-R-19-011
15-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
United States Environmental Protection Agency (U.S. EPA). 2005c. Drinking Water Criteria Document for
Brominated Trihalomethanes. EPA 822-R-05-011. November 15, 2005.
United States Environmental Protection Agency (U.S. EPA). 2007. Regulatory Impact Analysis of the
Proposed Revisions to the National Ambient Air Quality Standard. EPA 452/R-07-008. July 2007.
United States Environmental Protection Agency (U.S. EPA). 2008. Welcome to STORET, EPA's largest
computerized environmental data system. Available at http://www.epa.gov/storet/.
United States Environmental Protection Agency (U.S. EPA). 2008b. Final Ozone NAAQS Regulatory Impact
Analysis. EPA 452/R-08-003. March 2008.
United States Environmental Protection Agency (U.S. EPA). 2009a. Environmental Impact and Benefits
Assessment for Final Effluent Guidelines and Standards for the Construction and Development
Category. November, 2009. http://www.epa.gov/guide/construction/files/environment.pdf.
United States Environmental Protection Agency (U.S. EPA). 2009b. Integrated Science Assessment for
Particulate Matter (Final Report). U.S. Environmental Protection Agency, Washington, DC,
EPA/600/R-08/139F, 2009.
United States Environmental Protection Agency (U.S. EPA). 2009c. Software and User's Manuals,
http: //www .epa. gov/superfiind/lead/products .htm
United States Environmental Protection Agency (U.S. EPA). 2009d. Steam Electric Power Generating Point
Source Category: Final Detailed Study. EPA 821-R-09-008.
United States Environmental Protection Agency (U.S. EPA). 2010a. Guidelines for Preparing Economic
Analyses. EPA 240-R-10-001.
United States Environmental Protection Agency (U.S. EPA). 2010b. Integrated Risk Information System
(IRIS) Toxicological Review of Inorganic Arsenic (Cancer) (External Draft Review).
http://cfbub.epa.gov/ncea/iris drafts/recordisplav.cfm?deid=219111
United States Environmental Protection Agency (U.S. EPA). 2010c. Questionnaire for the Steam Electric
Power Generating Effluent Guidelines, OMB Control 2040-0281, May 20, 2010.
United States Environmental Protection Agency (U.S. EPA). 2011. Exposure Factors Handbook, 2011
Edition (Final). U.S. Environmental Protection Agency, Washington, DC. EPA-600-R-09-025F.
United States Environmental Protection Agency (U.S. EPA). 2013a. Benefit and Cost Analysis for the
Proposed Effluent Limitations Guidelines and Standards for the Steam Electric Power Generating
Point Source Category. EPA-821-R-13-004.
United States Environmental Protection Agency (U.S. EPA). 2013b. Integrated Science Assessment for Lead.
(EPA/600/R-10/075F). Research Triangle Park, NC.
United States Environmental Protection Agency (U.S. EPA). 2013b. Integrated Science Assessment of Ozone
and Related Photochemical Oxidants (Final Report). U.S. Environmental Protection Agency,
Washington, DC, EPA/600/R-10/076F, 2013.
EPA-821-R-19-011
15-10
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
United States Environmental Protection Agency (U.S. EPA). 2013c. Fiscal Year 2011 Drinking Water and
Ground Water Statistics. Washington, DC, EPA 816-R-13-003, March 2013
United States Environmental Protection Agency (U.S. EPA). 2015a. Benefit and Cost Analysis for the
Effluent Limitations Guidelines and Standards for the Steam Electric Power Generating Point Source
Category. EPA 821-R-15-005
United States Environmental Protection Agency (U.S. EPA). 2015b. Environmental Assessment for the
Effluent Limitations Guidelines and Standards for the Steam Electric Power Generating Point Source
Category (40 CFR 423). EPA 821-R-15-006
United States Environmental Protection Agency (U.S. EPA). 2016. Six-Year Review 3 Technical Support
Document for Disinfectants/Disinfection Byproducts Rules. Office of Water. EPA-810-R-16-012.
https://www.epa. gov/sites/production/files/2016-12/documents/81 Or 16012.pdf
United States Environmental Protection Agency (U.S. EPA). 2017a. Notes from Site Visit to Harris Treatment
Plant on December 12, 2017.
United States Environmental Protection Agency (U.S. EPA). 2017b. Chapter 3: Water Quality Criteria.
Water Quality Standards Handbook. EPA 823-B-17-001. 2017.
https://www.epa.gov/sites/production/files/2014-10/documents/handbook-chapter3.pdf
United States Environmental Protection Agency (U.S. EPA). 2018a. National Primary Drinking Water
Regulations, https://www.epa.gov/ground-water-and-drinking-water/national-primarv-drinking-
water-re gulations
United States Environmental Protection Agency (U.S. EPA). 2018b. National Recommended Water Quality
Criteria - Human Health Criteria Table, https://www.epa.gov/wqc/national-recommended-water-
quality-criteria-human-health-criteria-table
United States Environmental Protection Agency (U.S. EPA). 2018c. National Hydrography Dataset Plus
(NHDPlus). Available at
https://s3.ainazonaws.com/nhdpliis/NHDPliisV2l/Documentation/NHDPlusV2 User Guide.pdf.
United States Environmental Protection Agency (U.S. EPA). 2018d. Estimating the Benefit per Ton of
Reducing PM2.5 Precursors from 17 Sectors. Technical Support Document. February 2018.
United States Environmental Protection Agency (U.S. EPA). 2018f. Documentation for EPA's Power Sector
Modeling Platform v6 Using the Integrated Planning Model. May 2018
United States Environmental Protection Agency (U.S. EPA). 2019a. Supplemental Environmental
Assessment for Proposed Revisions to Effluent Limitations Guidelines and Standards for the Steam
Electric Power Generating Point Source Category (EA)
United States Environmental Protection Agency (U.S. EPA). 2019b. Supplemental Technical Development
Document for Proposed Revisions to Effluent Limitations Guidelines and Standards for the Steam
Electric Power Generating Point Source Category
United States Environmental Protection Agency (U.S. EPA). 2019c. Regulatory Impact Analysis for the
Proposed Revisions to the Effluent Limitations Guidelines and Standards for the Steam Electric
Power Generating Point Source Category (RIA)
EPA-821-R-19-011
15-11
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
United States Environmental Protection Agency (U.S. EPA). 2019d. Bromide case study memo.
United States Environmental Protection Agency (U.S. EPA). 2019e. Economic Analysis of the Final Rule to
Revise the TASCA Lead-Dust Hazard Standards. June 2019.
United States Environmental Protection Agency (U.S. EPA). 2019f. Regulatory Impact Analysis for Repeal
of the Clean Power Plan; Emission Guidelines for Greenhouse Gas Emissions from Existing Electric
Utility Generating Units; Revisions to Emission Guidelines Implementing Regulations. EPA-452/R-
19-003. June 2019
United States Environmental Protection Agency (U.S. EPA). 2019g. Screening-level Analysis of
Cardiovascular Health Benefits from Changes in Lead Exposure from Fish Tissue.
United States Environmental Protection Agency (U.S. EPA). 2019h. Economic Analysis of the Proposed
Lead and Copper Rule Revisions. October 2019.
United States Environmental Protection Agency (U.S. EPA). 2019i. Health Risk Reduction and Cost Analysis
of the Proposed Perchlorate National Primary Drinking Water Regulation. EPA 816-R-19-004.
United States Fish and Wildlife Service (U.S. FWS). 2006. National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation. FHW/06-NAT.
United States Fish and Wildlife Service (U.S. FWS). 2010. "North Florida Ecological Services Office:
Loggerhead Sea Turtle (Carettacaretta)." from
http://www.fws.gov/northflorida/seaturtles/turtle%20factsheets/loggerhead-sea-turtle.htm.
United States Fish and Wildlife Service (U.S. FWS). 2014a. "Environmental Conservation Online System."
Available at: http://ecos.fws.gov/
United States Fish and Wildlife Service (U.S. FWS). 2014b. "Critical Habitat Portal." Available
at: http://criticalhabitat.fws.gov/
United States Fish and Wildlife Service (U.S. FWS). 2016. National Survey of Fishing, Hunting, and
Wildlife-Associated Recreation.
United States Geological Survey (USGS). 2002. Enhanced River Reach File 2.0 (E2RF1). Available at
https://water.usgs.gOv/GIS/metadata/usgswrd/XML/erfl_2.xml#stdorder
United States Geological Survey (USGS). 2007. National Hydrography Dataset (NHD). Available at
http://nhd.usgs.gov/data.html
United States Geological Survey (USGS). 2009. RESIS II: An Updated Version of the Original Reservoir
Sedimentation Survey Information System (RESIS).
United States Geological Survey (USGS). 2018. National Hydrography Dataset. Available at
https://www.usgs.gov/core-science-svstems/ngp/national-hvdrographv/national-hvdrographv-
dataset?qt-science support page related con=0#qt-science support page related con
Vaughan, W.J. 1986. "The RFF Water Quality Ladder." In: Mitchell, R.C. and R.T. Carson. The Use of
Contingent Valuation Data for Benefit/Cost Analysis in Water Pollution Control, Final Report.
Washington: Resources for the Future. Appendix B.
EPA-821-R-19-011
15-12
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
15: Cited References
Villanueva, C. M., Fernandez, F., Malats, M., Grimalt, J. O., Kogevinas, M. A. 2003. A meta-analysis of
studies on individual consumption of chlorinated drinking water and bladder cancer. J. Epidemiol
Community Health. 57, 166-173.
Villanueva, C.M., K.P. Cantor, S. Cordier, J.J.K. Jaakkola, W.D. King, D.F. Lynch, S. Porru, and M.
Kogevinas. 2004. Disinfection byproducts and bladder cancer: a pooled analysis. Epidemiology
15(3):357-367.
Viscusi, K.W, J. Huber and J. Bell. 2008. "The Economic Value of Water Quality." Environmental and
Resource Economics 41: 169-187.
Walsh, P. J., J.W. Milon, and D.O. Scrogin. 2011. "The Spatial Extent of Water Quality Benefits in Urban
Housing Markets." Land Economics 87(4): 628-644.
Wathen, J.B., J. M. Lazorchak, A.R. Olsen, and A. Batt. 2015. "A national statistical survey assessment of
mercury concentrations in fillets of fish collected in the U.S. EPA national rivers and streams
assessment of the continental USA." Chemosphere 122(March 2015): 52-61.
Watson, K., Farre, M. J., & Knight, N. 2012. Strategies for the removal of halides from drinking water
sources, and their applicability in disinfection by-product minimisation: A critical review. Journal of
Environmental Management, 110(0), 276-298.
Whittington, D., & MacRae, D. (1986). The Issue of Standing in Cost-Benefit Analysis. Journal of Policy
Analysis and Management, 5(4): 665-682.
WHO. 2009. World Health Organization. Bromide in Drinking-Water. Background Document for
Development of WHO Guidelines for Drinking-Water Quality. WHO/HSE/WSH/09.01/6. Geneva,
Switzerland.
Wilcove, D. S. and L. L. Master. 2005. "How many endangered species are there in the United States?"
Frontiers in Ecology and the Environment 3(8): 414-420.
Williams, C.D., Burns, J.W., Chapman, A.D., Flewelling, L., Pawlowicz, M. and W. Carmichael. 2001.
Assessment of cyanotoxins in Florida's lakes, reservoirs and rivers. White paper produced for Florida
Harmful Algal Bloom Task Force.
Williams, J. D., M. L. Warren, et al. 1993. "Conservation Status of Freshwater Mussels of the United States
and Canada." Fisheries 18(9): 6-22.
Williams, J. E., J. E. Johnson, et al. 1989. "Fishes of North America Endangered, Threatened, or of Special
Concern: 1989." Fisheries 14(6): 2-20.
Williams, R., et al. 2000. An examination of fish consumption by Indiana recreational anglers: An onsite
survey. Technical Report 99-D-HDFW-2. Dept. of Forestry and Natural Resources, Purdue
University.
Woods & Poole Economics Inc. 2012. Complete Demographic Database. Washington, DC.
Zhu, M., E.F. Fitzgerald, K.H. Gelberg, S. Lin, and C.M. Druschel. 2010. "Maternal low-level lead exposure
and fetal growth." Environmental Health Perspectives 118(10): 1471-1475.
EPA-821-R-19-011
15-13
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix A: Changes to Benefits Analysis
Appendix A Changes to Benefits Methodology since 2015 Rule Analysis
The table below summarizes the principal methodological changes the EPA made to analyses of the benefits
of the regulatory options, as compared to the analyses of the 2015 rule. The benefits analysis methodology for
the 2015 rule is detailed in the BCA document (U.S. EPA, 2015a).
Table A-1: Changes to Benefits Analysis Since 2015 Final Rule
Benefits Category
Analysis Component
[2015 rule analysis value]
Changes to Analysis for regulatory options
[2019 rule analysis value]
General assumptions
Dollar year [all costs expressed in 2013
dollars].
Updated dollar year [2018],
Promulgation year [all costs and revenue
streams discounted back to 2015],
Updated promulgation year [2020],
Period of analysis [2019-2042],
Updated period of analysis [2021-2047],
Technology implementation years [2019-
2023],
Technology implementation years
constant across the options for a given
plant.
Updated technology implementation years
[2021-2028],
Technology implementation years vary
between options and plants.
Baseline is current conditions.
Baseline is 2015 ELG.
General pollutant
loadings and
concentrations
Affected reaches based on immediate
receiving reaches and flow paths in
medium-resolution NHDvl.
Updated immediate receiving reaches for
selected plants.
Affected reaches based on updated receiving
reaches and flow paths in medium-resolution
NHD v2.
Risk-Screening Environmental Indicators
(RSEI) modeling of toxics concentrations
in immediate and downstream reaches,
including [2012] TRI non-steam electric
power plant releases.
Transport and dilution calculations for
immediate and downstream reaches, including
[2016] TRI non-steam electric power plant
releases.
Pollutant concentrations based on mean
annual flows in NHDPIus vl.
Pollutant concentrations based on mean annual
flows in NHDPIus v2.
SPARROW modeling of nutrient and
sediment concentrations in receiving and
downstream reaches.
No change.
Assumes loading changes occur at mid-
point of technology implementation
period [2021],
Uses the annual average loadings for analysis
period [2021-2047], assuming that pre-
implementation loads are equal to current
loads.
Human health benefits from changes in exposure to trihalomethanes in drinking water
Public water systems
affected by bromide
discharges
Qualitative discussion.
See Section 3.3 for approach to modeling
changes in trihalomethane concentrations in
drinking water.
Lifetime changes in
incidence of bladder
cancer
Not addressed.
Lifetime risk model (See Section 4.3.3 for
approach to modeling changes in bladder
cancer incidence).
Monetization of
changes in incidence of
bladder cancer
Not addressed
Mortality valued using VSL (U.S. EPA, 2010a).
Morbidity valued based on COI (Greco et al,
2018).
EPA-821-R-19-011
A-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix A: Changes to Benefits Analysis
Table A-1: Changes to Benefits Analysis Since 2015 Final Rule
Benefits Category
Analysis Component
[2015 rule analysis value]
Changes to Analysis for regulatory options
[2019 rule analysis value]
Human health benefits from changes in pollutant exposure in recreationally- and subsistence-caught fish
Exposed populations
Population based on 2010 Census Data
Population for future years is based on
Woods & Poole (2012) forecasts for each
year from 2000 through 2040.
Population based on the 2016 American
Community Survey (U.S. Census Bureau, 2016)
Population for future years is based on U.S.
Census population projections for the United
States: 2017 - 2060.
Census block-focused analysis with [50
miles] travel distance.
No change
IQ losses in children
from lead exposure
Blood lead level - IEUBK.
No change
IQ losses - linking blood lead level to IQ
based on Lanphear et al. (2005).
Linking blood lead level to IQ based on Crump
et al. (2013)
Monetization (value of an IQ point) is
based on Salkever (1995) and Schwartz
(1994).
Used a modified Salkever (1995) IQ point value
(based on more recent data from the 1997
National Longitudinal Survey of Youth) (U.S.
EPA, 2019e); removed Schwartz (1994)
estimate; added sensitivity analysis using Lin et
al. (2018) value.
Cardiovascular disease
(CVD) in adults from
lead exposure
Blood lead level - Legget model.
EPA did not quantify and monetize this benefit
category given the small changes in exposure.
Linking blood lead level to CVD mortality
(Menke et al., 2006).
CVD quantification framework - lifetime
table approach.
Monetization - VSL.
IQ losses in infants from
mercury exposure
IQ losses - linking maternal mercury hair
content and subsequent childhood IQ loss
from Axelrad et al. (2007).
No change.
Monetization (value of an IQ point) is
based on Salkever (1995) and Schwartz
(1994).
Used a modified Salkever (1995) IQ point value
(based on more recent data from the 1997
National Longitudinal Survey of Youth) (U.S.
EPA, 2019e); removed Schwartz (1994)
estimate; added sensitivity analysis using Lin et
al. (2018) value.
Avoided cancer cases
from arsenic exposure
Main analysis - cancer slope factor (CSF)
based on incidences of skin cancer;
monetization - cost of illness (COI).
Main analysis: No change to the approach.
Sensitivity analysis - CSF for lung and
bladder cancer; monetization - VSL.
Sensitivity analysis: Did not perform since the
estimated change in pollutant load is small and
the estimated change in cancer cases is
essentially zero.
Non-market benefits from water quality improvements
Willingness-to-pay for
water quality
improvements
8-parameter water quality index toxics
subindex (arsenic, chromium, lead,
manganese, mercury, nickel, selenium,
and zinc.
9-parameter water quality index toxics
subindex (arsenic, chromium, copper, lead,
manganese, mercury, nickel, selenium, and
zinc).
Length weighted average AWQI for
reaches within [100 miles] of census
blocks.
No change.
EPA-821-R-19-011
A-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix A: Changes to Benefits Analysis
Table A-1: Changes to Benefits Analysis Since 2015 Final Rule
Benefits Category
Analysis Component
[2015 rule analysis value]
Changes to Analysis for regulatory options
[2019 rule analysis value]
Affected population consists of all
households in a given Census Block Group
(CBG). Households value all water quality
changes in a 100-mile radius.
No change to the approach.
Updated:
• Population - 2016 American Community
Survey (U.S. Census Bureau, 2016)
• Population growth - U.S. Census population
projections for the United States: 2017 -
2060
• Universe of the CBGs included in the analysis
to reflect changes in the universe of affected
reaches and changes in CBGs delineation
between 2010 and 2016.
Meta-regression model includes spatial
characteristics of the affected water
resources: size of the market, waterbody
characteristics (length and flow),
availability of substitute sites, land use
type in the abutting counties.
Variables characterizing the availability of
substitute site, size of the market, and land-use
were revised based on changes in the universe
of receiving reaches and CBGs included in the
analysis.
Effects on endangered
and threatened (T&E)
species
Categorical analysis based on habitat
overlap/proximity.
No change based on updated review of habitat
overlap.
Monetization based on meta-analysis of
willingness-to-pay to protect T&E species
(Richardson and Loomis 2009).
Qualitative discussion based on magnitude of
impacts.
Effects on groundwater
quality
Discussed qualitatively.
Not included in the analysis due to
promulgation of the CCR rule.
Air-related effects
Emissions changes
Emissions from changes in electricity
generation profile from 2015 IPM runs.
Emissions from changes in electricity
generation profile from 2018 IPM runs.
Transportation- and energy use-associated
emissions were updated to reflect new
technology basis for the options. Updates were
made to reflect new universe of facilities and
technology impacts.
Monetization
National average benefit-per-ton
estimates for S02 and NOx from Fann and
Fulcher (2012), single estimate.
Qualitative discussion.
Global social cost of carbon (SCC) value
from Interagency Working Group on
Social Cost of Carbon (IWGSCC 2013 a,b;
2015).
Domestic social cost of carbon (SC-C02) value
(see Appendix /).
Economic productivity
Impoundment releases
Reduced risk of impoundment releases
due to changes in the use of
impoundment.
Not included in the analysis due to
promulgation of the CCR rule.
Avoided cost of clean-up, natural resource
damages and transaction costs.
EPA-821-R-19-011
A-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix A: Changes to Benefits Analysis
Table A-1: Changes to Benefits Analysis Since 2015 Final Rule
Benefits Category
Analysis Component
[2015 rule analysis value]
Changes to Analysis for regulatory options
[2019 rule analysis value]
Changes in dredging
costs
Use SPARROW for estimating sediment
deposition.
No Change.
Navigational dredging locations from
USACE database (2013).
Reservoir locations from E2RF1 network
(SPARROW).
Cost of dredging based on USACE data
(2013).
Beneficial use of ash
Reduced disposal costs and avoided life-
cycle impacts from displaced virgin
material.
Qualitative discussion due to de minimis
changes in marketable ash tonnage.
Changes in groundwater
withdrawals
Increased availability of groundwater
resources.
No change (beyond updates to changes in
withdrawals).
Avoided cost of drinking water purchase.
Tourism, commercial
fisheries, property
values, surface water
withdrawals.
Qualitative discussion.
No change.
EPA-821-R-19-011
A-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Appenc WQI Calculation and Regional Subindices
B.1 WQI Calculation
The first step in the implementation of the WQI involves obtaining water quality levels for each parameter,
and for each waterbody, under both the baseline conditions and each option. Some parameter levels are field
measurements while others are modeled values.
The second step involves transforming the parameter measurements into subindex values that express water
quality conditions on a common scale of 0 to 100. The EPA used the subindex transformation curves
developed by Dunnette (1979) and Cude (2001) for the Oregon WQI for BOD, DO, and FC. For TSS, TN,
and TP concentrations, the EPA adapted the approach developed by Cude (2001) to account for the wide
range of natural or background nutrient and sediment concentrations that result from the variability in
geologic and other region-specific conditions, and to reflect the national context of the analysis. TSS, TN, and
TP subindex curves were developed for each Level III ecoregion (U.S. EPA, 2009a) using pre-compliance
(before the implementation of the 2015 rule) TSS, TN, and TP concentrations calculated in SPARROW at the
E2RF1 reach level.82'83'84 For each of the 85 Level III ecoregions intersected by the E2RF1 reach network, the
EPA derived the transformation curves by assigning a score of 100 to the 25th percentile of the reach-level
TSS concentrations in the ecoregion (i.e.. using the 25th percentile as a proxy for "reference" concentrations),
and a score of 70 to the median concentration. An exponential equation was then fitted to the two
concentration points following the approach used in Cude (2001).
For this analysis, the EPA also used a toxics-specific subindex curve based on the number of NRWQC
exceedances for toxics in each waterbody. National freshwater chronic NRWQC values are available for
arsenic, cadmium, chromium, copper, lead, mercury, nickel, selenium, and zinc. To develop this subindex
curve, the EPA used an approach developed by the Canadian Council of Ministers of the Environment
(CCME, 2001). The CCME water quality index is based on three attributes of water quality that relate to
water quality objectives: scope (number of monitored parameters that exceed water quality standard or
toxicological benchmark); frequency (number of individual measurements that do not meet objectives,
relative to the total number of measurements for the time period of interest) and amplitude (i.e., amount by
which measured values exceed the standards or benchmarks). Following the CCME approach, the EPA's
toxics subindex considers the number of parameters with exceedances of the relevant water quality criterion.
With regards to frequency, the EPA modeled long-term annual average concentrations in ambient water, and
therefore any exceedance of an NRWQC may indicate that ambient concentrations exceed NRWQC most of
82 The SPARROW model was developed by the USGS for the regional interpretation of water-quality monitoring data. The model
relates in-stream water-quality measurements to spatially referenced characteristics of watersheds, including contaminant sources
and factors influencing terrestrial and aquatic transport. SPARROW empirically estimates the origin and fate of contaminants in
river networks and quantifies uncertainties in model predictions. More information on SPARROW can be found at
http://water.usgs.gOv/nawqa/sparrow/FAQs/faq.html#l
83 The EPA's E2RF1 is a digital stream networks used in SPARROW models. This dataset extends over the continental United
States and includes approximately 62,000 stream reaches.
84 Following the approach the EPA used for the analysis of the Construction and Development Effluent Guidelines and Standards
(40 CFR Part 450) final rule in 2009 (74 FR 62995), the selected data exclude outlier TSS concentrations, defined as values that
exceed the 95th percentile based on the universe of all E2RF1 reaches modeled in SPARROW (U.S. EPA, 2009a). In the
Construction and Development ELG analysis, the USGS and the EPA had determined that these outlier values corresponded to
headwater reaches and were an artifact of the model rather than expected concentrations.
EPA-821-R-19-011
B-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
the time (assumed to be 100 percent of the time). The EPA did not consider amplitude, because if the annual
average concentration exceeds the chronic NRWQC then the water is impaired for that constituent and the
level of exceedance is of secondary concern. Using this approach, the subindex curve for toxics assigns the
lowest subindex score of 0 to waters where exceedances are observed for all nine of the toxics analyzed, and a
maximum score of 100 to waters where there are no exceedances. Intermediate values are distributed evenly
between 0 and 100.
Table B-l presents parameter-specific functions used for transforming water quality data into water quality
subindices for freshwater waterbodies for the six pollutants with individual subindices. Table B-2 presents the
subindex values for toxics. The equation parameters for each of the 85 ecoregion-specific TSS, TN, and TP
subindex curves are provided in the next section. The curves include threshold values below or above which
the subindex score does not change in response to changes in parameter levels. For example, improving DO
levels from 10.5 mg/L to 12 mg/L or from 2 mg/L to 3.3 mg/L would result in no change in the DO subbindex
score.
Table B-1: Freshwater Water Quality Subindices
Parameter
Concentrations
Concentration
Unit
Subindex
Dissolved Oxygen (DO)
DO saturation <100%
DO
DO < 3.3
mg/L
10
DO
3.3 < DO < 10.5
mg/L
-80.29+31.88xDO-1.401xDO2
DO
DO > 10.5
mg/L
100
100% < DO saturation < 275%
DO
NA
mg/L
100 x exp((DOsat - 100) x -1.197xl0"2)
275% < DO saturation
DO
NA
mg/L
10
Fecal Coliform (FC
FC
FC > 1,600
cfu/100 mL
10
FC
50 < FC < 1,600
cfu/100 mL
98 x exp((FC - 50) x -9.9178xl0"4)
FC
FC < 50
cfu/100 mL
98
Total Nitrogen (TN
a
TN
TN >TNio
mg/L
10
TN
TNioo ^ TN < TNio
mg/L
a x exp(TNxb); where a and b are ecoregion-
specific values
TN
TN < TNioo
mg/L
100
Total Phosphorus (TP)b
TP
TP > TPio
mg/L
10
TP
TPioo < TP < TPio
mg/L
a x exp(TPxb); where a and b are ecoregion-
specific values
TP
TP < TPioo
mg/L
100
Total Suspended Solids (TSS)C
TSS
TSS > TSSio
mg/L
10
TSS
TSSioo < TSS < TSSio
mg/L
a x exp(TSSxb); where a and b are ecoregion-
specific values
TSS
TSS < TSSioo
mg/L
100
EPA-821-R-19-011
B-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Table B-1: Freshwater Water Quality Subindices
Parameter
Concentrations
Concentration
Unit
Subindex
Biochemical Oxygen Demand, 5-day (BOD)
BOD
BOD >8
mg/L
10
BOD
BOD <8
mg/L
100 xexp(BODx-0.1993)
a. TN10 and TNIOO are ecoregion-specific TN concentration values that correspond to subindex scores of 10 and 100,
respectively. Use of 10 and 100 for the lower and upper bounds of the WQI subindex score follow the approach in Cude (2001)
b. TP10 and TP100 are ecoregion-specific TP concentration values that correspond to subindex scores of 10 and 100, respectively.
Use of 10 and 100 for the lower and upper bounds of the WQI subindex score follow the approach in Cude (2001)
c. TSS10 and TSS100 are ecoregion-specific TSS concentration values that correspond to subindex scores of 10 and 100,
respectively. Use of 10 and 100 for the lower and upper bounds of the WQI subindex score follow the approach in Cude (2001)
Source: EPA analysis using methodology in Cude (2001).
Table B-2: Freshwater Water Quality Subindex for Toxics
Number of Toxics with NRWQC
Subindex
Exceedances
0
100.0
1
88.9
2
77.8
3
66.7
4
55.6
5
44.4
6
33.3
7
22.2
8
11.1
9
0.0
The final step in implementing the WQI involves combining the individual parameter subindices into a single
WQI value that reflects the overall water quality across the parameters. The EPA calculated the overall WQI
for a given reach using a geometric mean function and assigned all WQ parameters an equal weight of 0.143
(l/7th of the overall score). Unweighted scores for individual metrics of a WQI have previously been used in
Cude (2001), CCME (2001), and Carruthers and Wazniak (2003).
Equation B-1 presents the EPA's calculation of the overall WQI score.
Equation B-1.
woi,=na'"'
7=1
WQIr = the multiplicative water quality index (from 0 to 100) for reach r
Q = the water quality subindex measure for parameter z
Wi = the weight of the z-th parameter (0.143)
n = the number of parameters (z. e., seven)
EPA-821-R-19-011 B-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
B.2 Regional Subindices
The following tables provide the ecoregion-specific parameters used in estimating the TSS, TN, or TP water
quality subindex, as follows:
- If [WQ Parameter] < WQ Parameter 100 Subindex =100
- If WQ Parameter 100 < [WQ Parameter] < WQ Parameter 10 Subindex = a exp(b [WQ Parameter])
- If [WQ Parameter] > WQ Parameter 10 Subindex = 10
Where [WQ Parameter] is the measured concentration of either TSS, TN, or TP and WQ Parameter i0, WQ
Parameter k>o, a, and b are specified in Table B-3 for TSS, Table B-4 for TN, and Table B-5 for TP.
Table B-3: TSS Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TSSioo
TSSio
10.1.2
Columbia Plateau
126.56
-0.0038
63
668
10.1.3
Northern Basin and Range
112.42
-0.0007
160
3,457
10.1.4
Wyoming Basin
123.36
-0.0010
220
2,513
10.1.5
Central Basin and Range
121.22
-0.0018
109
1,386
10.1.6
Colorado Plateaus
144.44
-0.0010
363
2,670
10.1.7
Arizona/New Mexico Plateau
126.76
-0.0004
668
6,349
10.1.8
Snake River Plain
146.39
-0.0027
142
994
10.2.1
Mojave Basin and Range
119.34
-0.0015
121
1,653
10.2.2
Sonoran Desert
112.39
-0.0002
567
12,097
10.2.4
Chihuahuan Desert
214.39
-0.0005
1,419
6,130
11.1.1
California Coastal Sage, Chaparral, and Oak Woodlands
127.97
-0.0012
205
2,124
11.1.2
Central California Valley
171.86
-0.0044
122
646
11.1.3
Southern and Baja California Pine-Oak Mountains
115.12
-0.0007
197
3,491
12.1.1
Madrean Archipelago
261.35
-0.0005
2,053
6,527
13.1.1
Arizona/New Mexico Mountains
120.98
-0.0004
477
6,233
15.4.1
Southern Florida Coastal Plain
116.95
-0.0405
4
61
5.2.1
Northern Lakes and Forests
157.76
-0.0233
20
118
5.2.2
Northern Minnesota Wetlands
154.99
-0.0186
24
147
5.3.1
Northern Appalachian and Atlantic Maritime Highlands
174.99
-0.0261
21
110
5.3.3
North Central Appalachians
245.15
-0.0176
51
182
6.2.10
Middle Rockies
144.64
-0.0038
98
703
6.2.11
Klamath Mountains
238.90
-0.0068
129
467
6.2.12
Sierra Nevada
185.36
-0.0116
53
252
6.2.13
Wasatch and Uinta Mountains
124.28
-0.0014
160
1,800
6.2.14
Southern Rockies
153.42
-0.0031
140
881
6.2.15
Idaho Batholith
184.23
-0.0142
43
205
6.2.3
Columbia Mountains/Northern Rockies
180.70
-0.0168
35
172
6.2.4
Canadian Rockies
396.62
-0.0308
45
119
6.2.5
North Cascades
240.95
-0.0193
46
165
6.2.7
Cascades
192.94
-0.0181
36
164
EPA-821-R-19-011
B-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Table B-3: TSS Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TSSioo
TSS10
6.2.8
Eastern Cascades Slopes and Foothills
178.82
-0.0145
40
199
6.2.9
Blue Mountains
148.35
-0.0037
107
729
7.1.7
Strait of Georgia/Puget Lowland
181.06
-0.0224
27
129
7.1.8
Coast Range
174.78
-0.0114
49
251
7.1.9
Willamette Valley
210.30
-0.0114
65
267
8.1.1
Eastern Great Lakes and Hudson Lowlands
144.62
-0.0104
36
257
8.1.10
Erie Drift Plain
133.08
-0.0037
78
700
8.1.2
Lake Erie Lowland
112.79
-0.0049
25
494
8.1.3
Northern Appalachian Plateau and Uplands
322.68
-0.0113
103
307
8.1.4
North Central Hardwood Forests
148.68
-0.0108
37
250
8.1.5
Driftless Area
117.97
-0.0012
141
2,057
8.1.6
S. Michigan/N. Indiana Drift Plains
191.44
-0.0143
46
206
8.1.7
Northeastern Coastal Zone
158.48
-0.0164
28
168
8.1.8
Maine/New Brunswick Plains and Hills
156.02
-0.0250
18
110
8.2.1
Southeastern Wisconsin Till Plains
121.34
-0.0042
46
594
8.2.2
Huron/Erie Lake Plains
145.17
-0.0058
65
461
8.2.3
Central Corn Belt Plains
187.95
-0.0033
191
889
8.2.4
Eastern Corn Belt Plains
235.18
-0.0030
282
1,053
8.3.1
Northern Piedmont
175.82
-0.0042
135
683
8.3.2
Interior River Valleys and Hills
149.68
-0.0013
303
2,081
8.3.3
Interior Plateau
220.47
-0.0037
217
836
8.3.4
Piedmont
224.11
-0.0048
169
648
8.3.5
Southeastern Plains
205.30
-0.0085
85
356
8.3.6
Mississippi Valley Loess Plains
492.49
-0.0048
333
812
8.3.7
South Central Plains
184.36
-0.0045
136
648
8.3.8
East Central Texas Plains
162.32
-0.0013
362
2,144
8.4.1
Ridge and Valley
186.83
-0.0063
99
465
8.4.2
Central Appalachians
166.76
-0.0062
82
454
8.4.3
Western Allegheny Plateau
183.67
-0.0032
190
910
8.4.4
Blue Ridge
216.16
-0.0087
89
353
8.4.5
Ozark Highlands
175.16
-0.0018
317
1,591
8.4.6
Boston Mountains
329.77
-0.0062
193
564
8.4.7
Arkansas Valley
283.25
-0.0040
261
836
8.4.8
Ouachita Mountains
212.77
-0.0048
157
637
8.4.9
Southwestern Appalachians
207.09
-0.0071
103
427
8.5.1
Middle Atlantic Coastal Plain
182.17
-0.0178
34
163
8.5.2
Mississippi Alluvial Plain
131.35
-0.0029
93
888
8.5.3
Southern Coastal Plain
138.62
-0.0144
23
183
8.5.4
Atlantic Coastal Pine Barrens
283.76
-0.0463
23
72
9.2.1
Aspen Parkland/Northern Glaciated Plains
136.43
-0.0005
640
5,226
9.2.2
Lake Manitoba and Lake Agassiz Plain
174.13
-0.0042
131
680
9.2.3
Western Corn Belt Plains
135.01
-0.0009
347
2,892
9.2.4
Central Irregular Plains
201.19
-0.0010
673
3,002
9.3.1
Northwestern Glaciated Plains
133.98
-0.0006
483
4,325
9.3.3
Northwestern Great Plains
130.60
-0.0004
636
6,424
9.3.4
Nebraska Sand Hills
289.85
-0.0066
162
510
EPA-821-R-19-011
B-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Table B-3: TSS Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TSSioo
TSS10
9.4.1
High Plains
125.61
-0.0005
507
5,061
9.4.2
Central Great Plains
156.84
-0.0005
925
5,505
9.4.3
Southwestern Tablelands
137.77
-0.0003
1,280
8,743
9.4.4
Flint Hills
270.93
-0.0009
1,084
3,666
9.4.5
Cross Timbers
134.97
-0.0006
523
4,337
9.4.6
Edwards Plateau
173.77
-0.0010
544
2,855
9.4.7
Texas Blackland Prairies
134.23
-0.0005
624
5,194
9.5.1
Western Gulf Coastal Plain
124.47
-0.0025
88
1,009
9.6.1
Southern Texas Plains/Interior Plains and Hills with
Xerophytic Shrub and Oak Forest
166.67
-0.0003
1,602
9,378
Table B-4: TN Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TNioo
TN io
10.1.2
Columbia Plateau
116.58
-0.663
0.23
3.70
10.1.3
Northern Basin and Range
126.97
-0.626
0.38
4.06
10.1.4
Wyoming Basin
124.89
-0.445
0.50
5.67
10.1.5
Central Basin and Range
116.66
-0.335
0.46
7.33
10.1.6
Colorado Plateaus
146.41
-0.588
0.65
4.56
10.1.7
Arizona/New Mexico Plateau
116.33
-0.286
0.53
8.58
10.1.8
Snake River Plain
129.93
-0.594
0.44
4.32
10.2.1
Mojave Basin and Range
136.69
-0.593
0.53
4.41
10.2.2
Sonoran Desert
117.99
-0.495
0.33
4.99
10.2.4
Chihuahuan Desert
104.20
-0.450
0.09
5.21
11.1.1
California Coastal Sage, Chaparral, and Oak Woodlands
123.22
-0.889
0.23
2.82
11.1.2
Central California Valley
126.07
-0.548
0.42
4.62
11.1.3
Southern and Baja California Pine-Oak Mountains
122.76
-0.564
0.36
4.45
12.1.1
Madrean Archipelago
130.61
-0.325
0.82
7.91
13.1.1
Arizona/New Mexico Mountains
141.64
-0.541
0.64
4.90
15.4.1
Southern Florida Coastal Plain
1000000
-29.36
0.33
0.39
5.2.1
Northern Lakes and Forests
141.98
-0.985
0.36
2.69
5.2.2
Northern Minnesota Wetlands
142.55
-0.781
0.45
3.40
5.3.1
Northern Appalachian and Atlantic Maritime Highlands
142.60
-0.854
0.42
3.11
5.3.3
North Central Appalachians
180.92
-0.897
0.66
3.23
6.2.10
Middle Rockies
136.51
-0.991
0.31
2.64
6.2.11
Klamath Mountains
140.34
-1.805
0.19
1.46
6.2.12
Sierra Nevada
143.02
-1.424
0.25
1.87
6.2.13
Wasatch and Uinta Mountains
129.75
-0.452
0.58
5.67
6.2.14
Southern Rockies
131.07
-0.660
0.41
3.90
6.2.15
Idaho Batholith
149.42
-1.775
0.23
1.52
6.2.3
Columbia Mountains/Northern Rockies
136.14
-1.599
0.19
1.63
6.2.4
Canadian Rockies
151.95
-2.098
0.20
1.30
6.2.5
North Cascades
155.86
-1.231
0.36
2.23
6.2.7
Cascades
143.07
-1.473
0.24
1.81
6.2.8
Eastern Cascades Slopes and Foothills
123.99
-1.070
0.20
2.35
6.2.9
Blue Mountains
125.19
-0.786
0.29
3.22
EPA-821-R-19-011
B-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Table B-4: TN Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TNioo
TN 10
7.1.7
Strait of Georgia/Puget Lowland
121.09
-0.723
0.26
3.45
7.1.8
Coast Range
136.15
-1.021
0.30
2.56
7.1.9
Willamette Valley
135.01
-0.809
0.37
3.22
8.1.1
Eastern Great Lakes and Hudson Lowlands
158.18
-0.563
0.81
4.90
8.1.2
Lake Erie Lowland
156.27
-0.380
1.18
7.23
8.1.3
Northern Appalachian Plateau and Uplands
431.78
-0.435
3.36
8.66
8.1.4
North Central Hardwood Forests
163.4
-0.599
0.82
4.66
8.1.5
Driftless Area
126.18
-0.272
0.85
9.32
8.1.6
S. Michigan/N. Indiana Drift Plains
130.25
-0.149
1.78
17.23
8.1.7
Northeastern Coastal Zone
125.75
-0.159
1.44
15.92
8.1.8
Maine/New Brunswick Plains and Hills
139.55
-0.553
0.60
4.77
8.1.10
Erie Drift Plain
148.99
-1.256
0.32
2.15
8.2.1
Southeastern Wisconsin Till Plains
134.85
-0.160
1.87
16.26
8.2.2
Huron/Erie Lake Plains
119.06
-0.091
1.91
27.22
8.2.3
Central Corn Belt Plains
135.57
-0.087
3.50
29.96
8.2.4
Eastern Corn Belt Plains
149.12
-0.122
3.28
22.15
8.3.1
Northern Piedmont
146.34
-0.314
1.21
8.55
8.3.2
Interior River Valleys and Hills
120.48
-0.131
1.43
19.00
8.3.3
Interior Plateau
146.39
-0.446
0.85
6.02
8.3.4
Piedmont
148.67
-0.637
0.62
4.24
8.3.5
Southeastern Plains
138.73
-0.727
0.45
3.62
8.3.6
Mississippi Valley Loess Plains
123.15
-0.379
0.55
6.62
8.3.7
South Central Plains
149.84
-0.706
0.57
3.83
8.3.8
East Central Texas Plains
136
-0.344
0.89
7.59
8.4.1
Ridge and Valley
158.11
-0.659
0.70
4.19
8.4.2
Central Appalachians
161.22
-0.907
0.53
3.07
8.4.3
Western Allegheny Plateau
125.25
-0.440
0.51
5.74
8.4.4
Blue Ridge
158.16
-0.777
0.59
3.55
8.4.5
Ozark Highlands
145.69
-0.513
0.73
5.22
8.4.6
Boston Mountains
168.59
-1.108
0.47
2.55
8.4.7
Arkansas Valley
135.4
-0.470
0.64
5.54
8.4.8
Ouachita Mountains
162.34
-0.942
0.51
2.96
8.4.9
Southwestern Appalachians
143.42
-0.645
0.56
4.13
8.5.1
Middle Atlantic Coastal Plain
123.43
-0.444
0.47
5.66
8.5.2
Mississippi Alluvial Plain
119.57
-0.310
0.58
8.00
8.5.3
Southern Coastal Plain
118.73
-0.701
0.24
3.53
8.5.4
Atlantic Coastal Pine Barrens
110.04
-0.482
0.20
4.98
9.2.1
Aspen Parkland/Northern Glaciated Plains
141.62
-0.086
4.06
30.82
9.2.2
Lake Manitoba and Lake Agassiz Plain
119.49
-0.082
2.18
30.25
9.2.3
Western Corn Belt Plains
129.28
-0.074
3.48
34.59
9.2.4
Central Irregular Plains
142.81
-0.184
1.93
14.45
9.3.1
Northwestern Glaciated Plains
120.91
-0.386
0.49
6.46
9.3.3
Northwestern Great Plains
125.65
-0.404
0.56
6.26
9.3.4
Nebraska Sand Hills
113.81
-0.324
0.40
7.51
9.4.1
High Plains
121.41
-0.161
1.21
15.51
9.4.2
Central Great Plains
129.36
-0.178
1.44
14.38
EPA-821-R-19-011
B-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Table B-4: TN Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TNioo
TN 10
9.4.3
Southwestern Tablelands
136.03
-0.413
0.74
6.32
9.4.4
Flint Hills
142.74
-0.343
1.04
7.75
9.4.5
Cross Timbers
130.87
-0.278
0.97
9.25
9.4.6
Edwards Plateau
141.98
-0.588
0.60
4.51
9.4.7
Texas Blackland Prairies
133.84
-0.243
1.20
10.68
9.5.1
Western Gulf Coastal Plain
106.22
-0.301
0.20
7.85
9.6.1
Southern Texas Plains/Interior Plains and Hills with
Xerophytic Shrub and Oak Forest
102.35
-0.374
0.06
6.22
Table B-5: TP Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TP ioo
TP io
10.1.2
Columbia Plateau
126.6
-3.83
0.06
0.66
10.1.3
Northern Basin and Range
147.4
-2.21
0.18
1.22
10.1.4
Wyoming Basin
165.9
-2.78
0.18
1.01
10.1.5
Central Basin and Range
143.8
-1.57
0.23
1.70
10.1.6
Colorado Plateaus
167.2
-2.54
0.20
1.11
10.1.7
Arizona/New Mexico Plateau
123.7
-0.78
0.27
3.21
10.1.8
Snake River Plain
168.7
-3.39
0.15
0.83
10.2.1
Mojave Basin and Range
140.8
-1.11
0.31
2.39
10.2.2
Sonoran Desert
139.9
-0.98
0.34
2.70
10.2.4
Chihuahuan Desert
122.9
-1.58
0.13
1.59
11.1.1
California Coastal Sage, Chaparral, and Oak Woodlands
132.9
-3.74
0.08
0.69
11.1.2
Central California Valley
125.1
-1.92
0.12
1.32
11.1.3
Southern and Baja California Pine-Oak Mountains
126.3
-2.14
0.11
1.19
12.1.1
Madrean Archipelago
212.0
-0.94
0.80
3.25
13.1.1
Arizona/New Mexico Mountains
140.6
-1.33
0.26
1.99
15.4.1
Southern Florida Coastal Plain
555.9
-306.0
0.01
0.01
5.2.1
Northern Lakes and Forests
157.9
-26.64
0.02
0.10
5.2.2
Northern Minnesota Wetlands
152.8
-16.37
0.03
0.17
5.3.1
Northern Appalachian and Atlantic Maritime Highlands
171.4
-21.87
0.02
0.13
5.3.3
North Central Appalachians
260.9
-21.53
0.04
0.15
6.2.10
Middle Rockies
157.8
-6.44
0.07
0.43
6.2.11
Klamath Mountains
189.0
-15.04
0.04
0.20
6.2.12
Sierra Nevada
205.2
-19.13
0.04
0.16
6.2.13
Wasatch and Uinta Mountains
142.6
-2.75
0.13
0.97
6.2.14
Southern Rockies
141.7
-5.46
0.06
0.49
6.2.15
Idaho Batholith
185.9
-21.89
0.03
0.13
6.2.3
Columbia Mountains/Northern Rockies
168.9
-17.88
0.03
0.16
6.2.4
Canadian Rockies
197.1
-27.87
0.02
0.11
6.2.5
North Cascades
289.6
-47.06
0.02
0.07
6.2.7
Cascades
227.9
-26.77
0.03
0.12
6.2.8
Eastern Cascades Slopes and Foothills
154.7
-10.55
0.04
0.26
6.2.9
Blue Mountains
141.6
-3.31
0.11
0.80
7.1.7
Strait of Georgia/Puget Lowland
165.3
-13.83
0.04
0.20
7.1.8
Coast Range
185.3
-14.77
0.04
0.20
EPA-821-R-19-011
B-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Table B-5: TP Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TPioo
TP io
7.1.9
Willamette Valley
159.5
-9.05
0.05
0.31
8.1.1
Eastern Great Lakes and Hudson Lowlands
148.0
-7.95
0.05
0.34
8.1.10
Erie Drift Plain
230.1
-9.61
0.09
0.33
8.1.2
Lake Erie Lowland
3,440.2
-8.89
0.40
0.66
8.1.3
Northern Appalachian Plateau and Uplands
317.2
-13.87
0.08
0.25
8.1.4
North Central Hardwood Forests
132.7
-4.91
0.06
0.53
8.1.5
Driftless Area
141.5
-2.26
0.15
1.17
8.1.6
S. Michigan/N. Indiana Drift Plains
184.3
-5.59
0.11
0.52
8.1.7
Northeastern Coastal Zone
174.0
-9.94
0.06
0.29
8.1.8
Maine/New Brunswick Plains and Hills
174.7
-28.94
0.02
0.10
8.2.1
Southeastern Wisconsin Till Plains
151.8
-3.59
0.12
0.76
8.2.2
Huron/Erie Lake Plains
141.2
-1.58
0.22
1.68
8.2.3
Central Corn Belt Plains
247.2
-2.67
0.34
1.20
8.2.4
Eastern Corn Belt Plains
223.4
-3.56
0.23
0.87
8.3.1
Northern Piedmont
196.0
-3.73
0.18
0.80
8.3.2
Interior River Valleys and Hills
161.0
-2.57
0.19
1.08
8.3.3
Interior Plateau
156.7
-3.62
0.12
0.76
8.3.4
Piedmont
197.7
-5.62
0.12
0.53
8.3.5
Southeastern Plains
223.4
-9.27
0.09
0.34
8.3.6
Mississippi Valley Loess Plains
177.2
-5.69
0.10
0.51
8.3.7
South Central Plains
168.0
-4.66
0.11
0.61
8.3.8
East Central Texas Plains
166.4
-1.68
0.30
1.68
8.4.1
Ridge and Valley
178.1
-6.41
0.09
0.45
8.4.2
Central Appalachians
225.7
-16.59
0.05
0.19
8.4.3
Western Allegheny Plateau
187.7
-8.37
0.08
0.35
8.4.4
Blue Ridge
174.1
-10.50
0.05
0.27
8.4.5
Ozark Highlands
152.7
-2.89
0.15
0.94
8.4.6
Boston Mountains
204.9
-7.36
0.10
0.41
8.4.7
Arkansas Valley
287.2
-5.79
0.18
0.58
8.4.8
Ouachita Mountains
158.5
-6.82
0.07
0.41
8.4.9
Southwestern Appalachians
169.7
-7.30
0.07
0.39
8.5.1
Middle Atlantic Coastal Plain
154.0
-6.82
0.06
0.40
8.5.2
Mississippi Alluvial Plain
141.3
-3.81
0.09
0.70
8.5.3
Southern Coastal Plain
144.7
-7.68
0.05
0.35
8.5.4
Atlantic Coastal Pine Barrens
126.8
-8.39
0.03
0.30
9.2.1
Aspen Parkland/Northern Glaciated Plains
156.1
-0.69
0.65
3.98
9.2.2
Lake Manitoba and Lake Agassiz Plain
132.2
-1.09
0.26
2.38
9.2.3
Western Corn Belt Plains
197.2
-1.68
0.40
1.77
9.2.4
Central Irregular Plains
201.0
-1.99
0.35
1.50
9.3.1
Northwestern Glaciated Plains
134.1
-1.65
0.18
1.58
9.3.3
Northwestern Great Plains
143.3
-1.27
0.28
2.10
9.3.4
Nebraska Sand Hills
185.0
-3.79
0.16
0.77
9.4.1
High Plains
153.1
-0.95
0.45
2.88
9.4.2
Central Great Plains
188.6
-1.18
0.54
2.49
9.4.3
Southwestern Tablelands
139.6
-0.97
0.34
2.71
9.4.4
Flint Hills
218.9
-2.35
0.33
1.31
EPA-821-R-19-011
B-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix B: WQI Calculation
Table B-5: TP Subindex Curve Parameters, by Ecoregion
ID
Ecoregion Name
a
b
TPioo
TP io
9.4.5
Cross Timbers
131.7
-0.78
0.35
3.31
9.4.6
Edwards Plateau
160.0
-1.38
0.34
2.00
9.4.7
Texas Blackland Prairies
149.6
-1.06
0.38
2.54
9.5.1
Western Gulf Coastal Plain
127.2
-1.86
0.13
1.36
9.6.1
Southern Texas Plains/Interior Plains and Hills with
Xerophytic Shrub and Oak Forest
104.2
-0.51
0.08
4.57
EPA-821-R-19-011
B-10
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Appendix C Additional Details on Modeling Change in Bladder Cancer Incidence
from Change in TTHM Exposure
C.1 Details on Life Table Approach
C. 1.1 Health Impact Function
Figure C-l shows the dependence between lifetime odds of bladder cancer and drinking water TTHM
concentration as reported by Villanueva et al. (2004). These data were used by Regli et al. (2015) to estimate
the log-linear relationship in Equation 4-1, which is also displayed in Figure C-l. As described in Chapter 4,
Regli et al. (2015) showed that, while the original analysis deviated from linearity, particularly at low doses,
the overall pooled exposure-response relationship for TTHM could be well-approximated by a linear slope
factor that predicted an incremental lifetime cancer risk of 1 in ten thousand exposed individuals (10~4) per
1 |_ig/L increase in TTHM.85
Figure C-1: Estimated Relationships between Lifetime Bladder Cancer Risk and TTHM Concentrations
in Drinking Water.
OR from Kogevinas et al
95% Confidence limits from Kogevinas et al
Linear function OR(THM4) = exp(THM4 * 0 00427)
THM4. ug/L
Source: Regli et al., 2015
The EPA used the Regli et al. (2015) relationship between the lifetime odds of bladder cancer and lifetime
TTHM exposure from drinking water to derive a set of age-specific health impact functions. A person's
lifetime TTHM exposure from drinking water by age a—denoted by xa—is defined as:
85 Regli et al (2015) addressed some of the limitations noted in the Hrudey et al. (2015) analysis. They suggested that the seeming
discrepancy between the slope factor derived from the pooled epidemiological data and that from animal studies was due
primarily to (1) potentially high human exposures to DBPs by the inhalation route, and (2) that trihalomethanes were acting as
proxies for other carcinogenic DBPs.
EPA-821-R-19-011
C-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Equation C-1. xa = TTHM; ,x0 = 0.
See Table C-1 at the end of this section for definitions of all variables used in the equations in this Appendix.
Assuming a baseline exposure of za and a regulatory option exposure of xa {i.e., exposure following
implementation of a regulatory option), the relative risk (RR) of bladder cancer by age a under the option
exposure relative to the baseline exposure can be expressed as:
Equation C-2. RR^.zJ = (|g)"'. (LR„ • gg - LRa + l)
where LRa is the lifetime risk of bladder cancer within age interval [0, a] (Fay et al. 2003) under baseline
conditions.
Combining Equation C-1 and Equation C-2 shows that the relative risk of bladder cancer by age a based on
Regli et al. (2015) depends only on the lifetime risk and on the magnitude of change in TTHM concentration
from baseline concentration, Axa = xa — za, but not on the baseline TTHM level:
/O(0)-e000427Xa\~^ ( O(0)-e000427Xa \
Equation C-3. ^Reglietal.(xa
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
implementation year (in this case 2021) and to alternative TTHM levels that reflect the impact of technology
implementation under each regulatory option starting in 2021.
To capture these effects while being consistent with the remainder of the cost-benefit framework, the EPA
modeled changes in health outcomes resulting from changes in exposure between 2021 and 2047. For these
exposures, the EPA modeled effects out to 2121 to capture the resultant lagged changes in lifetime bladder
cancer risk, but did not attribute changes in bromide loadings and TTHM exposures to the regulatory options
beyond 2047.87
The EPA tracks mortality and bladder cancer experience for a set of model populations defined by sex,
location, and age attained by 2021, which is denoted by A = 0,1,2,3,... 100. Each model population is
followed from birth (corresponding to calendar year 2021 — ^4) to age 100, using a one-year time step.
Below, we first describe the process for quantifying the evolution of model population A under the baseline
TTHM exposure assumptions. We then describe the process for quantifying the evolution of the population
under the regulatory option TTHM exposures. Finally, we describe the process for estimating the total
calendar year y-specific health benefits which aggregate estimates over all model populations (A =
0,1,2,3, ...100).
Evolution of Model Population A under Baseline TTHM Exposure
Given a model population A. for each current age a and calendar year y, the following baseline exposure
zay = ~T^f=d Baseline TTHMi y_a+( dependent quantities are computed:
• lc=o,a,y(za,y): The number of bladder cancer-free living individuals at the beginning of age a, in year
y;
• dc=o,a,y(za,y): The number of deaths among bladder cancer-free individuals aged a during the year
y;
• lc=i,a,y(za,y)'- The number of new bladder cancer cases among individuals aged a during the year y.
To compute each quantity above, the EPA makes an assumption about the priority of events that terminate a
person's existence in the pool of bladder cancer-free living individuals. These events are general population
deaths that occur with probability88 qc=o,a and new bladder cancer diagnoses that occur with probability ya.
which is approximated by age-specific annual bladder cancer incidence rate IRa ¦ 10"5. In the model, the
EPA assumes that the new cancer diagnoses occur after general population deaths and uses the following
recurrent equations for ages a > 0:89
87 This approach is equivalent to assuming that TTHM levels revert back to baseline conditions at the end of the regulatory option
costing period.
88 The model does not index the general population death rates using the calendar year, because the model relies on the most recent
static life tables.
89 The EPA notes that this is a conservative assumption that results in a lower bound estimate of the policy impact (with respect to
this particular uncertainty factor). An upper bound estimate of the policy impact can be obtained by assuming that new bladder
EPA-821-R-19-011
C-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Equation C-4.
1-C=0,a,y (^a,y) — lc=0,a—l,y—l i_za—l,y—l) — ^-C=0,a—l,y—li_za—l,y—l) ~ ^C=l,a-l,y-l^a-l,y-l)
Equation C-5. ^C=0,a,y(^za,y^) Qc=0,a ' ^c=0,a,y (^a,y)
Equation C-6. ^C=l,a,y(^a,y) ~ Ya ' (j-C=0,a,yi.za,y^) ~ ^-C=0,a,y (.za,y))
To initiate each set of recurrent equations, the EPA estimates the number of cancer-free individuals at age
a = 0, denoted by lc=Q 0y_^(z0y_^), that is consistent with the number of affected persons of age A in 2021,
denoted by P. To this end, Equation C-4, Equation C-5, and Equation C-6 are solved to find
'c=o,o,y-^(zo,y-^i) such that lc=o,A,202i(zA,202i) = P¦
Consistent with available bladder cancer survival statistics, the EPA models mortality experience in the
bladder cancer populations lc=i,a,y(za,y) as dependent on the age-at-onset a, disease duration k, and cancer
stage s (for bladder cancer there are four defined stages: localized, regional, distant, unstaged). Given each
age-specific share of new cancer cases lc=i,a,y(za,y) and age-specific share of new stage s cancers 8s=sa, the
EPA calculates the number of new stage s cancers occurring at age a in year y:
Equation C-7. ^S=s,a,y,o{.za,y^) — &S=s,a ' ^-C=l,a,y (_za,y)
For a model population aged A years in 2021 and cancer stage s, the EPA separately tracks 100 — A + 1 new
stage-specific bladder cancer populations from age-at-onset a to age 100.90 Next, a set of cancer duration k-
dependent annual death probabilities is derived for each population from available data on relative survival
rates91 rs=s a k and general population annual death probabilities qc=o,a+k as follows:
Equation C-8. qs=SAfc = 1 - r^'a+1 (l - qc=0,a+k)¦
rS=s,a,k
In estimating additional deaths in the cancer population in the year of diagnosis (i.e.. when k = 0), the EPA
accounts only for cancer population deaths that are in excess of the general population deaths. As such, the
estimate of additional cancer population deaths is computed as follows:
Equation C-9. ds=say 0[zay^ — (^C)s=s,a,0 ~ tfc=0,a) ' ^S=s,a,y,o(^a,y)'
diagnoses occur before general population deaths. In a limited sensitivity analysis, the EPA found that estimates generated using
this alternative assumption were approximately 5 percent larger than the estimates reported here.
90 In total, there are 4 ¦ (100 — A + 1) new cancer populations being tracked for each model population.
91 Note that rs=s a k is a multiplier that modifies the general probability of survival to age k to reflect the fact that the population
under consideration has developed cancer k years ago.
EPA-821-R-19-011
C-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
In years that follow the initial diagnosis year (i.e.. k > 0), the EPA uses the following recurrent equations to
estimate the number of people living with bladder cancer and the annual number of deaths in the bladder
cancer population:
Equation C-10. ^-S=s,a,y ,k{^a,y—fc) ^-S=s,a,y,k—fc) ^-S=s,a,y,k—
Equation C-11. ^-S=s,a,y ,k{^a,y—k^} Qs=s,a,k ' ^S=sfafyfk^afy—k^)m
Because the EPA is interested in bladder cancer-related deaths rather than all deaths in the bladder cancer
population, the EPA also tracks the number of excess bladder cancer population deaths (i.e.. the number of
deaths in the bladder cancer population over and above the number of deaths expected in the general
population of the same age). The excess deaths are computed as:
Equation C-12. es=s,a,y,ki^a,y—k) — Rs=s,a,k ' I,S=s,a,y,k(za,y—k) — Rc=0,a+k ' ^S=s,a,y,ki^a,y—k)
Evolution of Model Population A under the Regulatory Option TTHM Exposure
Under the baseline conditions when the change in TTHM is zero (i.e.. before 2021), the EPA approximates
the annual bladder cancer probability ya by age-specific annual bladder cancer incidence rate IRa ¦ 10~5. As
described in Section 4.3.3, current empirical evidence links TTHM exposure to the lifetime bladder cancer
risk, rather than annual bladder cancer probability. The EPA computes the TTHM-dependent annual new
bladder cancer cases under the regulatory option conditions, lc=i,a,y(xa,y)> in three steps. First, the EPA
recursively estimates LRay{zay), the lifetime risk of bladder cancer within age interval [0, a] under the
baseline conditions:
Equation C-13. LRay{zay) — o ^(z0 > 0 and LR0y_^z0y_^ — 0
Second, the result of Equation C-13 is combined with the relative risk estimate RR(xay, zay), based on Regli
etal. (2015):
Equation C-14. LRay(xay) = RR(xaiy,zay)LRaiy(zay)
This results in a series of lifetime bladder cancer risk estimates under the option conditions. Third, the EPA
computes a series of new annual bladder cancer case estimates under the option conditions as follows:
Equation C-15. ^C=l,a,y(xa,y) = (^^a+l,y+l(xa+l,y+l) — ^a,y (xa,y)) ' ^-C=0,0,y-A (z0,y-^l)
Health Effects and Benefits Attributable to Regulatory Options
To characterize the overall impact of the regulatory option in a given year y, for each model population
defined by age a in 2021, sex, and location, the EPA calculates three quantities: the incremental number of
EPA-821-R-19-011
C-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Appendix C: Bromide-related Benefits
new stage s bladder cancer cases (NCA y s), the incremental number of individuals living with stage s bladder
cancer (LCA y s), and the incremental number of excess deaths in the bladder cancer population (EDA y). The
formal definitions of each of these quantities are given below:
Equation C-16.
NCA,y,s = [0 < y - 2021 + A < 100] 1 (js=s,y-2021+A,y,o(,zy-2021+A,y) — ls=s,y-2021+A,0(Xy-2021+A,y))
Equation C-17.
Z100
[0
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Appendix C: Bromide-related Benefits
Table C-1: Health Risk Model Variable Definitions
Variable
Definition
tfs=s,a,k
Stage-specific probability of death in the bladder cancer population whose bladder cancer was
diagnosed at age a and they lived fc years after the diagnosis. Current age of these individuals is a +
fc.
d-S=s,a,y, 0 (?a,y )
The baseline number of deaths in the stage s cancer population in the year of diagnosis {i.e., when
k = 0), given the current age a and the corresponding year y.
?S=s,a,y,k a,y—k )
The baseline number of living with the stage s cancer in the fc-th year after diagnosis in year y,
given the cancer diagnosis at age a and the cumulative exposure through to that age and year y —
k.
d-s=s,a,y,k (%a,y—k )
The baseline number of deaths among those with the stage s cancer in the fc-th year after diagnosis
in year y, given the cancer diagnosis at age a and the cumulative exposure through to that age and
year y — fc.
&S=s,a,y,k (%a,y—k )
The baseline number of excess bladder cancer deaths {i.e., the number of deaths in the bladder
cancer population over and above the number of deaths expected in the general population of the
same age) among those with the stage s cancer in the fc-th year after diagnosis in year y, given the
cancer diagnosis at age a and the cumulative exposure through to that age and year y — fc.
LRa,y(za,y)
Recursive estimate of the lifetime risk of bladder cancer within age interval [0, a) under the
baseline conditions, given that age a occurs in year y
RR(xa,y, za,y)
Relative risk of bladder cancer by age a given that this age occurs in year y, baseline exposure zay
and option exposure xay
LRa,y (-^a.y)
Recursive estimate of the lifetime risk of bladder cancer within age interval [0, a) under the option
conditions, given that age a occurs in year y
N Q,y,s
The incremental number of new stage s bladder cancer cases in year y for the model population
aged A in 2021.
A,y,s
The incremental number of individuals living with stage s bladder cancer in year y for the model
population aged A in 2021.
EDA,y
The incremental number of excess in stage s bladder cancer population in year y for the model
population aged A in 2021.
C. 1.3 Detailed Input Data
As noted in Section 4.3.3, the EPA relied on the federal government data sources including EPA SDWIS,
ACS 2017 (U.S. Census Bureau, 2018), the Surveillance, Epidemiology, and End Results (SEER) program
database (National Cancer Institute), and the Center for Disease Control (CDC) National Center for Health
Statistics to characterize sex- and age group-specific general population mortality rates and bladder cancer
incidence rates used in model simulations. All of these data are compiled by the relevant federal agencies and
thus meet federal government data quality standards. These data sources are appropriate for this analysis
based on the standards underlying their collection and publication, and their applicability to analyzing health
effects of exposure to TTHM via drinking water. Table 4-6 in Section 4.3.3 summarizes the sex- and age
group-specific share of general population mortality rates and bladder cancer incidence. Table C-2 below
summarizes sex- and age group-specific distribution of bladder cancer cases over four analyzed stages as well
as onset-specific relative survival probability for each stage.
EPA-821-R-19-011
C-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Table C-2: Summary of Sex- and Age-specific Bladder Cancer Stage Distribution and Relative Survival Rates
Sex
Age
Stage
Stage-specific
proportion of
bladder
cancers*
Relative 0th
year survival
rate*
Relative 1st year
survival rate*
Relative 2nd year
survival rate*
Relative 3rd year
survival rate*
Relative 4th
year survival
rate*
Relative 5th
year survival
rate*
Male
Os
Localized
0.6667
1
0.9493
0.9199
0.8892
0.8823
0.8472
Male
Os
Regional
0.1424
1
0.8513
0.5967
0.5086
0.486
0.4459
Male
Os
Distant
0.1333
1
0.4269
0.2237
0.1679
0.1475
0.1475
Male
Os
Unstaged
0.0576
1
0.9662
0.8989
0.8989
0.8989
0.8989
Male
10s
Localized
0.6667
1
0.9493
0.9199
0.8892
0.8823
0.8472
Male
10s
Regional
0.1424
1
0.8513
0.5967
0.5086
0.486
0.4459
Male
10s
Distant
0.1333
1
0.4269
0.2237
0.1679
0.1475
0.1475
Male
10s
Unstaged
0.0576
1
0.9662
0.8989
0.8989
0.8989
0.8989
Male
20s
Localized
0.6667
1
0.9493
0.9199
0.8892
0.8823
0.8472
Male
20s
Regional
0.1424
1
0.8513
0.5967
0.5086
0.486
0.4459
Male
20s
Distant
0.1333
1
0.4269
0.2237
0.1679
0.1475
0.1475
Male
20s
Unstaged
0.0576
1
0.9662
0.8989
0.8989
0.8989
0.8989
Male
30s
Localized
0.6667
1
0.9493
0.9199
0.8892
0.8823
0.8472
Male
30s
Regional
0.1424
1
0.8513
0.5967
0.5086
0.486
0.4459
Male
30s
Distant
0.1333
1
0.4269
0.2237
0.1679
0.1475
0.1475
Male
30s
Unstaged
0.0576
1
0.9662
0.8989
0.8989
0.8989
0.8989
Male
40s
Localized
0.6745
1
0.9524
0.9173
0.8851
0.8692
0.8424
Male
40s
Regional
0.1581
1
0.8206
0.5944
0.4971
0.4599
0.4307
Male
40s
Distant
0.1161
1
0.4122
0.1944
0.134
0.1051
0.099
Male
40s
Unstaged
0.0513
1
0.909
0.8305
0.825
0.825
0.8194
Male
50s
Localized
0.6865
1
0.9471
0.899
0.8656
0.839
0.8141
Male
50s
Regional
0.1706
1
0.7793
0.5657
0.4738
0.426
0.4042
Male
50s
Distant
0.0997
1
0.3648
0.1599
0.0988
0.0688
0.0611
Male
50s
Unstaged
0.0432
1
0.8565
0.791
0.7595
0.7571
0.7338
Male
60s
Localized
0.7095
1
0.9347
0.8766
0.8369
0.8066
0.7798
Male
60s
Regional
0.1587
1
0.7563
0.5401
0.4609
0.4171
0.3864
Male
60s
Distant
0.0905
1
0.3318
0.1531
0.0924
0.0642
0.0587
Male
60s
Unstaged
0.0414
1
0.8309
0.7793
0.7358
0.7186
0.6887
Male
70s
Localized
0.7339
1
0.8885
0.8157
0.7647
0.7272
0.7011
Male
70s
Regional
0.1361
1
0.6912
0.4992
0.4245
0.3824
0.3484
Male
70s
Distant
0.0772
1
0.2845
0.1293
0.0778
0.0461
0.0412
Male
70s
Unstaged
0.0528
1
0.7137
0.6387
0.5944
0.5432
0.5097
EPA-821-R-19-011
C-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Table C-2: Summary of Sex- and Age-specific Bladder Cancer Stage Distribution and Relative Survival Rates
Sex
Age
Stage
Stage-specific
proportion of
bladder
cancers*
Relative 0th
year survival
rate*
Relative 1st year
survival rate*
Relative 2nd year
survival rate*
Relative 3rd year
survival rate*
Relative 4th
year survival
rate*
Relative 5th
year survival
rate*
Male
80s
Localized
0.7364
1
0.8254
0.7344
0.6783
0.633
0.6002
Male
80s
Regional
0.1202
1
0.6105
0.4347
0.369
0.3279
0.3007
Male
80s
Distant
0.0736
1
0.211
0.0919
0.0605
0.0326
0.0326
Male
80s
Unstaged
0.0698
1
0.5827
0.4887
0.4265
0.347
0.2969
Male
90s
Localized
0.7364
1
0.8254
0.7344
0.6783
0.633
0.6002
Male
90s
Regional
0.1202
1
0.6105
0.4347
0.369
0.3279
0.3007
Male
90s
Distant
0.0736
1
0.211
0.0919
0.0605
0.0326
0.0326
Male
90s
Unstaged
0.0698
1
0.5827
0.4887
0.4265
0.347
0.2969
Female
Os
Localized
0.5651
1
0.9159
0.849
0.832
0.8119
0.793
Female
Os
Regional
0.1818
1
0.7186
0.5379
0.4703
0.4362
0.4132
Female
Os
Distant
0.1572
1
0.3893
0.1676
0.1177
0.0295
0.0295
Female
Os
Unstaged
0.0958
1
0.866
0.8376
0.7778
0.7778
0.7778
Female
10s
Localized
0.5651
1
0.9159
0.849
0.832
0.8119
0.793
Female
10s
Regional
0.1818
1
0.7186
0.5379
0.4703
0.4362
0.4132
Female
10s
Distant
0.1572
1
0.3893
0.1676
0.1177
0.0295
0.0295
Female
10s
Unstaged
0.0958
1
0.866
0.8376
0.7778
0.7778
0.7778
Female
20s
Localized
0.5651
1
0.9159
0.849
0.832
0.8119
0.793
Female
20s
Regional
0.1818
1
0.7186
0.5379
0.4703
0.4362
0.4132
Female
20s
Distant
0.1572
1
0.3893
0.1676
0.1177
0.0295
0.0295
Female
20s
Unstaged
0.0958
1
0.866
0.8376
0.7778
0.7778
0.7778
Female
30s
Localized
0.5651
1
0.9159
0.849
0.832
0.8119
0.793
Female
30s
Regional
0.1818
1
0.7186
0.5379
0.4703
0.4362
0.4132
Female
30s
Distant
0.1572
1
0.3893
0.1676
0.1177
0.0295
0.0295
Female
30s
Unstaged
0.0958
1
0.866
0.8376
0.7778
0.7778
0.7778
Female
40s
Localized
0.5616
1
0.9197
0.8583
0.8387
0.8245
0.813
Female
40s
Regional
0.2096
1
0.7083
0.538
0.4485
0.4013
0.3835
Female
40s
Distant
0.1626
1
0.3715
0.1553
0.1064
0.0466
0.0401
Female
40s
Unstaged
0.0663
1
0.8791
0.8457
0.7731
0.7731
0.7427
Female
50s
Localized
0.588
1
0.9266
0.8667
0.8381
0.8184
0.8038
Female
50s
Regional
0.2248
1
0.7015
0.5234
0.4238
0.3682
0.352
Female
50s
Distant
0.1449
1
0.344
0.1406
0.091
0.0648
0.0587
EPA-821-R-19-011
C-9
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Table C-2: Summary of Sex- and Age-specific Bladder Cancer Stage Distribution and Relative Survival Rates
Sex
Age
Stage
Stage-specific
proportion of
bladder
cancers*
Relative 0th
year survival
rate*
Relative 1st year
survival rate*
Relative 2nd year
survival rate*
Relative 3rd year
survival rate*
Relative 4th
year survival
rate*
Relative 5th
year survival
rate*
Female
50s
Unstaged
0.0423
1
0.8572
0.8013
0.7539
0.7445
0.7036
Female
60s
Localized
0.638
1
0.902
0.8341
0.7981
0.7668
0.7424
Female
60s
Regional
0.1978
1
0.6912
0.499
0.41
0.3651
0.3417
Female
60s
Distant
0.1188
1
0.2937
0.117
0.0751
0.0509
0.0509
Female
60s
Unstaged
0.0453
1
0.7756
0.7048
0.6922
0.6669
0.6335
Female
70s
Localized
0.6745
1
0.809
0.7259
0.6782
0.6457
0.6188
Female
70s
Regional
0.157
1
0.5905
0.4232
0.3551
0.3215
0.3029
Female
70s
Distant
0.1072
1
0.2133
0.085
0.0615
0.0365
0.031
Female
70s
Unstaged
0.0613
1
0.5692
0.493
0.4627
0.4308
0.3993
Female
80s
Localized
0.6897
1
0.7242
0.6268
0.5644
0.5318
0.502
Female
80s
Regional
0.1276
1
0.4691
0.3327
0.2929
0.2632
0.2588
Female
80s
Distant
0.0983
1
0.1665
0.0739
0.0611
0.0438
0.0304
Female
80s
Unstaged
0.0845
1
0.3592
0.2713
0.2168
0.1873
0.1748
Female
90s
Localized
0.6897
1
0.7242
0.6268
0.5644
0.5318
0.502
Female
90s
Regional
0.1276
1
0.4691
0.3327
0.2929
0.2632
0.2588
Female
90s
Distant
0.0983
1
0.1665
0.0739
0.0611
0.0438
0.0304
Female
90s
Unstaged
0.0845
1
0.3592
0.2713
0.2168
0.1873
0.1748
Notes: * Single age-specific proportions and rates were aggregated up to the age groups reported in the table using the individual age-specific number of affected persons as weights.
EPA-821-R-19-011
C-10
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
C.2 Detailed Results from Analysis
The health impact model assumes that the proposed regulatory changes begin in 2021 and end by 2047 and
thus TTHM changes are in effect during this period. After 2047, TTHM levels return to baseline levels, i.e.,
ATTHM is zero. Due to the lasting effects of changes in TTHM exposure, the benefits of the policies after
2047 were included in the final calculations for each option. Table C-3 summarizes the health impact and
valuation results in millions of 2018 dollars for each proposed regulatory option, as shown graphically and
discussed in Section 4.4.
EPA-821-R-19-011
C-11
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Table C-3: Number of Adverse Health Effects Avoided Over Time Starting from 2021
Option
Evaluation decade
Totalc
2021-2030
2031-2040
2041-2050
2051-2060
2061-2070
2071-2080
2081-2090
2091-2100
2101-2110
2111-2121
Cancer morbidity cases avoided3
Option 1
0
-1
-1
0
0
0
0
0
0
0
-3
Option 2
42
63
72
42
42
38
29
16
2
-2
343
Option 3
48
71
81
47
47
43
33
17
2
-2
387
Option 4
94
140
161
94
93
86
65
34
5
-4
769
Excess cancer deaths avoided13
Option 2
0
0
0
0
0
0
0
0
0
0
-1
Option 2
13
24
29
17
17
16
13
8
2
0
139
Option 3
15
27
33
19
19
18
15
9
2
0
157
Option 4
30
53
65
39
38
36
29
18
4
-1
311
Annual value of morbidity avoided (million dollars)
Option 1
-$0.01
-$0.01
-$0.02
-$0.01
-$0.01
-$0.01
-$0.01
-$0.01
$0.00
$0.00
-$0.09
Option 2
$0.83
$1.38
$1.72
$1.20
$1.18
$1.11
$0.88
$0.52
$0.14
-$0.02
$8.95
Option 3
$0.95
$1.57
$1.95
$1.35
$1.33
$1.24
$0.99
$0.59
$0.16
-$0.02
$10.10
Option 4
$1.87
$3.09
$3.88
$2.70
$2.65
$2.47
$1.96
$1.16
$0.31
-$0.04
$20.05
Annual value of mortality avoided (million dollars)
Option 1
-$1.42
-$2.69
-$3.55
-$2.27
-$2.32
-$2.23
-$1.85
-$1.19
-$0.27
$0.06
-$17.75
Option 2
$152.78
$285.10
$361.88
$222.39
$232.55
$229.39
$192.14
$124.14
$28.21
-$5.81
$1,822.77
Option 3
$174.12
$324.01
$410.02
$250.50
$260.61
$257.16
$215.79
$139.07
$31.45
-$6.50
$2,056.22
Option 4
$342.88
$637.67
$813.85
$501.80
$520.59
$511.90
$428.15
$274.55
$61.79
-$12.83
$4,080.34
Notes:
a Number of TTHM-attributable bladder cancer cases that are expected to be avoided under the policy in the calendar time period.
b Number of excess deaths among the TTHM-attributable bladder cancer cases that are expected to be avoided under the policy in the calendar time period.
cTotal TTHM-attributable adverse health effects that are expected to be avoided between 2021 and 2121 as a result of the regulatory option changes in 2021-2047.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
C-12
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
C.3 Temporal Distribution of Benefits
Figure C-2 and Figure C-3 illustrate patterns of changes in benefits for the four regulatory options for the 100-
year simulation period of 2021 through 2121 based on the estimated cumulative annual value of morbidity
avoided and the estimated cumulative annual value of mortality, respectively (values are undiscounted). These
figures show the gradual increase in benefits for Options 2, 3, and 4 between 2021 and 2047, which continues
but at a reduced rate after 2047 until levelling off around 2101. As discussed in Section 4.4, benefits decrease
during the final decade for Options 2, 3, and 4. The magnitude of benefits associated with Option 1 are much
smaller and generally follow the inverse pattern when compared to Options 2, 3, and 4, due to the option
increasing bromide concentrations as compared to the baseline.
Figure C-2: Estimated Cumulative Annual Value of Cancer Morbidity Avoided, 2021-2121 (2018$
undiscounted).
U, $45
c
• Option 1
~ Option 2
¦ Option 3
Option 4
$35
$30
$25
$20
$15
$10
2021
2031
2041
2051
2061
2071
2081
2091
2101
2111
2121
-$5
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
C-13
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Figure C-3. Estimated Cumulative Annual Value of Mortality Avoided, 2021-2121 (2018$
undiscounted).
$4,500
VJ '
C
0
1
$4,000
$3,500
$3,000
$2,500
$2,000
$1,500
$1,000
$500
$0
-$500
Source: U.S. EPA Analysis, 2019.
C.4 Sensitivity Analysis Results
For FGD wastewater bromide loadings, the EPA developed three scenarios that reflect different assumptions
regarding bromide content: 1) lower bound loadings based solely on native bromide content in coal at all
plants, 2) upper bound loadings based on both native bromide content in coal and the use of bromide additives
and brominated activated carbon at all plants, and 3) best estimate loadings based on the EPA's estimates of
the native bromide content in coal and the most likely bromide usage for each plant (see Supplemental TDD).
In total, the EPA considered nine different loadings scenarios: the EPA's best estimate loadings for the
baseline, Option 1, Option 2, Option 3, and Option 4 (used in the analysis presented in Chapter 4); lower
bound loadings for baseline and Option 4; and upper bound loadings for baseline and Option 4. The next
section presents the results of scenarios based on the lower and upper bound loadings.
Section C.4.2 presents the results of scenarios based on alternative relationships between bromide
concentrations and TTHM concentrations changes.
C.4.1 Sensitivity to bromide loads
The EPA analyzed the sensitivity of the benefits to estimated bromide loadings from steam electric power
plants under lower bound and upper bound scenarios for regulatory Option 4. As detailed in the Supplemental
• Option 1
~ Option 2
¦ Option 3
A Option 4
2021
2031
2041
2051
2061
2071
2081
2091
2101
2111
2121
EPA-821-R-19-011
C-14
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
TDD, the lower bound scenario derives bromide loadings based solely on bromide levels occurring naturally
in coal whereas the upper bound scenario derives loadings based on both natural bromide content in coal and
bromide product additions for control of mercury emissions at all facilities (U.S. EPA, 2019b).
Results of this assessment yielded estimated annualized benefits over 27 years ranging from $25 million to
$215 million (3 percent discount rate) and from $16 million to $139 million (7 percent discount rate) for the
lower bound and upper bound scenarios, respectively.
C.4.2 Sensitivity to relationship between bromide and TTHM changes
As described in Section 4.3.2.3, the EPA used the relationship shown in Figure 4-2 to estimate the changes in
TTHM concentrations resulting from changes in bromide concentrations in source water as a result of the
regulatory options in comparison to the baseline. The median conversion factor used to develop the best-fit
curve reflects operating conditions for a diverse set of water treatment plants with varying treatment processes
and source water quality. The median conversion factor developed by Regli et al. (2015) was based on
monthly bromide-TTHM relationships and is used in this analysis to represent year-round conditions at PWS
potentially affected by steam electric power plant discharges. Long-term conditions are most relevant to
analyzing the relationship between TTHM and bladder cancer incidence.
To evaluate the sensitivity of the analysis to variability in the relationship between source water bromide level
changes and changes in treated water TTHM levels (such as those due to variability in PWS treatment
processes), the EPA also analyzed benefits using the 5th and 95th percentile estimates from Regli et al. (2015)
(Figure C-4). EPA determined that the TTHM values derived from the 5th and 95th percentile estimates are
useful for a sensitivity analysis, but notes that the conditions they reflect may be episodic and therefore less
likely to reflect long-term TTHM exposure and the resulting changes in bladder cancer incidence.
EPA-821-R-19-011
C-15
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix C: Bromide-related Benefits
Figure C-4: Modeled Sensitivity Analysis Relationship between Changes in Bromide Concentration
and Changes in TTHM Concentrations based on Regli et al. (2015).
• Median
~ 5th percentile
¦ 95th percentile
if ABr < 100 fig/L:
ATHM495th = -1.14xl0'7 ABr4
if ABr > 100 ug/L:
+ 3.58xl0'5 AB
r3 -4.51xl0r3 ABR2 + 3.82x10 J
ATHM435th
ABr
A THM4,5th =
17.5 +0.1 (ABr-100)
..J
r"'
if ABr < 100 /ig/L:
v
ATHM^ = -8.30xl0~8 ABr4 + 1.96x10 5 ABr3 -1.81x10 s1 ABR2 +
if ABr > 100 ng/L:
A THM4SM, = 5.80 + 0.022 (ABr-100)
1.26x1a1 ABr
ATHM450th
v
/
+
if ABr < 100 fig/L:
ATHM45th - 3.07xlC
if ABr > 100 ug/L:
8 ABr4 - 5.57x106 ABr3 + 2.61x10" ABR2 + 6
80xia3ABr
0 20 40 60 80 100 120 140 160 180 200
ABr (ug/L)
Source: U.S. EPA Analysis, 2019. Based on data in Regli et al. (2015).
Table C-4 summarizes the changes in bromide concentrations and associated changes in TTHM
concentrations, number of PWS, and populations affected for the 5th and 95th percentile sensitivity analyses
under Option 4.
Table C-4: Distribution of Estimated Changes in TTHM Concentration, Number of PWS and
Populations.
ABr range (ug/L)
ATTHM range (ng/L)a
Number of PWSb
Total population served
(million people)c
Option 4 - 5th Percentile
>0 to 10
6.16E-06 to 0.0879
626
24.6
10 to 30
0.0902 to 0.313
280
5.1
30 to 50
0.315 to 0.478
24
0.2
50 to 75
0.5 to 0.575
10
0.6
>75
0.606 to 3.8
29
0.9
Option 4 - 95th Percentile
>0 to 10
0.000347 to 3.38
626
24.6
10 to 30
3.45 to 8.28
280
5.1
30 to 50
8.31 to 11.4
24
0.2
50 to 75
11.8 to 13.9
10
0.6
EPA-821-R-19-011
C-16
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Appendix C: Bromide-related Benefits
Table C-4: Distribution of Estimated Changes in TTHM Concentration, Number of PWS and
Populations.
ABr range (ng/L)
ATTHM range (ng/L)a
Number of PWSb
Total population served
(million people)c
>75
14.9 to 37.5
29
0.9
a Nonzero changes in concentrations estimated under EPA's best estimate scenario using the 5th and 95th percentile relationships
between changes in bromide and changes in TTHM.
b Includes systems that are directly and indirectly affected by steam electric power plant discharges.
c Approximately 0.3 percent of the population served by PWS affected by bromide discharges from steam electric power plants
saw no change in bromide concentration under Option 4.
Source: U.S. EPA analysis, 2019.
Table C-5 summarizes results for the Br-TTHM relationship sensitivity scenarios. Total bladder cancer cases
avoided under Option 4 range from 97 to 2,417 and total cancer deaths avoided range from 39 to 978 for the
5th and 95th percentile estimates, respectively. Estimated annualized benefits associated with avoided cancer
cases and deaths range from $11 million to $265 million. These two bounds illustrate the range in benefits
associated with potential variability in TTHM formation.
Table C-5: Sensitivity of Estimated Bromide-related Benefits of Regulatory Option 4.
Changes in health
outcomes from TTHM
Benefits (million 2018$, discounted to 2020 at 3% and 7%)
exposure 2021-2047
Option
Br-TTHM
Total
Total
Total PV of avoided
Total PV of avoided
Annualized benefits
Relationship
bladder
mortality
morbidity avoided
over 27 years
cancer
cancer
deaths
avoided
3%
7%
3%
7%
3%
7%
cases
Discount
Discount
Discount
Discount
Discount
Discount
avoided
Rate
Rate
Rate
Rate
Rate
Rate
4
5th Percentile
97
39
$198.9
$86.7
$1.0
$0.4
$10.6
$6.8
4
95th Percentile
2,417
978
$4,977.9
$2,178.2
$24.9
$11.2
$265.0
$170.7
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
C-17
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix D: Water Quality and Fish Tissue
Appenc Derivation of Ambient Water and Fish Tissue Concentrations in
Receiving and Downstre. aches
This appendix describes the methodology the EPA used to estimate in-stream and fish tissue concentrations
under the baseline and each of the four regulatory options. The concentrations are used as inputs to estimate
the water quality changes and human health benefits of the regulatory options. Specifically, the EPA used in-
stream toxics concentrations to derive fish tissue concentrations used to analyze human health effects from
consuming self-caught fish (see Chapter J) and to analyze non-use benefits of water quality changes (see
Chapter 6). Nutrient and suspended sediment concentrations are used to support analysis of non-use benefits
from water quality changes (see Chapter 6).
The overall modeling methodology builds on data and methods described in the Supplemental EA and
Supplemental TDD for the regulatory options (U.S. EPA, 2019a; 2019b). The following sections discuss
calculations of the toxics concentrations in streams and fish tissue and nutrient and sediment concentrations in
streams.
D.1 Toxics
D. 1.1 Estimating Water Concentrations in each Reach
The EPA first estimated the baseline and post-compliance toxics concentrations in reaches receiving steam
electric power plant discharges and downstream reaches.
The D-FATE model (See Chapter 3) was used to estimate water concentrations. The model tracks the fate and
transport of discharged pollutants through a reach network defined based on the medium resolution NHD.92
The hydrography network represented in the D-FATE model consists of 10,315 reaches within 300 km of a
steam electric power plant, 10,284 of which are determined to be potentially fishable.93
The analysis involved the following key steps for the baseline and each of the four regulatory options:
• Summing plant-level loadings to the receiving reach. The EPA summed the estimated plant-level
annual average loads (see TDD) for each unique reach receiving plant discharges from steam electric
power plants in the baseline and four regulatory options. For a description of the approach EPA used
to identify the receiving waterbodies, see U.S. EPA (2019a).
92 The USGS's National Hydrology Dataset (NHD) defines a reach as a continuous piece of surface water with similar hydrologic
characteristics. In the NHD each reach is assigned a reach code; a reach may be composed of a single feature, like a lake or
isolated stream, but reaches may also be composed of a number of contiguous features. Each reach code occurs only once
throughout the nation and once assigned a reach code is permanently associated with its reach. If the reach is deleted, its reach
code is retired.
93 Reaches represented in the D-FATE model are those determined to be potentially fishable based on type and physical
characteristics. Because the D-FATE model calculates the movement of a chemical release downstream using flow data, reaches
must have at least one downstream or upstream connecting reach and have a non-negative flow and velocity. The D-FATE model
does not calculate concentrations for certain types of reaches, such as coastlines, treatment reservoirs, and bays; the downstream
path of any chemical is assumed to stop if one of these types of reach is encountered. Additionally, some types of reaches are
excluded from the set of fishable reaches, such as those designated as having Strahler Stream Order 1 in the NHDPlus, because
they do not have the flow rates and species diversity to support trophic level 3 and 4 species."
EPA-821-R-19-011
D-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix D: Water Quality and Fish Tissue
• Performing dilution and transport calculations. The D-FATE model calculates the concentration
of the pollutant in a given reach based on the total mass transported to the reach from upstream
sources and the EROM flows for each reach from NHDPlus v2. In the model, a plant is assumed to
release its annual load at a constant rate throughout the year. Each source-pollutant release is tracked
throughout the NHD reach network until the release has traveled 300 km (186 miles) downstream.
• Specifying concentrations in the water quality model. The D-FATE model includes background
data on estimated annual average pollutant concentrations to surface waters from facilities that
reported to the TRI in 2016. The EPA added background concentrations where available to
concentration estimates from steam electric power plant dischargers.
The EPA used the approach above to estimate annual average concentrations of ten toxics: arsenic, cadmium,
chromium VI, copper, lead, mercury, nickel, selenium, thallium, and zinc.
D. 1.2 Estimating Fish Tissue Concentrations in each Reach
To support analysis of the human health benefits associated with water quality improvements (see Chapter 4),
the EPA estimated concentrations of arsenic, lead, and mercury in fish tissue based on the D-FATE model
outputs discussed above.
The methodology follows the same general approach described in the Supplemental EA for estimating fish
tissue concentrations for receiving reaches (U.S. EPA, 2019a), but applies the calculations to the larger set of
reaches modeled using D-FATE, which include not only the receiving reaches analyzed in the EA, but also
downstream reaches. Further, the calculations use D-FATE-estimated concentrations as inputs, which account
not only for the steam electric power plant discharges, but also other major dischargers that report to TRI.
The analysis involved the following key steps for the baseline and each of the four regulatory options:
• Obtaining the relationship between water concentrations and fish tissue concentrations. The
EPA used the results of the Immediate Receiving Water (IRW) model (see Supplemental EA, U.S.
EPA 2019a) to parameterize the linear relationship between water concentrations in receiving reaches
and composite fish tissue concentrations (representative of trophic levels 3 and 4 fish consumed) in
these same reaches for each of the three toxics.
• Calculating fish tissue data for affected reaches. For reaches for which the D-FATE model
provides non-zero water concentrations (i.e.. reaches affected by steam electric power plants or other
TRI dischargers), the EPA used the relationship obtained in Step 1 to calculate a preliminary fish
tissue concentration for each pollutant.
• Imputing the fish tissue concentrations for all other modeled reaches. For reaches for which the
D-FATE model calculates water concentrations, the EPA added background fish tissue concentrations
based on the 10th percentile of the distribution of reported concentrations in fish tissue samples in the
National Listing Fish Advisory (NLFA) data94 (see Table D-l). The EPA found that the distribution
94 See https://fishadvisoryonline.epa.gov/general.aspx.
EPA-821-R-19-011
D-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix D: Water Quality and Fish Tissue
of these samples was consistent with values reported in Wathen et al (2015) and used the 10th
percentile as representative of background, "clean" reaches not affected by point source discharges.
• Validating and adjusting the fish tissue concentrations based on empirical data, if needed. The
EPA then applied the same method used to validate and adjust estimated fish tissue data in the IRW
model to ensure that the fish tissue concentrations calculated based on the D-FATE model outputs are
reasonable when compared to measured data. The approach involves applying order-of-magnitude
adjustments in cases where the preliminary concentrations are greater than empirical measurements
for a given reach or geographic area by an order of magnitude or more. The Supplemental EA
describes the methodology in greater detail.
The analysis provides background toxics-specific composite fish fillet concentrations for each reach modeled
in the D-FATE model. Total fish tissue concentrations (D-FATE modeled concentrations plus background
concentrations) are summarized in Table D-2.
Table D-1: Assumed Background Fish Tissue
Concentrations, based on 10th percentile
Parameter
Pollutant Concentration (mg/kg)
As
0.039
Hg
0.058
Pb
0.039
Source: U.S. EPA Analysis, 2019
Table D-2: Imputed and Validated Fish Tissue Concentrations by Regulatory Option
Fish fillet concentration (mg/kg)
Regulatory
Option
Arsenic
Lead
Mercury
Min
Max
Mean
Min
Max
Mean
Min
Max
Mean
Baseline
0.0390
0.0742
0.0390
0.0390
0.4919
0.0393
0.0580
4.2878
0.0628
Option 1
0.0390
0.0955
0.0390
0.0390
0.7650
0.0395
0.0580
5.5269
0.0657
Option 2
0.0390
0.0955
0.0390
0.0390
0.7650
0.0395
0.0580
54.694
0.0769
Option 3
0.0390
0.0955
0.0390
0.0390
0.7650
0.0395
0.0580
6.5401
0.0679
Option 4
0.0390
0.0955
0.0390
0.0390
0.7650
0.0395
0.0580
5.5269
0.0665
Source: U.S. EPA Analysis, 2019.
D.2 Nutrients and Suspended Sediment
The EPA used the USGS's SPARROW model to estimate nutrient and sediment concentrations in receiving
and downstream reaches. The calibrated, national models used for this analysis are the same as those used to
estimate in-stream concentrations of TN, TP and TSS in the Construction and Development (C&D) Industry
Category ELGs (see U.S. EPA, 2009a). The approach involved the following steps:
• Referencing the receiving reaches to E2RF1 reaches. The EPA overlaid the medium resolution
NHD and E2RF1 features in GIS to develop the crosswalk between the two hydrologic networks.
• Summing the loads for each E2RF1. The EPA summed the plant-level loadings over each E2RF1 in
the baseline and under each of the four regulatory options.
EPA-821-R-19-011
D-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix D: Water Quality and Fish Tissue
• Calculating the change in loading for each E2RF1. The EPA calculated the difference between the
baseline and post-compliance loadings under each of the four regulatory options.
• Specifying the change in loading in SPARROW. The national SPARROW models for nutrients do
not have an explicit explanatory variable for point source loadings in mass units. In the TN and TP
SPARROW models, point sources (e.g., wastewater treatment plants) are represented by a population
variable. The national calibrated models show contributions of 2.2514 kg TN/capita and
0.2319 kg TP/capita for point sources. The EPA used these calibrated loading factors to express the
load reductions obtained under each of the regulatory options into population-equivalent in
SPARROW. This population-equivalent loading was subtracted from the baseline population value
for each reach when running the SPARROW model. For the suspended sediment model, the EPA
used the same approach as used for the C&D ELG analysis, which involved adjusting the mass flux
attributed to the urban land explanatory variable in the model to subtract the change in loading
achieved under each option, under the assumption that steam electric power plant loadings are
implicitly accounted for in the urban land component of the model (see U.S. EPA, 2009a).
The model provides estimated annual average post-compliance concentrations in each E2RF1, which the EPA
compared with baseline conditions obtained directly from the national, calibrated model.
EPA-821-R-19-011
D-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix E: Georeferencing Intakes
Appendix E Georeferencing Surface Water Intakes to the Medium-resolution
Stream Network
The EPA used the following steps to assign PWS surface water intakes to waters represented in the medium-
resolution NHD Plus version 2 dataset and identify those intakes potentially affected by steam electric power
plant discharges.
1. Identify the closest (simple cartesian distance) medium-resolution NHD feature (including Flowline,
Area, and Waterbody) to each PWS intake.
2. If the closest feature to a given intake was an NHD Flowline, reference the intake to this flowline.
3. If the closest feature to a given intake was an NHD Area or Waterbody, consider the Flowlines
contained within or intersected by the Area/Waterbody.
a. If any of the Flowlines associated with the Area/Waterbody were on the flowpath
downstream from a steam electric power plant, select the Flowline within this set and closest
to the intake.
b. If none of the Flowlines were on the flowpath downstream from a steam electric power
plant, select the Flowline closest to the intake.
c. If there were no Flowlines associated with the Area/Waterbody, select the closest Flowline.
The EPA then compared the set of Flowline COMIDs identified in steps 2 and 3 to NHD COMIDs in the
downstream flowpath of steam electric power plant discharges. COMIDs that georeferenced directly to the
downstream flowpath received a "Category 1" designation. Intakes that were georeferenced to COMIDs
within 10 km of the downstream flowpath received a "Category 2" designation. The EPA included all intakes
within 10 km of the discharge flow path to account for cases where georeferencing did not select the correct
COMID based on uncertainty in the flow direction or stream network connectivity. For example, if a PWS
intake was located on a wide reach like the Mississippi River, the above methods may assign that intake to a
tributary COMID.
As discussed in Chapter 3, the EPA did not model complex waterbodies (e.g., Great Lakes) explicitly.
Therefore, the Agency reviewed all intakes within 50 miles of the plants discharging to the Great Lakes or
other non-modeled waterbodies to classify intakes that withdraw directly from the non-modeled waterbodies
as "Category 3". These intakes are excluded from the subsequent analysis.
Table E-l summarizes the intake categorization following the above steps.
Table E-1: Summary of Intakes Potentially Affected by Steam Electric Power Plant Discharges
Categorization
Number of Intakes
Category 1 (on flow path)
297
Category 2 (within 10 km of flow path) but not Category 1
313
Category 3 (on Great Lakes or other non-modeled waterbodies)
67
Total all categories
677
Source: U.S. EPA Analysis, 2019.
EPA-821-R-19-011
E-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix E: Georeferencing intakes
Figure E-l summarizes how the EPA subset Category 1 and 2 PWS intakes for a more targeted categorization
review.
Figure E-1: PWS Intakes Review Subset
G
610 unique
PWS fa cilities
"community
water systems"
with a
"permanent"
s-upply
georeferenceri
to a COM ID
with non-iero
changesin
bromide
concentrations
L
Source: U.S. EPA Analysis, 2019.
The EPA evaluated the "Category 2" PWS intakes further using spatial reference to any steam electric
downstream flow paths and SDWIS facility information, namely facility name.
The EPA excluded intakes from the benefits analysis if they were:
• on an upstream or visually unconnected body of water from the steam electric downstream flow path,
• did not sit on a visible body of water when looking at the topographical maps and/or orthophotos,
• had a PWS facility name indicating that it was not a surface water intake {i.e., included the word
"welF').95
The EPA recategorized intakes as Category 1 if they were:
• on the same NHD waterbody as the steam electric downstream flow path (prominent examples
include intakes on Lake Norman, Upper or Lower Potomac River, and Missouri River) or
• the PWS facility name in SDWIS corresponded with the named reach of the steam electric
downstream flow path.
Of the 271 Category 2 facilities that the EPA reviewed, 102 facilities were recategorized into Category
1. Therefore, the EPA included a total number of 349 PWS intakes96 in the human health benefits analysis.
This criterion resulted in the omission of only one facility in Tennessee.
— Only intakes with facility types categorized by SDWIS as "Intake" or "Reservoir" were retained in the human health benefits
analysis. One of the 349 PWS intakes (PWS ID IA9778045) was categorized as "Infiltration Gallery" and was thus not included,
bringing the total number of PWS intakes included in the analysis to 348 (see Table 4-1).
EPA-821-R-19-011
E-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Appendix F: Estimation of Exposed Population
Appendix F Estimation of Exposed Population for Fish Ingestion Pathway
The assessment uses the Census Block Group as the geographic unit of analysis, assigning a radial distance
(e.g., 50 miles) from the Census Block Group centroid. The EPA assumes that all modeled reaches within this
range are viable fishing sites, with all unaffected reaches viable substitutes for affected reaches within the area
around the Census Block Group.
By focusing on distance from the Census Block Group, rather than distances from affected reaches, each
household is only included in the assessment once, eliminating the potential for double-counting of
households that are near multiple affected waterbodies.
Figure F-l presents a hypothetical example focusing on two Census Block Groups (square at the center of
each circular area), each near five waterbodies with water quality changes under the regulatory options (thick
red lines).
Figure F-l. Illustration of Intersection of Census Block Groups and COMIDs.
0
9
0
o
I?
U
t?
I?
li
0
ii
0
a
0
I
ft' fi\r fi ft
Source: U.S. EPA (2015a).
Note that a similar approach is used to identify populations for the analysis of non-market benefits in Chapter
6. In that case, the circles represent the outer edge of the 100-mile buffer around each block group.
Highlighted in red are the affected NHD reaches under regulatory options for which baseline WQI and AWQI
would be estimated
EPA-821-R-19-011
F-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix G: IQ Sensitivity Analysis
Appendix G Sensitivity Analysis for IQ Point-based Human Health Effects
The EPA monetized the value of an IQ point based on the methodology from Salkever (1995). As a
sensitivity analysis of the benefits of lead and mercury exposure, the EPA used alternative, more conservative
estimates provided in Lin et al. (2018) which indicate that a one-point IQ reduction reduces expected lifetime
earnings by 1.39 percent (as compared to 2.63 percent based on Salkever (1995)). As noted in Sections 5.3
and 5.4, values of an IQ point used in the analysis of health effects in children from lead exposure are
discounted to the third year of life to represent the midpoint of the exposed children population, values of an
IQ point used in the analysis of health effects associated with in-utero exposure to mercury are discounted to
birth. Table G-l summarizes the estimated values of an IQ point based on Lin et al (2018), using 3 percent
and 7 percent discount rates.
Table G-1: Value of an IQ Point (2018$) based on
Expected Reductions in Lifetime Earnings
Discount Rate
Value of an IQ Point3 (2018$)
Value of an IQ point Discounted to Age 3
3 percent
$11,279
7 percent
$2,371
Value of an IQ point Discounted to Birth
3 percent
$10,322
7 percent
$1,936
a. Values are adjusted for the cost of education.
Source: U.S. EPA (2019e,2019h) analysis of data from Lin et al. (2018)
G.1 Health Effects in Children from Changes in Lead Exposure
Table G-2 shows the social welfare effects associated with changes in IQ losses from lead exposure via fish
consumption. The EPA estimated that all regulatory options lead to slight increases in lead exposure and, as a
result, forgone benefits. The total net change in IQ points over the entire population of children with changes
in lead exposure ranges from -11.07 points to 0.90 points. Annualized monetary values of changes in IQ
losses from differences in lead exposure, based on the Lin et al. (2018) IQ point value, range from -$4,950 to
$400 (3 percent discount rate) and from -$1,080 to $90 (7 percent discount rate).
Table G-2: Estimated Monetary Value of Changes in IQ Losses for Children Exposed to Lead
Annualized Value of Changes in IQ Point
Average Annual
Total Change in IQ Points,
Lossesab
Regulatory Option
Number of Affected
2021 to 2047 in All
(Thousands 2018$)
Children Oto 7C
Affected Children 0 to 7
3 Percent Discount
Rate
7 Percent Discount
Rate
Option 1
1,521,036
-3.58
-$1.6
-$0.35
Option 2
1,521,036
-11.07
-$4.95
-$1.08
Option 3
1,521,036
0.35
$0.16
$0.03
Option 4
1,521,036
0.90
$0.40
$0.09
a. Assumes that the loss of one IQ point results in the loss of 1.39 percent of lifetime earnings (following Lin et al., 2018 values from
U.S. EPA (2019e)).
b. Negative values represent forgone benefits.
c. The number of affected children is based on reaches analyzed across the four options. Some of the children included in this count
see no changes in exposure under some options.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
G-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix G: IQ Sensitivity Analysis
G.2 Heath Effects in Children from Changes in Mercury Exposure
Table G-3 shows the estimated changes in IQ point losses for infants exposed to mercury in-utero and the
corresponding monetary values, using a 3 percent and 7 percent discount rates. All regulatory options result in
a net increase in IQ losses and, as a result, in forgone benefits to society. Annualized monetary values of
increased IQ losses from changes in mercury exposure, based on the Lin et al. (2018) IQ point value, range
from -$0.17 million (Option 1) to -$1.54 million (Options 2 and 3) using a 3 percent discount rate.
Table G-3: Estimated Monetary Values from Changes in IQ Losses for Infants from Mercury Exposure
Regulatory Option
Number of Affected
Total Changes in IQ Losses,
2021 to 2047 in All
Affected Infants
Annualized Value of Changes in IQ Point
Lossesa b (Millions 2018$)
Infants per Yearc
3 Percent Discount
Rate
7 Percent Discount
Rate
Option 1
225,272
-411
-$0.17
-$0.03
Option 2
225,272
-3,785
-$1.54
-$0.30
Option 3
225,272
-3,777
-$1.54
-$0.30
Option 4
225,272
-2,021
-$0.81
-$0.16
a. Assumes that the loss of one IQ point results in the loss of 1.39 percent of lifetime earnings (following Lin et al., 2018 values from
U.S. EPA (2019e and 2019h)).
b. Negative values represent forgone benefits.
c. The number of affected children is based on reaches analyzed across the four options. Some of the children included in this count
see no changes in exposure under some options.
Source: U.S. EPA Analysis, 2019
EPA-821-R-19-011
G-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
Appenc ,ir . i Methodology for Estimati ' ¦ •
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
• Model 1 assumes that individuals' marginal WTP depends on the level of water quality, but not on
the magnitude of the water quality change specified in the survey. This restriction means that, the
meta-model satisfies the adding-up condition, a theoretically desirable property.
• Model 2 allows marginal WTP to depend not only on the level of water quality but also on the
magnitude of the water quality change specified in the survey. The model allows for the possibility
that marginal WTP for improving from 49 to 50 on the water quality index depends on whether
respondents were asked to value a total water quality change of 10, 20, or 50 points on a WQI scale.
This model provides a better statistical fit to the meta-data, but it satisfies the adding-up conditions
only if the same magnitude of the water quality change is considered (e.g., 10 points). To uniquely
define the demand curve and satisfy the adding-up condition using this model, the EPA treats the
water quality change variable as a methodological variable and therefore must make an assumption
about the size of the water quality change that would be appropriate to use in a stated preference
survey designed to value water quality changes resulting from the regulatory options. When the water
quality change is fixed at the mean of the meta-data, the predicted WTP is very close to the central
estimate from Model 1.
The EPA used the two meta-regression models in a benefit transfer approach that follows standard methods
described by Johnston et al. (2005), Shrestha et al. (2007), and Rosenberger and Phipps (2007). In particular,
literature on benefit transfer recommends selecting values for methodological variables included in the
regression equation with the goal of providing conservative WTP estimates, subject to consistency with
methodological guidance in the literature. The literature also recommends setting variables representing
policy outcomes and policy context (i.e., resource and population characteristics) at the levels that might be
expected from a regulation. The benefit transfer approach uses CBGs as the geographic unit of analysis.98 The
transfer approach involved projecting benefits in each CBG and year, based on the following general benefit
function:
Equation H-1.
In (MWTPYib) = Intercept + ^ (coefficient^) x (independent variable valuet)
Where
ln(MWTPr,B)
coefficient
independent
variable values
= The predicted natural log of marginal household WTP for a given year (7)
and CBG (B).
= A vector of variable coefficients from the meta-regression.
= A vector of independent variable values. Variables include baseline water
quality level ( WQI-BLy,b) and expected water quality under the regulatory
option ( WQI-PCy,b) for a given year and CBG.
A Census Block group is a group of Census Blocks (the smallest geographic unit for the Census) in a contiguous area that never
crosses a State or county boundary. A block group typically contains a population between 600 and 3,000 individuals. There are
217,740 block groups in the 2010 Census. See http://www.census.gov/geo/maps-data/data/tallies/tractblock.html.
EPA-821-R-19-011
H-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
Here, ln(MWTPYB) is the dependent variable in the meta-analysis—the log of approximated marginal WTP
per household, in a given CBG B for water quality in a given year Y." The baseline water quality level (WQI-
BLy.b) and expected water quality under the regulatory option (WQI-PCYb) were based on water quality at
waterbodies within a 100-mile buffer of the centroid of each CBG. A buffer of 100 miles is consistent with
Viscusi et al. (2008) and with the assumption that the majority of recreational trips would occur within a 2-
hour drive from home. Because marginal WTP is assumed to depend, according to Equation H-l, on both
baseline water quality level ( WQI-BLy,b) and expected water quality under the regulatory option ( WQI-PCy,b),
the EPA estimated the marginal WTP for water quality changes resulting from the regulatory options at the
mid-point of the range over which water quality was changed, WQIy,b = (1/2) ( WQI-BLy,b + WQI-PCYtB)).
In this analysis, the EPA estimated WTP for the households in each CBG for waters within a 100-mile radius
of that CBG's centroid. The EPA chose the 100 mile-radius because households are likely to be most familiar
with waterbodies and their qualities within the 100-mile distance. However, this assumption may be an
underestimate of the distance beyond which households have familiarity with and WTP for waterbodies
affected by steam electric power plant discharges and their quality. By focusing on a buffer around the CBG
as a unit of analysis, rather than buffers around affected waterbodies, each household is included in the
assessment exactly once, eliminating the potential for double-counting of households.100 Total national WTP
is calculated as the sum of estimated CBG-level WTP across all block groups that have at least one affected
waterbody within 100 miles. Using this approach, the EPA is unable to analyze the WTP for CBGs with no
affected waters within 100 miles. Appendix /''describes the methodology used to identify the relevant
populations.
In each CBG and year, predicted WTP per household is tailored by choosing appropriate input values for the
meta-analysis parameters describing the resource(s) valued, the extent of resource changes (i.e.. WQI- PCYb),
the scale of resource changes relative to the size of the buffer and relative to available substitutes, the
characteristics of surveyed populations (e.g., users, nonusers), and other methodological variables. For
example, the EPA assumed that household income (an independent variable) changes overtime, resulting in
household WTP values that vary by year.
Table H-l provides details on how the EPA used the meta-analysis to predict household WTP for each CBG
and year. The table presents the estimated regression equation intercept, variable coefficients (coefficient,) for
the two models, and the corresponding independent variable names and assigned values. The meta-regression
allows the Agency to forecast WTP based on assigned values for model variables that are chosen to represent
a resource change in the context of the regulatory options. The EPA assigned a value to each model variable
corresponding with theory, characteristics of the water resources, and sites potentially affected by the
regulatory options. This follows general guidance from Bergstrom and Taylor (2006) that meta-analysis
benefit transfer should incorporate theoretical expectations and structures, at least in a weak form.
99 To satisfy the adding-up condition, as noted above, the EPA normalized WTP values reported in the studies included in the meta-
data so that the dependent variable is MWTP per WQI point. This 'average' marginal WTP value is an approximation of the
MWTP value elicited in each survey scenario.
100 Population double-counting issues can arise when using "distance to waterbody" to assess simultaneous improvements to many
waterbodies.
EPA-821-R-19-011
H-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
In this instance, the EPA assigned six study and methodology variables, (thesis, volunt, nonparam,
non reviewed, lump sum, and WTP median) a value of zero. One methodological variable, outliers trim,
was included with an assigned value of 1. Because the interpretation of the study year variable (Lnyear) is
uncertain, the EPA gave the variable a value of 3.0796, which is the 75th percentile of the year values in the
meta-data. This value assignment reflects an equal probability that the variable represents a real time trend (in
which case its value should be set to the most recent year of the analysis) and spurious effects (in which case
its values should be set to the mean value from the meta-data). The choice experiment variable fee) was set to
1 to reflect recent trends in the use of choice experiments within the environmental valuation literature. Model
2 includes an additional variable, water quality change (In quality_ch), which as discussed above allows the
function to reflect differences in marginal WTP based on differences in the magnitude of changes presented to
survey respondents when eliciting values. To ensure that the benefit transfer function satisfies the adding-up
condition, this variable was treated as a demand curve shifter, similar to the methodological control variables,
and held fixed for the benefit calculations. To estimate low and high values of WTP for water quality changes
resulting from the regulatory options, the EPA estimated marginal WTP using two alternative settings of the
Inquality variable: AWQI = 5 units and AWQI = 50 units, which represent the low and high end of the range
of values observed in the meta-data.
All but one of the region and surveyed population variables vary based on the characteristics of each CBG.
For median household income, the EPA used CBG-level median household income data from the 2016
American Community Survey (5-year data) and used a stepwise autoregressive forecasting method to estimate
future annual state level median household income. The EPA set the variable nonusers only to zero because
water quality changes are expected to enhance both use and non-use values of the affected resources and thus
benefit both users and nonusers (a nonuser value of 1 implies WTP values that are representative of nonusers
only, whereas the default value of 0 indicates that both users and nonusers are included in the surveyed
population). The EPA set the variable river to 1 and mult type to 0 because the analysis focuses only on rivers
and streams. Other waterbody types (e.g., lakes and estuaries) are excluded from the analysis.
The geospatial variables corresponding to the sampled market and scale of the affected resources (In ar agr,
ln_ar ratio , sub proportion) vary based on attributes of the CBG and attributes of the nearby affected
resources. For all options, the affected resource is based on the 10,315 NHD reaches potentially affected by
steam electric power generating plant discharges under baseline conditions. The affected resource for each
CBG is the portion of the 10,315 reaches that fall within the 100-mile buffer of the CBG. Spatial scale is held
fixed across regulatory options. The variable corresponding to the sampled market (ln_ar ratio) is set to the
mean value across all CBGs included in the analysis of benefits from water quality changes resulting from the
regulatory options, and thus does not vary across affected CBGs.
Because data on specific recreational uses of the water resources affected by the regulatory options are not
available, the recreational use variables (swim use, gamejish, boat use) are set to zero, which corresponds to
"unspecified" or "all" recreational uses in the meta-data.101 Water quality variables (Q and Inquality_ch) vary
across CBGs and regulatory options based on the magnitude of the reach-length weighted average water
quality changes at affected resources within the 100-mile buffer of each CBG.
101 If a particular recreational use was not specified in the survey instrument, EPA assumed that survey respondents were thinking of
all relevant uses.
EPA-821-R-19-011
H-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
Table H-1: Independent Variable Assignments for Surface Water Quality Meta-Analysis
Variable Type
Variable
Coefficient
Assigned
Value
Explanation
Model 1
Model 2
Study
Methodology
and Year
intercept
-1.040
-6.14
Ce
0.377
0.423
1
Binary variable indicating that the study is a choice
experiment. Set to one to reflect that choice
experiments represent current state-of-art
methods in stated preference literature.
thesis
0.866
0.774
0
Binary variable indicating that the study is a thesis
or dissertation. Set to zero because studies
published in peer-reviewed journals are preferred.
Inyear
-0.412
-0.5
3.0796
Natural log of the year in which the study was
conducted {i.e., data were collected), converted to
an index by subtracting 1980. Set to the natural log
of the 75th percentile of the year index value for
studies in the metadata (21.7) to reflect
uncertainty in the variable interpretation. If the
variable represents a real time trend, the
appropriate value should reflect the most recent
year of the analysis. If it represents spurious
effects, the values should reflect the mid-point
from meta-data. Both interpretations are equally
probable.
volunt
-1.390
-1.184
0
Binary variable indicating that WTP was estimated
using a payment vehicle described as voluntary as
opposed to, for example, property taxes. Set to
zero because hypothetical voluntary payment
mechanisms are not incentive compatible
(Mitchell and Carson 1989).
outlierstrim
-0.367
-0.291
1
Binary variable indicating that outlier bids were
excluded when estimating WTP. Set to one
because WTP estimates that exclude outlier bids
are preferable.
nonparam
-0.408
-0.39
0
Binary variable indicating that regression analysis
was not used to model WTP. Set to zero because
use of the regression analysis to estimate WTP
values is preferred.
nonreviewed
-0.709
-0.871
0
Binary variable indicating that the study was not
published in a peer-reviewed journal. Set to zero
because studies published in peer-reviewed
journals are preferred.
lumpsum
0.843
0.773
0
Binary variable indicating that the study provided
WTP as a one-time, lump sum or provided annual
WTP values for a payment period of five years or
less. Set to zero to reflect that the majority of
studies from the meta-data estimated an annual
WTP, and to produce an annual WTP prediction.
EPA-821-R-19-011
H-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
Table H-1: Independent Variable Assignments for Surface Water Quality Meta-Analysis
Variable Type
Variable
Coefficient
Assigned
Value
Explanation
Model 1
Model 2
wtpmedian
-0.161
-0.151
0
Binary variable indicating that the WTP measure
from the study is the median. Set to zero because
only average or mean WTP values in combination
with the number of affected households would
mathematically yield total benefits if the
distribution of WTP is not perfectly symmetrical.
Region and
Surveyed
Population
northeast
1.180
0.593
Varies
Binary variable indicating that the affected
population is located in a Northeast U.S. state,
defined as ME, NH, VT, MA, Rl, CT, and NY. Set
based on the state in which the CBG is located.
central
0.561
0.726
Varies
Binary variable indicating that the affected
population is located in a Central U.S. state,
defined as OH, Ml, IN, IL, Wl, MN, IA, MO, ND, SD,
NE, KS, MT, WY, UT, and CO. Set based on the
state in which the CBG is located.
south
1.400
1.563
Varies
Binary variable indicating that the affected
population is located in a Southern U.S. state,
defined as NC, SC, GA, FL, KY, TN, MS, AL, AR, LA,
OK, TX, and NM. Set based on the state in which
the CBG is located.
nonusersonly
-0.586
-0.54
0
Dummy variable indicating that the sampled
population included nonusers only; the alternative
case includes all households. Set to zero to
estimate the total value for aquatic habitat
changes for all households, including users and
nonusers.
Inincome
0.333
0.96
Varies
Natural log of median household income values
assigned separately for each CBG. Varies by year
based on the estimated income growth in future
years.
Sampled
Market and
Affected
Resource
multtype3
-0.827
-0.63
0
Binary variable indicating that multiple waterbody
types are affected {e.g., river and lakes). Set to
zero because calculations are based exclusively on
rivers.
River
-0.079
-0.174
1
Binary variable indicating that rivers are affected.
Set to one because calculations are based
exclusively on stream miles. The EPA did not
estimate water quality changes for other
waterbody types {e.g., lakes and estuaries).
swim use
-0.234
-0.27
0
Binary variables that identify studies in which
swimming, gamefish, and boating uses are
specifically identified. Since data on specific
recreational uses of the reaches affected by steam
electric power plant discharges are not available,
set to zero, which corresponds to all recreational
uses.
Gamefish
0.233
-0.01
0
boatuse
-0.725
-0.32
0
EPA-821-R-19-011
H-6
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
Table H-1: Independent Variable Assignments for Surface Water Quality Meta-Analysis
Variable Type
Variable
Coefficient
Assigned
Explanation
Model 1
Model 2
Value
Natural log of the proportion of the affected
resource area which is agricultural based on
National Land Cover Database, reflecting the
nature of development in the area surrounding the
resource. Used Census county boundary layers to
Inaragr
-0.271
-0.413
Varies
identify counties that intersect affected resources
within the 100-mile buffer of each CBG. For
intersecting counties, calculated the fraction of
total land area that is agricultural using the
National Land Cover Dataset (NLCD). The ln_ar_agr
variable was coded in the metadata to reflect the
area surrounding the affected resources.
The natural log of the ratio of the sampled area
(sa area) relative to the affected resource area
(defined as the total area of counties that intersect
Inarratio
-0.034
-0.057
1.238
the affected resource(s)) (ar_total_area). Set to
the mean value from the CBG's with 100-mile
buffers containing waters affected by the
regulatory options.
The size of the affected resources relative to
available substitutes. Calculated as the ratio of
subproportion
1.100
0.607
Varies
affected reaches miles to the total number of
reach miles within the buffer that are the same
order(s) as the affected reaches within the buffer.
Its value can range from 0 to 1.
Water Quality
Because marginal WTP is assumed to depend on
both baseline water quality and expected water
quality under the regulatory option, this variable is
set to the mid-point of the range of water quality
Q
-0.015
-0.004
Varies
changes due to the regulatory options, WQIy,b =
(1/2)(WQI-BLy,b + WQI-PCy,b)- Calculated as the
length-weighted average WQI score for all
potentially affected COM IDs within the 100-mile
buffer of each CBG.
Inqualitych
NA
-0.746
ln(5) or
ln(50)
Ln_quality_ch was set to the natural log of
AWQI=5 or AWQI=50 for high and low estimates of
the marginal WTP, respectively.
a. The meta-data includes six waterbody categories (1) river and stream, (2) lake, (3) all freshwater, (4) estuary, (5) river and lake, (6)
salt pond/marshes, Variable multi-type takes on a value of 1 if the study focused on waterbody categories (3) and (6). The EPA notes
that the overall effect of this variable should be considered in conjunction with the regional dummies (e.g., a study of the Lake
Okeechobee basin in Florida) and that only eight percent of all observations in the meta-data fall in the multiple waterbody
categories.
The central estimates for total WTP results shown in Table 6-2 are closer to the low estimates than the high
estimates for each regulatory option. The EPA tested several different functional forms for Model 2 and found
that the model has the highest explanatory power (R-squared) when water quality change is included in
logged form. This implies that water quality change has a nonlinear effect on marginal WTP (MWTP). In
particular, small initial increases in the scale of the water quality change scenario have a larger effect on
EPA-821-R-19-011
H-7
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix H: WTP Estimation Methodology
MWTP than subsequent increases. Therefore, the central estimate of MWTP (based on a water quality change
scenario of approximately 20 units) is closer to the low MWTP estimate (based on a water quality change
scenario of 50) than to the high MWTP estimate (based on a water quality change scenario of 5). In addition,
when Model 2 is used in a benefits transfer application with a water quality change of +20, the mean of the
meta-data, the results are very close to the results of Model 1. The EPA presents the results as a range because
a water quality change of +5 is closer to the size of water quality changes projected to result from the
regulatory options than the +20 analog to the central estimate, while the +50 represents the upper end of water
quality changes in existing surveys (and the lower end of the sensitivity benefits range).
EPA-821-R-19-011
H-8
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix I: WTP Estimation Methodology
Append: Uncertainty Associated wtah Estimate "? ^ ! Social Cost of Carbon
The methodology used to develop interim domestic SC-CO2 estimates and uncertainty associated with the
interim SC-CO2 values are the same as described in the RIA for the Affordable Clean Energy (ACE) final rule
(see U.S. EPA, 2019f). This appendix applies the methodology to the analysis of the climate benefits of
changes in CO2 emissions under the steam electric ELG regulatory options.
1.1 Overview of Methodology Used to Develop Interim Domestic SC-CO2 Estimates
The domestic SC-CO2 estimates rely on the same ensemble of three integrated assessment models (IAMs) that
were used to develop the global SC-CO2 estimates (DICE 2010, FUND 3.8, and PAGE 2009)102 used in the
benefits analysis of the 2015 ELG (see U.S. EPA, 2015a). The three IAMs translate emissions into changes in
atmospheric greenhouse concentrations, atmospheric concentrations into changes in temperature, and changes
in temperature into economic damages. The emissions projections used in the models are based on specified
socio-economic (GDP and population) pathways. These emissions are translated into atmospheric
concentrations, and concentrations are translated into warming based on each model's simplified
representation of the climate and a key parameter, equilibrium climate sensitivity. The effect of the changes is
estimated in terms of consumption-equivalent economic damages. As in the estimation of SC-CO2 estimates
used in the 2015 benefits analysis (U.S. EPA, 2015a), three key inputs were harmonized across the three
models: a probability distribution for equilibrium climate sensitivity; five scenarios for economic, population,
and emissions growth; and discount rates.103 All other model features were left unchanged. Future damages
are discounted using constant discount rates of both 3 and 7 percent, as recommended by OMB Circular A-4.
The domestic share of the global SC-CO2 - i. e., an approximation of the climate change impacts that occur
within U.S. borders - are calculated directly in both FUND and PAGE. However, DICE 2010 generates only
global SC-CO2 estimates. Therefore, EPA approximated U.S. damages as 10 percent of the global values from
the DICE model runs, based on the results from a regionalized version of the model (RICE 2010) reported in
Table 2 ofNordhaus (2017).104
The steps involved in estimating the social cost of CO2 are as follows. The three integrated assessment models
(FUND, DICE, and PAGE) are run using the harmonized equilibrium climate sensitivity distribution, five
socioeconomic and emissions scenarios, constant discount rates described above. Because the climate
sensitivity parameter is modeled probabilistically, and because PAGE and FUND incorporate uncertainty in
other model parameters, the final output from each model run is a distribution over the SC-CO2 in year t
based on a Monte Carlo simulation of 10,000 runs. For each of the IAMs, the basic computational steps for
calculating the social cost estimate in a particular year t are:
102 The full models names are as follows: Dynamic Integrated Climate and Economy (DICE); Climate Framework for Uncertainty,
Negotiation, and Distribution (FUND); and Policy Analysis of the Greenhouse Gas Effect (PAGE).
103 See the summary of the methodology in the 2015 Clean Power Plan docket, document ID number EPA-HQ-OAR-2013-0602-
37033, "Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866, Interagency
Working Group on Social Cost of Carbon (May 2013, Revised July 2015)". See also National Academies (2017) for a detailed
discussion of each of these modeling assumptions.
104 Nordhaus, William D. 2017. Revisiting the social cost of carbon. Proceedings of the National Academy of Sciences of the United
States, 114(7): 1518-1523.
EPA-821-R-19-011
1-1
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix I: WTP Estimation Methodology
1.) calculate the temperature effects and (consumption-equivalent) damages in each year resulting from
the baseline path of emissions;
2.) adjust the model to reflect an additional unit of emissions in year t;
3.) recalculate the temperature effects and damages expected in all years beyond t resulting from this
adjusted path of emissions, as in step 1; and
4.) subtract the damages computed in step 1 from those in step 3 in each model period and discount the
resulting path of marginal damages back to the year of emissions. In PAGE and FUND step 4 focuses
on the damages attributed to the US region in the models. As noted above, DICE does not explicitly
include a separate US region in the model and therefore, EPA approximates U.S. damages in step 4 as
10 percent of the global values based on the results of Nordhaus (2017).
This exercise produces 30 separate distributions of the SC-CO2 for a given year, the product of 3 models, 2
discount rates, and 5 socioeconomic scenarios. Following the approach used by the IWG, the estimates are
equally weighted across models and socioeconomic scenarios in order to reduce the dimensionality of the
results down to two separate distributions, one for each discount rate.
1.2 Treatment of Uncertainty in Interim Domestic SC-CO2 Estimates
There are various sources of uncertainty in the SC-CO2 estimates used in this BCA. Some uncertainties
pertain to aspects of the natural world, such as quantifying the physical effects of greenhouse gas emissions
on Earth systems. Other sources of uncertainty are associated with current and future human behavior and
well-being, such as population and economic growth, GHG emissions, the translation of Earth system
changes to economic damages, and the role of adaptation. It is important to note that even in the presence of
uncertainty, scientific and economic analysis can provide valuable information to the public and decision
makers, though the uncertainty should be acknowledged and when possible taken into account in the analysis
(Institute of Medicine, 2013). OMB Circular A-4 also requires a thorough discussion of key sources of
uncertainty in the calculation of benefits and costs, including more rigorous quantitative approaches for higher
consequence rules. This section summarizes the sources of uncertainty considered in a quantitative manner in
the domestic SC-CO2 estimates.
The domestic SC-CO2 estimates consider various sources of uncertainty through a combination of a multi-
model ensemble, probabilistic analysis, and scenario analysis. We provide a summary of this analysis here;
more detailed discussion of each model and the harmonized input assumptions can be found in the 2017
National Academies report. For example, the three IAMs used collectively span a wide range of Earth system
and economic outcomes to help reflect the uncertainty in the literature and in the underlying dynamics being
modeled. The use of an ensemble of three different models at least partially addresses the fact that no single
model includes all of the quantified economic damages. It also helps to reflect structural uncertainty across
the models, which is uncertainty in the underlying relationships between GHG emissions, Earth systems, and
economic damages that are included in the models. Bearing in mind the different limitations of each model
and lacking an objective basis upon which to differentially weight the models, the three integrated assessment
models are given equal weight in the analysis.
Monte Carlo techniques were used to run the IAMs a large number of times. In each simulation the uncertain
parameters are represented by random draws from their defined probability distributions. In all three models
the equilibrium climate sensitivity is treated probabilistically based on the probability distribution from Roe
EPA-821-R-19-011
I-2
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix I: WTP Estimation Methodology
and Baker (2007) calibrated to the IPCC AR4 consensus statement about this key parameter.105 The
equilibrium climate sensitivity is a key parameter in this analysis because it helps define the strength of the
climate response to increasing GHG concentrations in the atmosphere. In addition, the FUND and PAGE
models define many of their parameters with probability distributions instead of point estimates. For these
two models, the model developers' default probability distributions are maintained for all parameters other
than those superseded by the harmonized inputs (i.e., equilibrium climate sensitivity, socioeconomic and
emissions scenarios, and discount rates). More information on the uncertain parameters in PAGE and FUND
is available upon request.
For the socioeconomic and emissions scenarios, uncertainty is included in the analysis by considering a range
of scenarios selected from the Stanford Energy Modeling Forum exercise, EMF-22. Given the dearth of
information on the likelihood of a full range of future socioeconomic pathways at the time the original
modeling was conducted, and without a basis for assigning differential weights to scenarios, the range of
uncertainty was reflected by simply weighting each of the five scenarios equally for the consolidated
estimates. To better understand how the results vary across scenarios, results of each model run are available
in the docket for the ACE final rule (Docket ID EPA-HQ-OAR-2017-0355).
The outcome of accounting for various sources of uncertainty using the approaches described above is a
frequency distribution of the SC-CO2 estimates for emissions occurring in a given year for each discount rate.
Unlike the approach taken for consolidating results across models and socioeconomic and emissions
scenarios, the SC-CO2 estimates are not pooled across different discount rates because the range of discount
rates reflects both uncertainty and, at least in part, different policy or value judgements; uncertainty regarding
this key assumption is discussed in more detail below. The frequency distributions reflect the uncertainty
around the input parameters for which probability distributions were defined, as well as from the multi-model
ensemble and socioeconomic and emissions scenarios where probabilities were implied by the equal
weighting assumption. It is important to note that the set of SC-CO2 estimates obtained from this analysis
does not yield a probability distribution that fully characterizes uncertainty about the SC-CO2 due to impact
categories omitted from the models and sources of uncertainty that have not been fully characterized due to
data limitations.
Figure 1-1 presents the frequency distribution of the domestic SC-CO2 estimates for emissions in 2030 for
each discount rate. Each distribution represents 150,000 estimates based on 10,000 simulations for each
combination of the three models and five socioeconomic and emissions scenarios. In general, the distributions
are skewed to the right and have long right tails, which tend to be longer for lower discount rates. To highlight
the difference between the impact of the discount rate on the SC-CO2 and other quantified sources of
uncertainty, the bars below the frequency distributions provide a symmetric representation of quantified
variability in the SC-CO2 estimates conditioned on each discount rate. The full set of SC-CO2 results through
2050 is available in the docket for the ACE final rule (Docket ID EPA-HQ-OAR-2017-0355).
105 Specifically, the Roe and Baker distribution for the climate sensitivity parameter was bounded between 0 and 10 with a median of
3 °C and a cumulative probability between 2 and 4.5 °C of two-thirds.
EPA-821-R-19-011
I-3
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs Appendix I: WTP Estimation Methodology
m
o
o
v>
c
o
jo co
=5 d
£
CO
o
CM
d
o
d
7% Average = $1
3% Average = $8
Qdt
_L
I
Discount Rate
~ 7%
~ 3%
5tn - 95tri Percentile
of Simulations
l l l
0
T"
I I I
12
iii r~i r
16 20 24
T"
I I
32
T"
4 8 12 16 20 24 28 32 36
Interim U.S. Domestic Social Cost of Carbon in 2030 [2016$ / metric ton C02]
40
Figure 1-1: Frequency Distribution of Interim Domestic SC-CO2 Estimates for 2030 (in 2016$ per metric
ton C02)
As illustrated by the frequency distributions in Figure 1-1, the assumed discount rate plays a critical role in the
ultimate estimate of the social cost of carbon. This is because CO2 emissions today continue to impact society
far out into the future, so with a higher discount rate, costs that accrue to future generations are weighted less,
resulting in a lower estimate. Circular A-4 recommends that costs and benefits be discounted using the rates
of 3 percent and 7 percent to reflect the opportunity cost of consumption and capital, respectively. Circular A-
4 also recommends quantitative sensitivity analysis of key assumptions1"6, and offers guidance on what
sensitivity analysis can be conducted in cases where a rule will have important intergenerational benefits or
costs. To account for ethical considerations of future generations and potential uncertainty in the discount rate
over long time horizons, Circular A-4 suggests "further sensitivity analysis using a lower but positive
discount rate in addition to calculating net benefit using discount rates of 3 and 7 percent" (page 36) and notes
that research from the 1990s suggests intergenerational rates "from 1 to 3 percent per annum" (OMB 2003).
We consider the uncertainty in this key assumption by calculating the domestic SC-CO2 based on a
2.5 percent discount rate, in addition to the 3 and 7 percent used in the main analysis. Using a 2.5 percent
discount rate, the average domestic SC-CO2 estimate across all the model runs for emissions occurring over
2020-2045 ranges from $10 to $14 per metric ton of CO2 (in 2018 dollars). In this case the forgone domestic
climate benefits in 2025 are $25 million and $4 million under Options 2and 4, respectively; by 2035, the
100 "If benefit or cost estimates depend heavily on certain assumptions, you should make those assumptions explicit and carry out
sensitivity analyses using plausible alternative assumptions." (OMB 2003, page 42).
EPA-821-R-19-011
I-4
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix I: WTP Estimation Methodology
estimated forgone benefits increase to $30 million and $14 million under Options 2 and 4, respectively; and
by 2045, the estimated forgone benefits are $37 million and $20 million under Options 2 and 4, respectively.
In addition to the approach to accounting for the quantifiable uncertainty described above, the scientific and
economics literature has further explored known sources of uncertainty related to estimates of the SC-CO2.
For example, researchers have published papers that explore the sensitivity of IAMs and the resulting SC-CO2
estimates to different assumptions embedded in the models (see, e.g., Hope (2013), Anthoff and Tol (2013),
and Nordhaus (2014)). However, there remain additional sources of uncertainty that have not been fully
characterized and explored due to remaining data limitations. Additional research is needed in order to expand
the quantification of various sources of uncertainty in estimates of the SC-CO2 (e.g., developing explicit
probability distributions for more inputs pertaining to climate impacts and their valuation). On the issue of
intergenerational discounting, some experts have argued that a declining discount rate would be appropriate to
analyze impacts that occur far into the future (Arrow et al., 2013). However, additional research and analysis
is still needed to develop a methodology for implementing a declining discount rate and to understand the
implications of applying these theoretical lessons in practice. The 2017 National Academies report also
provides recommendations pertaining to discounting, emphasizing the need to more explicitly model the
uncertainty surrounding discount rates over long time horizons, its connection to uncertainty in economic
growth, and, in turn, to climate damages using a Ramsey-like formula (National Academies 2017). These and
other research needs are discussed in detail in the 2017 National Academies' recommendations for a
comprehensive update to the current methodology, including a more robust incorporation of uncertainty.
1.3 Forgone Global Climate Benefits
In addition to requiring reporting of impacts at a domestic level, OMB Circular A-4 states that when an
agency "evaluate[s] a regulation that is likely to have effects beyond the borders of the United States, these
effects should be reported separately" (OMB, 2003; page 15).107 This guidance is relevant to the valuation of
damages from CO2 and other GHGs, given that GHGs contribute to damages around the world independent of
the country in which they are emitted. Therefore, in this section we present the forgone global climate benefits
in 2030 from this proposed rulemaking using the global SC-CO2 estimates corresponding to the model runs
that generated the domestic SC-CO2 estimates used in the main analysis. The average global SC-CO2 estimate
across all the model runs for emissions occurring over 2025-2045 range from $6 to $13 per metric ton of CO2
emissions (in 2018 dollars) using a 7 percent discount rate, and $55 to $76 per metric ton of CO2 emissions
(in 2018 dollars) using a 3 percent discount rate. The domestic SC-CO2 estimates presented above are
approximately 18 percent and 14 percent of these global SC-CO2 estimates for the 7 percent and 3 percent
discount rates, respectively.
Applying these estimates to the forgone CO2 emission reductions results in estimated forgone global climate
benefits in 2025 of $14.9 million and $2.3 million under Options 2 and 4, respectively, using a 7 percent
107 While Circular A-4 does not elaborate on this guidance, the basic argument for adopting a domestic only perspective for the central
benefit-cost analysis of domestic policies is based on the fact that the authority to regulate only extends to a nation's own residents
who have consented to adhere to the same set of rules and values for collective decision-making, as well as the assumption that most
domestic policies will have negligible effects on the welfare of other countries' residents (EPA 2010a; Kopp et al. 1997; Whittington
et al. 1986). In the context of policies that are expected to result in substantial effects outside of U.S. borders, an active literature has
emerged discussing how to appropriately treat these impacts for purposes of domestic policymaking (e.g., Gayer and Viscusi 2016,
2017; Anthoff and Tol, 2010; Fraas et al. 2016; Revesz et al. 2017). This discourse has been primarily focused on the regulation of
GHGs, for which domestic policies may result in impacts outside of U.S. borders due to the global nature of the pollutants.
EPA-821-R-19-011
I-5
-------
BCAfor Proposed Revisions to the Steam Electric Power Generating ELGs
Appendix I: WTP Estimation Methodology
discount rate; the forgone benefits increase to $126.3 million and $19.3 million under options 2 and 4,
respectively, using a 3 percent discount rate. By 2045, the estimated forgone global climate benefits are
$34.8 million and $18.8 million, for options 2 and 4, respectively, using a 7 percent discount rate. Using a 3
percent discount rate, the estimated forgone benefits increase to $198.8 million and $107.2 million, for
options 2 and 4, respectively.
Under the sensitivity analysis considered above using a 2.5 percent discount rate, the average global SC-CO2
estimate across all the model runs for emissions occurring over 2025-2045 ranges from $80 to $105 per
metric ton of CO2 (2018 dollars); in this case the forgone global climate benefits in 2025 are $185.7 million
and $28.4 million under options 2 and 4, respectively; by 2045, the forgone global benefits in this sensitivity
case increase to $276.6 million and $149.1 million, for options 2 and 4, respectively.
EPA-821-R-19-011
I-6
------- |