United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-453/R-98-004a
February 1998
Air
& EPA
Study of Hazardous Air Pollutant
Emissions from Electric Utility Steam
Generating Units — Final Report to
Congress
Volume 1.
-------
ACKNOWLEDGMENT S
The U.S. Environmental Protection Agency would like to
acknowledge numerous technical authors for their significant
contributions to the development of this report. Many of the primary
authors are listed below. The EPA would also like to thank the
numerous individuals from various organizations for their invaluable
review, comments, and inputs for the development of this report.
These individuals are too numerous to list here, but they include
representatives from many organizations, including: U.S. EPA work
group members and other EPA staff; other Federal Agencies; Electric
Power Industry; State and local Agencies; Environmental Groups;
scientific peer reviewers from academia and consulting firms; and
other public reviewers. In addition, the EPA would like to thank the
Research Triangle Institute, EC/R Incorporated, and S. Cohen &
Associates for the valuable technical support provided for the
development of this report.
U.S. EPA AUTHORS
Charles L. French, Risk and Exposure Assessment Group
Office of Air Quality Planning and Standards
William H. Maxwell, Combustion Group
Office of Air Quality Planning and Standards
Warren D. Peters, Air Quality Modeling Group
Office of Air Quality Planning and Standards
Glenn E. Rice, National Center for Environmental Assessment - Cincinnati
Office of Research and Development
0. Russell Bullock, Atmospheric Sciences Modeling Division
National Oceanic and Atmospheric Administration
on assignment to the U.S. EPA National Exposure Research Laboratory
Amy B. Vasu, Risk and Exposure Assessment Group
Office of Air Quality Planning and Standards
Robert Hetes, Office of Air Quality Planning and Standards
Albert Colli, Office of Radiation and Indoor Air
Christopher Nelson, Office of Radiation and Indoor Air
Bradford F. Lyons, Oak Ridge National Laboratory (non-EPA)
Additional EPA staff that contributed to this report include:
Ted Palma, Mike Dusetzina, George Duggan, Dianne Byrne, Stan Durkee,
Martha Keating, Neal Nelson, and others.
-------
TABLE OF CONTENTS
Volume 1
Section
Glossary xxv
ES-1
1.0
2.0
INTRODUCTION
1
1
1
1
1
.1
.2
.2 .
.3
.4
LEGISLATIVE MANDATE
CAA PROVISIONS AND STUDIES RELATED TO THIS STUDY . .
. 1 Nonattainment Provisions
1.2.2 Acid Deposition Control
1.2.3 New Source Performance Standards
1.2.4 Hazardous Air Pollutants
1.2.5 Other Studies
OVERVIEW AND APPROACH OF ELECTRIC UTILITY HAP STUDY .
REFERENCES
CHARACTERIZATION OF THE INDUSTRY
2
2
2
2
2
2
2
.1
.2
.3
.4
.5
.6
.7
INDUSTRY BACKGROUND
FOSSIL-FUEL-FIRED ELECTRIC UTILITY STEAM -GENE RAT ING
UNITS
2.2.1 Types of Electric Utility Facilities ....
2.2.2 Types of Ownership
DESIGN OF ELECTRIC UTILITY UNITS
2.3.1 Furnace Types
2.3.2 Bottom Types
2.3.3 Cogeneration
2.3.4 Combined- Cycle Systems
PARTI CULATE MATTER CONTROL
2.4.1 Mechanical Collectors
2.4.2 Electrostatic Precipitators
2.4.3 Particulate Matter Scrubbers
2.4.4 Fabric Filters
2.4.5 Comparison of Particle Collectors
SULFUR DIOXIDE CONTROL
2.5.1 Precombustion Control: Fuel Options ....
2.5.2 Postcombustion Control: Flue Gas Scrubbing
for S02 Control
NOX CONTROL
2.6.1 Combustion Control
2.6.2 Postcombustion Control
UTILITY INDUSTRY AFTER IMPLEMENTATION OF 1990
AMENDMENTS
2.7.1 Industry Growth
2.7.2 Title I and Title IV, Phase I and Phase II,
1-1
1-1
1-2
1-2
1-3
1-5
1-6
1-6
1-8
1-10
2-1
2-1
2-3
2-3
2-4
2-5
2-5
2-9
2-9
2-9
2-9
2-10
2-12
2-12
2-13
2-13
2-14
2-15
2-17
2-21
2-23
2-24
2-25
2-26
Compliance Strategy Impact 2-31
2.7.3 Compliance Strategy Impacts of Other
Activities 2-33
2.8 REFERENCES 2-34
111
-------
TABLE OF CONTENTS (continued)
Volume 1
Section
3.0 EMISSION DATA GATHERING AND ANALYSIS 3-1
3.1 LITERATURE REVIEW AND BACKGROUND 3-1
3.2 POLLUTANTS STUDIED 3-3
3.3 DESCRIPTION OF EMISSION TEST PROGRAMS 3-3
3.4 DEVELOPMENT OF HAP EMISSION TOTALS 3-4
3.4.1 Program Operation 3-4
3.4.2 Data Sources 3-4
3.4.3 Operational Status of Boilers 3-8
3.4.4 Trace Element Concentration in Fuel .... 3-8
3.4.5 HC1 and HF Concentration in Fuel 3-9
3.4.6 Emission Modification Factors for
Inorganic HAPs 3-10
3.4.7 Acid Gas HAPs 3-11
3.4.8 Organic HAPs 3-12
3.4.9 Model Estimates for the Year 2010 3-12
3.5 SELECTED ESTIMATED NATIONWIDE HAP EMISSIONS 3-14
3.6 COMPARISON OF EFP ESTIMATES WITH TEST DATA 3-14
3.7 CHARACTERISTIC PLANT EMISSIONS 3-16
3.8 UNCERTAINTY ANALYSIS OF EMISSION FACTOR PROGRAM . . . 3-19
3.9 REFERENCES 3-20
4.0 INTRODUCTION FOR THE HEALTH HAZARD RISK ANALYSIS 4-1
4.1 INTRODUCTION AND BACKGROUND 4-1
4.1.1 Principles of Risk Assessment 4-1
4.1.2 U.S. EPA Risk Assessment Guidelines .... 4-3
4.1.3 Risk Assessment Council (RAC) Guidance ... 4-3
4.1.4 NAS Report Science and Judgement in Risk
Assessment 4-3
4.1.5 SPC's Guidance for Risk Characterization . . 4-4
4.2 GENERAL APPROACH AND METHODS FOR THE UTILITY HEALTH
HAZARD RISK ASSESSMENT 4-4
4.3 HEALTH EFFECTS DATA: HAZARD IDENTIFICATION AND DOSE
RESPONSE 4-5
4.3.1 Hazard Identification for Carcinogens ... 4-5
4.3.2 General Discussion of Dose Response . . . . 4-5
4.3.3 Dose-Response Evaluation for Carcinogens . . 4-7
4.3.4 Long-Term Noncancer Health Effects Data . . 4-8
4.3.5 Short-Term Noncancer Health Effects Data . . 4-9
4.3.6 Summary of Health Effects Data Sources . . . 4-9
4.4 METHODOLOGY FOR ESTIMATING INHALATION EXPOSURE FOR
LOCAL ANALYSIS 4-9
4.4.1 Emissions Characterization 4-9
4.4.2 Atmospheric Fate and Transport 4-10
4.4.3 Characterization of Study Population .... 4-10
4.4.4 Exposure Calculations 4-11
IV
-------
TABLE OF CONTENTS (continued)
Volume 1
Section
4.5 METHODOLOGY FOR ESTIMATING QUANTITATIVE INHALATION
RISKS 4-11
4.5.1 Estimating Cancer Inhalation Risks 4-11
4.5.2 Individual Cancer Risk 4-12
4.5.3 Population Cancer Risk 4-12
4.5.4 Distribution of Individual Risk within a
Population 4-13
4.5.5 Aggregate Inhalation Cancer Risk 4-13
4.5.6 Estimating Noncancer Inhalation Risks . . . 4-13
4.5.7 Inhalation Hazard Quotient (HQ) 4-14
4.5.8 Total Risk for Noncancer Effects 4-14
4.5.9 Direct Inhalation Exposure and Risk
Default Options 4-15
4.6 REFERENCES 4-17
5.0 SCREENING ASSESSMENT TO DETERMINE PRIORITY HAPS 5-1
5.1 MODELING DESCRIPTION 5-1
5.2 SCREENING CRITERIA 5-1
5.3 INHALATION SCREENING ASSESSMENT FOR COAL-FIRED
UTILITIES 5-2
5.4 INHALATION SCREENING ASSESSMENT FOR OIL- AND GAS-
FIRED UTILITIES 5-2
5.5 MULTIPATHWAY SCREENING ANALYSIS FOR NONRADIONUCLIDE
HAPS 5-13
5.5.1 Overview 5-13
5.5.2 Prioritization of HAPs for Multipathway
Exposure Assessment 5-14
5.6 SELECTION OF HAPS FOR FURTHER ANALYSIS 5-18
5.7 LIMITATIONS OF SCREENING ASSESSMENT 5-19
5.8 REFERENCES 5-20
6.0 INHALATION RISK ASSESSMENT 6-1
6.1 BASELINE ASSESSMENT OF INHALATION EXPOSURES AND
RISKS FOR 13 PRIORITY POLLUTANTS 6-1
6.1.1 Baseline Inhalation Risks for Coal-Fired
Utilities for Priority HAPs 6-2
6.1.2 Baseline Inhalation Risks for Oil-Fired
Utilities 6-7
6.1.3 Baseline Risks from Gas-Fired Utilities . . 6-12
6.2 DISTINGUISHING BETWEEN URBAN AND RURAL LOCATIONS . . 6-12
6.3 INHALATION RISK ESTIMATES FOR THE YEAR 2010 6-15
6.4 ASSESSMENT OF POTENTIAL RISKS DUE TO SHORT-TERM
EXPOSURE 6-15
6.4.1 Methodology 6-17
6.4.2 Results 6-18
6.5 OVERLAPPING PLUMES/DOUBLE COUNTING 6-18
v
-------
TABLE OF CONTENTS (continued)
Volume 1
Section
6.6 ASSESSMENT OF EXPOSURE DUE TO LONG-RANGE TRANSPORT . 6-20
6.6.1 History and Background Information 6-20
6.6.2 RELMAP Modeling Approach for Particulate
Metals 6-21
6.6.3 Model Parameterizations 6-23
6.6.4 Exposure and Risk Estimates 6-25
6.7 DISCUSSION OF BACKGROUND EXPOSURES 6-43
6.7.1 Arsenic 6-43
6.7.2 Chromium, Nickel, Manganese, and HC1 . . . . 6-45
6.8 CHROMIUM SPECIATION UNCERTAINTY AND IMPACT ON RISK
ESTIMATES 6-45
6.9 ISSUES WITH ARSENIC CANCER UNIT RISK ESTIMATE AND
IMPACT ON INHALATION RISK ESTIMATES 6-47
6.10 NICKEL SPECIATION UNCERTAINTY AND IMPACT ON RISK
ESTIMATES 6-47
6.10.1 Alternative Analysis for Estimating
Population Risks 6-49
6.11 POTENTIAL INCREASED DIOXIN EMISSIONS FROM UTILITIES
WITH ELECTROSTATIC PRECIPITATORS 6-50
6.12 DISCUSSION OF UNCERTAINTY AND ASSUMPTIONS FOR DOSE-
RESPONSE ASSESSMENT FOR CARCINOGENS 6-50
6.12.1 Default Options 6-52
6.12.2 Models, Methods, and Data 6-53
6.12.3 Discussion of Uncertainty in lUREs 6-56
6.12.4 Variability in Cancer Dose-Response
Assessment 6-57
6.13 PRELIMINARY QUANTITATIVE UNCERTAINTY AND VARIABILITY
ANALYSIS FOR INHALATION EXPOSURE AND RISK
ASSESSMENT 6-58
6.13.1 Introduction 6-58
6.13.2 Approach to Quantitative Uncertainty
Analysis 6-59
6.13.3 Discussion of Results of the Quantitative
Uncertainty Analysis 6-70
6.14 QUALITATIVE DISCUSSION OF ADDITIONAL UNCERTAINTIES . 6-72
6.14.1 Uncertainty Using lUREs 6-72
6.14.2 Residence Time and Activity Patterns .... 6-72
6.15 REFERENCES 6-74
7.0 MERCURY ASSESSMENT 7-1
7.1 OVERVIEW 7-1
7.1.1 The Mercury Cycle 7-1
7.1.2 Atmospheric Processes 7-6
7.1.3 Terrestrial and Aquatic Fate of Mercury . . 7-13
7.2 MERCURY HEALTH EFFECTS 7-16
7.2.1 Toxicokinetics 7-16
7.2.2 Biological Effects 7-17
vi
-------
TABLE OF CONTENTS (continued)
Volume 1
Section
7.3
7 .4
7.5
7.6
7.7
7.8
7.9
7.2.3 Sensitive Subpopulations
7.2.4 Interactions
7.2.5 Hazard Identification/Dose-Response
Assessment
7.2.6 Ongoing Research
7.2.7 Research Needs
MERCURY CONCENTRATIONS IN BIOTA
MEASUREMENT DATA NEAR UTILITIES
MODEL FRAMEWORK
7.5.1 Models Used
7.5.2 Modeling of Long-Range Fate and Transport
of Mercury
7.5.3 Modeling the Local Transport of Mercury in
the Atmosphere
7.5.4 Modeling Mercury in a Watershed
7.5.5 Exposure Modeling
RESULTS
CONCLUSIONS
DISCUSSION OF POTENTIAL CONCERNS OF MERCURY
EMISSIONS FROM UTILITIES
REFERENCES
8.0 QUALITATIVE MULTI PATHWAY ASSESSMENT FOR LEAD AND CADMIUM
8.1
8.2
8.3
8.4
8.5
BACKGROUND
LEAD COMPOUNDS
CADMIUM COMPOUNDS
OVERALL SUMMARY
REFERENCES
7-18
7-18
7-18
7-19
7-19
7-21
7-22
7-22
7-23
7-23
7-28
7-29
7-34
7-35
7-40
7-45
7-46
8-1
8-1
8-1
8-3
8-16
8-17
9.0 MULTIPATHWAY EXPOSURE AND RISK ASSESSMENT FOR
RADIONUCLIDES 9-1
9.1 SUMMARY OF RADIONUCLIDE ANALYSIS 9-1
9.1.1 Natural Radionuclide Content in Fossil
Fuels: Coal 9-2
9.1.2 Natural Radionuclide Content in Fossil
Fuels: Natural Gas 9-2
9.1.3 Natural Radionuclide Content in Fossil
Fuels: Oil 9-3
9.1.4 Radionuclide Emissions from Fossil-Fueled
Plants 9-4
9.1.5 Summary of CAP-93 Model 9-5
9.1.6 Estimates of Population Health Risks .... 9-8
9.2 RADIONUCLIDE UNCERTAINTY ANALYSIS 9-10
9.3 SUMMARY FINDINGS 9-12
9.4 REFERENCES 9-17
VII
-------
TABLE OF CONTENTS (continued)
Volume 1
Section
10.0
SCREENING LEVEL ASSESSMENT OF MULT I PATHWAY EXPOSURES AND
RISKS
10.1
10.2
10.3
10.4
10.5
10.6
10.7
TO ARSENIC EMISSIONS
PURPOSE AND SCOPE
10.1.1 Rational and Usefulness of Model Plant
Approach
BACKGROUND INFORMATION ON ARSENIC
10.2.1 Forms of Arsenic in the Environment . .
10.2.2 Sources of Arsenic
10.2.3 Arsenic in the Atmosphere
10.2.4 Arsenic in Water
10.2.5 Arsenic in Sediments
10.2.6 Arsenic in Soil
10.2.7 Arsenic in Terrestrial Plants
10.2.8 Arsenic in Aquatic Plants
10.2.9 Arsenic in Terrestrial Animals
10.2.10 Arsenic in Fish
10.2.11 Speciation of Arsenic in Food Products .
10.2.12 Arsenic Near Anthropogenic Sources . . .
SUMMARY OF MODELS AND APPROACH
10.3.1 Source Classes Considered and Model
Plant Approach
10.3.2 Atmospheric Transport Modeling
10.3.3 Indirect Exposure Modeling
10.3.4 Determination of Background Values . . .
MODELING RESULTS
10.4.1 Air Modeling Results/Comparison with
Measured Data
10.4.2 Indirect Exposure Modeling
HAZARD IDENTIFICATION AND DOSE-RESPONSE FOR
ARSENIC
10.5.1 Introduction
10.5.2 Cancer Effects of Arsenic
10.5.3 Noncancer Effects of Arsenic
RISK CHARACTERIZATION
10.6.1 Discussion of Cancer Risk Assessment
Results
10.6.2 Discussion of the Noncancer Risk
Assessment Results
CONCLUSIONS
10-1
10-1
10-1
. . 10-2
. . 10-2
10-4
. . 10-4
10-4
10-6
10-6
. . 10-7
. . 10-7
. . 10-7
10-7
10-10
10-10
10-10
10-12
10-12
10-12
10-19
10-21
10-21
10-23
10-34
10-34
10-34
10-36
10-36
10-37
10-43
10-43
10.7.1 Contribution of Arsenic Emissions from
Utilities to Concentrations in
Environmental Media and Biota 10-43
10.7.2 Determination of Dominant Pathways of
Potential Exposure to Anthropogenic
Arsenic Emissions 10-43
Vlll
-------
TABLE OF CONTENTS (continued)
Volume 1
Section Page
UNCERTAINTIES AND LIMITATIONS 10-44
10.8.1 Limitations and Uncertainties for the
Multipathway Exposure Modeling 10-44
10.8.2 Limitations and Uncertainties for the
Risk Characterization 10-44
10.9 RESEARCH NEEDS 10-45
10.10 REFERENCES 10-46
11.0 A MULTIPATHWAY SCREENING-LEVEL ASSESSMENT FOR DIOXINS/
FURANS 11-1
11.1 INTRODUCTION 11-1
11.2 LONG-RANGE TRANSPORT MODELING 11-6
11.3 RISK ASSESSMENT METHODOLOGY 11-10
11.3.1 Emissions Sources 11-10
11.3.2 Local Air Dispersion Modeling 11-10
11.3.3 Exposure Modeling and Risk Calculation . . 11-14
11.4 DISPERSION, EXPOSURE, AND RISK RESULTS 11-18
11.5 UNCERTAINTY AND SENSITIVITY ANALYSIS 11-21
11.5.1 Model Elasticity 11-22
11.5.2 Fish Consumption Pathway Sensitivity
Analysis 11-23
11.5.3 Sensitivity Analysis of Plume Impaction . 11-25
11.6 SUMMARY OF RESULTS 11-27
11.7 REFERENCES 11-30
12 . 0 LITERATURE REVIEW ON THE POTENTIAL IMPACTS OF HYDROGEN
CHLORIDE AND HYDROGEN FLUORIDE EMISSIONS 12-1
12.1 OVERVIEW 12-1
12.2 SUMMARY OF FINDINGS 12-1
12.2.1 Hydrogen Chloride 12-1
12.2.2 Hydrogen Fluoride 12-3
12.3 REFERENCES 12-6
13 . 0 ALTERNATIVE CONTROL STRATEGIES FOR HAZARDOUS AIR
POLLUTANT EMISSIONS REDUCTIONS 13-1
13.1 PRECOMBUSTION CONTROLS 13-1
13.1.1 Fuel Switching 13-1
13.1.2 Coal Cleaning 13-5
13.1.3 Coal Gasification 13-11
13.2 COMBUSTION CONTROL 13-12
13.3 POSTCOMBUSTION CONTROL 13-22
13.3.1 Particulate Phase Controls 13-22
13.3.2 Vapor Phase Controls 13-28
13.3.3 Acid Gas Control 13-28
13.3.4 Carbon Adsorption 13-31
13.4 ALTERNATIVE CONTROLS 13-31
IX
-------
TABLE OF CONTENTS (continued)
Volume 1
Section
13.5
13.6
13.7
POLLUTANT TRADEOFFS
13.5.1 HAP Increase/Decrease
13.5.2 Water/Solid Waste Considerations
AVAILABLE CONTROL TECHNOLOGY AND STRATEGIES FOR
MERCURY CONTROL
13.6.1 Pre-Combustion Strategies
13.6.2 Post -Combust ion Strategies
REFERENCES
Page
13-32
13-32
13-34
13-36
13-36
13-37
13-48
14.0 SUMMARY OF RESULTS, TECHNICAL FINDINGS, AND RESEARCH
NEEDS
14.1
14.2
14.3
14.4
14.5
14.6
14.7
INDUSTRY GROWTH AND HAP EMISSIONS
INHALATION RISK ASSESSMENT
MERCURY
DIOXINS AND ARSENIC
RADIONUCLIDE ANALYSIS
ALTERNATIVE CONTROL STRATEGIES
AREAS FOR FURTHER RESEARCH AND ANALYSIS
14.7.1 Emissions Data for Dioxins
14.7.2 Speciation of Nickel and Chromium ....
14.7.3 Multipathway Risk Assessment
14.7.4 Long-range Transport Exposures
14.7.5 Mercury Issues
14.7.6 Projections to the Year 2010
14.7.7 Ecological Risks
14.7.8 Criteria Pollutant and Acid Rain Programs
14.7.9 Short-term Emissions
14-1
. 14-1
14-1
14-3
14-5
14-6
14-6
. 14-8
. 14-8
. 14-8
. 14-8
. 14-8
14-8
. 14-9
14-9
14-10
14-10
X
-------
TABLE OF CONTENTS (Continued)
Volume 2
Section
Page
Appendix A Median Emission Factors, Determined from Test
Report Data, and Total 1990, 1994, and 2010
Emissions, Projected with the Emission Factor
Program A-1
Appendix B Matrix of Electric Utility Steam-Generating Units
and Emission Test Sites B-l
Appendix C Listing of Emission Modification Factors for Trace
Elements Used In the Individual Boiler Analysis . . C-l
Appendix D Discussion of the Methodology Used to Develop
Nationwide Emission Totals D-l
Appendix E Health Effects Summaries: Overview E-l
Appendix F Documentation of the Inhalation Human Exposure
Modeling for the Utility Study F-l
Appendix G Data Tables for Dioxin Multipathway Assessment ... G-l
Appendix H Literature Review of the Potential Impacts of
Hydrogen Chloride and Hydrogen Fluoride H-l
Appendix I Mercury Control Technologies 1-1
XI
-------
LIST OF TABLES
Table
ES-1. Nationwide Utility Emissions for Thirteen Priority HAPs ES-5
ES-2. Estimated Emissions for Nine Priority HAPs from
Characteristic Utility Units (1990; tons per year) . . . ES-6
ES-3. Summary of High-End Inhalation Cancer Risk Estimates
from Local Analysis for Coal-Fired Utilities for
the Year 1990 ES-11
ES-4. Summary of High-End Inhalation Cancer Risk Estimates
Based on Local Analysis for Oil-Fired Utilities for
the Year 1990 ES-12
ES-5. Summary of High-End Inhalation Risk Estimates Due to
Local and Long-Range Transport ES-13
2-1. Comparison of Particulate Matter Collection Systems . . 2-14
2-2. Distribution of S02 Control Technologies in 1994 .... 2-18
2-3. Distribution of NOX Control by Fuel Burned, by Unit,
in 1994 2-26
2-4. Fuel Use in the Electric Utility Industry by Fuel Type,
Quadrillion Btu/yr 2-29
3-1. Assigned Chloride and HC1 Concentrations in Coal,
by State of Coal Origin 3-11
3-2. Average Higher Heating Values of Coal 3-13
3-3. Selected Nationwide HAP Emissions 3-15
3-4. Comparison of Utility Boiler Emissions from EFP
Estimates and from Tests 3-17
3-5. Emissions from a Characteristic Coal-Fired Electric
Utility Plant (1994) 3-18
3-6. Emissions from a Characteristic Oil-Fired Electric
Utility Plant (1994) 3-19
3-7. Emissions from a Characteristic Natural Gas-Fired
Electric Utility Plant (1994) 3-19
4-1. Weight-of-Evidence (WOE) Classification 4-6
5-1. Inhalation Screening Assessment for Carcinogenic HAPs
from Coal-Fired Utilities for Which Quantitative Cancer
Risk Estimates Were Available 5-3
5-2. Inhalation Screening Assessment for Noncancer Effects of
HAPs Emitted from Coal-Fired Utilities for Which
Inhalation Reference Concentrations Are Available . . . 5-5
5-3. Inhalation Screening Assessment for HAPS Emitted from
Coal-Fired Utilities for Which No EPA-Verified Health
Benchmarks Are Available (Comparison of Highest
Modeled Air Concentration to Various Non-EPA Health
Benchmarks) 5-7
XII
-------
LIST OF TABLES (continued)
Inhalation Screening Assessment for Carcinogenic HAPS
from Oil-Fired Utilities for Which Quantitative Cancer
Risk Estimates Were Available 5-9
5-5. Inhalation Screening Assessment for Noncancer Effects
of HAPS Emitted from Oil-Fired Utilities for Which
EPA-Verified Inhalation Reference Concentrations Are
Available 5-11
5-6. Inhalation Screening Assessment for HAPS Emitted from
Oil-Fired Utilities for Which No EPA-Verified Health
Benchmarks Are Available (Comparison of Highest Modeled
Concentration to Various Non-EPA Health Benchmarks) . . 5-12
Inhalation Screening Assessment for HAPS Emitted from
Gas-Fired Utilities 5-13
Thirteen HAPs Selected from the Hazard-Based
Multipathway Ranking (shown in order of ranking), and
the Overall and Individual Criterion Scores Assigned
to Each 5-16
5-9. Comparison of Cancer and Noncancer Effects Benchmarks
and Emissions Estimates for 13 Selected HAPs 5-17
5-10. Pollutants Considered Priority for Further Analysis
Based on Results of Screening Assessment 5-19
6-1. Summary of High-End Risk Estimates from Chronic
Inhalation Exposure by HAP for 424 U.S. Coal-Fired
Utilities Based on the Baseline Inhalation Risk
Assessment 6-3
6-2. Summary of High-End Estimates of Population Exposed at
Various Levels of Inhalation Risk or Greater by HAP:
Coal-Fired Utilities 6-7
6-3. Summary of the High-End Risk Estimates from Inhalation
Exposure for Priority HAPs for 137 U.S. Oil-Fired
Utilities Based on the Baseline Risk Assessment .... 6-8
6-4. Summary of High-End Estimates of Population Exposed
Through Inhalation at Various Levels of Risk or Greater
from Oil-Fired Utilities 6-12
6-5. Summary of High-End Inhalation Risk Estimates for
Gas-Fired Utilities 6-13
6-6. Comparison of High-End Inhalation Cancer Risk Estimates
Based on (1) HEM Modeling Using Urban Default Assumption
and (2) HEM Modeling Using Urban vs. Rural Distinction . 6-14
6-7. Comparison of High-End Inhalation Noncancer Risk
Estimates Based on (1) HEM Modeling Using Urban Default
Assumption and (2) HEM Modeling Using Urban vs. Rural
Distinction 6-14
6-8. Estimated High-End Inhalation Cancer Risks for the
Year 2010 Compared to 1990 for Coal- and Oil-Fired
Utilities 6-16
Xlll
-------
LIST OF TABLES (continued)
Table
6-9. Estimated High-End Inhalation Noncancer Risks for
Coal-Fired Utilities for the Year 2010 Compared to
the Year 1990 6-16
6-10. Noncancer Reference Exposure Levels (Acute) from CAPCOA 6-17
6-11. Sample Stack Parameters for Typical Utility Plant . . . 6-18
6-12. Stack and Emission Values Input to TSCREEN 6-19
6-13. Results of the TSCREEN Model 6-19
6-14. Comparison of Risk Estimates for Single-Count Versus
Double-Count Runs to Assess the Impact of Overlapping
Plumes 6-21
6-15. Windspeeds Used for Each Pasquill Stability Category in
CARB Subroutine Calculations 6-25
6-16. Roughness Length Used for Each Land-Use Category in
CARB Subroutine Calculations 6-25
6-17. RELMAP Predicted Air Concentrations 6-26
6-18. Predicted Exposure and High-End Risk Estimates Based on
RELMAP Modeling of Particulate Metal Emissions from All
Oil- and Coal-Fired Utilities in the United States. . . 6-41
6-19. Summary of the High-End Estimates of the Inhalation Risk
Estimates Due to Local and Long-Range Transport .... 6-42
6-20. Chromium Speciation Analysis for Coal-Fired Utilities:
Inhalation Risk Estimates due to Chromium Emissions
Based on Various Assumptions of Percent Chromium VI . . 6-46
6-21. Chromium Speciation Analysis for Oil-Fired Utilities:
Inhalation Risk Estimates due to Chromium Based on
Various Assumptions of Percent Chromium VI 6-46
6-22. High-End Arsenic Inhalation Risk Estimates: Comparison
of Results Using the EPRI, EPA-Verified, and Canadian
lUREs 6-48
6-23. Nickel from Oil-Fired Utilities: Inhalation Cancer Risk
Estimates Based On Various Assumptions of Speciation and
Cancer Potency 6-50
6-24. Comparison of Nickel Exposure to Various Noncancer
Health Benchmarks 6-52
6-25. Summary of Basic Parameters Used in the Inhalation Risk
Assessment for Electric Utilities 6-61
6-26. Summary of Results for Monte Carlo Simulation of HAP
Emissions (kg/year) from Oil-Fired Plant No. 29 .... 6-66
6-27. Distribution of MIR: Plant No. 29: Comparison of FCEM
and SGS Concentration Data 6-71
7-1. Best Point Estimates of National Mercury Emission Rates
by Category 7-8
7-2. Summary of U.S. EPA Hazard Identification/Dose-Response
Assessment for Methylmercury 7-20
7-3. Models Used to Predict Mercury Air Concentrations,
Deposition Fluxes, and Environmental Concentrations . . 7-24
xiv
-------
LIST OF TABLES (continued)
Table
7-4. Mercury Emissions Inventory Used in the RELMAP Modeling
(Based on the 1994-95 Estimates) 7-24
7-5. Process Parameters for Model Plants 7-30
7-6. Percentiles of the Methylmercury Bioaccumulation Factor 7-34
7-7. Fish Consumption Rates for Columbia River Tribes .... 7-36
7-8. Daily Fish Consumption Rates Among Adults in the
Columbia River Tribes 7-36
7-9. Fish Consumption Rates Used in This Study 7-36
7-10. Model Results for Eastern Site, RELMAP 50th Percentile
(utilities only) 7-37
7-11. Model Results for Western Site, RELMAP 50th Percentile
(utilities only) 7-37
7-12. Predicted Exposure Results for Eastern Site, RELMAP 50th
Percentile (utilities only) 7-38
7-13. Predicted Exposure Results for Western Site, RELMAP 50th
Percentile (utilities only) 7-38
8-1. Concentration of Lead in Various Food Products 8-4
8-2. Concentration of Cadmium in Various Food Products . . . 8-12
9-1. Utilization and Radionuclide Content by Coal Rank . . . 9-3
9-2. Estimates of Average Radionuclide Concentrations in
42 Residual Fuel Oil Samples 9-5
9-3. Average Annual Radionuclide Emissions per Operating
Boiler Unit and per Billion Kilowatt-Hour Electricity
Generated 9-6
9-4. Frequency Distribution of Lifetime Fatal Cancer Risks
for All Plants 9-13
9-5. Plants with the Highest Estimated Maximum Individual Risk
(MIR) 9-13
9-6 Average Background Radiation Doses (effective dose
equivalent excluding inhaled radon pogeny) 9-15
9-7 Average Annual Background Exposures Due To Radon Pogency 9-15
10-1. Common Arsenic Compounds, and Classification by Valence
State and Organic/Inorganic 10-3
10-2. National Arsenic Atmospheric Emission Estimates by
Source Category 10-5
10-3. Reported Arsenic Air Concentrations 10-5
10-4. Measured Arsenic Deposition Rates 10-6
10-5. Measured Arsenic Concentrations in Plants 10-8
10-6. Measured Arsenic Concentrations in Meat and Other Animal
Products 10-9
10-7. Total Arsenic Concentrations in Freshwater Fish in the
United States 10-10
xv
-------
LIST OF TABLES (continued)
Table
10-8. Percentage of Inorganic Arsenic Compared to Total
Arsenic in Selected Foods 10-11
10-9. Environmental Concentrations near Facilities 10-11
10-10. Summary of Model Plants and Emission Rates Used for
the Assessment. 10-13
10-11. Summary of Human Exposure Scenarios Considered. . . . 10-17
10-12. Default Values of Scenario-Independent Exposure
Parameters 10-17
10-13. Values for Scenario-Dependent Exposure Parameters. . . 10-18
10-14. Fraction of Arsenic Emissions Predicted to Be
Deposited Within 50 km in an Arid Site. 10-22
10-15. Fraction of Arsenic Emissions Predicted to Be
Deposited Within 50 km in a Humid Site. 10-22
10-16. RELMAP Air Modeling Results 10-23
10-17. Predicted Surface Water and Benthic Sediment
Concentrations for the Hypothetical Water Bodies. . . 10-24
10-18. Modeled Arsenic Concentrations. 10-24
10-19. Predicted Total Arsenic Exposure for Hypothetical
Receptors 10-25
10-20. Predicted Total Inorganic Arsenic Exposure for
Hypothetical Receptors. 10-25
10-21. Watershed Air Concentration 10-27
10-22. Watershed Deposition Rate 10-27
10-23. Surface Water Concentration 10-28
10-24. Untilled Soil Concentration 10-28
10-25. Grain Concentration 10-29
10-26. Inorganic Arsenic Intake via Ingestion for Pica Child 10-30
10-27. Inorganic Arsenic Intake via Ingestion for Subsistence
Farmer Adult 10-30
10-28. Inorganic Arsenic Intake via Ingestion for Subsistence
Farmer Child 10-31
10-29. Inorganic Arsenic Intake via Ingestion for Subsistence
Fisher Adult 10-31
10-30. Inorganic Arsenic Intake via Ingestion for Subsistence
Fisher Child 10-32
10-31. Inorganic Arsenic Intake via Inhalation for Pica Child 10-32
10-32. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Adult and Subsistence Fisher Adult 10-33
10-33. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Child and Subsistence Fisher Child. 10-33
10-34. Inorganic Arsenic Intake, Predicted Cancer Risk, and
Noncancer Hazard Quotient (HQ) for Pica Child .... 10-38
10-35. Inorganic Arsenic Intake, Predicted Cancer Risk, and
Noncancer Hazards for Subsistence Farmer Adult .... 10-38
10-36. Inorganic Arsenic Intake, Predicted Cancer Risk, and
Noncancer Hazards for Subsistence Farmer Child .... 10-39
xvi
-------
LIST OF TABLES (continued)
Table
10-37. Inorganic Arsenic Intake, Predicted Cancer Risk, and
Noncancer Hazards for Subsistence Fisher Adult .... 10-39
10-38. Inorganic Arsenic Intake, Predicted Cancer Risk, and
Noncancer Hazards for Subsistence Fisher Child .... 10-40
10-39. Inorganic Arsenic Intake via Inhalation and Predicted
Cancer Risks for Pica Child 10-40
10-40. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Adult and Subsistence Fisher Adult
and Predicted Cancer Risks 10-41
10-41. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Child and Predicted Cancer Risks . 10-41
11-1. Congener-Specific Emissions Rates for Model
Plants (kg/yr) 11-11
11-2. Process Parameters for Model Plants 11-11
11-3. Naming Scheme for Eight Model Plants 11-12
11-4. Process-Specific Depletion Parameters 11-13
11-5. Other Emissions Source Information 11-13
11-6. Summary of Receptor Scenarios and Pathways 11-17
11-7. The 16 Hypothetical Scenarios Included in the
Screening Level, Model Plant, Dioxin Multipathway
Exposure and Risk Assessment 11-17
11-8. Summary of Predicted Cancer Risks from the Screening
Level Multipathway Assessment for Model Plants, for
16 Hypothetical Scenarios 11-20
11-9. Fish Consumption Pathway Sensitivity Analysis Inputs
and Results 11-26
11-10. Maximum Dispersion Modeling Locations and
Concentrations 11-28
11-11 Dispersion Modeling Concentrations at Specified
Distances 11-29
13-1. Comparison of Average Concentrations of Trace Elements
in Utility Fuels 13-2
13-2. Trace Element Reductions Achieved Through Conventional
Coal Cleaning 13-9
13-3. Emissions from an Air-Blown, Fixed-Bed Gasifier . . . 13-13
13-4. Comparison of Electric Utility Emissions Before and
After Application of NOX Control or Application of
Greater NOX Control on a Unit With Lesser NOX Control
(Ib/trillion Btu) 13-14
13-5. Comparison of Wet Bottom vs. Dry Bottom Electric
Utility Boilers Metallic HAP Emissions, Trace
Element Removal, and Trace Metal Concentrations in
Feed Coal 13-16
13-6. Descriptive Statistics for HAP Removal Efficiencies
Shown in Figures 13-7 and 13-8 13-23
xvi i
-------
LIST OF TABLES (continued)
Table
13-7. Descriptive Statistics for HAP Removal Efficiencies
Shown in Figures 13-9 and 13-10
13-8. Descriptive Statistics for HAP Removal Efficiencies
Shown in Figures 13-11 and 13-12
13-9. Descriptive Statistics for HAP Removal Efficiencies
Shown in Figures 13-13 and 13-14
13-10. Particulate Metallic HAP Removal Percentage from ESPs
and FFs (Excluding Mercury) .
13-11. Descriptive Statistics for HAP Removal Efficiencies
Shown in Figures 13-15 and 13-16
13-12. Descriptive Statistics for HAP Removal Efficiencies
Shown in Figures 13-17 and 13-18
13-13. Qualitative Effects of Different Control Strategies
on Air Emissions of HAPs
13-14. Comparison of Typical Uncontrolled Flue Gas Parameters
at Utilities and MWCs
13-24
13-25
13-26
13-27
13-29
13-30
13-33
13-42
xvi 11
-------
LIST OF FIGURES
Figure Page
ES-1. Summary of the utility air toxics study ES-3
ES-2. Number of coal-fired utilities posing various levels
of maximum individual risks (by levels of MIR) . . . ES-9
ES-3. Number of oil-fired utilities posing various levels
of maximum individual risks (by levels of MIR) . . ES-10
2-1. Fossil fuel use in the utility industry in 1994 . . 2-2
2-2. Unit types in the utility industry by fuel type
in 1994 2-8
2-3. Particulate control in the utility industry by fuel
type in 1994 2-11
2-4. S02 control in the utility industry in 1994 (coal-
fired boilers only) 2-16
2-5. Nitrogen oxide control in the utility industry by
fuel type in 1994 2-22
2-6. Fuel use in the utility industry by fuel type in
1990 and projections for the year 2010 2-27
2-7. Projected use of fuels by 2010 for utility industry 2-30
3-1. Trace elements in coal 3-5
3-2. Trace elements in oil and natural gas 3-6
3-3. Organic emissions 3-7
6-la. Number of coal-fired utilities posing various levels
of maximum individual risks (by levels of MIR) . . . 6-5
6-lb. Number of coal-fired utilities posing various levels
of maximum individual risk (by levels of MIR) . . . 6-6
6-2a. Number of oil-fired utilities posing various levels
of maximum individual risk (by levels of MIR) . . . 6-10
6-2b. Number of oil-fired utilities posing various levels
of maximum individual risk (by levels of MIR) . . . 6-11
6-3. Results of the RELMAP modeling analysis from 1990
emissions estimates for arsenic from coal utilities:
predicted air concentration of arsenic,
units: ng/m3 6-27
6-4. Results of the RELMAP modeling analysis from 1990
emissions estimates for arsenic from oil utilities:
predicted air concentration of arsenic,
units: ng/m3 6-28
6-5. Results of the RELMAP modeling analysis from 1990
emissions estimates for arsenic from coal and oil
utilities: predicted air concentration of arsenic,
units: ng/m3 6-29
6-6. Results of the RELMAP modeling analysis from 1990
emissions estimates for cadmium from coal utilities:
predicted air concentration of cadmium,
units: ng/m3 6-30
xix
-------
LIST OF FIGURES (continued)
Figure Page
6-7. Results of the RELMAP modeling analysis from 1990
emissions estimates for cadmium from oil utilities:
predicted air concentration of cadmium,
units: ng/m3 6-31
6-8. Results of the RELMAP modeling analysis from 1990
emissions estimates for cadmium from coal and oil
utilities: predicted air concentration of cadmium,
units: ng/m3 6-32
6-9. Results of the RELMAP modeling analysis from 1990
emissions estimates for chromium from coal utilities:
predicted air concentration of chromium,
units: ng/m3 6-33
6-10. Results of the RELMAP modeling analysis from 1990
emissions estimates for chromium from oil utilities:
predicted air concentration of chromium,
units: ng/m3 6-34
6-11. Results of the RELMAP modeling analysis from 1990
emissions estimates for chromium from coal and oil
utilities: predicted air concentration of chromium,
units: ng/m3 6-35
6-12. Results of the RELMAP modeling analysis from 1990
emissions estimates for nickel from coal utilities:
predicted air concentration of nickel,
units: ng/m3 6-36
6-13. Results of the RELMAP modeling analysis from 1990
emissions estimates for nickel from oil utilities:
predicted air concentration of nickel,
units: ng/m3 6-37
6-14. Results of the RELMAP modeling analysis from 1990
emissions estimates for nickel from coal and oil
utilities: predicted air concentration of nickel,
units: ng/m3 6-38
6-15. Estimates of annual cancer incidence due to
inhalation exposure to HAP emissions from oil-fired
electric utilities based on the local analysis using
alternative UREs for nickel (as % of nickel
subsulfide) 6-51
6-16. Depiction of combining component uncertainty
distributions (i.e., emissions, dispersion, and
exposure-response) into an overall distribution of
uncertainty (e.g., MIR) 6-65
6-17. Summary of results of Monte Carlo simulation of HAP
emissions from oil-fired plant no. 29 6-67
7-1. Comparison of estimated current and pre-industrial
mercury budgets and fluxes 7-3
xx
-------
LIST OF FIGURES (continued)
Figure Page
7-2. Total modeled mercury deposits from wet and dry
deposition from coal utilities based on 1994 emissions
estimates as modeled with RELMAP,
units: //g/m2/yr 7-25
7-3. Total modeled mercury deposits from wet and dry
deposition from oil utilities based on 1994
emissions estimates as modeled with RELMAP,
units: //g/m2/yr 7-26
7-4. Total modeled mercury deposits from wet and dry
deposition from coal and oil utilities based on 1994
emissions estimates as modeled with RELMAP,
units: //g/m2/yr 7-27
7-5. Overview of the IEM-2M watershed modules 7-31
7-6. Configuration of hypothetical water body and
watershed relative to local source 7-32
8-1. Results of the RELMAP modeling analysis from 1990
emissions estimates for lead from coal utilities:
predicted air concentration of lead, units: ng/m3 . 8-5
8-2. Results of the RELMAP modeling analysis from 1990
emissions estimates for lead from oil utilities:
predicted air concentration of lead, units: ng/m3 . 8-6
8-3 Results of the RELMAP modeling analysis from 1990
emissions estimates for lead from coal and oil
utilities: predicted air concentration of lead,
units: ng/m3 8-7
8-4. Predicted lead wet and dry deposition from coal
utilities based on 1990 emissions estimates as
modeled with RELMAP, units: //g/m2/yr 8-8
8-5. Predicted lead wet and dry deposition from oil
utilities based on 1990 emissions estimates as
modeled with RELMAP, units: //g/m2/yr 8-9
8-6. Predicted lead wet and dry deposition from coal
and oil utilities based on 1990 emissions
estimates as modeled with RELMAP,
units: //g/m2/yr 8-10
8-7. Predicted cadmium wet and dry deposition from coal
utilities based on 1990 emissions estimates as
modeled with RELMAP, units: //g/m2/yr 8-13
8-8. Predicted cadmium wet and dry deposition from oil
utilities based on 1990 emissions estimates as
modeled with RELMAP, units: //g/m2/yr 8-14
8-9. Predicted cadmium wet and dry deposition from coal
and oil utilities based on 1990 emissions estimates
as modeled with RELMAP, units: //g/m2/yr 8-15
xxi
-------
LIST OF FIGURES (continued)
Figure Page
10-1. The generalized geochemical cycle for arsenic . . . 10-3
10-2. Predicted arsenic wet and dry deposition from coal
utilities based on 1990 emissions estimates as
modeled with RELMAP, units: //g/m2/yr 10-14
10-3. Predicted arsenic wet and dry deposition from oil
utilities based on 1990 emissions estimates as
modeled with RELMAP, units: //g/m2/yr 10-15
10-4. Predicted arsenic wet and dry deposition from coal
and oil utilities based on 1990 emissions estimates
as modeled with RELMAP, units: //g/m2/yr 10-16
10-5. Location of waterbody considered within watershed 10-20
11-1. Results of the RELMAP modeling analysis from 1990
emissions estimates for total dioxin (incl TEQ
factors) from coal and oil utilities: predicted
air concentration of total dioxin, units: attograms
(10"18 grams)/m3 11-7
11-2. Predicted total dioxin (incl TEQ factors) wet
deposition from coal and oil utilities based on
1990 emissions estimates as modeled with RELMAP,
units: picograms (10~12 grams)/m3 11-8
11-3. Predicted total dioxin (incl TEQ factors) dry
deposition from coal and oil utilities based on 1990
emissions estimates as modeled with RELMAP,
units: picograms (10~12 grams)/m3 11-9
11-4. Flow chart of multipathway processes 11-15
11-5. Risk model sensitivity to changes in meat and fish
factors 11-24
11-6. Risk model sensitivity to changes in soil loss rate
constant 11-24
11-7. Sensitivity of predicted risk to the subsistence
fisher to changes in parameter values 11-26
13-1. Relation between the concentrations of mercury and
Sulfur in 153 samples of coal shipments 13-4
13-2(a-g). Relation between concentration of selected trace
Elements and sulfur in modified USGS data 13-6
13-3. Coal gasification combined cycle technology . . . 13-12
13-4 (a-c) . Average coal-fired boiler emissions, trace metal
removal, and average trace element concentration
in feed coal vs. bottom type (bituminous and
subbituminous coal) 13-19
13-5 (a-c) . Average coal-fired boiler emissions, trace metal
removal, and average trace element concentration
in feed coal vs. bottom type (bituminous coal-fired
only) 13-20
xxi i
-------
LIST OF FIGURES (continued)
Figure Page
13-6 (a-c). Average coal-fired boiler emissions, trace metal
removal, and average trace element concentration
in feed coal vs. bottom type (subbituminous
coal-fired only) 13-21
13-7. Removal of metallic HAPs by electrostatic
precipitators (cold-side, coal) (includes, arsenic,
beryllium, cadmium, chromium, lead, and manganese) 13-23
13-8. Removal of mercury by electrostatic precipitators
(cold-side, coal) 13-23
13-9. Removal of metallic HAPs by electrostatic
precipitators (hot-side, coal) (includes, arsenic,
beryllium, cadmium, chromium, lead, and manganese) 13-24
13-10. Removal of mercury by electrostatic precipitators
(hot-side, coal) 13-24
13-11. Removal of metallic HAPs by an electrostatic
precipitator (oil) (includes, arsenic, lead, and
nickel) 13-25
13-12. Removal of mercury by an electrostatic
precipitator (oil) 13-25
13-13. Removal of metallic HAPs by a fabric filter(coal)
(includes, arsenic, beryllium, cadmium, chromium,
lead, and manganese) 13-26
13-14. Removal of mercury by a fabric filter (coal) . . . 13-26
13-15. Removal of metallic HAPs by an FGD (coal) (includes,
arsenic, beryllium, cadmium, chromium, lead, and
manganese) 13-29
13-16. Removal of mercury by an FGD (coal) 13-29
13-17. Removal of metallic HAPs by a spray dryer
adsorber/fabric filter (coal) (includes, arsenic,
beryllium, cadmium, chromium, lead, and manganese) 13-30
13-18 Removal of mercury by a spray dryer adsorber/
fabric filter (coal) 13-30
xxi 11
-------
-------
Glossary
AALG
AC
AECDP
AFBC
APCD
ACGIH
ARD
ATSDR
BAF
BAP
BBF
BCF
BFB
BOO
BSAF
CAA
CAP
CAPCOA
CARB
CCT
CEDF
CFB
CLD
CNS
CSF
DNA
DOC
DOE
DSM
E
ECTC
ambient air level goal
activated carbon
Advanced Emissions Control Development Program (Babcock
& Wilcox)
atmospheric fluidized-bed combustor
air pollution control device
American Conference of Government Industrial Hygienists
Acid Rain Division (EPA)
Agency for Toxic Substances and Disease Registry
bioaccumulation factor
benzo[a]pyrene
biased burner firing
bioconcentration factor
bubbling fluidized bed
burners out of service
biota sediment accumulation factor
Clean Air Act
Clean Air Act Assessment Package
California Air Pollution Control Officers Association
California Air Resource Board
clean coal technology
Clean Environment Development Facility (Babcock &
Wilcox)
circulating fluidized bed
certainly lethal dose
central nervous system
cancer slope factor
deoxyribonucleic acid
dissolved organic carbon
Department of Energy
demand side management
(ratio of) exposure
Environmental Control Test Center (EPRI)
xxv
-------
EEI Edison Electric Institute
EFP emission factor program
EIA Energy Information Administration (DOE)
E-LIDS™ Enhanced Limestone Injection Dry Scrubbing
EMF emission modification factor
ENAMAP Eastern North American Model of Air Pollution
EPA Environmental Protection Agency
EPRI Electric Power Research Institute
ESP electrostatic precipitator
EURMAP European Regional Model of Air Pollution
FBC fluidized-bed combustor
FCEM field chemical emissions monitoring
FETC Federal Energy Technology Center
FF fabric filter
FGD flue gas desulfurization
FGR fluidized gas recirculation
FTIR Fourier transform infrared
GEIA Global Emissions Inventory Activity
GI gastrointestinal
GIS Geographic Information System
HAP hazardous air pollutant
HEC human equivalent concentration
HEM Human Exposure Model
HI hazard index
HQ hazard quotient
IARC International Agency for Research on Cancer
ICRP International Commission on Radiological Protection
IDLH immediately dangerous to life and health
IEM indirect exposure methodology
IGCC integrated gasification combined cycle
IPP independent power producer
IRIS Integrated Risk Information System
IRP Inerts Ranking Program
ISC3 Industrial Source Complex Version 3
ISCLT2 Industrial Source Complex Long Term Version 2
xxvi
-------
ISCLT3 Industrial Source Complex Long Term Version 3
ISCST3 Industrial Source Complex Short Term Version 3
IURE inhalation unit risk estimate
LADD lifetime averaged daily dose
LC50/LD50 lethal concentration/dose that kills 50% of test animals
LCUB large coal-fired utility boiler
LET linear energy transfer
L/G liquid-to-gas
LOAEL lowest-observed-adverse-effect level
LRT long-range transport
MACT maximum achievable control technology
MCL maximum contaminant level
MCLG maximum contaminant level goal
MCUB medium coal-fired utility boiler
MDL minimum detectable level
ME I maximally exposed individual
MFP monofluorophosphate
MIR maximum individual risk
MLE maximum likelihood estimate
MRL minimal risk level
MWC municipal waste combustor
MWe megawatts electric
MWI medical waste incinerator
NAAQS national ambient air quality standards
NAPAP National Acid Precipitation Assessment Program
NAREL National Air and Radiation Environmental Laboratory
NAS National Academy of Science
NASN National Air Surveillance Network
NGM nested grid model
NIEHS National Institute of Environmental Health Sciences
NIOSH National Institute for Occupational Safety and Health
NIST National Institute of Standards and Technology
NMHC nonmethane hydrocarbon
NOAA National Oceanographic and Atmospheric Administration
NOAEL no-observed-adverse-effect level
xxvi i
-------
NOEC
NRC
NSPC
NSPS
NTP
NWS
OAQPS
OAR
OFA
ORD
ORIA
OSF
OSHA
OTAG
CURE
PAB
PAH
PBPK
PCB
PCDD
PCDF
PEL
PFBC
PM
PSCCO
PUR PA
RAG
RBC
RCRA
REL
RELMAP
RfC
RfD
SAB
SAMSON
no-observed-effect concentration
National Research Council
Northern States Power Company
new source performance standards
National Toxicology Program
National Weather Service
Office of Air Quality Planning and Standards (EPA)
Office of Air and Radiation (EPA)
overfire air
Office of Research and Development (EPA)
Office of Radiation and Indoor Air (EPA)
off-stoichiometric firing
Occupational Safety and Health Administration
Ozone Transport Assessment Group
oral unit risk estimate
Pollutant Assessment Branch (EPA)
polycyclic aromatic hydrocarbon
physiologically based pharmacokinetic
polycyclic biphenyl
polychlorinated dibenzo-p-dioxin
polychlorinated dibenzofuran
permissible exposure limit
pressurized fluidized-bed combustor
particulate matter
Public Services Company of Colorado
Public Utility Regulatory Policies Act
Risk Assessment Council (EPA)
risk based concentration
Resource Conservation and Recovery Act
reference exposure level
Regional Lagrangian Model of Air Pollution
reference concentration
reference dose
Science Advisory Board
Solar and Meteorological Surface Observation Network
xxvi11
-------
SBS small boiler simulator
SCR selective catalytic reduction
SCRAM-BBS Support Center for Regulatory Air Models Bulletin Board
System
SCUB small coal-fired utility boiler
SDA spray dryer absorber
SE standard error
SGU steam generating unit
SIP State Implementation Plan
SNCR selective noncatalytic reduction
SPC Science Policy Council (EPA)
STAR STability ARray
TCDD tetrachlorodibenzo-p-dioxin
TE trace element
TEF toxicity (or toxic) equivalency factor
TEQ toxicity equivalent
TLV threshold limit value
TOC total organic compounds
tpy tons per year
TRI Toxics Release Inventory
TSP total suspended particulate
TTN Technology Transfer Network
UARG Utility Air Regulatory Group
UDI Utility Data Institute
UNDEERC University of North Dakota Energy and Environmental
Research Center
URE unit risk estimate
USGS U.S. Geological Survey
USPHS U.S. Public Health Service
VOC volatile organic compound
WHO World Health Organization
WL working level
WLM working level month
WOE weight of evidence
xxix
-------
EXECUTIVE SUMMARY
ES.l LEGISLATIVE MANDATE
In section 112(n)(1)(A) of the Clean Air Act, as amended (the
Act), Congress directs the United States Environmental Protection
Agency (EPA) to:
"... perform a study of the hazards to public health
reasonably anticipated to occur as a result of emissions by
electric utility steam generating units of ... [hazardous
air pollutants] ... after imposition of the requirements of
this Act."
Section 112(a)(8) of the Act defines an "electric utility
steam-generating unit" as "any fossil-fuel-fired combustion unit of
more than 25 megawatts electric (MWe) that serves a generator that
produces electricity for sale." A unit that cogenerates steam and
electricity and supplies more than one-third of its potential electric
output capacity and more than 25 MWe output to any utility power
distribution system for sale is also considered an electric utility
steam-generating unit (i.e., utility unit).
Section 112(n)(1)(A) also requires that:
• The EPA develop and describe alternative control strategies
for hazardous air pollutants (HAPs) that may warrant
regulation under section 112; and
• The EPA proceed with rulemaking activities under section 112
to control HAP emissions from utilities if EPA finds such
regulation is appropriate and necessary after considering
the results of the study.
ES.2 REGULATORY DETERMINATION
This report does not contain a determination as to whether or not
regulations to control HAP emissions from utility units are
appropriate and necessary. The Agency has deferred the regulatory
determination until a later date.
ES.3 OVERVIEW APPROACH TO COMPLETING THE STUDY
The study included numerous separate and interrelated analyses.
First, HAP emissions test data were gathered from 52 utility units
(i.e., boilers), including a range of coal-, oil-, and natural
gas-fired utility units. Second, the emissions test data along with
facility specific information (e.g., boiler type, control device, fuel
usage) were used to estimate HAP emissions from all 684 utility plants
in the United States (U.S.). Third, a screening level hazard/risk
assessment was completed to prioritize the HAPs for further analyses.
Fourth, various priority HAPs were analyzed for inhalation and
ES-1
-------
multipathway exposures and risks and other potential impacts. In
addition, potential control strategies were analyzed for the priority
HAPs. The overall summary of the study is presented in Figure ES-1.
This report presents the findings of the study. The primary
components of this report are: (1) a description of the industry;
(2) an analysis of emissions data; (3) an assessment of hazards and
risks due to inhalation exposures to 67 HAPs; (4) assessments of risks
due to multipathway (inhalation plus non-inhalation) exposures to four
HAPs (radionuclides, mercury, arsenic, and dioxins); and (5) a
discussion of alternative control strategies.
The study was based primarily on two scenarios: (1) 1990 base
year emissions; and (2) 2010 emissions. In addition, emissions for
1994 were estimated using the most recent data. The 1990 scenario was
chosen since that was the year the Amendments to the Act were passed
and was the latest year for which utility operational data were
available at the time the study was initiated. The 2010 scenario was
selected to meet the section 112 (n) (1) (A) mandate to evaluate hazards
"after imposition of the requirements of the Act." Primarily, this
meant assessing the hazards after the acid rain program is in place.
The 2010 scenario also included estimated changes in HAP emissions
resulting from projected trends in fuel choices and projected
increases in electric power demands. However, the effects of other
on-going or potential activities that were not factored into the 2010
projections (e.g., industry restructuring, new ozone and particulate
matter [PM] standards, global climate change programs) may result in
the 2010 projections being either underestimated or overestimated.
ES.4 EMISSIONS DATA ANALYSIS
A total of 684 utility plants (i.e., utilities) were identified
as meeting the criteria for the study in 1990 in the U.S. These
utilities are fueled primarily by coal (59 percent of total units),
oil (12 percent), or natural gas (29 percent). Many plants have two
or more units and several plants burn more than one type of fuel
(e.g., contain both coal- and oil-fired units). In 1990, there were
426 plants that burned coal as one of their fuels, 137 plants that
burned oil, and 267 plants that burned natural gas.
Emission estimates for the years 1990, 1994, and 2010 were based
on emissions test data from 52 units obtained from extensive emission
tests by the Electric Power Research Institute (EPRI), the Department
of Energy (DOE), the Northern States Power Company, and the EPA. The
testing program was designed to test a wide range of facility types
with a variety of control scenarios; therefore, the data are
considered generally representative of the industry. However, there
are uncertainties in the data because of the small sample sizes for
specific boiler types and control scenarios.
ES-2
-------
Figure ES-1. Overall Structure of Utility Air Toxics Study Analyses
Emissions Data Analysis
- Test Data from 52 Units
- 67 HAPs Identified
- Emissions Estimates for 684 Plants in U.S.
HCIand HF
- Literature Review
of Potential Impacts
Screening Assessment
to Prioritize HAPs
-14 HAPs identified
Alternative Control
Strategies Analysis
- Mercury
- Metals
- Dioxins
- HCI/HF
Inhalation Exposure/Risk Assessment
M
CO
I
co
Local Analysis
- Within 50 km of Each Plant
- Modeled 67 HAPs from All
Plants with Local Model
(HEM)
Local plus Long-range
Transport
- Beyond 50 km to U.S. Borders
- Regional Model (RELMAP)
- Modeled 4 HAPs from All
Plants (Ni, As, Cd, Cr)
Estimated Air
Concentration in
U.S. from All
Plants
Multipathway Exposure/Risk Assessment
Local Model (CAP 93)
- Radionuclides only
- All Plants
- MIR 3E-05
- Cancer Incidence 0.3
cases/yr (coal & oil)
Risk Assessment Results
- Ni, As, Cr, Rads Primary HAPs
- Coal Incidence up to 0.2 cases/yr
- Oil Incidence up to 0.5 cases/yr
- Highest Predicted Cancer MIR
for Ni up to 5E-05
- Most Risks Predicted
to be < 1E-06
Risk Assessment Results
- Coal Incidence Increased 7 Fold,
to 1.3 cases/yr
- Oil Incidence up to 0.5 cases/yr
- No Change in Predicted Cancer
MIRs
up
Regional Model (RELMAP)
- All Plants
-7 HAPs (Hg, As, Dioxins,
Cd, Cr, Ni, Pb)
Estimated Air
Concentration and
Depostion in U.S.
from All Plants
MAPS
Representative Plant Analyses
- 4 Model Plants
-3HAPs(Hg, As, Dioxins)
- Local Dep. Model (ISC3)
-Food Web Model (I EM)
- Hypothetical Exposure Scenarios
Arsenic - Results
- Background Dominates
Exposure
- Highest Predicted Cancer
MIR up to 1E-04 ('pica child")
- Most Risks Predicted
to be < 1E-05
Mercury - Results
- Fish was Dominant
Exposure Pathway
- Effect of Concern is
Devel. Neuro. Toxicity
- Plausible Link
Dioxins - Results
- Highly Toxic at Low Levels
- Highest Predicted Cancer
MIRupto2E-04(coal)
- Majority of Risks Predicted
to be < 1 E-05
-------
These test data provided the basis for estimating average annual
emissions for each of the 684 plants. A total of 67 of the 188 HAPs
listed in section 112 of the Act were identified in the emissions
testing program as potentially being emitted by utilities. Tables
ES-1 and ES-2 present estimated emissions for, respectively, a subset
of priority HAPs for 1990, 1994, and 2010, and for a set of
characteristic boilers for 1994.
Although the EPA used average annual emissions estimates in
assessing long-term exposures to individual HAPs on a national basis,
emissions test data were not available for each utility in the U.S.
Therefore, estimates for individual plants are particularly uncertain.
Based on an uncertainty analysis, the average annual emissions
estimates are expected to be roughly within a factor of plus or minus
three of actual annual emissions. However, even this uncertainty
analysis had limitations. For example, the uncertainty analysis did
not include data on potential upsets or unusual operating conditions;
therefore, the range of uncertainty could be greater.
ES.5 GENERAL APPROACH TO EXPOSURE AND RISK ASSESSMENT
Most of the risk assessment focused on inhalation exposure. All
67 HAPs were assessed for inhalation exposures, at least at a
screening level. For many of the 67 HAPs, inhalation exposure is
believed to be the dominant exposure pathway. However, for HAPs that
are persistent and/or bioaccumulate, and are toxic by ingestion (or
are radioactive), the non-inhalation exposure pathways could be more
important. Based on a screening and prioritization assessment, which
is described below, the EPA identified four high priority HAPs
(radionuclides, mercury, arsenic, dioxins) to assess for non-
inhalation exposures. In addition, cadmium and lead were identified
as next highest priority. Multipathway assessments are presented for
radionuclides, mercury, arsenic, and dioxins. The other two HAPs
(lead and cadmium) were examined qualitatively for their potential for
multipathway hazards.
ES.6 SCREENING ASSESSMENT
As outlined in Figure ES-1, EPA initially conducted a screening
assessment that considered inhalation and non-inhalation exposure
routes for all 67 HAPs to identify priority HAPs for more detailed
assessment. To screen for inhalation exposures, the EPA used the
Human Exposure Model (HEM) to model the 67 HAPs from all 684 utility
plants utilizing generally conservative assumptions (i.e., assumptions
that are more likely to overestimate rather than underestimate risks)
to estimate inhalation risks for maximally exposed individuals (MEIs).
ES-4
-------
Table ES-1. Nationwide Utility Emissions for Thirteen Priority HAPs3
HAP
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Nickel
Hydrogen chloride
Hydrogen fluoride
Acrolein
Dioxinsd
Formaldehyde
Nationwide HAP emission estimates (tons per year)b
Coal
1990
61
7.1
3.3
73
75
164
46
58
143,000
20,000
25
0.000097
35
1994
56
7.9
3.2
62
62
168
51
52
134,000
23,000
27
0.00012
29
2010
71
8.2
3.8
87
87
219
60
69
155,000
26,000
34
0.00020
45
Oil
1990
5
0.46
1.7
4.7
11
9.3
0.25
390
2,900
140
NM
1 x 10'5
19
1994
4
0.4
1.1
3.9
8.9
7.3
0.2
320
2,100
280
NM
9x10'6
9.3
2010
3
0.23
0.9
2.4
5.4
4.7
0.13
200
1,500
73
NM
3x10'6
9.5
Natural gas
1990
0.15
NMC
-
-
0.43
-
0.0015
2.2
NM
NM
NM
NM
36
1994
0.18
NM
-
-
0.47
-
0.0017
2.4
NM
NM
NM
NM
39
2010
0.25
NM
-
-
0.68
-
0.024
3.5
NM
NM
NM
NM
57
M
CD
I
Ul
a Radionuclides are the one priority HAP not included on this table because radionuclide emissions are measured in different units (i.e., curies per year) and,
therefore, would not provide a relevant comparison to the other HAPs shown. Radionuclide emissions are presented in chapter 9.
b The emissions estimates in this table are derived from model projections based on a limited sample of specific boiler types and control scenarios. Therefore,
there are uncertainties in these numbers (see section ES.4 for discussion).
c NM = Not measured.
d These emissions estimates were calculated using the toxic equivalency (TEQ) approach, which is based on the summation of the emissions of each congener
after adjusting fortoxicity relative to 2,3,7,8-tetrachlorodibenzo-p-dioxin (i.e., 2,3,7,8-TCDD).
-------
Table ES-2. Estimated Emissions for Nine Priority HAPs from
Characteristic Utility Units (1994; tons per year)a
Fuel:
Unit size (MWe):
Arsenic
Cadmium
Chromium
Lead
Mercury
Hydrogen chloride
Hydrogen fluoride
Dioxins0
Nickel
Coal
325
0.0050
0.0023
0.11
0.021
0.05
190
14
0.00000013
NC
Oil
160
0.0062
0.0014
0.0062
0.014
0.0012
9.4
NC
0.000000023
1.7
Natural gas
240
0.0003
NCb
NC
NC
NC
NC
NC
NC
0.004
a There are uncertainties in these numbers. Based on an uncertainty analysis, the EPA predicts that the emissions
estimates are generally within a factor of roughly three of actual emissions.
b NC = Not calculated.
c See footnote d of Table ES-1.
If the MEI risk was above a minimum measure (e.g., exposure greater
than one-tenth the inhalation reference concentration [RfC] a or cancer
risk greater than 1 chance in 10 million), then the HAP was chosen for
more study. For non-inhalation exposures, the 67 HAPs were
prioritized by considering five criteria: (1) persistence;
(2) tendency to bioaccumulate; (3) toxicity; (4) emissions quantity;
and (5) radioactivity.
Based on this screening assessment, a total of 14 HAPs were
identified as priority. Twelve HAPs (arsenic, beryllium, cadmium,
chromium, manganese, nickel, hydrogen chloride [HC1] , hydrogen
fluoride [HF], acrolein, dioxins, formaldehyde, and radionuclides)
were identified as priority pollutants for further study based on
potential for inhalation exposures and risks. Four of these 12 HAPs
(arsenic, cadmium, dioxins, and radionuclides) plus 2 additional HAPs
(mercury and lead) were considered priority for multipathway
exposure); of these 6 HAPs, 4 (arsenic, mercury, dioxins, and
radionuclides) were identified as the highest priority to assess for
The RfC is an estimate (with uncertainty spanning perhaps an order of
magnitude) of the daily inhalation exposure of the human population
(including sensitive subgroups) that is likely to be without appreciable
risk of deleterious effects during a lifetime.
ES-6
-------
multipathway exposures and risks. Overall, a total of 14 of the 67
HAPs were considered priority. The other 53 HAPs were not evaluated
beyond the screening assessment.
ES.7 INHALATION RISK ASSESSMENT -- LOCAL ANALYSIS
The EPA estimated inhalation exposures and risks due to
dispersion of HAP emissions within 50 kilometers (km) of each of the
684 plants (i.e., local analysis). For 13 of the 14 priority HAPs,
the HEM was used; for radionuclides, the Clean Air Act Assessment
Package-1993 (CAP-93) model was used. The HEM exposure modeling
conducted for the inhalation risk assessment was very similar to the
modeling conducted for the screening assessment. The same default
options and same input data were used. However, there is one
important difference. For the inhalation risk assessment, a
distinction was made between urban and rural locations. If a plant is
located in an urban area, it was modeled using the urban mode (i.e.,
dispersion is assumed to be characteristic of emissions emitted by a
facility in an urban location where there are buildings nearby).
Dispersion of the pollutant plume in an urban area is expected to
exhibit greater turbulence because of heat transfer and obstacles
(i.e., large buildings). If a plant is located in a rural location,
it was modeled using the rural mode (i.e., dispersion is assumed to be
characteristic of a facility located in a rural location). In the
screening assessment, all plants were modeled using the urban default
because using the urban default typically leads to more conservative
(i.e., higher) estimates of human exposures, which is appropriate for
a screening assessment. However, using the urban and rural
distinction is believed to reflect more realistic conditions.
The cancer risks for all gas-fired plants were well below one
chance in one million (i.e., < 1 x 10"s) and no noncancer hazards were
identified. Therefore, gas-fired plants are omitted from the
following discussions.
In cases where data were missing or incomplete, the EPA had to
make various assumptions. A few of these assumptions are more likely
to overestimate risks. Other assumptions used are likely to
underestimate risks. Based on an uncertainty analysis conducted for
this study, it is estimated that these assumptions taken together lead
to a reasonable high-end estimate (i.e., conservative, but within the
bounds of reasonable estimates) of the risks due to inhalation
exposure within 50 km of plants. Within the limits of current
scientific information, this approach is, therefore, most likely to
overestimate health risks for these pollutants. The uncertainty
analysis suggests that the most likely estimated inhalation MIRs
(i.e., central tendency MIRs) may be roughly 2 to 10 times lower than
the high-end MIRs presented below. The average individual risks due
to inhalation exposure to utility HAP emissions for the total exposed
U.S. population (roughly 200,000,000 people) are predicted to be
roughly 100 to 1000 times lower than the high-end inhalation MIRs.
ES-7
-------
ES. 7.1 Inhalation Cancer Risks for Coal-Fired Utilities Based on
Local Analysis (1990)
The vast majority of coal-fired plants (424 of the 426 plants)
are estimated to pose lifetime cancer risks (i.e., increased
probability of an exposed person getting cancer during a lifetime) of
less than 1 x 10"s due to inhalation exposure to utility HAP emissions.
Only two of the 426 plants are estimated to potentially pose
inhalation risks greater than 1 x 10"s (see Figure ES-2).
The increased lifetime cancer MIR due to inhalation exposure to
coal-fired utility HAP emissions, based on the local analysis, is
estimated to be no greater than 3 x I0~6. Arsenic and chromium are the
HAPs contributing most to the inhalation risks (see Table ES-3). All
other HAPs, including radionuclides, were estimated to present
inhalation risks less than 1 x 10"s for coal-fired units.
The cancer incidence in the U.S. due to inhalation exposure to
HAPs (including radionuclides) from all 426 coal-fired plants based on
the local analysis is estimated to be no greater than approximately
0.2 cancer case per year (cases/yr), or 1 case every 5 years.
However, as described in later sections, the consideration of long-
range dispersion of HAPs (beyond 50 km) results in increased estimates
for cancer incidence.
ES.7.2 Inhalation Cancer Risks for Oil-Fired Utilities Based on Local
Analysis (1990)
The majority of the oil-fired plants (125 of the 137 plants) are
estimated to pose inhalation cancer MIRs less than 1 x I0~6. However,
up to 11 of the 137 oil-fired plants are estimated to potentially
present inhalation MIRs above 1 x 10"s (see Figure ES-3). Nickel,
arsenic, radionuclides, and chromium are the primary contributors to
these cancer risks.
For oil-fired utilities, the highest contribution to the MIRs is
from nickel. However, there are substantial uncertainties with the
nickel risk estimates. Nickel is emitted in several different forms
(e.g., nickel oxides, soluble nickel, sulfidic nickel) and the health
effects of these different forms vary, and for some forms are unknown
or uncertain. Nickel subsulfide (which is one of the possible forms
of sulfidic nickel) is a known human carcinogen and appears to be the
most carcinogenic form based on available data. Based on limited
data, 3 to 26 percent of the nickel emissions are believed to be
sulfidic nickel. It is not known how much of the sulfidic nickel
emissions are nickel subsulfide. Several other nickel species (e.g.,
nickel oxides) are also potentially carcinogenic but the potencies are
not known.
ES-8
-------
Figure ES-2. Number of Coal-Fired Utilities Posing Various Levels of
Maximum Individual Risks (By Levels of MIR)
M
CD
I
C
(5
a.
•o
SI
E
3
z
All carcinogenic non-radionuclide HAPs
Maximum individual risk
1 E-8 to 1 E-7
1 E-7 to 1 E-6 ^H 1 E-6 to 1 E-5
Note: The high-end maximum individual risks (MIRs) are presented as exponents in this figure. For example, an increased cancer risk of one chance in one
million (i.e., 1 x 10~6) is shown as 1E-6 in this Figure. The figure shows that 91 plants are estimated to pose an MIR between 1 x 107 and 1 x 106.
-------
Figure ES-3. Number of Oil-Fired Utilities Posing Various Levels of
Maximum Individual Risks (By Levels of MIR)
M
CD
I
60 -
50 -
1 30
'o
fe
S2
= 20 -
All carcinogenic non-radionuclide HAPs
52
1E-8 to 1E-7 1E-7 to 1 E-6
Maximum individual risk
1 E-6 to 1 E-5
1 E-5 to 1 E-4
1 E-8 to 1 E-7 kVX/d 1 E-7 to 1 E-6 1 E-6 to 1 E-5 ^H 1 E-5 to 1 E-4
Note: The MIRs are presented as exponents in this figure. For example, an increased cancer risk of 1 x 10 "6is shown as 1E-6 here. The figure shows there are
9 oil-fired plants with estimated MIRs between 1 x 10~6and 1 x 10~5
-------
Table ES-3. Summary of High-End Inhalation Cancer Risk Estimates
from Local Analysis for Coal-Fired Utilities for the Year 1990
HAP
Arsenic
Chromium
Total" (Aggregate of HAPs)
Highest
Cancer MIRa
2x10'6
1 x10'6
3x10'6
Population with lifetime risk
>1 xlO-6
850
110
850
Number plants with
MIR>1 xlO-6
2
1
2
a Estimated lifetime maximum individual risk (MIR) due to inhalation exposure for the "highest risk" coal-fired plant.
Based on an uncertainty analysis, these estimates are considered reasonable high-end estimates (see section
ES.7.4 for discussion).
b Estimated risk due to inhalation of the aggregate of HAPs assuming additivity of risk for 26 individual carcinogenic
HAPs.
To evaluate the range of potential risks due to nickel emissions,
the EPA estimated risks using various assumptions for nickel cancer
potency (presented in chapter 6). For example, assuming the nickel
mix is 50 percent as carcinogenic as nickel subsulfide, the highest
inhalation cancer MIR due to the aggregate of HAP emissions from the
highest risk oil-fired utility plant is estimated to be 6 x 10 "5.
Assuming the nickel mix is 10 percent as carcinogenic as nickel
subsulfide, the highest inhalation cancer MIR due to the aggregate of
HAP emissions from the highest risk oil-fired utility plant is
approximately 3 x 10"5. The values in Table ES-4 and Figure ES-3 are
based on the conservative assumption that the nickel mix is 50 percent
as carcinogenic as nickel subsulfide.
Estimated risks due to inhalation exposure for a subset of HAPs
based on the local analysis are presented in Table ES-4. All other
HAPs analyzed were estimated to pose inhalation cancer risks below
1 x IQ-6 for all 137 oil-fired plants.
The cancer incidence in the U.S. due to inhalation exposure to
HAP emissions (including radionuclides) from all 137 oil-fired
utilities, based on the local analysis, is estimated to be no greater
than 0.5 cancer case/yr.
ES.7.3 Inhalation Cancer Risks Based on Long-Range Transport
In addition to the above analyses, the EPA conducted long-range
transport analyses to assess emissions dispersion and exposures on a
national scale for 1990. The Regional Lagrangian Model of Air
Pollution (RELMAP) was used to estimate the dispersion of HAP
emissions from the facility stack out to the borders of the
continental U.S. This is in contrast to the HEM, which estimates
dispersion and air concentrations within 50 km of the source.
ES-11
-------
Table ES-4. Summary of High-end Inhalation Cancer Risk Estimates
Based on Local Analysis for Oil-Fired Utilities for the Year 1990
HAP
Nickel"
Arsenic
Radionuclides
Chromium
Cadmium
Total0 (aggregate)
Highest MIRa
5x1Q-5
1 x1Q-5
1 x1Q-5
5x10'6
2x10'6
6x1Q-5
Population with lifetime risk
> 1 x 10-6
110,000
2,400
2,400
2,300
45
110,000
Number plants with MIR
>1 xlO-6
11
2
2
1
1
11
a Estimated lifetime maximum individual risk (MIR) due to inhalation exposure for the "highest risk" oil-fired plant.
Based on an uncertainty analysis, these estimates are considered reasonable high-end estimates (see section
ES.7.4 for discussion).
b The estimates for nickel and total HAPs are based on the assumption that the mix of nickel compounds is 50
percent as carcinogenic as nickel subsulfide.
c Estimated risk due to inhalation of the aggregate of HAPs assuming additivity of risk for 14 individual carcinogenic
HAPs.
The RELMAP modeling was conducted for all coal- and oil-fired
utilities, but was limited to mercury, cadmium, chromium, arsenic,
nickel, lead, and dioxins. Only inhalation exposures to the
carcinogenic HAPs are discussed in this section. Deposition and
multipathway concerns are discussed elsewhere in this report. The
long-range transport modeling indicates that the local HEM analysis
alone does not account for a substantial percentage of the population
exposures due to coal-fired utility emissions. A comparison of the
HEM results to the RELMAP results indicates a significant portion of
emissions disperse further than 50 km, as would be expected for these
HAPs, which are mostly fine particulate substances emitted from
elevated stacks.
The RELMAP results for arsenic, cadmium, chromium, and nickel
(which are emitted mainly as PM) were used to estimate the potential
long-range transport inhalation exposures for other carcinogenic HAPs.
Using this methodology, the highest cancer incidence due to inhalation
exposure to HAPs from coal-fired utilities considering both local and
long-range transport is estimated to be up to 1.3 cases/yr, which is
about 7 times greater than the incidence estimated in the local
analysis alone. The cancer incidence for oil-fired utilities did not
change (see Table ES-5).
ES-12
-------
Table ES-5. Summary of High-End Inhalation Risk Estimates Due to
Local and Long-Range Transport
LOCAL IMPACTS (dispersion within 50 km of each utility plant) d
Pollutant
Radionuclides
Nickel3
Chromium
Arsenic
Cadmium
All Others"
Total0
OIL-FIRED PLANTS
Maximum individual
risk (MIR)
1 x1Q-5
5x1Q-5
5x10'6
1 x1Q-5
2x10'6
8x10'7
6x1Q-5
Annual increased
cancer Incidence
0.2
0.2
0.02
0.04
0.005
0.005
0.5
COAL-FIRED PLANTS
Maximum individual risk
(MIR)
2x10'8
7x10-7
1 x10-6
2x10-6
2x10-7
8x10-7
3x10-6
Annual increased
cancer incidence
0.1
0.005
0.02
0.05
0.0006
0.004
0.2
LOCAL PLUS LONG-RANGE IMPACTS (dispersion from utility emission points to borders of continental U.S.)
Pollutant
Radionuclides
Nickel3
Chromium
Arsenic
Cadmium
All Others"
Total0
OIL-FIRED PLANTS
Maximum individual
risk (MIR)
1 x 1Q-5
5x1Q-5
5x10'6
1 x1Q-5
2x10'6
8x10'7
6x1Q-5
Annual increased
cancer incidence
0.2
0.2
0.02
0.05
0.006
0.006
0.5
COAL-FIRED PLANTS
Maximum individual risk
(MIR)
Not estimated
1 x10'8
2x10-6
3x10-6
3x10-7
1 x10-6
4x10-6
Annual increased
cancer incidence
0.7
0.038
0.15
0.37
0.005
0.028
1.3
a Assumes that the nickel mixture is 50 percent as carcinogenic as nickel subsulfide.
b Estimated risks due to exposure to all remaining HAPs analyzed (i.e., excluding nickel, arsenic, chromium,
cadmium, and radionuclides).
c Aggregate risk (risk due to inhalation exposure to all carcinogenic HAPs, assuming additivity of risks).
d There are uncertainties associated with these risk estimates. See sections ES.7.4 for discussion.
A comparison between the HEM local dispersion results and the
long-range transport modeling results indicates that long-range
transport is much less important for the MIR than it is for cancer
incidence. For example, the MIR from the local analyses for coal-
fired utilities (i.e., inhalation risk of 3 x 1Q~6) is predicted to
ES-13
-------
increase by roughly 10 to 20 percent to about 4 x 10~s when ambient
concentrations are added from long-range transport of arsenic from all
other utilities in the continental U.S. For oil-fired utilities, the
long-range transport of HAPs has no impact on the highest inhalation
MIR because of the remote location of the two highest risk oil-fired
plants.
ES.7.4 Uncertainties with the Inhalation Cancer Risk Assessment
There are several areas of uncertainty in the inhalation risk
assessment including: (1) the impacts of long-range transport;
(2) the emissions and health effects of different forms of chromium
and nickel; (3) the use of a linear non-threshold high-to-low dose
extrapolation model for estimating cancer risks at low exposure
concentrations; (4) the impacts of episodic releases resulting from
upsets or unusual operating conditions; (5) how residence times and
activity patterns impact the exposures; (6) the impacts on sensitive
subpopulations; (7) the impacts of background exposures; and (8) the
risk of complex pollutant mixtures.
The uncertainty analysis indicates that the inhalation cancer
MIRs and incidence estimates presented above are reasonable high-end
estimates of the risks due to inhalation exposure within 50 km of each
plant. That is, the estimates are considered generally conservative
(i.e., predicted to be roughly the 90th to 95th percentile). The
uncertainty analysis suggests that the most likely estimated
inhalation MIRs (i.e., central tendency MIRs) may be roughly 2 to 10
times lower than the high-end MIRs presented above. The average
individual risks due to inhalation exposure to utility HAP emissions
for the total exposed U.S. population (roughly 200,000,000 people) are
predicted to be roughly 100 to 1,000 times lower than the high-end
inhalation MIRs.
ES.7.5 Summary of the Inhalation Cancer Risks
For the majority of utility plants (approximately 671 of the 684
plants), the estimated inhalation cancer risks due to HAP emissions
are less than 1 x 10~s. However, several plants (2 coal plants and up
to 11 oil plants) are estimated to potentially pose inhalation cancer
risks above 1 x 10"s. One oil plant is estimated to pose a high-end
inhalation cancer MIR of up to 6 x 10"5. Based on the assessment, no
greater than 1.8 cancer cases/yr are estimated to occur in the U.S.
due to inhalation exposure to HAP emissions from all coal- and oil-
fired utilities. Further research and evaluation may be needed to
more comprehensively assess the inhalation cancer risks, especially to
reduce the uncertainties associated with the nickel risk estimates.
ES.7.6 Inhalation Noncancer Risks
The EPA also assessed noncancer risks (i.e., health effects other
than cancer) due to short- and long-term inhalation exposure.
Manganese, HC1, HF, and acrolein were found to be the four HAPs of
highest potential concern for noncancer effects.
ES-14
-------
Based on modeling HAPs for all 684 plants with the HEM, estimated
long-term ambient HAP concentrations were generally 100 to 10,000
times below the RfC or similar benchmark. The highest estimated long-
term ambient HAP concentration was 10 times below the RfC.
Using a short-term air dispersion model that considers all
reasonable meteorological conditions, EPA modeled maximum one-hour
concentrations for three HAPs (HC1, HF, and acrolein). The highest
short-term exposure was 140 times below the acute reference level.
ES.8 MERCURY MULTIPATHWAY ASSESSMENT
ES.8.1 Background Discussion for Mercury
Mercury cycles in the environment as a result of natural and
human (anthropogenic) activities. The amount of mercury mobilized and
released into the biosphere has increased since the beginning of the
industrial age. Most of the mercury in the atmosphere is elemental
mercury vapor, which circulates in the atmosphere for up to a year,
and hence can be widely dispersed and transported thousands of miles
from likely sources of emission. After it deposits, mercury commonly
is emitted back to the atmosphere either as a gas or associated with
particles, to be re-deposited elsewhere. As it cycles between the
atmosphere, land, and water, mercury undergoes a series of complex
chemical and physical transformations, many of which are not
completely understood.
Mercury is a persistent element and bioaccumulates in the food
web. Mercury accumulates most efficiently in the aquatic food web.
Predatory organisms at the top of the food web generally have higher
mercury concentrations. Nearly all of the mercury that accumulates in
fish tissue is methylmercury. Inorganic mercury, which is less
efficiently absorbed and more readily eliminated from the body than
methylmercury, does not tend to bioaccumulate.
Fish consumption dominates the pathway for human and wildlife
exposure to methylmercury. The EPA's 1997 Mercury Study Report to
Congress supports a plausible link between anthropogenic releases of
mercury from industrial and combustion sources in the U.S. and
methylmercury in fish. However, these fish methylmercury
concentrations also result from existing background concentrations of
mercury (which may consist of mercury from natural sources, as well as
mercury which has been re-emitted from the oceans or soils) and
deposition from the global reservoir (which includes mercury emitted
by other countries). Given the current scientific understanding of
the environmental fate and transport of this element, it is not
possible to quantify how much of the methylmercury in fish consumed by
the U.S. population is contributed by U.S. emissions relative to other
sources of mercury (such as natural sources and re-emissions from the
global pool). As a result, it cannot be assumed that a change in
total mercury emissions will be linearly related to any resulting
change in methylmercury in fish, nor over what time period these
changes would occur. This is an area of ongoing study.
ES-15
-------
ES.8.2 Methylmercury Health Effects
Epidemics of mercury poisoning following high-dose exposures to
methylmercury in Japan and Iraq demonstrated that neurotoxicity is the
health effect of greatest concern when methylmercury exposure occurs
to the developing fetus. Dietary methylmercury is almost completely
absorbed into the blood and distributed to all tissues including the
brain; it also readily passes through the placenta to the fetus and
fetal brain. The reference dose (RfD) is an amount of methylmercury,
which when ingested daily over a lifetime is anticipated to be without
adverse health effects to humans, including sensitive subpopulations.
At the RfD or below, exposures are expected to be safe. The risk
following exposures above the RfD is uncertain, but risk increases as
exposures to methylmercury increase.
Extrapolating from the high-dose exposures that occurred in the
Iraq incident, the U.S. EPA derived a RfD for methylmercury of 0.1
microgram per kilogram body weight per day (ug/kg bw/day). While the
U.S. EPA was advised by scientific reviewers to employ this RfD for
this analysis, new data are emerging. Currently ongoing are two large
epidemiology studies in the Seychelle Islands and in the Faroe Islands
that were designed to evaluate childhood development and neurotoxicity
in relation to fetal exposures to methylmercury in fish-consuming
populations. Because of various limitations and uncertainties in all
of the available data, the U.S. EPA and other Federal agencies intend
to participate in an interagency review of the human data on
methylmercury, including the most recent studies from the Seychelle
Islands and the Faroe Islands. The purposes of this review are to
refine the estimates of the level of exposure to mercury associated
with subtle neurological endpoints and to further consensus between
all of the Federal agencies. After this process, the U.S. EPA will
determine if a change in the RfD for methylmercury is warranted.
(Note: see the 1997 EPA Mercury Study Report to Congress for further
discussion and assessment of mercury health effects and public health
impacts).
ES . 8.3 Mercury Multipathway Exposure Assessment
Mercury was considered highest priority for multipathway exposure
analysis. To assess the transport and deposition of mercury emissions
from utilities and to estimate concentrations in environmental media
and biota, three modeling efforts were undertaken: (1) long-range
modeling, (2) local scale modeling, and (3) modeling of environmental
concentrations. The RELMAP was used to predict long-range dispersion
and deposition across the U.S. For the local analysis, a model
designed to predict deposition of HAPs within 50 km, the Industrial
Source Complex Version 3 (ISC3) air dispersion model, was used. Next,
the EPA's Indirect Exposure Model Version 2M (IEM-2M) was used to
estimate mercury environmental concentrations and human exposures.
Hypothetical exposure scenarios were evaluated for four model plants
(a large coal-fired, a medium coal-fired, a small coal-fired, and a
medium oil-fired utility boiler). The analysis included three types
of plant locations: (1) rural (agricultural), (2) near lakes
ES-16
-------
(lacustrine), and (3) urban. Three human fish consumption scenarios
were considered.
The modeling provided information on whether local and/or long-
range transport of mercury is significant in a variety of scenarios.
The models indicate that most of the mercury from utilities is
transported further than 50 km from the source. The fate and
transport models provided an assessment of potential inhalation and
ingestion exposures.
ES.8.4 Summary of Mercury Assessment Results for Utilities
Recent estimates of annual total global mercury emissions from
all sources (natural and anthropogenic) are about 5,000 to 5,500 tons
per year (tpy). Of this total, about 1,000 tpy are estimated to be
natural emissions and about 2,000 tpy are estimated to be
contributions through the natural global cycle of re-emissions of
mercury associated with past anthropogenic activity. Current
anthropogenic emissions account for the remaining 2,000 tpy. Point
sources such as fuel combustion; waste incineration; industrial
processes (e.g., chlor-alkali plants); and metal ore roasting,
refining, and processing are the largest point source categories on a
world-wide basis.
For the year 1994, coal-fired utilities were estimated to emit
approximately 51 tpy of mercury in the U.S., which is estimated to be
33 percent of the 158 tpy of airborne anthropogenic emissions of
mercury in the U.S. If one assumes that current anthropogenic
activity represents between 40 and 75 percent of the total airborne
emissions (anthropogenic plus other emissions [e.g., natural
emissions]), one can calculate that U.S. utilities emit roughly 13 to
26 percent of the total (natural plus anthropogenic) airborne
emissions of mercury in the U.S.
Given the global estimates of 5,000 to 5,500 tpy (which are
highly uncertain), U.S. anthropogenic mercury emissions are estimated
to account for roughly 3 percent of the global total, and U.S.
utilities are estimated to account for roughly 1 percent of total
global emissions.
A computer simulation of long-range transport of mercury
emissions from all U.S. sources conducted for the EPA's 1997 Mercury
Study Report to Congress suggests that about one-third (~ 52 tons) of
the 158 tpy of U.S. anthropogenic emissions are deposited, through wet
and dry deposition, within the lower 48 States. The remaining two-
thirds (~ 107 tons) is transported outside of U.S. borders where it
diffuses into the global reservoir. In addition, the computer
simulation suggests that another 35 tons of mercury from the global
reservoir is deposited for a total deposition of roughly 87 tpy in the
U.S. Although this type of modeling is uncertain, the simulation
suggests that about three times as much mercury is being added to the
global reservoir from U.S. sources as is being deposited from it.
What is not uncertain is that additional emissions to air will
ES-17
-------
contribute to levels in the global reservoir and deposition to water
bodies.
Long-range transport modeling conducted as part of this Utility
Study predicts that approximately 30 percent (15 tpy) of the utility
mercury emissions deposit in the continental U.S. The estimated
annual deposition rates resulting from utility mercury emissions range
from 0.5 to greater than 10 micrograms per square meter. Long-range
transport modeling also predicts that the highest deposition occurs in
the eastern half of the U.S., particularly areas such as southeastern
Great Lakes and Ohio River Valley, central and western Pennsylvania,
large urban areas in the eastern U.S. (e.g., Washington, B.C., New
York City) and various locations in the vicinity of large coal-fired
utilities. Based on the limited available receptor monitoring data,
the RELMAP model seems to be accurate within a factor of plus or minus
2. That is, the RELMAP model seems to over- and underestimate mercury
values within a factor of two and appears to be relatively unbiased in
its predictions.
The modeling assessment in conjunction with available scientific
knowledge, supports a plausible link between anthropogenic mercury
emissions and mercury found in freshwater fish. As noted above, there
are many sources of mercury emissions worldwide, both natural and
anthropogenic. The coal-fired utilities are one category of the
mercury sources.
Mercury is considered the highest priority for multipathway
analyses because it is an environmentally persistent, toxic element.
Mercury is deposited to soil and terrestrial vegetation but at levels
that do not result in human exposures likely to be detrimental to
health through terrestrial exposure pathways. However, in its
methylated form mercury bioaccumulates in the food web (especially the
aquatic food web). Modeling results suggest that most of the mercury
emitted to the atmosphere is deposited more than 50 km away from the
source, especially sources that have tall stacks. As stated above,
the modeling assessment from the Mercury Study in conjunction with
available scientific knowledge, supports a plausible link between
anthropogenic mercury emissions and mercury found in freshwater fish.
Additional emissions to air will contribute to levels in the global
reservoir and deposition to water bodies. As a result, mercury
emissions from utility units may add to the existing environmental
burden.
At this time, the available information, on balance, indicates
that utility mercury emissions are of sufficient potential concern for
public health to merit further research and monitoring. The EPA
recognizes that there are substantial uncertainties that make it
difficult to quantify the magnitude of the risks due to utility
mercury emissions, and that further research and/or evaluation would
be needed to reduce these uncertainties. Remaining questions include
the following: (1) what is the quantitative relationship between a
change in U.S. mercury emissions and the resulting change in
ES-18
-------
methylmercury levels in fish; (2) what are the actual consumption
patterns and estimated methylmercury exposures of the subpopulations
of concern; (3) what are the actual mercury levels in a statistically
valid and representative sample of the U.S. population and susceptible
subpopulations; (4) what exposure levels are likely to result in
adverse health effects; (5) what affects the formation of
methylmercury in waterbodies and its bioaccumulation in fish; (6) how
much mercury is emitted from natural sources and past anthropogenic
sources; and (7) how much mercury is removed during coal cleaning and
other ongoing practices for pollution control. New data that could
reduce some of the uncertainties are likely to become available in the
next several years, and EPA plans to review and consider these data,
as appropriate, in future decisions.
Regarding potential methods for reducing mercury emissions, the
EPA has not identified any demonstrated add-on control technologies
currently in use in the U.S. that effectively remove mercury from
utility emissions. (However, there may be add-on control technologies
used in other source categories that effectively reduce mercury
emissions.) Based on available data, total mercury removal by
existing PM control devices on coal-fired utilities varies
considerably, ranging from 0 to 82 percent removal (with a median
efficiency of 15 percent removal) for cold-side electrostatic
precipitators (ESPs), and from 0 to 73 percent removal (with a median
efficiency of 8 percent removal) for fabric filters. Also, hot-side
ESPs exhibited no mercury control. Existing flue gas desulfurization
(FGD) units exhibit limited mercury control, ranging from 0 to 62
percent removal, with a median removal of 23 percent. The mercury
control efficiency of FGD units is a function of several factors
including temperature, plant configuration, and type of coal. Pilot-
scale studies have shown that mercury removal can be enhanced through
the use of activated carbon injection. However, the limited results
to date utilizing carbon injection are inconsistent and more data and
research are needed. Other various pollution prevention strategies,
such as coal cleaning, have shown some effectiveness in reducing
utility emissions of mercury. Conventional coal cleaning removes, on
average, approximately 21 percent of the mercury contained in the
coal. Also, fuel switching, such as switching from coal to natural
gas, would result in decreased emissions of mercury.
ES.9 SCREENING LEVEL MULTIPATHWAY ASSESSMENT FOR ARSENIC
Arsenic is a naturally occurring element found normally, in
various concentrations, in soil. In addition, arsenic can also be
naturally present in other media (e.g., various food sources and
water). Arsenic levels have been measured in a variety of foods.
Even though shellfish and other marine foods contain the greatest
concentrations of total arsenic, much of the arsenic present in fish
and shellfish exists in the less toxic organic form. Other food
products, such as meats, rice, and cereals, contain higher
percentages, and often higher total amounts, of inorganic arsenic,
which is the form of primary toxicological concern.
ES-19
-------
Arsenic is also naturally present in trace amounts in coal and
oil. When coal or oil are burned, some of this naturally occuring
arsenic is released to the atmosphere. The quantity of arsenic
released from any utility plant is dependent on many factors including
the concentration of arsenic in the fuel, control device efficiency,
and other factors.
Utilities emit about 62 tpy of arsenic nationwide, about 3 to 4
percent of the total anthropogenic arsenic emissions in the U.S.
Because of its chemical and physical characteristics, arsenic emitted
to the atmosphere may be transported to other environmental media
(soil or water), thus allowing non-inhalation exposures to occur.
ES.9.1 Exposure Modeling
It was not possible to model every utility plant for arsenic
multipathway exposures. Therefore, a screening level model plant
approach was used. Four model plants (i.e., a large coal-fired, a
medium coal-fired, a small coal-fired, and a medium oil-fired utility
boiler) were designed to characterize typical utility plants. In
taking the model plant approach, it was realized that there would be a
great deal of uncertainty surrounding the predicted fate and transport
of arsenic as well as the exposures. However, the assessment was
useful for estimating potential risks due to utility arsenic
emissions. Three models were used to predict environmental arsenic
concentrations and exposure: the RELMAP, the ISC3, and the Indirect
Exposure Model Version 2 (IEM-2). These models were used to predict
the fate and transport of arsenic emissions and to estimate human
exposures to arsenic through multiple exposure routes, including food
consumption, water ingestion, and inhalation. Three basic exposure
scenarios were considered: a subsistence farmer (adult and child), a
subsistence fisher (adult and child), and a pica child (i.e., a child
that ingests significant quantities of soil). These scenarios were
considered because they represent possible high-end scenarios for
exposure to arsenic.
ES.9.2 Health Effects of Arsenic
Inhalation exposure to inorganic arsenic has been strongly
associated with lung cancer in humans. Human exposure to inorganic
arsenic, via ingestion, has been associated with an increased risk of
several types of cancer, including skin, bladder, liver, and lung
cancers. Oral exposure to inorganic arsenic has also been associated
with noncancer effects, including effects to the central nervous
system, cardiovascular system, liver, kidney, and blood.
ES . 9.3 Approach for Estimating Screening Level Arsenic Risks
Increased cancer risks were estimated for each hypothetical
scenario, for the four model plants, each of which was placed in two
different hypothetical locations (i.e., an eastern humid site and a
dry western site). For each of the exposure scenarios, except for the
pica child, it is assumed that the hypothetical person is exposed for
30 years. For the pica child, it is assumed that exposure occurs for
ES-20
-------
7 years. Risks were estimated by multiplying the estimated intakes of
arsenic by the EPA's cancer potency factor for arsenic.
ES.9.4 Screening Level Arsenic Risk Assessment Results
The results of the screening level multipathway arsenic exposure
assessment provide an indication of the potential hazards and risks
that may occur due to emissions from a utility plant. However, the
results are not applicable to any particular plant. There are
uncertainties and limitations to the analysis.
Exposures to inorganic arsenic due to background levels and due
to emissions from the model utility boilers were predicted to be
mainly through the ingestion of grains. Exposure to inorganic arsenic
through the ingestion of fish was not predicted to be a major pathway
of exposure because there is considerable evidence that little of the
total arsenic in fish tissue is inorganic arsenic. Soil ingestion is
the major route of exposure to inorganic arsenic for the pica child.
ES.9.4.1 Arsenic Cancer Risks. The cancer risks due to
multipathway exposures to inorganic arsenic, as estimated in the model
plant analysis using hypothetical scenarios, due to utility emissions
alone (no background) were estimated to range from 4 x 10 "7 to 1 x 10 "4.
The highest estimated risk (1 x 10~4) was for a pica child assumed to
be living at the point of maximum deposition. The arsenic emissions
from the large coal-fired model utility boiler at the eastern humid
site were estimated to pose this highest risk for the pica child.
When the risk from background exposure (2 x 10 ~4) is added to the
maximum risk from utility exposure, the risk for the pica child is
estimated to be up to 3 x 10~4. The "pica child" is considered a high-
end, conservative scenario.
Background exposures were estimated to dominate the exposures and
risks in all scenarios. When considering only the arsenic emissions
from the model utility units (not including background), in all
scenarios it was the large coal-fired unit that was estimated to pose
the greatest multipathway risks and the medium coal-fired unit was
estimated to pose the next highest risks. The small coal-fired unit
and the oil-fired unit were estimated to present lower risks.
ES.9.4.2 Uncertainty Discussion. There are uncertainties
associated with the cancer risk estimates from arsenic. The analysis
was based on model plants and hypothetical constructs; therefore, the
results are not applicable for any specific utility plant. Further
analyses are needed to better characterize the risks posed by arsenic
emissions from utilities. A few uncertainties are discussed here.
Exposure to arsenic through the ingestion of tap or well water
was not included in this assessment. The exposure modeling assessment
was based on a model plant analysis, hypothetical scenarios, and
incorporated data with varying degrees of uncertainty. Also, there
are uncertainties associated with the health effects data for arsenic.
ES-21
-------
For example, the animal ingestion studies have not clearly shown an
association between arsenic ingestion exposure and cancer.
ES.10 DIOXIN SCREENING LEVEL MULTIPATHWAY ASSESSMENT
The highest MEI inhalation cancer risk due to dioxin emissions
from any utility plant based on the HEM analysis (described in section
ES.7) was estimated to be 1 x 10~7. The EPA estimates that coal-fired
utilities emit 0.2 pounds per year (Ib/yr) of dioxin (toxic
equivalents, TEQ) and that oil-fired utilities emit 0.01 Ib/yr. These
estimates combined are roughly 1 percent of the nationwide
anthropogenic dioxin emissions. However, dioxin emissions data were
only available for twelve utility plants and 42 percent of the
measurements were below the minimum detection limit. Moreover,
dioxins are not part of the naturally occuring fossil fuel. They are
formed in highly complicated reactions which may occur with unknown
frequency during combustion. Therefore, the emissions data for
dioxins from utilities, which are the basis of exposure modeling, are
considered more uncertain than the emissions data for many of the
other HAPs.
For the screening level multipathway analysis, the transport,
deposition, multipathway exposures, and human cancer risks were
assessed for utility emissions of polychlorinated dibenzo-p-dioxins
(PCDDs) and polychlorinated dibenzofurans (PCDFs), collectively
referred to as dioxins. Atmospheric deposition of dioxin emissions
can be important because dioxins tend to persist in the environment
and bioaccumulate in the food web. Environmental persistence and
bioaccumulation, coupled with carcinogenic effects at very low levels,
make multipathway exposure an important consideration for dioxins.
ES.10.1 Methods
The basic approach for estimating screening level multipathway
exposures to dioxins was similar to the methods described above for
mercury and arsenic. However, there were some differences. The EPA's
ISCST3 model was used to predict deposition and air concentrations of
dioxins within 50 km of each of four model plants. Model plants were
selected to represent both large and small coal- and oil-fired
utilities. A modified version of the IBM spreadsheet model was used
to estimate environmental concentrations, exposures to the
environmental concentrations for 16 hypothetical human scenarios, and
the resulting cancer risks. Pathways assessed include inhalation,
dermal contact with soil, and ingestion of water, soil, fish, plants,
and animals.
ES.10.2 Results
Since the analysis was based on model plants, using hypothetical
scenarios, the results are not applicable to any specific plant and
contain substantial uncertainties about the risks due to dioxin
emissions. Total modeled screening level lifetime cancer risks
related to multipathway exposure to dioxins for the four-model plant
analysis ranged from 1 x 10"10 to 2 x 10"4. The results of this
ES-22
-------
analysis indicate that the exposures and risks due to fish consumption
are the highest of all pathways considered. The highest modeled
result of 2 x 10"4 lifetime cancer risk was obtained for the subsistence
fisher exposure scenario. In all modeled scenarios, the non-
inhalation exposures were at least one order of magnitude larger than
the inhalation exposures, thus demonstrating the potential
significance of including multipathway exposure analysis in the risk
assessments for pollutants that are environmentally persistent and
tend to bioaccumulate. Also, unlike the results for arsenic, modeled
exposures to dioxins for each pathway exceed the background exposure
estimates for dioxins.
ES.10.3 Uncertainty Discussion
Several sensitivity analyses were completed for the screening
level multipathway assessment of utility dioxin risks to assess the
reasonableness of the results. The assumptions with the greatest
impact on the predicted risk to the subsistence fisher were those made
about the biota-sediment accumulation factor. This sensitivity
analysis suggests that the modeling results are reasonable for a
screening level analysis.
ES.ll MULTIPATHWAY ASSESSMENT FOR RADIONUCLIDES
Radionuclide emissions from utilities may result in human
exposure from multiple pathways including: (1) external radiation
exposure from radionuclides suspended in air or deposited on the
ground, and (2) internal exposure from the inhalation of airborne
contaminants or ingestion of contaminated food. The CAP-93 model was
used to estimate multipathway exposures and risks due to radionuclide
emissions to humans within 50 km of all 684 utilities. However, this
assessment did not use site-specific data for the non-inhalation
exposure analysis, but rather relied on various generic assumptions
and general input data.
Based on the CAP-93 modeling, 667 of the 684 plants are estimated
to pose multipathway risks less than 1 x 10"5. The highest estimated
multipathway radiation exposure for the MEI due to radionuclide
emissions from utilities was predicted to be 1.5 millirems (mRems) per
year, which is estimated to pose an increased cancer risk of 3 x 10 "5.
Seventeen plants (13 coal- and 4 oil-fired plants) were estimated to
pose multipathway risks between 1 x 10 "5 and 3 x 10"5. The estimated
cancer incidence in the U.S., due to emissions and dispersion of
radionuclides within 50 km of each utility, is estimated to be 0.3
cancer deaths/yr. The cancer incidence appears to be mostly due to
inhalation exposure. The non-inhalation exposures contribute only
slightly to the incidence. The non-inhalation exposure pathways have
a greater impact on the MEIs, especially for coal-fired plants.
The risks due to exposure to radionuclides from utilities are
substantially lower than the risks due to natural background
radiation. The average exposure to natural background radiation
(excluding radon) for the U.S. population has been estimated to be
ES-23
-------
roughly about 100 mRems per year, which is about 67 times higher than
the highest exposure due to utility radionuclide emissions.
ES.12 QUALITATIVE MULTIPATHWAY EXPOSURE ASSESSMENT
The EPA recognizes that non-inhalation exposure pathways could be
important for additional HAPs that are persistent and tend to
bioaccumulate. A few additional HAPs that were not modeled for
multipathway exposures are discussed below.
ES.12.1 Cadmium and Lead
Cadmium emissions from the vast majority of plants (683 of the
684 plants) are estimated to pose inhalation risks less than 10 ~6, and
the highest modeled air concentration of lead was 200 times below the
national ambient air quality standard (NAAQS). Cadmium and lead are
persistent, may bioaccumulate, and are toxic by ingestion. However,
since the emission quantities and inhalation risks are relatively low,
the EPA does not plan to conduct future evaluations of multipathway
exposures of cadmium and lead from utilities.
ES.12.2 Nickel and Chromium
Nickel and chromium were not considered to be priority for non-
inhalation exposures. At relatively high oral doses, nickel and
chromium do cause noncancer toxicity. However, there are considerable
uncertainties about the noncancer toxicity of nickel and chromium at
relatively low ingestion doses (below the toxic threshold). Also, it
is uncertain whether they pose a carcinogenic risk by ingestion.
Hence, EPA does not plan to assess multipathway exposures for nickel
and chromium for utilities.
ES.13 POTENTIAL IMPACTS OF HYDROGEN CHLORIDE AND FLUORIDE
No exceedances of the health benchmarks (e.g., RfCs) for HC1 or
HF were identified in the inhalation exposure assessment. However,
emissions of HCL and HF may contribute to acid deposition and, to a
lesser extent to PM fine and visibility problems. To the extent that
these emissions may contribute to such problems, they could be
addressed through other Titles of the Act.
ES.14 ALTERNATIVE CONTROL AND PREVENTION STRATEGIES
There are numerous potential alternative control strategies for
reducing HAPs. These include precombustion controls (e.g., fuel
switching, coal switching, coal cleaning, coal gasification),
combustion controls, post combustion controls (e.g., PM controls, S02
controls), and approaches that prevent pollution by improving
efficiency in supply (e.g., promoting energy efficiency in combustion)
or demand (e.g., demand side management [DSM] , pollution prevention,
energy conservation). The degree of feasibility, cost, and
effectiveness of each of these potential control technologies varies.
For example, coal cleaning tends to remove at least some of all the
trace metals, with lead concentrations being removed to the greatest
ES-24
-------
extent (averaging approximately 55 percent removal) and mercury being
removed the least (averaging approximately 21 percent). Existing PM
controls tend to effectively remove the trace metals (with the
exception of mercury) while FGD units remove trace metals less
effectively and exhibit more variability. Fuel switching (e.g.,
switching from coal to natural gas) could result in substantial
reductions in HAP emissions. There are few existing data that show
the HAP reduction effectiveness of DSM, pollution prevention, and
energy conservation. These control strategies need to be examined
further for technical and economic considerations.
ES.15 OTHER ISSUES AND FINDINGS
ES.15.1 Emissions and Risks for the Year 2010
In addition to the 1990 analysis, the EPA also estimated
emissions and inhalation risks for the year 2010. There are
substantial data gaps and uncertainties in the projections to the year
2010. However, the approach utilized is reasonable given the
limitations of data to complete such projections.
Based on EPA's assessment for this report, HAP emissions from
coal-fired utilities are predicted to increase by 10 to 30 percent by
the year 2010. Predicted changes that were included in the 2010
emissions projections include the installation of scrubbers for a
small number of facilities, the closing of a few facilities, and an
increase in fuel consumption of other facilities. However, based on
EPA's exposure modeling analysis for the year 2010, the inhalation
risks in 2010 for coal-fired utilities are estimated to be roughly
equivalent to the 1990 inhalation risks. For oil-fired plants,
emissions and inhalation risks are estimated to decrease by 30 to 50
percent by the year 2010. Multipathway risks for 2010 were not
assessed. Utilization of add-on controls to comply with the acid rain
program are not expected to significantly impact on HAP emissions due
to their limited numbers and limited HAP control efficiency
improvement. However, if additional actions are taken to reduce
emissions of criteria pollutants, acid rain precursors, or global
warming compounds (e.g., use of fuel switching or add-on controls to
reduce SOX, NOX, and/or carbon dioxide emissions), these actions could
result in reductions in HAP emissions. For example, analyses
performed to assess compliance with the revised NAAQS for ozone and PM
indicate that mercury emissions in 2010 may be reduced by
approximately 16 percent (11 tpy) over those projected in this report.
Other potential (but unknown) actions (e.g., repowering,
restructuring) may have a significant impact on HAP emissions;
however, these unknowns were not included in the 2010 projection.
ES.15.2 Peer Review
Draft versions of Chapters 1 through 9 and 13 of this report and
draft technical support documents were reviewed by many non-EPA
scientists representing industry, environmental groups, academia, and
other parties. Chapters 10, 11, and 12 are new chapters produced in
response to major comments from the reviewers. EPA held a scientific
ES-25
-------
peer review meeting and also a public meeting in July 1995 to obtain
comments from reviewers. In February, April, and September 1996, all
sections of the draft report underwent additional review by EPA, State
and local Agencies, and other Federal Agencies. Additional review
occurred during 1997. The EPA has revised the report, as appropriate,
based on the reviewers' comments. However, there were several
comments that could not be fully addressed because of limitations in
data, methods, and resources. In addition, there were some comments
that EPA did not agree with. Also, the new chapters (10 to 12) have
only undergone a limited review. Draft versions of this report, along
with all the comments received, have been submitted to the public
docket (A-92-55) at the following address: U.S. EPA, Air and Radiation
Docket and Information Center, mail code 6102, 401 M Street, S.W.,
Washington, D.C. 20460; telephone number (202) 260-7548. Materials
are available for public review at the docket center or copies may be
mailed (for a fee) on request by calling the above number.
ES.15.3 Industry Report
If alternative methods and assumptions were used to study the HAP
emissions from utilities, the results would likely be somewhat
different. To assess the impact of using alternative assumptions and
methods, it is useful to compare the EPA study with a similar study
completed by the EPRI.
The EPRI prepared a report, entitled "Electric Utility Trace
Substances Synthesis Report," (November 1994) that paralleled the
EPA's study. Many of the same emissions data were used and similar
risk assessment methods were utilized. The EPRI study concluded that
cancer inhalation risks are below 1 x 1Q~6 for all utilities, and
noncancer inhalation risks are well below Federal threshold levels for
all utilities. Population inhalation risks were determined by the
EPRI to be insignificant (less than 0.1 cancer case/year). Case
studies at four plants found that multimedia risks, including mercury,
are below levels of concern.
The EPRI's risk estimates are generally similar to, but in
several cases lower than, those of EPA. Differences between the
studies include: (1) EPA's use of a higher unit risk factor for
arsenic; (2) EPA's assumption that nickel was carcinogenic (EPRI
assumed nickel was not carcinogenic); (3) EPA's evaluation of exposure
beyond 50 km to all locations in the U.S. (EPRI did not attempt this
analysis); (4) EPRI's radionuclide analysis was based on several model
plants, while the EPA evaluated every plant in the U.S.; and (5) the
EPRI assumed that chromium emissions were five percent chromium VI
(the carcinogenic form), while EPA assumed that 11 percent (for coal-
fired plants) and 18 percent (for oil-fired plants) were chromium VI.
In addition, the EPRI mercury multimedia study considered only the
local impact from four plants (not worst-case) and did not include
potential impacts of total nationwide utility mercury emissions and
contributions to total environmental loadings.
ES-26
-------
ES.15.4 Potential Environmental Impacts Not Included in Study
There are other potential environmental issues associated with
utilities not assessed in this report. These include: (1) the
impacts of criteria pollutants (S02, NOX, PM, carbon monoxide, and
ozone) or acid rain precursors (S02 and NOX) , which are studied and
regulated under other sections of the Act; (2)an assessment of
ecological impacts of HAPs; (3) the impacts of carbon dioxide
emissions and climate; and (4) the impacts resulting from
restructuring, mining, drilling, solid waste disposal, transmission,
transportation, or other activities associated with electric power
generation. These issues and potential impacts were not assessed
because they were considered beyond the scope of this study as
mandated by Section 112(n) of the Act.
ES.15.5 Link to Particulate Matter
Arsenic, cadmium, chromium, lead, nickel, radionuclides, and
several other HAPs are emitted primarily as PM. Consequently, these
HAPs may contribute to PM emissions and PM health concerns, especially
from poorly controlled coal-fired units and uncontrolled oil-fired
units (about two-thirds of oil-fired units are uncontrolled for PM).
Impacts for PM were not addressed in this study, but are being studied
under Title I of the Act. If additional controls of PM emissions are
utilized, this could result in reductions in HAP emissions.
ES.16 OVERALL TECHNICAL SUMMARY AND CONCLUSIONS
Based on available information and current analyses, the EPA
believes that mercury from coal-fired utilities is the HAP of greatest
potential concern and merits additional research and monitoring.
There are uncertainties regarding the extent of risks due to mercury
exposures including those from utility emissions. Further research
and evaluation are needed to gain a better understanding of the risks
and impacts of utility mercury emissions. In addition, further
research and evaluation of potential control technologies and
strategies for mercury are needed.
For a few other HAPs, there also are still some remaining
potential concerns and uncertainties that may need further study.
First, the screening multipathway assessments for dioxins and arsenic
suggest that these two HAPs are of potential concern (primarily from
coal-fired plants); however, further evaluations and review are needed
to better characterize the impacts of dioxins and arsenic emissions
from utilities. Second, nickel emissions from oil-fired utilities are
of potential concern, but significant uncertainties still exist with
regards to the nickel forms emitted from utilities and the health
effects of those various forms. The impacts due to HAP emissions from
gas-fired utilities are negligible based on the results of this study;
therefore, the EPA feels that there is no need for further evaluation
of the risks of HAP emissions from natural gas-fired utilities.
ES-27
-------
ES.17 AREAS FOR FURTHER RESEARCH AND ANALYSIS
There are many uncertainties and data gaps described throughout
this report. This section summarizes several important areas in which
further research or scientific work may be needed.
ES.17.1 Emissions Data for Dioxins
Emissions data for dioxin compounds were available from less than
12 utility plants. Many of the measurements were near the detection
limits. Therefore, there are greater uncertainties with the dioxin
emissions than for the other HAPs. Research may be needed to gain a
better understanding of the dioxin emissions from utilities and the
dioxin formation, if any, in various utility boiler types (e.g., units
with cold-side or hot-side ESPs).
ES.17.2 Speciation of Nickel
There are significant uncertainties regarding the forms of nickel
emitted from oil-fired utilities and their associated health effects.
Research would be useful to determine the emissions quantities of
various nickel forms and the health effects of various nickel forms.
ES.1.7.3 Multipathway Risk Assessment
Further work may be needed to better characterize the risks due
to multipathway exposure to certain HAPs (e.g., arsenic and dioxins).
ES.17.4 Local, Regional, and Long-range Transport Exposures
Further modeling and evaluation may be needed to better
characterize the impacts of local, regional, and long-range transport
of HAPs from utilities.
ES.17.5 Mercury
There are numerous areas regarding mercury that may need further
research, study, or evaluation. A few potential areas for further
study include the following:
(1) additional data on mercury content of various types of coal;
(2) improved methods for measuring mercury levels in water;
(3) the impact of reducing mercury emissions from coal-fired
facilities on the bioaccumulation of mercury in fish;
(4) statistically valid and reliable estimates of methylmercury
exposure levels in the U.S. population and susceptible
subpopulations, as measured in human hair;
(5) the occupational, dietary and behavioral factors that affect
mercury exposures for people who are determined to be
exposed above a threshold of concern;
(6) the human health and environmental benefits that would be
expected by reducing mercury emissions from U.S. utilities;
ES-28
-------
(7) control technologies or pollution prevention options that
are available, or will be available, that could potentially
reduce mercury emissions and what are the costs of those
options;
(8) how do other regulations, programs and activities (e.g.,
acid rain program, electricity restructuring, NAAQSs, and
climate change) affect mercury emissions; and
(9) additional data on mercury emissions (e.g., how much is
emitted from various types of units, how much is divalent vs
elemental mercury, and how do factors such as control device,
fuel type, and plant configuration affect emissions and
speciation).
Several additional uncertainties and potential areas for further
research on mercury are discussed in other sections of this report.
ES.17.6 Projections to the Year 2010
There are significant uncertainties and unknowns in the emissions
and risk projections made to the year 2010 (e.g., impact of
electricity restructuring; impact of State efforts to regulate such
restructuring; impact of any climate change abatement initiatives).
Research and evaluation in these areas may be needed.
ES.17.7 Ecological Risks
The effects of HAPs on wildlife, endangered species, and
terrestrial and aquatic ecosystems were not evaluated in this study.
Although not mandated by section 112(n)(1)(A), further evaluation of
ecological risks due to HAP emissions would be needed to fully
evaluate the impacts of utility HAP emissions.
ES.17.8 Criteria Pollutant and Acid Rain Programs
Further evaluation is needed to assess the impacts of the Acid
Rain and Criteria Pollutant programs (e.g., impact of revisions to the
PM-fine and ozone NAAQS; impact of Ozone Transport Assessment Group
[OTAG] activities) on HAP emissions, especially for mercury.
ES-29
-------
1.0 INTRODUCTION
This chapter presents an introduction to the study of hazardous
air pollutant (HAP) emissions from electric utility steam-generating
units (i.e., utilities). The chapter is divided into three main
sections: the legislative mandate that requires this report, the
provisions of the 1990 amendments to the Clean Air Act (CAA or the
Act) related to this study, and an overview of the utility study and
its approach to meeting the provisions of the Act.
1.1 LEGISLATIVE MANDATE
In section 112(n)(1)(A) of the Act, Congress directs the U.S.
Environmental Protection Agency (EPA) to:
"... perform a study of the hazards to public health
reasonably anticipated to occur as a result of emissions by
electric utility steam generating units of ... [HAPs] ...
after imposition of the requirements of this Act."
Section 112(a)(8) of the Act defines an "electric utility
steam-generating unit" as "any fossil-fuel-fired combustion unit of
more than 25 megawatts electric (MWe) that serves a generator that
produces electricity for sale." A unit that cogenerates steam and
electricity and supplies more than one-third of its potential electric
output capacity and more than 25 MWe output to any utility power
distribution system for sale is also considered an electric utility
steam-generating unit.
Section 112(n)(1)(A) also requires that:
• Results of this study be presented in a report to Congress
by November 1993;
• The EPA develop and describe alternative control strategies
for HAPs that may warrant regulation under section 112; and
• The EPA proceed with rulemaking activities under section 112
to control HAP emissions from utilities if it determines
from the study that such regulation is appropriate and
necessary.
Section 112(n)(1)(A) does not include a requirement to analyze
the cost(s) of alternative control strategies in the study.
Therefore, no cost analyses (e.g., control costs, economic, cost-
benefit) have been performed as a part of this study. These analyses
would be conducted as part of the rulemaking process should EPA
determine that regulations are appropriate and necessary.
The EPA began work in 1991 to develop and collect the information
and data needed to prepare this study of HAP emissions from electric
utilities. At that time, only a small amount of reliable data on HAP
1-1
-------
emissions from utilities were available. In October 1996, the Agency
published a three-volume report, Study of Hazardous Air Pollutant
Emissions from Electric Utility Steam Generating Units--Interim Final
Report. (EPA-453/R-96-013). This final report incorporates
additional analyses and includes more recent data on emissions,
control technologies, and health effects.
This report discusses the possible impact of pollution controls
required by other Federal regulations or sections of the Act,
estimates which HAPs are present in utility unit emissions, and
estimates exposures and risk to humans from the emission of these
HAPs.
1.2 CAA PROVISIONS AND STUDIES RELATED TO THIS STUDY
The CAA contains several provisions relating to electric
utilities that will impact the industry well into the future.
Environmental regulations implementing many of these requirements are
now in effect; others have been established since the date of the last
report; and others are under development.
This section summarizes the major provisions of the Act affecting
electric utilities and their relevance to this study. These include
nonattainment provisions, acid deposition control programs, and new
source performance standards (NSPS) discussed in sections 1.2.1
through 1.2.3. The development of regulations for HAP under section
112 of the Act and other related studies required by section 112 are
discussed in sections 1.2.4 and 1.2.5, respectively.
1.2.1 Nonattainment Provisions
Title I of the Act includes requirements for attaining and
maintaining the national ambient air quality standards (NAAQS).
Section 108 of the Act directs EPA to identify certain pollutants
which may reasonably be anticipated to endanger public health and
welfare. Section 109 directs the Administrator to establish primary
and secondary NAAQS for the identified pollutants. Under section 110
of the Act and related provisions, States are primarily responsible
for ensuring attainment and maintenance of the ambient standards. The
EPA has established NAAQS for six criteria pollutants: ozone (03) ,
carbon monoxide (CO), particulate matter (PM), lead, sulfur dioxide
(S02) , and nitrogen oxides (NOX) under Title 40, Part 50 of the Code of
Federal Regulations (40 CFR Part 50). Electric utilities are
significant emitters of S02 and NOX; NOX emissions from electric
utilities account for about one-third of nationwide emissions.1
Electric utilities also emit other criteria pollutants such as PM as
well as air toxics.
The EPA issued revised NAAQS for 03 and PM on July 18, 1997
(Federal Register, volume 62, page 38856 [62 FR 38856]). The new
rules strengthened the primary standard for 03, added standards for PM
less than 2.5 microns in size (PM2 5) to supplement the PM10 primary
standard, and revised secondary standards.2 As part of this
1-2
-------
rulemaking, the EPA also proposed rules requiring States to develop
programs to reduce regional haze.
To achieve the new standards, EPA has developed an integrated
strategy that will require reductions in NOX and S02 as well as
volatile organic compounds (VOC) and PM. Nitrogen oxide is a
precursor to the formation of ground-level 03, and S02 is a precursor
to the formation of PM in the atmosphere. Electric utilities will be
affected as States reduce emissions to meet the new standards. One
EPA study predicts that nationwide NOX reductions ranging from 25 to 90
percent, depending on the particular State or non-attainment area,
will be needed to attain the revised 03 ambient standard.3
The Regulatory Impact Analysis for the revised standards assumes
that much of the needed emission reductions would be achieved through
the Acid Deposition Program for S02 and NOX discussed in section 1.2.2
and through the revised NSPS discussed in section 1.2.3. According to
the analysis, S02 and NOX emissions from utilities will be reduced by
approximately 40 and 50 percent, respectively, by the year 2010.4
These analyses also estimate a 16 percent reduction (approximately 11
tons per year) in utility mercury emissions (in 2010) as a result of
compliance with the revised NAAQS (primarily related to the impact of
the S02 strategy to meet the PM NAAQS).5
In a related action, EPA proposed rules requiring 22 States and
the District of Columbia to submit State Implementation Plans (SIPs)
that address the regional transport of ground-level 03. The proposed
rule would decrease ozone transport in the eastern half of the United
States by reducing NOX emissions. Under the proposed rule, States may
reduce emissions from sources they choose, although utility and large
nonutility point sources are expected to be affected. Implementation
of the proposed rule would reduce total emissions of NOX by 35 percent
based on analyses by EPA and the Ozone Transport Assessment Group
(OTAG). The EPA estimates that this action will bring areas into
attainment with the revised 03 standard without additional local
controls. Many of these States are expected to reduce NOX emissions by
participating in the cap-and-trade program discussed in section I.2.2.6
1.2.2 Acid Deposition Control
Title IV of the Act sets as its primary goal the reduction of
annual S02 emissions by 10 million tons below 1980 levels. To achieve
these reductions, the law requires a two-phase tightening of the
restrictions placed on fossil-fuel power plants (i.e., utilities).
Phase I of EPA's S02 Program (40 CFR Parts 72 through 75) began in 1995
and affects 263 units at 110 mostly coal-burning electric utility
plants in 21 States. An additional 182 units joined Phase I as
substitution or compensating units, bringing the total of Phase I
units to 445. Emissions data indicate that 1995 S02 emissions at these
units nationwide were reduced by almost 40 percent below their
required level of 8.7 million tons.7 The second phase begins in the
year 2000 and covers an additional 1,600 boilers. The EPA believes
the 10 million ton goal will be met before the year 2010. To reduce
1-3
-------
S02 emissions, an affected source may: (1) install flue gas scrubbers,
(2) switch to a fuel that contains less sulfur, or (3) purchase
emission allowances. The control option a utility selects to comply
with the S02 reduction requirements may also have an effect on HAP
emissions.
Under the S02 program, affected units are allocated allowances
based on their historic fuel consumption and a specific emissions
rate. Each allowance permits a unit to emit one ton of S02 per year.
For each ton discharged, one allowance is retired. Allowances may be
bought, sold, or banked and are tracked through a computerized system.
However, no source can emit at a level violating Federal or State
limits set under Title I of the Act. Sources also must obtain a
permit and meet continuous emission monitoring requirements for S02,
NOX, and carbon dioxide (C02), as well as volumetric flow and opacity
monitoring requirements.
Section 407 of the Act establishes the NOX Emission Reduction
Program with the goal of reducing emissions by 2 million tons from
1980 levels. Like the S02 emission reduction program, the NOX program
is implemented in two phases beginning in 1996 and 2000. Under Phase
I rules (40 CFR Part 76), approximately 277 dry-bottom wall-fired
boilers and tangentially-fired boilers (Group I) must meet applicable
annual average emission rates of 0.45 pound per million British
thermal units (lb/MMBtu) and 0.50 lb/MMBtu, respectively, by January 1,
1996. Utilities can meet the limits by installing low-NOx burner
technology or other combustion control technology or by averaging
emissions among several units. An affected unit also may obtain an
alternative emission limit under specified conditions. Implementation
of Phase I will decrease annual NOX emissions by over 400,000 tons per
year (tpy) between 1996 and 1999 (60 FR 18751, April 13, 1995) . 8'9
The EPA issued final rules implementing Phase II of the program
in late 1996 (61 FR 67112, December 19, 1996). In these rules, EPA
determined that more effective low NOX burner technology is available
to establish more stringent standards for Phase II, Group I boilers
than those established for Phase I. Emission limits for Group II
boilers (wet bottom, cyclones, cell burners, and vertically-fired
boilers) were also established based on NOX control technologies
comparable in cost to low NOX burners (selective catalytic reduction
[SCR]). Selective catalytic reduction is a commercially available
flue gas treatment technology that injects ammonia into the flue gas
in the presence of a catalyst. The catalyst promotes reactions that
convert NOX to nitrogen and water. By the year 2000, the Phase II rule
(affecting 775 units) will achieve an additional reduction of 1.17
million tons of NOX per year. Phase I and Phase II together are
estimated to decrease nationwide annual NOX emissions by 2.06 million
tpy beginning in the year 2000.10
The final Phase II rule includes an option allowing a State or
group of States to petition EPA to accept an emissions cap-and-trade
program as a substitute for compliance with the Group 2 limits and
1-4
-------
additional reductions required for Group 1 boilers. The petition may
be granted if the Administrator finds that alternative compliance
through the cap-and-trade program will achieve lower total NOX
emissions from Group 1 and Group 2 boilers than if the new limits were
applicable. The Phase I limits established in 1995 would apply to
Group 1 boilers in a cap-and-trade program. This provision is
expected to affect boilers located in the OTAG region which contains
about 87 percent of the units covered by the Phase II rule."
In related developments, the EPA is currently developing a model
cap-and-trade program to facilitate NOX emission reductions from large
stationary sources choosing to participate. The Agency intends to
propose the rule in early 1998 and finalize the action in conjunction
with the ozone transport rulemaking in September 1998. 1:L
1.2.3 New Source Performance Standards
Section 111 of the Act requires the development of NSPS for newly
constructed or modified affected facilities. Section 403 of the Act,
as amended, revised the definition of the term "standard of
performance" to mean:
"...a standard for emissions of air pollutants which reflects the
degree of emission reduction achievable through the application
of the best system of emission reduction which (taking into the
cost of achieving such reduction and any nonair quality health
and environmental impact and energy requirements) the
Administrator determines has been adequately demonstrated."
New source performance standards currently provide the major
regulatory authority for the control of air emissions from utilities.
Fossil-fuel-fired steam generating units greater than 73 MW heat input
that were constructed or modified after August 17, 1971, are subject
to requirements of 40 CFR Part 60, Subpart D; units constructed or
modified after September 18, 1987, are subject to 40 CFR Part 60,
Subpart Da. These rules define "fossil fuel" as "natural gas,
petroleum, coal, and any form of solid, liquid, or gaseous fuel
derived from such material for the purpose of creating useful heat."
Fossil fuels include coal (bituminous, subbituminous, anthracite,
lignite), oil (Nos. 2, 4, and 6), and natural gas. Subparts D and Da
include limits for emissions of S02, NOX, and PM based primarily on the
use of scrubbers or low sulfur coal, combustion modification
techniques (overfire air, low excess air, and reduced heat release
rate), and PM control devices. Provisions also are included for the
use of continuous opacity monitoring systems and continuous emission
monitoring systems for S02 and NOX and oxygen (02) or C02.
Section 407 of the Act requires EPA to revise the NSPS for NOX
emissions from utility and nonutility units to reflect improvements in
emission reduction methods. The EPA proposed revisions to the NOX
limit for utility units (i.e., boilers) in 40 CFR Part 60, Subpart Da
(regardless of fuel type) based on coal-firing and the performance of
SCR control technology, in combination with combustion controls (62 FR
1-5
-------
36947; July 9, 1997) . Thus, units can meet the proposed standards by
using clean fuels such as natural gas or by installing more effective
control systems. The proposed rule also revised the emission limit to
incorporate an output-based format that will encourage unit operating
efficiency and pollution prevention. The EPA estimates that about
43,600 tons of NOX per year would be emitted from 17 new utility
boilers expected to be constructed over the next 5 years. The
proposed revised standards would reduce these emissions by about
25,800 tpy.12
The NSPS program results indirectly in the control of some HAPs.
For example, NSPS that limit emissions of PM will also control HAPs
that are PM or that condense onto the PM in the affected gas streams.
Furthermore, the use of S02 scrubbers (currently on about 14 percent of
the units) will also control some vapor-phase HAPs, such as hydrogen
chloride (HC1) and hydrogen fluoride (HF), in addition to providing
some control of mercury.
1.2.4 Hazardous Air Pollutants
Section 112(d) of the Act requires that EPA promulgate
regulations for the control of HAPs listed in section 112 (b) of the
Act from both new and existing major sources. A "major" source means
a source that:
"... emits or has the potential to emit, considering
controls, 10 tons per year or more of any HAP or 25 tons per
year or more of any combination of HAPs."
Regulations developed under section 112(d) must reflect the
maximum degree of reduction in emissions of HAP that is achievable,
taking into consideration the cost of achieving the emissions
reduction, and any non-air quality health and environmental reduction
and energy requirements. This level of control is commonly known as
the maximum achievable control technology (MACT). For new sources,
MACT standards cannot be less stringent than the emission control that
is achieved in practice by the best-controlled similar source. The
MACT standards for existing sources cannot be less stringent than the
average emission limitation achieved by the best-performing 12 percent
of existing sources for categories and subcategories with 30 or more
sources, or the best-performing 5 sources for categories or
subcategories with fewer than 30 sources. Section 112(d) also
provides that the Administrator may distinguish among classes, types,
and sizes of sources within a source category when establishing
standards. Regulations for the control of HAP emissions from
utilities will be developed under this authority if such regulations
are determined to be necessary and appropriate.
1.2.5 Other Studies
The 1990 amendments to section 112 of the Act also mandate five
other related studies: (1) the mercury study, (2) the National
Institute of Environmental Health Sciences (NIEHS) health effects of
mercury study, (3) the National Academy of Sciences (NAS) risk
1-6
-------
assessment methodologies study, (4) the Great Waters study, and
(5) the Presidential Risk Commission.
1.2.5.1 Mercury Study. Section 112(n)(1)(B) requires the EPA to
complete a study of mercury emissions from utilities, municipal waste
combustion units, and other sources, including area sources, by
November 15, 1994. The study is to consider the rate and mass of
mercury emissions, the health and environmental effects of such
emissions, technologies that are available to control such emissions,
and the costs of such technologies. The EPA just recently (December
19, 1997) published the Final Mercury Study Report to Congress. The
3-year delay was necessary to allow sufficient time for data
gathering, analyses, writing, and extensive peer review. The Mercury
Study is closely related to this Utility Study because utilities are
the largest anthropogenic source of mercury emissions. Utilities
(primarily coal-fired utilities) are estimated to emit approximately
33 percent of the airborne anthropogenic mercury in the United States.
Several analyses and conclusions contained in the Mercury Study are
applicable to utilities, and are discussed in Chapter 7 of this
report.
1.2.5.2 NIEHS Health Effects of Mercury Study. Under section
112(n)(1)(c), the NIEHS is required to perform a study identifying the
threshold level of mercury exposure that would not adversely affect
human health. A report on the NIEHS study was published in 1993. 13
1.2.5.3 NAS Risk Assessment Methodologies Study. In January
1995, the NAS finalized a report14 on the risk assessment methodologies
used by the EPA. The results of the NAS study were consulted to help
develop the methodologies for the risk assessment portions of this
study.
1.2.5.4 The Great Waters Study. In response to section 112(m),
the EPA finalized a report in May 1994 on the atmospheric deposition
of pollutants to the "Great Waters," namely, the Great Lakes,
Chesapeake Bay, Lake Champlain, and coastal waters.15 The pollutants
of concern to the Great Waters study that are emitted from utilities
include lead, cadmium, dioxins, and, in particular, mercury. The
report discussed the following:
• The contribution of atmospheric deposition to pollutant
loadings in these waters
• Environmental and public health effects of atmospheric
pollution that is deposited to these waters
• Sources of pollutants deposited to these waters.
The May 1994 report noted that the Great Waters are polluted by
HAPs that originate from local and distant sources; however, more data
are needed to identify sources of the pollutants. The recommendations
of the May 1994 Great Waters report were: (1) the EPA should strive
1-7
-------
to reduce emissions of the pollutants of concern through
implementation of the Act; (2) a comprehensive approach should be
taken, both within the EPA and with other agencies, to reduce and
preferably prevent pollution in air, water, and soil; and (3) the EPA
should continue to support research for emissions inventories, risk
assessment, and regulatory benefits assessment.
Following the first Report to Congress, the EPA published the
"Final Water Quality Guidance for the Great Lakes System" required by
section 118(c) (2) of the Clean Water Act (60 FR 15366, March 23,
1995). This guidance document established minimum water quality
criteria, methodologies, policies, and procedures for the Great Lakes
System. States and Tribes in the Great Lakes Basin were required to
incorporate these provisions into their water quality standards and
National Permit Discharge Elimination System permit programs by March
1997. In the guidance, EPA recognized that non-point sources of
mercury, particularly by air deposition, are the most significant
remaining contributors of mercury to the Great Lakes System. The EPA
followed the guidance with the "Final Water Quality Guidance for the
Great Lakes System Draft Mercury Permitting Strategy," released for
public comment in June 1997 (62 FR 31025) . The final permitting
strategy will be finalized in the near future.
The second report to Congress on the atmospheric deposition of
pollutants to the Great Waters was completed in June 1997. The report
confirmed, and provided additional support for, the findings of the
first Report to Congress that persistent and bioaccumulative toxic
pollutants and excessive nitrogen can adversely affect the
environmental condition of the Great Waters. Electric utilities and
mobile sources are identified, in modeling studies and emission data,
as major contributors of NOX to the Chesapeake Bay and its watershed. 1S
1.2.5.5 Presidential Risk Commission. In section 303 of Title
III of the 1990 amendments to the Act, Congress directed that the
President form a Commission whose mandate would be to "...make a full
investigation of the policy implications and appropriate uses of risk
assessment and risk management in regulatory programs under various
Federal laws to prevent cancer and other chronic human health effects
which may result from exposure to hazardous substances." This
Commission has issued the report in two volumes. Volume 1 entitled,
"Framework for Environmental Health Risk Assessment," was issued in
February 1997. Volume 2 entitled, "Risk Assessment and Risk
Management in Regulatory Decision-Making," was issued in April 1997.
1.3 OVERVIEW AND APPROACH OF ELECTRIC UTILITY HAP STUDY
This report is the result of the work of government and
nongovernment personnel. Emissions testing and emission estimation
issues were discussed among numerous branches within the EPA and among
representatives of industry, the Electric Power Research Institute
(EPRI), and the Department of Energy (DOE). In particular, EPRI, DOE,
and the EPA coordinated their utility emissions testing to cover more
1-8
-------
plant configurations and obtain as much information as possible for
the assessment. Portions of this report, and the data and
methodologies utilized, were reviewed by numerous scientific experts
within and outside the Agency. Outside reviewers included
representatives from industry, other Federal agencies, State and local
agencies, academia, and environmental organizations.3
The report is organized as follows. The electric utility
industry is described in Chapter 2, including the types of fossil
fuels, boilers, and air pollution control devices in use in the year
1990, as well as changes in control devices and fuel usage expected
for the year 2010. Chapter 3 describes emissions testing conducted
since 1990, the determination of emission modification factors (EMFs)
from test reports, and the estimation of emissions for several
characteristic units using a computer emission factor program.
Chapter 4 introduces the health hazard assessment. The screening risk
assessment used to determine the priority HAPs is described in
Chapter 5. Chapter 6 discusses the inhalation route for HAP exposure,
while Chapters 7 through 11 address multipathway exposures to mercury,
lead, cadmium, radionuclides, arsenic, and dioxins. Chapter 12
discusses the potential impacts of HC1 and HF. Alternative control
strategies for HAP emissions reductions are given in Chapter 13.
Chapter 14 presents the conclusions of the study. Additional
supporting material is provided in the appendices.
Reviewers provided comment through a variety of venues (e.g., EPA Work
Group, scientific peer review, Federal interagency review, public
comment period). However, participation by a reviewer did not imply
agreement with the methodology or conclusions presented by the EPA. All
comments were considered during revision of the document.
1-9
-------
1.4 REFERENCES
1. U.S. Environmental Protection Agency. Nitrogen Oxides: Impacts
on Public Health and the Environment. Office of Air and
Radiation. Washington, DC. August 1997. p. 15. Report
downloaded from EPA's Technology Transfer Network.
2. U.S. Environmental Protection Agency. Fact Sheets for EPA's
Revised Ozone Standard and Revised Particulate Matter Standards.
Office of Air and Radiation. Washington, DC. July 17, 1997.
3. Reference 1. pp. 44-46.
4. U.S. Environmental Protection Agency. Regulatory Impact Analysis
for the Particulate Matter and Ozone National Ambient Air Quality
Standards and Proposed Regional Haze Rule. Appendix A:
Emissions and Air Quality and Appendix H: Economic Impact
Analysis and Supporting Information. Office of Air Quality
Planning and Standards. Research Triangle Park, NC. July 16,
1997.
5. Reference 4. p. 12-77.
6. U.S. Environmental Protection Agency. Fact Sheet: Proposed Rule
for Reducing Regional Transport of Ground-Level Ozone (Smog).
October 10, 1997.
7. U.S. Environmental Protection Agency. Fact Sheet: Acid Rain
Program Overview. Washington, DC. Taken from EPA web site on
November 14, 1997.
8. U.S. Environmental Protection Agency. Fact Sheet: Phase I of
the NOX Reduction Program. Acid Rain Program. Washington, DC.
Taken from EPA web site on November 14, 1997.
9. Reference 1. p. 140.
10. U.S. Environmental Protection Agency. Nitrogen Oxides Reduction
Program Final Rule for Phase II (Group 1 and Group 2 Boilers).
Acid Rain Program. Washington, DC. Taken from EPA web site on
November 14, 1997.
11. U.S. Environmental Protection Agency. Fact Sheet: EPA's Model
NOX Trading Rule Development. Acid Rain Program. Washington, DC.
1997.
12. U.S. Environmental Protection Agency. Fact Sheet: Proposed
Revision of Standards of Performance for Nitrogen Oxide Emissions
from New Fossil-Fuel Fired Steam Generating Units and Proposed
Revisions to Reporting Requirements for Standards of Performance
for New Fossil Fuel-Fired Steam Generating Units. July 1, 1997.
Downloaded from EPA's Technology Transfer Network.
1-10
-------
13. U.S. Department of Health and Human Services, National Institutes
of Health. Report to Congress on Methylmercury. NIEHS, Research
Triangle Park, NC. 1993.
14. National Academy of Sciences, National Research Council. Science
and Judgement in Risk Assessment. Washington, DC. 1994.
15. U.S. Environmental Protection Agency. Deposition of Air
Pollutants to the Great Waters: First Report to Congress. EPA-
453/R-93-055. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. May 1994.
16. U.S. Environmental Protection Agency. Deposition of Air
Pollutants to the Great Waters: Second Report to Congress. EPA-
453/R-97-011. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. June 30, 1997, pp. i-vi, 1-14.
1-11
-------
2.0 CHARACTERIZATION OF THE INDUSTRY
This chapter presents a characterization of the fossil-fuel-fired
utility industry. This is provided as a brief background for those
who may not be familiar with the industry. In addition, some
components of the process itself (e.g., type of boiler, method of
firing, type of emission control) may impact on the generation or
emissions of HAPs. These process components are introduced in this
chapter and their impact on HAPs is discussed in chapter 13. The
chapter is divided into seven main sections: background of the
industry, types and ownership of utilities, utility furnace design, PM
control, S02 control, NOX control, and a projected characterization of
the utility industry after implementation of the 1990 amendments to
the Act. All of the sections except the last describe the utility
industry as it existed in 1994.a The last section projects conditions
that are expected to exist in 2010, after the amendments are fully
implemented.
2.1 INDUSTRY BACKGROUND
An electric utility steam-generating unit is defined (section
112(a) (8) of the Act) as any fossil-fuel-fired combustion unit of more
than 25 megawatts electric (MWe) that serves a generator producing
electricity for sale. It can also be defined as a unit that
cogenerates steam and electricity and supplies more than one-third of
its potential electric output capacity and more than 25 MWe output to
any utility power distribution system for sale.
Fossil fuel-fired electric utility steam-generating units are
fueled primarily by coal, oil, or natural gas. Figure 2-1 shows the
1994 distribution of fossil fuels burned by the electric utility
industry by unit (i.e., individual boiler) and by total megawatts.1
Coal-fired boilers account for the largest portion of the industry by
number of units (1,026 units, 61 percent), representing 68 percent of
the industry's total megawatts. Gas-fired boilers make up 30 percent
of the industry's units (493 units) and account for 23 percent of the
total megawatts.b Oil-fired boilers account for 9 percent of the units
1994 was chosen as the analysis year for this final report because that is
the last year for which complete, plant-specific data were available from
the Utility Data Institute (UDI), the same data source as was used for the
1990 analyses and upon which the 2010 projections are based. The UDI
database is based on the same data as that used by the Energy Information
Administration (which may have more recent information) but, because of
varying uses and means of data verification, the data sets may not agree
completely.
It should be noted that the 1994 data for the number of individual
combined-cycle turbine systems were not available. However, the number of
plants (34, all gas-fired) and their total number of megawatts (10,047.87)
was available, and these totals were included in Figure 2-1. This would
have the effect of biasing the gas-fired boiler unit numbers lower than
normal, but would have no effect on the gas-fired boiler MW numbers.
2-1
-------
61%
68%
30%
23%
9%
By unit
9%
By megawatt
Coal
Gas
Oil
Figure 2-1. Fossil fuel use in the utility industry in 1994.1
-------
(149 units) and represent 9 percent of the megawatts. Fossil-fuel-
fired electric utility steam generating units accounted for 81 percent
of the total industry in 1994.1 Other fuels utilized include biomass
at 0.4 percent and "other" (including nuclear, geothermal,
hydroelectric, etc.) at 17.9 percent. Units less than 25 MWe (the
defining limit set by the Act) comprised 0.7 percent of the industry.
2.2 FOSSIL-FUEL-FIRED ELECTRIC UTILITY STEAM-GENERATING UNITS
This section describes the two basic types of utility facilities
and the types of ownership in the industry.
2.2.1 Types of Electric Utility Facilities
There are two basic types of facilities in the utility industry:
conventional utility power facilities and cogeneration facilities.
Although both types of facilities share similar designs, their major
difference is that conventional utility power facilities produce their
power solely for commercial power production whereas cogeneration
facilities produce their power primarily for an industrial purpose and
sell excess steam or electricity equal to more than one-third of their
potential electric output capacity and more than 25 MWe output to any
utility power distribution system.
Conventional facilities consist of units that produce heat in a
boiler to make high-pressure steam, which in turn powers units that
produce electricity through a combined cycle turbine system or a steam
turbine (see section 2.3.4). In both systems, the steam is recycled
without being used for any other purpose. Conventional facilities
account for most of the utility steam-generating units in the United
States. In 1994, there were 1,668 conventional utility steam-
generating units in the United States, with 1,026 burning coal of some
type. The total output was 464.8 gigawatts electrical (GWe). 1
Cogeneration is defined as the simultaneous production of power
(usually electricity) and another form of useful thermal energy
(usually steam or hot water) from a single fuel-consuming process.2
Cogeneration facilities can also consist of units that produce heat in
a boiler to make high-pressure steam that powers a steam turbine to
produce electricity or units that produce electricity through a
combined-cycle turbine system. Because of their primary uses as
industrial power and steam sources, however, they normally are too
small to fit the regulatory definition of a utility boiler. There
were 218 fossil-fuel-fired cogeneration facilities rated at 25 MWe or
greater that provided at least one-third of their excess power to a
grid operating in the United States as of 1990. These cogeneration
facilities consist of coal-, oil-, and gas-fired steam turbines and
combined-cycle turbines that provide 21,053 MWe of capacity. This
megawatt capacity was less than 5 percent of the total conventional
utility capacity in 1990 and was made up of only 54 coal-fired plants
(providing 5,098 MWe of capacity) and 12 oil-fired plants (providing
756 MWe of capacity). Thus, the electrical capacity of the coal- and
2-3
-------
oil-fired cogeneration facilities represented less than 1.2 percent of
total utility capacity in 1990.3
2.2.2 Types of Ownership
There are four basic types of electric power ownership in the
utility industry: publicly owned utility companies, Federal power
agencies, rural electric cooperatives, and investor-owned utility
generating companies. Publicly owned utilities are not-for-profit and
are operated by municipalities, counties, States, or other bodies such
as public utility districts. Federal power agencies are Federal
government agencies that provide electric power, usually to rural or
remote areas. Rural electric cooperatives are private, not-for-profit
corporations owned by their members who are also the customers they
serve; the cooperatives are not a part of the municipal government.
Investor-owned utility generating companies are owned by their
investors and sell electricity to make a profit.4
The oldest and largest companies (based on total megawatts
electric capacity) are the investor-owned utilities. Although
numbering only approximately 244 separate companies, investor-owned
utilities provided 74.3 percent of kilowatt hour (kwh) generation of
electric power to the Nation in 1994. Publicly owned utility
companies, which consist of approximately 2,020 separate companies,
represent 10.8 percent of the Nation's electric power supply. The
10 Federal power agencies generate 9.2 percent of the Nation's
electric power supply. Rural electric cooperatives, numbering
approximately 931 separate companies, provides 5.7 percent of the
Nation's electric power supply. These utilities maintain jointly
owned electric power grids to which electric power is supplied and
then sold to other utilities, industries, and individual customers.5
One of the fastest growing areas of the electric utility industry
has been nonutility generators. Nonutility generating units are
generally smaller than other utility units, of newer design
technology, and built to fill a specific need for power in their
geographic area(s). Nonutility generating units are usually privately
owned (although some are sponsored by larger publicly or investor-
owned utilities) and sell their power to private customers and the
jointly owned electric power grids.
Ownership of nonutility generators can be further divided into
ownership by:
2-4
-------
• Units that cogenerate steam and electricity (qualifying
facilities6) ;c
• Small power producers (<80 MWe) that generate electricity
primarily from a renewable source; and
• Other nonutility generators (e.g., independent power
producers [IPPs], units that cogenerate steam and
electricity [nonqualifying facilities], and other commercial
and industrial units).
In the last few years, the electric utility industry has
undergone a large restructuring brought on by the impending
deregulation of the United States electric utility industry. Under
this deregulation, consumers will be able to buy their electric power
from any supplier willing to provide power in their area, thus
breaking up the virtual monopolies that certain power companies and
agencies have had since electric power was first provided.
Many older performing units are being sold or shut down for the
sake of efficiency so that electric utility companies have an
inventory of newer, more efficient units with modern pollution
controls. Also, some companies are buying units in States where they
traditionally have never supplied power in the past. In the next
decade, these reorganizations will substantially change the makeup of
electric power ownership in the utility industry.
2.3 DESIGN OF ELECTRIC UTILITY UNITS
This section contains a summary of unit designs used in the
utility industry. Hazardous air pollutants are either formed during
combustion or introduced into the combustion unit (e.g., trace
constituents in the fuel). Thus, the design and operation of a unit
may impact on the generation and emission of HAPs.
2.3.1 Furnace Types
Utility furnace-fired boilers can be divided into five basic
firing types: stoker-, cyclone-, tangential-, and wall-fired boilers
and fluidized-bed combustors (FBCs).
2.3.1.1 Stoker-Fired Boilers. Stoker firing is one of the
oldest furnace firing methods still in use. In this process, fuel is
deposited on a moving or stationary grate or spread mechanically or
pneumatically from points usually 10 to 20 feet above the grate.7 The
A qualifying facility, under the Public Utility Regulatory Policies Act
(PURPA), Sections 292.303 and 292.305, may buy or sell energy to the
local utility or indirectly to other utilities. The local utility is
obligated to purchase or sell the energy at a price that is "just,
reasonable, and in the public interest" and does not "discriminate
against any qualifying facility in comparison to rates for sales to
other customers served by the electric utility."
2-5
-------
process utilizes both the combustion of fine coal powder in air and
the combustion of larger particles that fall and burn in the fuel bed
on the grate.8 Because of their design, stokers are used only for
smaller furnaces firing coal.
2.3.1.2 Cyclone-Fired Boilers. Cyclone firing uses several
water-cooled horizontal burners that produce high-temperature flames
that circulate in a cyclonic pattern. The burner design and placement
cause the ash to become a molten slag that is collected below the furn
ace. Because of this slagging system, cyclone-firing furnaces are
almost exclusively coal-fired; however, some units can fire oil.9
2.3.1.3 Tangential-Fired Boilers. Tangential-fired boilers are
based on the concept of a single flame envelope and project both fuel
and combustion air from the corners of the furnace. The flames are
directed on a line tangent to a small circle lying in a horizontal
plane at the center of the furnace. This action produces a fireball
that moves in a cyclonic motion and expands to fill the furnace.10
Tangential-fired boilers can fire coal, gas, or oil.
2.3.1.4 Wall-Fired Boilers. Wall-fired boilers are
characterized by rows of burners on the wall(s) of the furnace. The
two basic forms of wall-fired furnaces are single wall (having burners
on one wall) or opposed (having burners on walls that face each
other). Circular register burners and cell burners are types of
burner configurations found in single-wall or opposed-wall-fired
units. A circular register burner is a single burner mounted in the
furnace wall, separated from other burners so that it has a separate,
distinct flame zone. Cell burners are several circular register
burners grouped closely together to concentrate their distinct flame
zones. This use of a distinct flame zone is in contrast to the
fireball effect created by the tangentially fired furnace.11
Wall-fired boilers can fire coal, gas, or oil.
2.3.1.5 Fluidized-bed Combustors. In a typical FBC, combustion
occurs when coal, together with inert material (e.g., sand, silica,
alumina, or ash) and/or a sorbent such as limestone, are suspended
through the action of primary combustion air distributed below the
combustor floor.12 "Fluidized" refers to the state of the bed of
material (fuel or fuel and inert material [or sorbent]) as gas passes
through the bed. As the gas flow rate is increased, the forces on the
particles become just sufficient to cause buoyancy. The gas cushion
between the solids allows the particles to move freely, giving the bed
a liquid-like characteristic.13
Fluidized-bed combustors can be further divided into circulating
fluidized-bed (CFB) and bubbling fluidized-bed (BFB) steam generators.
The main difference between these two types is the state of
fluidization, which in turn depends mainly on the bed particle
diameter and fluidizing velocity. The CFB combustors have relatively
high velocities and fine bed particle size, whereas the BFB combustors
have relatively low velocities and coarse bed-particle size.14'15
2-6
-------
Most FBCs are of the atmospheric fluidized-bed combustor (AFBC)
type, which, as the name suggests, operate at atmospheric pressure. A
newer type of FBC is the pressurized fluidized-bed combustor (PFBC).
These combustors are physically smaller (yet maintain the same
megawatt capacity as equivalent AFBCs), operate at 10 to 20 times
atmospheric pressure, and incorporate a gas turbine in their power
production cycle. Because of these features, PFBCs offer a
potentially significant gain in overall thermal efficiency over
AFBCs.1S
2.3.1.6 Distribution of Furnace Types. Figure 2-2 shows the
1994 distribution of furnace types by fuel in the utility industry by
unit and by total megawatts.1 Wall-fired designs account for the
largest portion of the coal-fired units by number of units (48.8
percent), which represents 48.3 percent of the coal-fired units' total
megawatts. The second and third most common designs are the
tangential-fired and cyclone-fired units. Tangential firing is used
in 41.2 percent of the units (43.3 percent of the total megawatts),
and cyclone firing is used in 8.5 percent of the units (8 percent of
the total megawatts). Stoker-fired boilers and FBCs account for about
1.5 percent of designs among the coal-fired units (0.4 percent of the
total coal-fired megawatts). Wall-fired designs represent the largest
portion of gas- and oil-fired units by number of units (66 percent),
which represents 62.1 percent of the total megawatts. The second most
common design is the tangential-fired unit. Tangential-fired units
represent 28.5 percent (31 percent of the total megawatts) of the gas-
and oil-fired units, and combined-cycle gas turbine units account for
about 5.3 percent (6.9 percent of the total megawatts) of designs for
gas- and oil-fired units.d There is one known cyclone-fired unit
fueled by oil. This unit represents 0.2 percent (0.1 percent of the
total megawatts) of the gas- and oil-fired units.
2.3.1.7 Effects of Furnace Type on HAP Emissions. Many of the
organic HAPs leaving a furnace in the gas stream are produced in the
combustion zone and succeeding parts of the gas path. Factors
expected to affect the types and quantities of HAPs produced and
emitted include temperature, residence time, fuel characteristics,
firing scheme, bottom-ash and/or fly-ash partitioning, and adsorption
onto ash. By comparison, essentially all elemental HAPs leaving the
furnace enter with the fuel. The proportion of elemental HAPs in the
gas stream depends primarily on the bottom-ash and/or fly-ash
partitioning and adsorption onto ash. For both cases, furnace type
appears to influence the HAPs that leave the furnace and continue to a
control device or stack. Chapter 13 provides a discussion, from
It should be noted that the 1994 data for the number of individual
combined-cycle turbine systems were not available. However, the number
of plants (34, all gas-fired) and their total number of megawatts
(10,047.87) was available, and these totals were included in Figure 2-2.
This would have the effect of biasing the gas-fired boiler unit numbers
lower than normal, but would have no effect on the gas-fired boiler MW
numbers.
2-7
-------
0.7%
0.3%
I
00
41.2%
43.3%
48.3%
5.3C
0.8C
8.5%
0.1%
8.0%
Coal by Unit
Coal by Megawatt
66.0%
6.9%
28.5%
Gas and Oil by Unit
0.2%
62.1%
31.0%
Gas and Oil by Megawatt
0.1%
Tangential Fired
Wall Fired
Cyclone Fired
Stoker Fired
Fluidized Bed
Combined-Cycle
Gas Turbine
Figure 2-2. Unit types in the utility industry by fuel type in 1994.1
-------
limited data, suggesting that, for example, organic HAP emissions are
increased as furnace conditions are changed. Similarly for elemental
HAPs, chapter 13 shows data suggesting that, for example, cyclone
boilers emit some elemental HAPs at lower rates than tangential
boilers, and tangential boilers emit at lower rates than cyclone
boilers for other HAPs. Although tentative, furnace type
characterizations as related to HAP emissions are used for the
Nationwide emission factor program described later in this report.
Appendix D describes the construction of the program and the manner in
which HAP emissions are assigned to each furnace type.
2.3.2 Bottom Types
There are two types of furnace bottoms, wet and dry. The type of
bottom used depends on the type of fuel to be burned and on the
engineering requirements of the furnace. Wet-bottom furnaces sweep
the flame across the furnace floor at all firing rates to maintain the
ash in a molten state. Because of the ash handling and temperature
limitations of wet-bottom boilers, dry-bottom furnaces are the only
type currently used in new furnace construction.
In dry-bottom boilers, the ash reaches the melting point but
cools when deposited on the furnace walls; thus, it can be removed in
a dry state. This type of bottom is used in furnaces with tilting
fuel nozzles. It can provide a wider steam temperature control range
and can burn coals with widely varying ash characteristics.17
2.3.3 Cogeneration
Units that cogenerate steam and electricity can be classified as
topping or bottoming systems. Topping systems produce electricity
first, and all or part of their exhaust heat is subsequently used in
an industrial process. A bottoming system uses waste heat from a
boiler (or other high-temperature thermal process) to run a steam
turbine and/or generating unit to produce electricity. 18~20
2.3.4 Combined-Cycle Systems
The use of one source of hot gas to produce electricity by the
means of two separate thermal cycles and associated turbines is known
as combined cycle. An example would be a combustion gas turbine's
exhaust gas used to create steam for a steam turbine. Only systems
that incorporate a steam turbine as one of the two cycles are
considered in this study. Simple-cycle gas turbines with waste gas
vented directly to the atmosphere are not considered. Combined-cycle
systems consisting of a gas turbine with exhaust gases serving a heat
recovery steam generator are considered if they otherwise meet the
definition of an electric utility steam generating unit.
2.4 PARTICULATE MATTER CONTROL
This section describes the four major types of PM controls used
on utility boilers: mechanical collectors, electrostatic
precipitators (ESPs), particle scrubbers, and fabric filters (FFs).
2-9
-------
Figure 2-3 illustrates the 1994 distribution of PM control by fuel in
the utility industry by unit and by total MW.1
In 1994, ESPs accounted for the largest portion of the PM control
technology used on coal-fired units by number of units (91 percent)
and by total megawatts (91 percent). The second most common control
technology was the FF (also referred to as a baghouse). Fabric
filters were used on 7 percent of the coal-fired units (6 percent of
total megawatts). Particle scrubbers were used on 2 percent of the
coal-fired units (approximately 3 percent of the total megawatts).
Uncontrolled units represented the largest portion of the oil-
fired units (56 percent) and accounted for 48 percent of the oil-fired
industry's total MWs. Electrostatic precipitators were used on
22 percent of the oil-fired units or at 27.4 percent of the MWe
capacity of the oil-fired industry. Mechanical controls (cyclones)
were used on 21 percent of the oil-fired units (24.5 percent of the
total MWs). There is one known oil-fired unit controlled by a fabric
filter. This unit represents 1 percent (0.1 percent of the total
megawatts) of the oil-fired units. Gas-fired units had no PM
controls.1
As PM is formed during the combustion process and moves through
the boiler system, HAPs can be condensed or adsorbed on particle
surfaces. Although most particles are formed in the 3-micrometers
(urn) to 50-um range21 (on a mass basis), HAPs tend to concentrate
preferentially on particles smaller than about 7 urn, and especially on
those around 0.3 urn.22 Because of this preferential concentration,
high collection efficiency for fine particles is an important factor
in evaluating HAP control from PM collection devices. Each of the
four major control devices is described here, along with its method of
operation and collection efficiency by particle size. Much of the
efficiency data by particle size originates from extensive studies
performed by the EPA expressly for the purpose of comparing field
performance of FFs, ESPs, and particle scrubbers applied to combustion
sources. Special care was taken to provide accurate measurements for
particles smaller than about 10 urn (PM10) .
2.4.1 Mechanical Collectors
Mechanical collectors are the oldest, simplest, and least
efficient of the four types of PM control devices. The collectors
used for utility boilers are generally in the form of groups of
cylinders with conical bottoms (multicyclones). Particles in the
entering gas stream are hurled to the outside of the cylinder by
centrifugal force and are discharged at the bottom of the cone.
Collection efficiency for a typical multicyclone may be about 70 to
75 percent for 10-urn particles, but may drop to less than 20 percent
for 1-um particles.23 Thus, the multicyclone would be the least
effective of the four devices discussed here for reducing HAPs emitted
into the atmosphere as small particles or attached to small particles
(and was assumed to have no control effect on HAPs in the
calculations) .24
2-10
-------
91%
91%
21%
22%
Coal by Unit
1%
Oil by Unit
56%
Coal by Megawatt
24.5%
27.4%
0.1%
Oil by Megawatt
6%
48%
ESPs
FFs (Baghouses)
Particlute Scrubber
Mechanical
No Control
Figure 2-3. Particulate control in the utility industry by fuel type in 1994.1
-------
2.4.2 Electrostatic Precipitators
Electrostatic precipitators have been used on boilers for about
80 years, can be designed for high efficiencies (>99 percent, but at
the cost of increased unit size), and are the most frequently used PM
control devices on utility boilers. They operate by imparting an
electrical charge to incoming particles, then attracting the particles
to oppositely charged plates for collection. The collected particles
are periodically dislodged in sheets or agglomerates by rapping the
plates. Particle removal in an ESP depends largely on the electrical
resistivity of the particles being collected. An optimum value exists
for any ash; above and below this value, particles become less
effectively charged and collected. Coal that contains a moderate to
high amount of sulfur (more than about 3 percent) produces an easily
collected fly ash. Low-sulfur coal produces a high-resistivity fly
ash that is difficult to collect. Resistivity of the fly ash can be
changed by operating the boiler at a different temperature or by
conditioning the particles upstream of the ESP with sulfur trioxide,
sulfuric acid, water, sodium, or ammonia. In addition, efficiency is
not uniform for all particle sizes. For coal fly ash, particles
larger than about 1 to 8 urn and smaller than about 0.3 urn (not to be
confused with total PM) are typically collected with efficiencies from
95 to 99.9 percent.25 Particles near the 0.3 urn size are in a charging
transition region that reduces collection efficiency.26 These
particles have been shown to have lower collection efficiency (about
80 to 95 percent). However, for particles in the 1- to 8-um size
range, the reasons for poorer collection efficiency are not as well
understood. There is often a penetration peak in this size range.27
If these particles escape capture by the ESP, boiler emissions are
likely to show an increase in smaller particles that may be enriched
with HAPs.28 As mentioned above, ESPs can be designed to control
particulate emissions to high efficiencies. On a total mass basis,
these efficiencies can be equivalent to those of FFs. However, on a
fine particulate basis, the ESP may not be quite as effective as an
FF. Because designing for higher overall efficiencies in an ESP
requires increasing the size (and cost) of the device, past practice
has been to design to meet regulatory requirements. Further study is
required to determine the capabilities of ESPs for higher overall HAP
removal compared to other control systems.
2.4.3 Particulate Matter Scrubbers
The use of wet scrubbers for PM collection has three distinct
disadvantages: high energy consumption when high efficiency is
required, the presence of a wet effluent to be disposed of, and
difficulty in obtaining high collection efficiencies for fine
particles. Scrubbers operate by shattering streams of water into
small droplets that collide with and trap PM contained in the flue gas
or by forcing the flue gas into intimate contact with water films.
The particle-laden droplets or water films coalesce and are collected
in a sump at the bottom of the scrubber. The three common types of
scrubbers for fly ash control are venturi, preformed spray, and moving
bed. Venturi scrubbers, the type most commonly used for utility
systems, transport particle-laden flue gas through a constriction at
2-12
-------
which violent mixing takes place. Water is introduced either at, or
upstream of, the constriction. Preformed spray scrubbers are usually
vertical cylinders with flue gas passing upward through droplets
sprayed from nozzles near the top of the unit. Moving-bed scrubbers
have an upper chamber in which a bed of low-density spheres (often
plastic) is irrigated by streams of water from above. Gas passing
upward through the bed agitates the wetted spheres, which continually
expose fresh liquid surfaces for particle transfer. Particle
collection efficiency in scrubbers is generally size and energy
dependent. Although some scrubbers collect particles at high
efficiency with low energy consumption, venturi scrubbers are normally
energy intensive compared to ESPs or FFs. Particles larger than a few
micrometers can be collected with efficiencies greater than 99
percent, but, at sizes smaller than about 1 or 2 //m, efficiency may be
reduced to less than 50 percent.29 Because of this low collection
efficiency, the emission of HAP-laden particles from scrubbers is
expected to be greater than for ESPs. However, water in the scrubber
may remove water-soluble HAPs.30
2.4.4 Fabric Filters
Fabric filters have been used on utility boilers for about
20 years. They are inherently efficient and are effective when
high-efficiency PM collection is required. Unlike ESPs, their size is
not a strong function of desired efficiency. They must be designed
and operated carefully to ensure that the fabric tubes (bags) inside
the collector are not damaged or destroyed by adverse operating
conditions. Fabric filters collect PM by placing a fabric barrier in
the flue gas path. Gas passes freely through the fabric, but
particles are trapped and retained for periodic removal. Data from a
small utility boiler show collection efficiencies not lower than 99.6
percent across all particle sizes from 0.3 //m to about 10 //m (the
range of the measuring equipment) . 31 Because of its high collection
efficiency for small particles, the baghouse should be particularly
effective for removing particles that have been enriched with HAPs.32'33
However, further study is required to determine if baghouses can
remove significantly greater quantities of HAPs than are removed by
other control systems.
2.4.5 Comparison of Particle Collectors
Table 2-1 compares the characteristics and capabilities of the
four particle collection devices described. Fabric filters and ESPs
appear to provide the highest mass collection efficiency for fly ash.
Fabric filters appear to be the best of the four devices for capturing
small particles that may be enriched with HAPs. Examination of
Tables 13-6 and 13-9, which compare HAP removal by cold-side ESPs and
FFs on utility boilers, also suggests that FFs may be more effective.
However, further study is required for confirmation.
2-13
-------
Table 2-1. Comparison of Particulate Matter Collection Systems3
Collector
Multicyclone
ESPs
Particle scrubber
FFs
Typical mass
efficiency, %
70-90
99-99.7
95-99
99-99.9
Efficiency at
0.3 urn, %
0-15
80-95
30-85
99-99.8
Energy consumption, in. H,O
Collector System
4-10
0.5-1
2-70
5-10
7-13
3.5-4
5-73
8-13
ESPs = Electrostatic precipitators.
FFs = Fabric filters.
2.5 SULFUR DIOXIDE CONTROL
Sulfur dioxide emissions are controlled through either
(1) precombustion measures, namely, the combustion of fuels that
contain lesser amounts of sulfur; (2) combustion measures, such as an
FBC system that combusts coal and limestone (or an inert material);
and (3) postcombustion measures, such as the use of flue gas scrubbing
(or flue gas desulfurization [FGD]) devices.
Precombustion measures may include the use of compliance fuels
(fuels having characteristics, such as low sulfur content, that allow
the user to comply with emission limitations solely by switching to
the fuel) to meet State implementation plans (SIPs) or NSPS. The use
of SIP or NSPS fuels means that the sulfur content in the fuel is
sufficiently low that add-on controls or postcombustion controls are
not required. As of 1994, all oil- and gas-fired units burn
compliance fuel, whereas approximately 85 percent of the coal-fired
units burn compliance coal.1
Compliance coal may be obtained through the mining of
lower-sulfur coals, coal washing, and/or coal blending. (Because coal
washing is reviewed in section 2.5.1, it is only briefly mentioned
here.) Most bituminous coals are cleaned in order to meet customer
specifications on sulfur, ash, and heating content. In the process of
cleaning, the sulfur and ash content of the coal are reduced, while
the heating content may be increased. Consequently, less of the
cleaned coal, containing less sulfur, is needed to achieve a given
heating rate. Compliance coal may also be obtained through coal
blending, in which higher-sulfur coals are blended with lower-sulfur
coals.
Combustion measures control emissions of S02 from six coal-fired
units, representing a total capacity of 815.6 MWe. These units are
FBCs and control S02 in the combustion zone by using limestone as a
sorbent.1
2-14
-------
Figure 2-4 shows S02 control devices used in coal-fired utilities
in 1994 based on the number of units and total MW capacity.1 As shown
in Figure 2-4, 15 percent of the units, representing about 22 percent
of the coal-fired generation capacity, used postcombustion flue gas
scrubbing to comply with S02 regulations. A wet FGD was used at
approximately 14 percent of the units (approximately 21 percent of the
coal-fired total electric capacity), whereas a spray dryer adsorber/FF
(SDA/FF, also called a dry scrubber) system was used at approximately
1 percent of the coal-fired units (approximately 1 percent of the
coal-fired total electric capacity).
Sulfur dioxide emission standards for utility steam generators
vary according to the size, age, and location of a facility. Existing
boilers are regulated by SIPs. Plants built after 1971 are subject to
NSPS S02 emission limits of 1.2 pounds per million British thermal
units heat input (lb/MMBtu). Plants built after 1978 are additionally
required to reduce their S02 emissions by 70 to 90 percent.
The extent of postcombustion S02 control used by the utility
industry will increase in response to Title IV of the 1990 amendments
to the Act, which require S02 reduction in two phases. The likely mix
of S02 control approaches that will be used to comply with the Phase I
and Phase II requirements is discussed in section 2.7.
2.5.1 Precombustion Control: Fuel Options
By using coal with an appropriately low sulfur content,
85 percent of the coal-based utility units currently comply with S02
emission limits. Compliance coals may be mined from the ground or may
be obtained by cleaning or blending mined coal.
Physical coal cleaning typically involves (1) size reduction and
screening, (2) gravity separation of coal from sulfur-bearing mineral
impurities, and (3) dewatering and drying.35 Approximately 77 percent
of the eastern and midwestern bituminous coal shipments are subjected
to some physical cleaning process.36 Subbituminous and lignite coals
are not routinely cleaned.37'38 The primary purpose of physical
cleaning has been to remove ash; coal cleaning has the consequence of
increasing the heating value of the coal and reducing the sulfur
content in the coal.39 Bituminous coals from the eastern United
States, cleaned with a 1.6 specific gravity separation, were found to
provide reductions of 48 percent ash, 65 percent pyritic sulfur, 43
percent total sulfur, and 48 percent S02 emissions at a Btu recovery
rate of 94 percent.40
As with sulfur, many trace elements may be both organically bound
and present as a part of a mineral in the same coal. Thus, physical
coal cleaning has the potential to remove some of the trace elements
associated with the mineral matter. Recent experimental studies
showed significant reductions of a number of trace elements.35'41 The
reduction percentages were found to depend on the type of coal and the
trace element's nature within the coal. For a few trace elements, an
enrichment effect was observed for some of the coal samples; however,
2-15
-------
I
I-1
CTl
85%
78%
14%
By Unit
21%
By Megawatt
No Control Device
FGD Scrubber
SDA/FF Baghouse
Figure 2-4. S02 control in the utility industry in 1994 (coal-fired boilers only)?
-------
when expressed on a Btu basis, physical cleaning will always reduce,
to some extent, the amount of trace elements present in coal. The
effectiveness of coal cleaning in reducing concentrations of trace
elements in coal is discussed in section 13.1.2.
2.5.2 Postcombustion Control: Flue Gas Scrubbing for S02 Control
According to the 1995 compilation of the Edison Electric
Institute's (EEI) Power Statistics database (examining 1994 data),
scrubbers were installed on 152 boiler units (out of about
1,026 coal-fired units in the United States) with a total rated
capacity of 70,458 MWe.1 Table 2-2 lists the different types of
scrubbing installations used in United States utility power plants.
As shown in Table 2-2, wet limestone/lime slurry scrubbing represents
the most prevalent scrubber type with almost 80 percent of the total
flue gas scrubbing capacity.1
2.5.2.1 Wet Limestone. In a wet limestone scrubber, flue gas
containing S02 is brought into contact with a limestone-water slurry.
The S02 is absorbed into the slurry and reacts with limestone to form
an insoluble sludge. The sludge, mostly calcium sulfite hemihydrate
and gypsum, is usually disposed of in a pond specifically constructed
for the purpose .42
The two common absorber designs include fixed packing and
horizontal or vertical spray towers, with spray towers being the most
prevalent. The absorber must be constructed of materials that resist
corrosion, erosion, and scaling. To reduce corrosion and erosion
problems, a scrubber is located downstream of a PM collection device.
A flue gas cooler and humidifier are used to cool the flue gases,
generally to 50° C (122° F) , prior to absorption. The size and number
of scrubber modules are directly related to boiler size, load
fluctuations, and system availability and compliance requirements.
Auxiliary equipment includes a demister to remove entrained
droplets from the scrubber outlet gas, a heat exchanger system to
reheat the outlet gas prior to exhaust, a slurry preparation system,
and a disposal system for the large quantities of sludge produced.
Sludge disposal needs can be very site specific and depend upon the
local climate and soil conditions.43
The basic wet limestone scrubbing process is simple and well
established. Limestone sorbent is cheap and generally locally
available in the United States. The S02 removal efficiencies of
existing wet limestone scrubbers range from 31 to 97 percent, with an
average of 78 percent.1 Operating parameters affecting S02 removal
efficiency include liquid-to-gas ratio, pH of the scrubbing medium,
and the ratio of calcium sorbent to S02. Periodic maintenance is
needed because of scaling, erosion, and plugging problems.
2-17
-------
Table 2-2. Distribution of S02 Control Technologies in 19941
Scrubber type
Wet limestone
Wet lime
Dry lime/SDA
Sodium carbonate
Dual-alkali
Wellman-Lord
Mag-Ox
Dry aqueous carbonate
No. of boiler units
70
44
15
9
6
4
3
1
Installed FGD capacity
(MWe)
35,101
21,172
5,615
3,181
2,267
1,779
895
450
Total
Total percent of
installed FGD capacity
(%)
49.8
30.0
8.0
4.5
3.2
2.5
1.3
0.7
100.0
FGD = Flue gas desulfurization.
SDA = Spray dryer adsorber.
Recent advancements include the use of additives or design
changes to promote S02 absorption or to reduce scaling and
precipitation problems. Gypsum can now be recovered as a salable
byproduct. Extensive operating experience has increased industry
confidence in designing larger, more reliable limestone scrubber
modules. In 1994, wet limestone scrubbers were used at 70 units, or
at 35,101 MWe of the total scrubbing capacity.1
2.5.2.2 Wet Lime. In a wet lime scrubber, flue gas containing
S02 is contacted with hydrated lime-water slurry; the SO, is absorbed
into the slurry and reacts with hydrated lime to form an insoluble
sludge. The hydrated lime provides greater alkalinity (higher pH) and
reactivity than limestone.44
Wet lime scrubbing is a proven technology; considerable operating
experience has been gained in 44 utility units.1 These units
represented 21,172 MWe of the total scrubbing capacity in 1994. The
S02 removal efficiencies of existing wet lime scrubbers range from 30
to 95 percent. Recent advances include the use of additives to
improve performance, reduce scaling problems, and produce a salable
gypsum byproduct. Lime scrubbing processes require appropriate
disposal of large quantities of waste sludge.
2.5.2.3 Dry Lime/Spray Dryer Adsorber. This process produces
dry reaction waste products for easy disposal. In this process, flue
gas at air preheater outlet temperatures of 121° to 177° C (250° to
2-18
-------
350° F) is contacted with fine spray droplets of hydrated lime slurry
in a spray dryer vessel. The S02 is absorbed in the slurry and reacts
with the hydrated lime reagent to form solid calcium sulfite and
calcium sulfate as in a wet lime scrubber.45 The water is evaporated
by the heat of the flue gas. The dried solids are entrained in the
flue gas, along with fly ash, and are collected in a PM collection
device. Most of the S02 removal occurs in the spray dryer vessel
itself, although some additional S02 capture has also been observed in
downstream PM collection devices, especially baghouses.
The primary operating parameters affecting S02 removal are the
calcium-reagent-to-sulfur stoichiometric ratio and the approach to
saturation in the spray dryer. To increase overall sorbent
utilization, the solids collected in the spray dryer and the PM
collection device may be recycled. The S02 removal efficiencies of the
existing lime spray dryer systems range from 60 to 85 percent.1 Spray
dryers were used at 15 units and constituted 5,615 MWe of scrubbing
capacity in 1994.
2.5.2.4 Wet Sodium Carbonate. Flue gas scrubbing with sodium
carbonate solution minimizes the operation and maintenance problems
related to lime and/or limestone slurry scrubbers. However, the
process uses a reagent that is relatively expensive unless it can be
found as a byproduct from another process or as a locally mined
material (trona). There were nine units (in 1994) using wet sodium
carbonate scrubbing in the United States, representing 3,181 MWe of
the total scrubbing capacity.1 Waste products of this process include
sodium sulfite and sodium sulfate.
Due to the higher solubility and greater reactivity of the sodium
carbonate compared to lime and/or limestone, a smaller size scrubber
can be used. The primary operating parameters are liquid-to-gas ratio
and the reagent stoichiometric ratios used. Sorbent utilizations are
high. The S02 removal efficiencies reported for this process range
from 75 to 90 percent.1 The soluble reaction products must be treated
before disposal. The treated flue gas is demisted and reheated before
exhausting through a stack.
2.5.2.5 Dual Alkali. A dual alkali system combines the
operational advantage of a sodium-based solution scrubbing system with
the economic advantage of a lime and/or limestone-based system. As
practiced in the United States, a dual (or double) alkali system uses
a sodium sulfite solution to absorb S02 from flue gas and to form
sodium bisulfite. The spent sorbent is reacted with lime to
precipitate calcium sulfite and to regenerate the active sodium
sulfite sorbent.46 The precipitated calcium salts are separated and
dewatered for disposal. The treated flue gas is demisted and reheated
before it is exhausted through a stack.
The dual alkali process has been installed (1994) on six boiler
units in the United States with a combined capacity of 2,267 MWe. The
S02 removal efficiencies at these plants range from 77 to 93 percent.1
2-19
-------
This process also requires appropriate disposal of large quantities of
waste calcium salts. Recent advances in this process include forced
oxidation of calcium sulfite to a salable gypsum byproduct, which
reduces the waste disposal load.
2.5.2.6 WeiIman-Lord. In the Wellman-Lord process, S02 from the
flue gas is absorbed in a sodium sulfite solution to form sodium
bisulfite as in the dual-alkali process. The spent sorbent is,
however, thermally regenerated by reversing the absorption reaction.
Regenerated sodium sulfite crystals are dissolved and returned to the
absorber. The concentrated, stripped S02 stream is converted to
salable sulfuric acid, elemental sulfur, or liquid S02. 47 The treated
flue gas is demisted and reheated before it is exhausted through a
stack. The Wellman-Lord process has been installed on four United
States boiler units with a combined capacity of 1,779 MWe (1994), with
S02 removal efficiencies ranging from 65 to 74 percent.1
2.5.2.7 Magnesium Oxide. Similar to Wellman-Lord, the magnesium
oxide (MAG-OX) FGD process is regenerable. The S02 in the flue gas is
absorbed by a magnesium oxide slurry, and the resulting magnesium
sulfite is calcined to regenerate magnesium oxide that is slurried and
recycled back to the absorber. The S02-rich gas produced in the
regeneration step is processed further to produce a salable product
such as sulfuric acid or elemental sulfur.48
Because of the high-temperature regeneration step at 800° to
1,000° C (1,472° to 1,832° F), energy requirements for this process are
high. However, due to the regenerative nature of the process, reagent
and disposal costs are small. Scrubber plugging and scaling problems
are reduced compared to a limestone scrubbing system. The corrosion
and/or erosion problems related to a slurry operation are still
significant. The magnesium oxide process has been installed on three
boiler units in the United States with a combined capacity of 895 MWe
(1994) . The S02 removal efficiencies at these plants range from 85 to
94 percent.1
2.5.2.8 Dry Aqueous Carbonate. In the dry aqueous carbonate
process, the flue gas is contacted with an aqueous sodium carbonate
solution in a spray dryer. The sodium carbonate reacts with and
removes S02 from the flue gases, then the solution is evaporated to
dryness by the hot flue gases. The dry reaction products (sodium
sulfite, sodium sulfate, and unreacted sodium carbonate) are removed
from the flue gases by passage through multicyclones and an ESP.
Subsequent processing of the reaction products with crushed coal
yields regenerated sodium carbonate and hydrogen sulfide gas. The
sodium carbonate is recycled to the spray absorber, and hydrogen
sulfide gas is converted to salable sulfur.49 Only one unit, of 450
MWe capacity, uses the dry aqueous carbonate system for FGD (1994); it
has a 70 percent S02 removal efficiency.1
2-20
-------
2 . 6 NOX CONTROL
This section provides a brief review of the formation and control
of NOX emissions, as well as the general types of NOX control used in
the utility industry. Detailed information on the formation and
control of NOX can be found in four major technical documents.50"53
Figure 2-5 shows NOX control approaches used in 1994 based on the
number of units and total MW capacity.1 Around 67 percent of
coal-fired plants, representing about 50.4 percent of the coal-fired
MW capacity, had no NOX control, whereas around 33 percent of the
units, representing about 49.6 percent of the coal-fired MW capacity,
used some kind of NOX control. Approximately 72 percent of the gas-
and oil-fired units, with about 61 percent of the MW capacity, did not
use NOX control, whereas approximately 28 percent of the units,
representing about 39 percent of the gas- and oil-fired MW capacity,
used some kind of NOX control. The gas- and oil-fired portion of
Figure 2-5 does not contain data from combined-cycle turbine systems.
The chemical species nitrogen dioxide (N02) and nitric oxide
(NO) are collectively called NOX. In general, NOX from combustion
consists of about 95 percent NO and 5 percent N02; however, NOX is
reported as N02.54 Nitrogen oxides are primarily formed during fossil
fuel combustion in one of two ways: (1) oxidation of nitrogen in the
combustion air to give thermal NOX, or (2) oxidation of nitrogen
contained in the fuel to give fuel NOX. There is a third form of NOX,
namely prompt NOX, that is formed by the reaction of nitrogen and
hydrocarbons in the fuel, but prompt NOX has a lifetime of several
microseconds.55 Thermal NOX is the predominant form during the
combustion of fuels that contain relatively little fuel-bound nitrogen
(such as natural gas and distillate oil). Both thermal and fuel NOX
are formed during the combustion of fuels that contain fuel-bound
nitrogen (such as residual oil and coal).5S Fuel switching, then, may
yield reduced NOX emissions.
The formation of NOX in coal-fired units depends on factors such
as the type of boiler, type of burner, and facility operation.57 Any
of these factors that increase temperature or residence time at high
temperature will promote NOX formation.58 In general, cyclone and other
wet-bottom boilers have relatively higher NOX emissions, with an
approximate range of 1 to 2 Ib/MMBtu, than do dry-bottom boilers,
which have an approximate range of 0.4 to 1.5 Ib/MMBtu.59 With regard
to the type of burner, wall-fired wet-bottom boilers have relatively
higher NOX emissions with an approximate range of 1.6 to 2 Ib/MMBtu,
wall-fired dry-bottom boilers have moderate NOX emissions with an
approximate range of 0.5 to 1.45 Ib/MMBtu, and tangential-fired dry-
bottom boilers have relatively lower NOX emissions at approximately
0.4 to 0.9 Ib/MMBtu.so Because of their low combustion temperatures,
an FBC's thermal NOX is essentially zero. Design features such as
staged combustion can significantly reduce fuel NOX, leading to low NOX
2-21
-------
67%
50.4%
33%
49.6%
Coal by Unit
Coal by Megawatt
72%
61%
Uncontrolled
Controlled
39%
Gas and Oil by Unit
Gas and Oil by Megawatt
Figure 2-5. Nitrogen oxide control in the utility industry by fuel type in 1994 .1
-------
The reduction of NOX emissions is important for controlling acid
rain and ozone formation.62 Techniques used to reduce NOX formation
include those for combustion and postcombustion control. Combustion
control techniques regulate the amount of combustion air and may also
control the flame temperature at different stages of the combustion
process; postcombustion control involves the removal of NOX from the
flue gas.63 More than one form of combustion control may be used for a
given unit.
2.6.1 Combustion Control
Control can be achieved through staged combustion (also called
air staging). With staged combustion, the primary combustion zone is
fired with most of the air needed for complete combustion. The
remaining air needed is introduced into the products of the incomplete
combustion in a second combustion zone. Air staging lowers the peak
flame temperature, thereby reducing thermal NOX, and reduces the
production of fuel NOX by reducing the oxygen available for combination
with the fuel nitrogen.64 Staged combustion may be achieved through
low NOX burners, overfire air (OFA), off-stoichiometric firing (OSF),
selective or biased burner firing (BBF), and burners-out-of-
service (BOOS) .65 Each of these methods requires modifying equipment
or operating conditions so that a fuel-rich condition exists near the
burners. In cyclone boilers, combustion occurs with a molten ash
layer and the combustion gases flow to the main furnace; this design
precludes the use of low NOX burners and air staging.66
Low NOX burners may be used in coal-, oil-, and gas-fired boilers
to lower NOX emissions by about 25 to 55 percent.67 Overfire air may be
used as a single NOX control technique, with NOX reductions of 15 to
50 percent.68'69 When OFA is combined with low NOX burners, reductions
of up to 60 percent may result.70 The actual NOX reduction achieved
with a given control technique may vary from site to site.71
Just as the combustion air to the primary combustion zone may be
reduced, part of the fuel may be diverted to create a secondary flame
with fuel-rich conditions downstream of the primary combustion zone.
This combustion technique is termed reburn and involves injecting
10 to 20 percent of the fuel after the primary combustion zone and
completing the combustion with OFA.72 The fuel injected downstream is
not necessarily the same as that used in the preliminary combustion
zone. In most applications of reburn, the primary fuel is coal and
the reburn fuel is natural gas. Natural gas reburn has been
successfully demonstrated in several field tests in the United States
and abroad.73'74 Reburn with other fuels, primarily coal, is currently
under development, as are improvements in the process.75
Other ways to reduce NOX formation by reducing peak flame
temperature include using flue gas recirculation (FGR), reducing
amounts of OFA, injecting steam or water into the primary combustion
zone, and increasing spacing between burners.76 By using FGR to return
part of the flue gas to the primary combustion zone, the flame
temperature and the concentration of oxygen in the primary combustion
2-23
-------
zone are reduced. Flue gas recirculation is usually used with natural
gas and distillate oil combustion. The peak temperature may also be
reduced in natural gas and distillate fuel oil combustion units by
reducing the amount of combustion air that is preheated; however, the
unit efficiency will also be reduced.
Temperatures may also be reduced in the primary combustion zone
by increasing the spacing between burners for greater heat transfer to
heat-absorbing surfaces.77 Another combustion control technique
involves reducing the boiler load. In this case, the formation of
thermal NOX generally decreases directly with decreases in heat release
rate; however, reducing the load may cause poor air and fuel mixing
and increase carbon monoxide (CO) and soot emissions.78
2.6.2 Postcombustion Control
Postcombustion control involves the removal of NOX from the flue
gas downstream of the combustion zone and is achieved either by
reducing NOX emissions only (selective noncatalytic reduction [SNCR])
or by reducing combined emissions of CO, hydrocarbons, and NOX
(selective catalytic reduction [SCR]).79 Postcombustion control had,
up to 1994, seen limited use in new coal-fired units with the
application concentrated in California, where SCR is used at
cogeneration plants and with gas-fired turbines and where SNCR is used
at FBCs, two pulverized coal-fired units, and a gas-fired unit
boiler.80 Since 1994, SCR has been installed on seven utility boiler
units, five of which are cogeneration units.81
With SCR, ammonia or another reducing agent is diluted with air
or steam, and the mixture is injected into the flue gas upstream of a
vanadium, titanium, platinum, or zeolite catalyst bed. The NOX is
reduced to molecular nitrogen on the catalyst surface.82 Selective
catalytic reduction units provide up to 70 to 90 percent NOX reduction83
and are usually located between the economizer outlet and air heater
flue-gas inlet, where temperatures are 230° to 400° C (450° to 750° F).84
Selective noncatalytic reduction is currently achieved
commercially in one of two ways: THERMAL DeNOxe, an Exxon process, or
NOXOUT®, an EPRI process. THERMAL DeNOxe reduces NOX to nitrogen
through injection of ammonia into the air-rich flue gas. NOXOUT®
achieves NOX reduction by injecting urea into the oxygen-rich and/or
high-temperature convection part of the boiler.85
The necessity of using nitrogen-based reagents requires SCR and
SNCR systems to closely monitor and control the rate of reagent
injection. If injection rates are too high, NOX emissions may increase
(in SNCR systems), and stack emissions of ammonia may also occur in
concentrations of 10 to 50 ppm. A portion (usually around 5 percent)
of the NO reduction by SNCR systems is due to transformation of NO to
N20, which is a global warming gas.
2-24
-------
Table 2-3 presents a general breakdown of utility industry NOX
control usage according to the 1995 EEI power statistics database
(1994 data)-1 As shown in Table 2-3, most of the utility industry has
no NOX control; 64 percent of the dry-bottom coal-fired boiler units,
87 percent of the wet-bottom coal-fired boiler units, 76 percent of
the oil-fired boiler units, 70 percent of the gas-fired boiler units,
and 100 percent of the combined-cycle turbine units had no NOX control
in 1994 (see Note b in Table 2-3). Units that had NOX control
equipment used various types of staged combustion techniques,
including low-NOx burners, OFA, OSF, BBF, and BOOS. Staged combustion
control was used in 33 percent of the dry-bottom coal-fired units,
11 percent of the wet-bottom coal-fired units, 24 percent of the oil-
fired units, and 30 percent of the gas-fired boiler units. Table 2-3
also shows that approximately 3 percent of the dry-bottom coal-fired
units and 2 percent of the wet-bottom coal-fired units had boiler
design as a NOX control method.
2.7 UTILITY INDUSTRY AFTER IMPLEMENTATION OF 1990 AMENDMENTS
This section describes the changes in the utility industry
expected during the 1990-2010 time frame. The effect of planned
generation capacity growth on the fuel use and technologies that will
be used for steam and power generation is discussed in section 2.7.1.
Title IV of the Act requires the utility industry to reduce S02
emissions in two phases. The effect of S02 control measures likely to
be used to comply with the Phase I and Phase II requirements on the
overall mix of utility S02 control technology is discussed in
section 2.7.2. For the purposes of this analysis, the projected
compliance date for Phase II was determined to be 2010. This year was
chosen after discussions with Agency, nonagency, and industry sources
concerning possible delays written into Title IV of the Act.
Title IV also contains other provisions that will affect utility
responses to regulations. These revisions include topics such as
permitting, monitoring, enforcement, repowering, and penalties.
Although these provisions affect the manner in which the utility
industry will respond to regulations, they are generally subsidiary to
emissions estimates based on fuel usage. These provisions are not
discussed further here.
Since the Interim Final Report to Congress, the EPA has obtained
and analyzed current information and future projections on energy
production (by fuel) in the electric utility industry.8S It appears
that the fuel usage projections listed below are being met and are
proceeding toward what the EPA (using the Acid Rain Division's [ARD]
model projections) predicted in the Interim Final Report to Congress.
Therefore this section and the 2010 projected emissions were not
changed in this Final Report.
2-25
-------
Table 2-3.
in 1994 !
Distribution of NOX Control by Fuel Burned, by Unit,
Fuel
Coal, boiler bottom
Dry
Wet
Oil
Gas
Combined-cycle turbine
Percent NOX control3
None
64 (46)
87 (85)
76 (66)
70 (60)
100(100)b
Staged combustion
33 (49)
11 (12)
24 (34)
30 (40)
-
Boiler design
3(5)
2(1)
-
-
-
a Values listed in parentheses are the percent distribution by MWe for each type of fuel.
b To cool combustion gases, steam or water may be injected with the fuel, with the air, or directly into the
combustion zone. This technique is used for gas-fired turbines due to the relatively low efficiency penalty
(typically 1 percent).87 However, this technique is not used for utility boilers because it has a high efficiency
penalty (about 10 percent).87 Steam or water injection was present in the 1990 utility data but seems to missing
from the 1994 utility data set. In the 1990 data, approximately 36 percent of the combined-cycle turbine units used
steam or water injection for NOX control, whereas only approximately 2 percent of the boilers reported using this
technique. The EPA believes that this technique is still being used but the companies that were surveyed
neglected to include this in the responses.
2.7.1 Industry Growth
The publicly owned utility companies, Federal power agencies,
rural electric cooperatives, and investor-owned utility generating
companies are projected to increase their new generating capacity in
service or scheduled for service in the 1990-2010 time frame by
750 billion kwh, from 1,940 to 2,690 billion kwh.88 These and other
projections for utility industry configuration and growth were taken
from a study titled, Economic Analysis of the Title IV Requirements of
The 1990 Clean Air Act Amendments, produced for the ARD of the EPA's
Office of Air and Radiation (OAR) by ICF Resources Incorporated. This
single projection is used by the Office of Air Quality Planning and
Standards (OAQPS) in this report to maintain consistency with the ARD.
Figure 2-6 compares utility fossil fuel consumption, on a Btu
basis, for 1990 and projected use for 2010 (publicly owned utility
companies, rural electric cooperatives, investor-owned utility
generating companies only) .89 On this basis, the predominant fossil
fuel both in 1990 and projected for 2010 is coal, at approximately 81
percent of the total industry fossil-fuel usage (22 quadrillion Btu/yr
[Quads] in 2010) . Oil and gas consumption in 1990 were, respectively,
6 and 13 percent of the total industry fossil-fuel usage on a Btu/yr
basis. For 2010, oil consumption was projected to decrease to
2 percent (0.6 Quads), and gas consumption was projected to increase
to 17 percent (4.5 Quads) on a Btu/yr basis for the total industry
fossil-fuel usage.90 Based on the ARD model projections, coal
consumption in 2010 is expected to be the same percentage of the total
utility fossil-fuel usage as in 1990 (81 percent).
2-26
-------
81%
81%
13%
6%
Utility Fuel Usage in 1990
(Btu/yr)
17%
2%
Utility Fuel Usage in 2010
(Btu/yr)
1990
2010
Fuel
Coal
Oil
Gas
Btu/yr
1.7x1016
1.2x1015
2.8 x1015
Coal
Gas
Oil
Fuel
Coal
Oil
Gas
Btu/yr
2.2 x1016
6.2 x1014
4.5 x1015
Figure 2-6. Fuel use in the utility industry by fuel type
in 1990 and projections for the year 2010.899°
-------
Recent projections by the United States Department of Energy's
(DOE) Energy Information Administration (EIA) were reviewed to
determine the current validity of this projected fuel use scenario
(see Table 2-4).a6 Based on the 1998 Annual Energy Outlook, coal use
in the electric utility industry in 2010 is projected to be 21.34
Quads (73 percent of total utility fossil-fuel usage), oil use 0.35
Quads (1 percent), and natural gas use 7.38 Quads (25 percent). The
EIA projections include consideration of issues related to decreased
electricity production from nuclear power, lower coal prices, lower
capital costs for coal-fired generating technologies, higher
electricity demand, and industry restructuring. Although the share of
coal generation declines in the 1998 estimate relative to earlier
projections (primarily due to restructuring considerations), the
projection for coal-fired fuel use in 2010 is higher than the 1997
projection and is essentially the same as that of EPA's ARD. Thus,
the Agency has not changed its projections related to emissions in
2010.
Figure 2-7 shows the projected growth of each utility fuel
between 1990 and 2010.90 Between 1990 and 2010, fuel consumption is
projected to change as follows: coal will increase by 29 percent, oil
will decrease by 48 percent, and natural gas will increase by
61 percent. Based on the recent EIA projection noted above, coal
consumption will increase by 26 percent, oil consumption will decrease
by 71 percent, and natural gas consumption will increase by 164
percent.
The projected increase or decrease in nationwide fuel consumption
noted above has been apportioned to only those units projected to be
in existence in 2010. The actual increased consumption (coal) would,
in most cases, be distributed among new units (existing units not
being able to increase their capacity factors to account for the
majority of the growth). These new units could be of various sizes
and be located at new or existing sites. However, since the Agency
can estimate neither the size nor the location of the new units, the
increased consumption has been allocated to existing units (in 2010)
for the analyses. This allocation is believed reasonable because
(1) many new units would be built on the site of existing utility
facilities (thus, "co-locating" the emissions) and (2) the analyses
are based on total fuel used (rather than on capacity factor, etc.).
The decrease in oil consumption could result in (or result from)
units being retired or in a decrease in capacity factor, or a mix.
The decreased consumption has been allocated among those oil-fired
units EPA believes will be operative in 2010.
Any new units built to accommodate the increased consumption
would be required to comply with all applicable NSPS and State and
local regulations. However, for the purposes of the analyses the
controls currently in use on the unit were considered to be in place
for the same unit with increased fuel consumption.
2-28
-------
Table 2-4. Fuel Use in the Electric Utility Industry by Fuel Type, Quadrillion Btu/yr
Fuel
Coal
Oil
Gas
EPA, projection
1990a
17
1.2
2.8
2010a
22
0.62
4.5
EIA, actual
1990b
16.19
1.25
2.88
1991b
16.03
1.18
2.86
1992b
16.21
0.95
2.83
1993b
16.79
1.05
2.74
1994b
16.90
0.97
3.05
1995b
16.99
0.66
3.28
1996b
17.93
0.73
2.80
EIA, projection
201 Oc
Reference
case
19.91
0.57
7.09
Forecast ranges
19.31 -21.09
0.56-0.6
6.37-7.89
201 Od
21.34
0.35
7.38
a Utility Data Institute.91
b U.S. Department of Energy, Energy Information Agency92
c Annual Energy Outlook. 1997.93
d Annual Energy Outlook. 1998.94
-------
I
co
o
Years
2010
Gas
Oil
Coal
Figure 2-7. Projected use of fuels by 2010 for utility industry. 89
-------
The Agency projects that 135 units will be retired during the
period from 1990 to 2010. These units have been removed from the
2010 analyses.
2.7.2 Title I and Title IV, Phase I and Phase II, Compliance
Strategy Impact
Phase I and Phase II requirements of the Acid Rain Program
establish a cap on the national, annual S02 emissions. To achieve
compliance with the requirements, utilities may use one or any
combination of the following strategies (among others) at any given
unit: (1) install flue gas scrubbers, (2) switch to a fuel that
contains less sulfur, and (3) reduce the capacity factor of the
Phase I unit to the extent that the unit is in compliance and provide
plans for replacing the reduced capacity. This reduction can be
accomplished by either: (1) energy conservation, (2) improved boiler
efficiency, (3) use of a designated sulfur-free (nuclear or hydro, but
not natural gas-fired) replacement, (4) use of a Phase II compensating
unit, or (5) purchase of emission allowances.
The Phase I requirements affected 261 generating units (435 with
substitution or compensating generating units) . 95 The 174 substitution
or compensating generating units are not included in the following
discussion. Examining the method used by the 261 Phase I generating
units to comply with the provisions, the following was found:96
• 53 percent (136 units accounting for 59 percent of the 1995
S02 emission reductions) switched to a coal that contains
less sulfur
• 27 percent (83 units accounting for 9 percent of the 1995
S02 emission reductions) purchased additional emission
allowances
• 16 percent (27 units accounting for 28 percent of the 1995
S02 emission reductions) installed flue gas scrubbers
• 2 percent (7 units accounting for 2 percent of the 1995 S02
emission reductions) were retired
• 2 percent (8 units accounting for 2 percent of the 1995 S02
emission reductions) either repowered using new boiler
technology, or switched to natural gas or low sulfur oil.
Each of the 27 units known to be installing scrubber units was
modeled with the scrubber unit in place for the 2010 scenario. The
EPA modeled the remaining 234 units by increasing their coal
consumption in proportion to ICF Resources, Inc. (2010) projections.
These increases were also weighted by the expected increased use of
western, low-sulfur coal.
Under Phase II of the Acid Rain Program, an additional
approximately 1,600 generating units will be covered by the year 2000.
2-31
-------
Although industry projections suggest an additional 25 units (at 10
plants) will install scrubbers to comply with Phase II, the EPA
believes that these units will comply with Phase II requirements by
using alternate methods.97 This assumption is based on several factors
including: (1) the increased availability of low-sulfur coal at
favorable prices; (2) the introduction of processes that reduce S02
emissions by 20 to 50 percent through partial cleaning of higher
sulfur coal, which allows for a variety of coal types to be utilized
(although the impact of these processes as Title IV control options is
uncertain at this time); and (3) the increased age and small size of
the affected units, giving the utility companies little incentive to
spend large amounts of capital on installing scrubbers.98
Many utility units will be implementing NOX controls to comply
with both Title I and IV requirements. This control may involve
switching from coal- or oil-firing to natural gas-firing (for at least
a portion of the year), improved combustion controls, or installation
of low-NOx burners, among other activities. If a fuel switch was
known, that switch was accounted for in the 1990 versus 2010 analysis.
No change in a unit's burner configuration (i.e., "old" versus new
low-NOx) was included in the 2010 scenario. The impact of low-NQ,
burner installation is discussed in chapter 13.
Under the Acid Rain Program, the rules for NOX control (40 CFR
Part 7b) require tangential-fired and dry-bottom wall-fired boilers
subject to Phase I S02 reduction requirements to meet annual average
NOX emission limits of 0.45 lb/MMBtu and 0.50 lb/MMBtu, respectively,
by January 1, 1996. Utilities can meet the Title I and IV
requirements by installing low-NOx burner technology or by averaging
emissions among several units.
Since the Interim Report to Congress, additional rules (e.g.,
revised NAAQS, revised NSPS, and Acid Rain Program for NOX and SOX)
have been promulgated that could pressure the electric utility
industry to consider options beyond those considered earlier. These
include additional fuel switching or the adoption of SCR or SNCR
rather than the addition of low-NOx burners because of tightening NOX
emission standards. Since limited data were available assessing the
HAP removal potential of SCR and SNCR, they were not addressed in the
2010 program output. Also, additional FGD units may be installed to
comply with the revised PM NAAQS (which impact on sulfate rather than
traditional PM). However, as no area has been determined to be in
non-compliance yet, no units have been planned and such installations
were not addressed in the 2010 analyses. To the extent that
additional FGD units are installed (for NSPS or NAAQS compliance) the
overall effect could be lower HAP emissions in 2010.
Under Phase II of the Acid Rain Program, the EPA established NOX
emission limits for all other boilers, including wet-bottom wall-fired
boilers and cyclones, by January 1, 1997; affected units must be in
compliance by January 1, 2000. 99 EPA also revised the emission limits
for dry-bottom wall-fired boilers and tangential-fired Phase II units.
2-32
-------
Particulate control devices may also need to be upgraded at
individual utility units to account for the different ash qualities of
any new coal being utilized to comply with S02 requirements or to
account for installation of low-NOx burners. In late 1993, the Utility
Data Institute (UDI) conducted a particulate control equipment survey
to identify those utility facilities that were either in the process
of upgrading their PM removal equipment or had definite plans to do so
in the near future.97 The survey was mailed to 286 utilities and
received a 68 percent response. No information was received for 831
units; 1,215 units indicated that no PM control equipment
modifications were planned. Modification plans were received for
132 units. The data received were analyzed for any potential impact
on HAP emissions.100 From the data, it appears that the modifications
are being made strictly to account for differences in ash quality as
coals are switched and not to effect an overall increase in PM control
efficiency. Therefore, for the 2010 scenario analysis, it has been
assumed that no change in PM control efficiency will occur since the
actual reported values do not vary significantly. In addition, this
assumption will account for any future degradation in PM control
performance. The validity of this assumption is borne out by
indications that some utility units are experiencing emissions
increases (as evidenced by continuous emission monitor excess emission
reports) following switches to lower sulfur coal and/or installation
of low-NOx burners.101 It is not known how transient these excess
emissions will be.
2.7.3 Compliance Strategy Impacts of Other Activities
Other activities, not directly related to CAA mandates, will also
impact on electric utility industry control strategies and emissions.
These include developments related to electricity industry
restructuring, such the Federal Energy Regulatory Commission's (FERC)
Open Access Rule (Order 888) finalized in April 1996, changes in the
energy production from other sources (e.g., accelerated retirement of
nuclear plants, increases in the use of biomass), overall national
demand for electricity, relative differences in fuel costs, and any
future "global warming" abatement initiatives. Responses to these
activities (i.e., specific control strategies) were not included in
the analyses for this report. As can be seen from the EIA projections
presented earlier, it is expected that, over the long-term, natural
gas will increase its share of the fossil-fuel generation. However,
it should be noted (see Table 2-4) that between 1995 and 1996 (the
first year of industry restructuring), coal consumption for
electricity generation increased by approximately one quad while
natural gas consumption decreased by approximately 0.5 quad.
Preliminary data for 1997 indicate that coal use continues to increase
while natural gas use has leveled off.102 It is not known what factors
are involved in this trend (e.g., nuclear outages, relative fuel
prices, seasonal weather conditions) but some parties believe that
restructuring is playing a role.103 In addition, it is not known how
this short-term trend may ultimately factor into long-term
projections.
2-33
-------
2.8 REFERENCES
1. Utility Data Institute. EEI Power Statistics Data Base.
Washington, DC. 1995. (excluding cogeneration facilities)
2. Singer, J. G. (ed.). Combustion, Fossil Power. Fourth Edition.
Combustion Engineering, Inc., Windsor, CT. 1991. p. 1-12.
3. Utility Data Institute. EEI Non-Utility Power Plant Data Base.
Washington, DC. 1993.
4. National Energy Foundation. Getting to Know Public Power. Salt
Lake City, UT. 1993.
5. U.S. Electric Utility Statistics. Public Power. American Public
Power Association, Washington, DC. January-February 1997.
pp. 53, 54.
6. §292.202, Subpart B-Qualifying Cogeneration and Small Power
Production Facilities. Part 292-Regulations under Sections 201
and 210 of the Public Utility Regulatory Policies Act of 1978
with regard to Small Power Production and Cogeneration. CFR 18,
Parts 280 to 399, revised as of April 1, 1992.
7.
8.
9.
Ref.
Ref.
2,
2,
Babcock
Wilcox.
10.
11.
12.
Ref.
Ref.
2,
2,
Elliot,
P.
P.
&
. 4-
. 12
29.
-15.
Wilcox. Steam, Its Generation and Use. Babcock &
New
P.
P.
T.
. 12
. 12
. C.
McGraw-Hill,
13.
14.
15.
16.
17.
18.
Ref.
Ref.
Ref.
Ref.
Ref.
Ref.
2,
2,
2,
2,
2,
2,
P.
P.
P.
P.
P.
P.
. 9-
. 9-
. 9-
. 9-
. 7-
. 1-
York. 1978. p. 10-3.
-4 .
-2 .
Standard Handbook of Powerplant Engineering.
Inc., New York. 1989. p. 4.50.
2 .
4 .
5.
30.
6.
13.
2-34
-------
19. Federal Energy Regulatory Commission. Cogeneration.
Washington, DC. 1985. pp. 3, 5.
20. Ref. 2, p. 1-16.
21. Damle, A. S., D. S. Ensor, and M. B. Ranade. Coal Combustion
Aerosol Formation Mechanisms: A Review. Aerosol Science and
Technology. Volume 1, No. 1. 1982. pp. 119-132.
22. Tumati, P. R., and M. S. Devito. Retention of Condensed/Solid
Phase Trace Elements in an Electrostatic Precipitator. Presented
at International Air Toxics Conference, Washington, DC. Electric
Power Research Institute, Palo Alto, CA. November 4-6, 1991.
p. 20.
23. U.S. Environmental Protection Agency. Control Techniques for
Particulate Emissions from Stationary Sources -- Volumes 1 and
2. EPA-450/3-81-005a, b. Office of Air Quality Planning and
Standards, Research Triangle Park, NC. 1982. p. 4.2-23.
24. Leith, D., and D. Mehta. Cyclone Performance and Design.
Atmospheric Environ. Volume 7. 1973. pp. 527-549.
25. Ref. 23, Vol. 1, pp. 4.3-20 and 4.3-22.
26. Buonicore, A. J., and W. T. Davis (eds.). Air Pollution
Engineering Manual. Air and Waste Management Association, Van
Nostrand Reinhold, New York. 1992. p. 95.
27. Ref. 23, Vol. 1, pp. 4.3-14 to 4.3-23.
28. White, H. J. Industrial Electrostatic Precipitation. Addison-
Wesley, Reading, MA. 1963.
29. Ref. 23, Vol. 1, pp. 4.5-22 to 4.5-29.
30. Calvert, S., J. Goldshmid, D. Leith, and D. Mehta. Wet Scrubber
System Study, Vol I: Scrubber Handbook. NTIS No. PB 213-016.
U.S. Environmental Protection Agency, Research Triangle Park, NC.
1972 .
31. Ref. 23, Vol. 1, p. 4.4-12.
32. McKenna, J. D., and J. H. Turner. Fabric Filter - Baghouses I,
Theory, Design, and Selection (A Reference Text). ETS, Inc.,
Roanoke, VA. 1989.
33. Donovan, R. P. Fabric Filtration for Combustion Sources:
Fundamentals and Basic Technology. Marcel Dekker, New York.
1985.
2-35
-------
34. Memorandum from J. H. Turner, RTI, to J. D. Cole, RTI.
March 17, 1993. Table presenting a comparison of particulate
matter collection systems.
35. Khoury, D. L. Coal Cleaning Technology. Noyes Data Corporation,
Park Ridge, NJ. 1981. p. 24.
36. Akers, D., C. Raleigh, G. Shirey, and R. Dospoy. The Effect of
Coal Cleaning on Trace Elements, Draft Report, Application of
Algorithms. Prepared for EPRI by CQ, Inc. February 11, 1994.
37. Letter from Burke, F. P., of CONSOL Inc., to W. H. Maxwell, EPA.
May 28, 1993. Use of USGS data in estimating the emissions of
air toxics.
38. Cavallaro, J. A., A. W. Deurbrouck, R. P. Killmeyer, and W.
Fuchs. Sulfur and Ash Reduction Potential and Selected Chemical
and Physical Properties of United States Coals, Executive
Summary. U.S. DOE Report DOE/PETC/TR-91/6 (DE91015938) .
Pittsburgh, PA. September 1991. pp. 7-9.
39. Ref. 35, p. 21.
40. Ref. 38, p. 1.
41. Akers, D. J. Coal Cleaning: A Trace Element Control Option.
Paper presented at EPRI Symposium on Managing Hazardous Air
Pollutants: State of the Art. Washington, DC. November 1991.
42. Satriana, M. New Developments in Flue Gas Desulfurization
Technology. Noyes Data Corporation, Park Ridge, NJ. 1981.
p. 10.
43. Ref. 42, pp. 13-14.
44. Ref. 42, p. 15.
45. Ref. 42, pp. 180-181.
46. Ref. 42, pp. 85-88.
47. Ref. 42, pp. 155-161.
48. Ref. 42, pp. 142-146.
49. Ref. 42, pp. 75-78.
50. Ref. 2, pp. 4-30 to 4-34, 12-6 to 12-8, 15-64 to 15-68.
2-36
-------
51. U.S. Environmental Protection Agency. Evaluation and Costing of
Nox Controls for Existing Utility Boilers in the NESCAUM Region.
EPA-453/R-92-010. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. December 1992.
52. U.S. Environmental Protection Agency. Summary of Nox Control
Technologies and Their Availability and Extent of Application.
EPA-450/3-92-004. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. February 1992. pp. 1-1 to 3-25.
53. 1991 Joint Symposium on Stationary Combustion Nox Control,
EPA/EPRI. March 25-28. 1991.
54. Ref. 2, p. 4-30.
55. Ref. 53, p. 2-3.
56. Radian Corporation. Combustion Modification NOX Controls for
Wall-Fired and Tangential-Fired Boilers. EPA contract No. 68-DO-
0125. Prepared for Air and Radiation Division, U.S.
Environmental Protection Agency. July 1991. p. 1.
57. Ref. 56, p. 1.
58. Low-N0x Combustion Retrofit Projects, Pittsburgh Energy
Technology Center. Pittsburgh, PA. Review, 6. Summer 1992.
p. 9.
59. Ref. 58, p. 9, Figure 2. Values represent emissions before the
1993 new source performance standards.
60. Ref. 58, p. 9.
61. Ref. 2, p. 9-4.
62. Ref. 2, p. 4-30.
63. Ref. 53, p. 2-1.
64. Ref. 56, p. 5.
65. Ref. 53, p. 2-4.
66. Ref. 26, p. 241.
67. Ref. 52, pp. 4-4, 4-25.
68. Ref. 52, p. 4-4.
69. Ref. 53, p. 3-19.
70. Ref. 52, p. 4-4.
2-37
-------
71. Ref. 53, p. 3-8.
72. Ref. 53, p. 2-8.
73. Borio, R., R. Lewis, D. Steen, and A. Lookman. Long-Term NOX
Emissions Results with Natural Gas Returning on a Coal-Fired
Cyclone Boiler. Presented at the EPA/EPRI Joint Symposium on
Stationary Combustion NOX Control, Bal Harbor, Florida. May 24-
27, 1993.
74. LaFlesh, R. C., R. D. Lewis, R. E. Hall, V. R. Kolter, and Y. M.
Mospan. Three-Stage Combustion (Reburning) Test Results from a
300-MWe Boiler in the Ukraine. Presented at the EPA/EPRI Joint
Symposium on Stationary Combustion NOX Control, Bal Harbor,
Florida. May 24-27, 1993.
75. Ref. 56, p. 2.
76. Ref. 53, pp. 2-7 to 2-14.
77. Ref. 53, p. 2-14.
78. Ref. 53, pp. 2-14, 2-15.
79. Ref. 53, p. 2-16.
80. Ref. 52, p. 4-32.
81. U.S. Department of Energy. Control of Nitrogen Oxide Emissions:
Selective Catalytic Reduction (SCR). Clean Coal Technology,
Topical Report Number 9. July 1997. pp. 14-15.
82. Ref. 53, p. 2-16.
83. Ref. 52, p. 4-35.
84. Ref. 52, pp. 4-35, 4-37.
85. Ref. 53, p. 2-17.
86. U.S. Department of Energy - Energy Information Agency (EIA).
Early Release of the Annual Energy Outlook 1998. Obtained from
EIAs website, "http://www.eia.doe.gov/oiaf/aeo98/
earlyrel.html" Washington, DC. November 1997. Figure 5. p. 5.
87. Ref. 53, p. 2-10.
88. ICF Resources Incorporated. Economic Analysis of the Title IV
Requirements of the 1990 Clean Air Act Amendments. Prepared for
the U.S. EPA, Office of Air and Radiation, Acid Rain Division.
Washington, DC. February 1994. p. 8.
2-38
-------
89. Utility Data Institute. EEI Power Statistics Data Base.
Washington, DC. 1992. (excluding cogeneration facilities)
90. Ref. 88, p. C-l.
91. Utility Data Institute. Interim Final Utility Toxics Report,
Chapter 2, Reference 1; ICR Resources, Incorporated. Interim
Final Utility Toxics Report, Chapter 2.
92. U.S. Department of Energy, Energy Information Agency. Table 2.6.
Energy Input at Electric Utilities. Obtained from EIAs website,
"ftp://ftp.eia.doe.gov/pub/energy.overview/ monthly.energy/
mer2-6" Washington, DC. January 1998.
93. Annual Energy Outlook. 1997. DOE/EIA-0383(97). December 1996.
Table B2, p. 126.
94. Annual Energy Outlook. 1998. EIA. December 1997. Table 2.
95. U.S. Department of Energy - Energy Information Agency (EIA). The
Effects of Title IV of the Clean Air Act Ammendments of 1990 on
Electric Utilities: An Update. DOE/EIA-0582(97) Office of Coal,
Nuclear, Electric and Alternate Fuels. Washington, DC. March
1997. p. 5.
96. Ref. 95, Table 2. p. 6.
97. Letter from Zeugin, L. B., Hunton & Williams, to W. H. Maxwell,
EPA. April 11, 1994. UDI Particulate Control Equipment Survey.
98. J. Makansi. New Supply Lines, Tools Keep Coal Switching in
Forefront Power. May 1994. p. 41.
99. Federal Register, 57(228), 1992. p. 55633.
100. Memorandum from W. H. Maxwell, EPA, to Electric Utility Air
Toxics Project Files. August 18, 1995. Comparison of PM control
efficiencies: new or upgraded vs. old units.
101. Letter from Schultz, D., EPA (Region 5), to W. H. Maxwell, EPA.
July 18, 1995. Future year (2010) projection of utility HAP
emissions.
102. Ref. 94.
103. Sullivan, Edward, Environmental Commissioner of the State of
Maine, et al., to Browner, Carol, EPA/AX. Letter and
attachments. Comments on industry restructuring. January 16,
1998.
2-39
-------
3.0 EMISSION DATA GATHERING AND ANALYSIS
3.1 LITERATURE REVIEW AND BACKGROUND
Prior to the beginning of this study, the Agency conducted a
literature search of available nonradionuclide HAP emission and
control information and assessed the usefulness of these data. Much
of the data had been gathered over an extended time period using a
wide variety of dated, and sometimes ill-defined, sampling and
analytical techniques. Many of these techniques, including the method
for mercury, have since been replaced with more accurate methods. The
data in the literature exhibited extensive variability in the reported
concentrations of HAPs in emissions (sometimes varying by several
orders of magnitude). There was often insufficient documentation of
the techniques and assumptions used to distinguish the reliable data
from the unreliable data.
In addition, many of these literature data were gathered at
laboratory or pilot-scale installations or from utility units that did
not reflect the configurations of the current utility unit population.
Again, there was often insufficient documentation of the design
parameters or process operating conditions to assess the validity of
the data or the impact of the process operating conditions on the
nonradionuclide emissions.
Because of these deficiencies, the Agency was not able to use the
prior existing data (prior to 1990) in control strategy analysis or to
project the data for nationwide application in the health hazard
assessment. The EPRI and DOE conducted major test programs during the
period beginning in 1990 to obtain HAP data from the utility industry
and coordinated these programs and test methodology processes with the
EPA. These new data from field testing became available for this
report beginning in late 1993 .
To obtain the necessary nonradionuclide chemical HAP emission
test data, two avenues were followed. The first was to pursue
cooperation with industry and DOE test programs, and the second
involved Agency testing for HAP organics. The EPRI performed 36
emissions tests at 34 locations of their member companies for
approximately 25 of the 189 HAPs listed in section 112(b) of the Act.
Of these locations, test reports were available for 29 sites (and
particulate control data for one additional site) in time for
inclusion in the health hazard assessment. These tests encompassed
coal-, oil-, and gas-fired boilers of several firing types and
emission control technologies. Emission test sites were selected
based on industry utilization (e.g., the largest percentage of coal-
fired units are dry bottom, use bituminous coal, with ESPs for PM
control). This approach allowed the acquisition of data for the
broadest spectrum of the utility industry in the most cost-effective
manner. Some of the EPRI emission test sites were DOE Clean Coal
Technology (CCT) sites, which provided for the acquisition of HAP data
before and after installation of controls for NOX, an important element
3-1
-------
in the acid rain program (under Title IV of the Act). This
information will be helpful in determining the implications of the
acid rain program on HAP emissions. In the test programs, samples
were collected before and after each emission control device when
feasible. The Northern States Power Company (NSPC) also provided
eight test reports from five of its coal-fired plants for testing
performed from 1990 to 1992.
The DOE, through what is now its Federal Energy Technology Center
(FETC), initiated contract activities in mid-1992 for a HAP emission
sampling program at eight coal-fired utility units. The contracts
were awarded in early 1993 and the sampling was completed at seven
sites over the period from June to December of 1993 . The DOE program
was similar in nature and scope to that of EPRI, although the number
of facilities evaluated was much smaller under the DOE program. The
timing of the DOE program was such that the data were available
concurrently with those from the EPRI studies and could be analyzed
for this report.
The EPA was involved with the design and test method selection
for both the EPRI and the DOE test programs. The Agency also cofunded
a field validation of several mercury emission test methods at a coal-
fired utility boiler with EPRI, including those methods that measure
the various species of mercury that may be emitted from a utility
boiler.1
For the EPRI program, the Agency independently developed a matrix
of the industry and established that the types of plants selected for
the EPRI program were the same types that would have been selected for
an EPA emission test program. For the DOE program, the Agency had the
opportunity to provide input into the type of plants that should be
selected. All emission test reports from both programs, and from
individual company tests, were reviewed by the EPA for completeness,
adherence to accepted sampling and analytical techniques, and proper
unit operations (typical information missing from the existing
literature-based database). The Agency provided support for the
onsite quality assurance/quality control activities performed during
several of the DOE emission tests.
The EPA also completed the initial development of the Fourier
transform infrared (FTIR) spectrometry field testing system and system
validation for real-time, simultaneous measurement of approximately
120 gaseous organic HAPs. Validation tests for the FTIR at a coal-
fired unit were conducted in February 1993. 2 The FTIR system was
utilized in emission testing by the EPA at five utility sites. The
FTIR system is a lower-cost and much more flexible measurement
technology than those currently available for sources of organic HAPs.
To examine the magnitude of HAP emissions from utility units, the EPA
conducted emissions testing with FTIR as a screening level analysis.
Since few HAPs were detected with FTIR testing at these five utility
sites, the EPA decided not to use the FTIR test results to estimate
HAP emissions from utility units at this time.
3-2
-------
For inclusion in this report, a total of 58 tests were conducted
at 52 sites using FTIR and conventional sampling and analysis methods
from the EPRI, the DOE, the NSPC, and the EPA. Although 58 test
reports were received by EPA in time for inclusion in this study,
4 contained data that could not be used in the emission factor program
(EFP) described in section 3.4. These reports were excluded because
measurements were not made between the boiler and the PM control
device. This exclusion resulted in a test containing only a fuel
analysis and stack emission numbers, which did not allow analysis of
control system effectiveness. For draft versions of this report,
emissions were estimated for 1990 based on the 48 usable tests then
available. Estimates for 1994 are based on the 54 usable tests.
Emission estimates for 1994 use the same modeling procedure as for
1990, but with minor revisions as noted hereafter.
Data reliability and the precision and accuracy of the analytical
techniques for each test were addressed by the individual test
contractors in their test reports. Where the contractor had major
concerns about the quality of the data or found gross departures from
expected precision or accuracy of a particular test analysis, the EPA
refrained from using the data in its computations.
3.2 POLLUTANTS STUDIED
As many as possible of the 189 HAPs listed in section 112 (b) were
included in this study. Table A-l (Appendix A) lists the organic HAPs
that were detected at least once in the utility test data (excluding
FTIR-detected data), the estimated nationwide HAP emissions in 1990
and 1994, and the projected nationwide emissions for 2010.
The Agency's Office of Radiation and Indoor Air (ORIA) has
completed a report on radionuclide emissions from the utility
industry. The results of this study, along with their impact on
public health, are included in chapter 9.
3.3 DESCRIPTION OF EMISSION TEST PROGRAMS
At the beginning of this study, the utility industry was
characterized. Through the use of the EEI Power Statistics Database
(1991) from the UDI,3 a matrix was developed showing the ranking of
utility unit configurations from the most to the least prevalent.
Table B-l (Appendix B) shows these unit configurations down to a unit
type that accounts for only 1.05 percent of the fossil-fuel-fired unit
megawatts in the United States (plus any additional unit types tested
that were below this cutoff). The matrix was then used only as a
guide to gather data on the largest number of unit configurations
possible with the available resources by targeting the most prevalent
unit types. It should be noted that the totals in Table B-l were
taken from the 1991 EEI Power Statistics Database and do not correlate
with the 1994 industry statistics given in chapter 2.
3-3
-------
The emission test reports used in this study were produced for
various government agencies as well as for nongovernment and industry
groups (discussed above). Although various test contractors performed
this emission testing, certain specific testing protocols were
followed. Table B-2 (Appendix B) provides a list of all the sites
that were available for this report and were tested under the DOE, the
NSPC, the EPRI, and the EPA test programs. The table also shows the
type of fuel burned and the emission controls applied to the boiler
system. In some cases, the controls are pilot-scale units applied to
a slipstream from the boiler flue gas system. The contractor who
tested the boiler and the date of the test report are also given.
3.4 DEVELOPMENT OF HAP EMISSION TOTALS
To estimate emissions of HAPs from fossil-fuel-fired electric
utility units (>25 MWe), the EPA developed the EFP. This program
incorporates unit configuration data from individual units as well as
emission testing data to compute estimated emissions. An explanation
of the program and several assumptions about the data and how they
were used are described here.
3.4.1 Program Operation
Emissions of HAPs considered in this study consist of two types:
trace elements and organic compounds. Trace elements exist in the
fuel when fired, whereas the organic HAPs are mostly formed during
combustion and postcombustion processes. Different programming
methods are required for handling the two types of HAPs. Program
diagrams for modeling trace element emissions are shown in Figure 3-1
for coal and Figure 3-2 for oil and gas. The two figures differ only
in treatment of the fuel before trace elements reach the boiler.
Figure 3-3 shows the program diagrams for modeling organic HAP
emissions.
3.4.2 Data Sources
The EFP was built to accept data from two sources. The first
source of data is a data input file containing plant configurations,
unit fuel usage, and stack parameters. This input file was based on
the UDI/EEI Power Statistics database (1991 and 1994 editions) and an
extract from Production Costs, U.S. Gas Turbine and Combined-Cycle
Power Plants (for 1994 estimates). These databases were composed of
responses from electric utilities to the yearly updated DOE Energy
Information Administration (EIA) Form EIA-767.
The second data file is an emissions modification factor (EMF)
database. This database contains information from emissions tests
conducted by EPRI, DOE, and the electric utility industry.
The program first searches the input file for the type of fuel
burned and the amount of fuel consumed per year in an individual unit.
If the fuel type is coal, the EFP then looks for the coal's State of
origin. Origin is important because the trace elements in coal are
addressed by coal type (bituminous, subbituminous, and lignite) and
3-4
-------
UDI/EEI plant configuration
information
USGS coal data
(by State and coal type)
USGS coal data
(by State and coal type)
Apply boiler TE emission
modification factor
What is
particulate
matter (PM) control
type?*
Apply the PM control TE
emission modification factor
Apply SO control TE emission
modification factor
kg/yr of specific trace element
exiting unit stack
Figure 3-1. Trace elements in coal.
*Taken from UDI/EEI data.
3-5
-------
UDI/EEI plant configuration
information
Oil
Natural gas
Used fuel oil No. 6 (residual)
for all oil types
Trace elements in oil taken
from plant testing
Trace elements in natural gas
taken from plant testing
(only two sets of data)
Used a denisty of 8.2 Ib/gal for
feed rate calculation
Trace elements (TE) to boiler
Apply boiler TE emission
modification factor
What is
particulate
^matter (PM) contrq]^
type?*
Apply the PM control TE
emission modification factor
Apply S02 control TE emission
modification factor
kg/yr of specific trace element
exiting unit stack
Figure 3-2. Trace elements in oil and natural gas.
'Taken from UDI/EEI data.
3-6
-------
Oil
Obtain unit's fuel
consumption
Obtain unit's fuel
consumption
Obtain unit's fuel
consumption
For individual HAPs,
find the median
Ib/trillion Btu emission
factor for a specific
HAP1
For individual HAPs,
find the median
Ib/trillion Btu emission
factor for a specific
HAP2
For individual HAPs, find
the geometric mean
kg/10 9 cu ft emission
factor for a specific
HAP
Individual fuel
consumption x
emission factor x heat
content of 150,000
Btu/gal
Individual fuel
consumption x
emission factor
Individual fuel
consumption x
emission factor x
higher heating value
for bituminous coal
(12,688 Btu/lb coal)
Individual fuel
consumption x
emission factor x
higher heating value
forsubbituminous
coal (9,967 Btu/lb
Individual fuel
consumption x
emission factor x
higher heating value
for lignite coal (6,800
Btu/lb coal)
Convert into kg/yr
stack emission for
HAP
'Only oil-fired units were used to obtain these emission factors.
'Only coal-fired units were used to obtain these emission factors.
'Only gas-fired units were used to obtain these emission factors.
Figure 3-3. Organic emissions.
3-7
-------
State of origin in the United States Geological Survey (USGS)
database, which contains analyses of 3,331 core and channel samples of
coal. The samples come from either the top 50 economically feasible
coal seams in the United States during 1990, or from seams associated
with 1991 coal receipts for electric utility plants.
3.4.3 Operational Status of Boilers
The operational status of units was taken from the UNIT_90.dbf
file of the EEI/UDI Power Statistics database (1991 edition addressing
1990 data) or the similar file for 1994 data. Only units that were
listed as either operational or on standby were used in the EFP. It
was found that 151 units were listed as being on standby in the
EEI/UDI Power Statistics database but were actually on indefinite
standby and, thus, did not emit any HAPs. These units were excluded
from the nationwide emissions totals in Appendix A. Other units
listed on indefinite standby (i.e., no fuel burned) were excluded from
1994 emission estimates.
Only coal-fired, oil-fired, and natural gas-fired units were
included in the EFP. This decision was made because units using these
fuels make up an overwhelming majority of the fossil-fuel-fired electric
utility units with a capacity of >25 MWe.
Anthracite was disregarded as a fuel because of the limited
number of units burning this type of coal.4 Four units burning
anthracite coal were assigned to burn bituminous coal for program
computations.
The 1990 EEI/UDI database had a number of gaps in the fuel
consumption data. Some of these gaps were filled by data supplied
voluntarily by the industry. To address the remaining gaps, the
available data were plotted and point-slope equations were fit to
estimate fuel consumption.5 These equations involved plotting
nameplate megawatts (modified to take into account the unit's capacity
factor) against fuel usage. If the fuel usage and the unit capacity
factor in 1990 were not given, 1989 fuel consumption data were used.
If 1989 data were not available, the geometric mean of the 1980-1988
EEI fuel consumption data was used. When all other options had been
tried unsuccessfully, an average fuel consumption of units rated
within ±5 MW of the unit with unknown fuel usage was used. Similar
problems in the 1994 UDI/EEI database were solved by using 1990 data
where possible and by similar methods to those stated above when not
possible.
Utility units may burn coal that originated from several States;
however, in the EFP each coal-fired unit was assigned a single State
of coal origin.6 The State of origin used in the EFP was the State
that contributed the highest percentage of the unit's coal.
3.4.4 Trace Element Concentration in Fuel
The USGS database contains concentrations of trace elements in
coal that were extracted from the ground but does not include analyses
3-8
-------
of coal shipments. The concentrations of trace elements in coal in
the ground and in coal shipments to utilities may differ because, in
the process of preparing a coal shipment, some of the mineral matter
in coal may be removed. Since approximately 77 percent of the eastern
and midwestern bituminous coal shipments are cleaned 7 to meet customer
specifications on heat, ash, and sulfur content, a coal cleaning
factor was applied to most bituminous coals in the EFP.8 Two
exceptions were bituminous coals from Illinois and Colorado, for which
analyses were on an as-shipped basis representative of the coal to be
fired.
For a unit that burned bituminous coal, the feed rate in
kilograms per year (kg/yr) of trace elements to the boiler was
determined from the average trace element concentration in the coal, a
coal cleaning factor, and the annual fuel consumption rate. No coal
cleaning factors were applied to lignite and subbituminous coals. See
Appendix D for listings of trace elements in coal, coal cleaning
factors, and equations (Nos. 1 and 2 in Table D-2) used in the EFP.
Oil-fired organic HAP exit concentration calculations included a
150,000-Btu/gallon (gal) heating value for oil. An oil density of
8.2 Ib/gal was also used.
An emission rate for each organic HAP emitted from gas-fired
units was extracted from the test reports. There were only two test
reports on gas-fired units that analyzed organic HAPs, and a geometric
mean emission rate of each observed organic HAP was used. This rate
in kilogram HAP/109 cubic feet was then multiplied by the unit's gas
consumption to obtain a kilogram HAP/year stack emission rate of each
specific HAP.
3.4.5 HC1 and HF Concentration in Fuel
To obtain hydrogen chloride (HC1) or hydrogen fluoride (HF)
emissions from the boiler, emission factors were derived by performing
mass balances for chloride and fluoride, then converting these
balances to the equivalent levels of HC1 or HF throughout the boiler
system.9 For example, for each Ib/hr of chloride in the feed coal at
one of the test sites, 0.63 Ib/hr of HC1 was found in the gas stream
leaving the boiler. Similarly for HF, the boiler emissions were 0.64
Ib/hr for each Ib/hr of fluoride in the coal. For ease of
programming, the HC1 and HF emissions were addressed starting in the
fuel. This programming was done by multiplying the chloride and
fluoride concentrations in the fuel constituents by 0.63 or 0.64,
respectively. The resulting numbers allowed direct conversion into
boiler emissions that could be further modified for systems with PM
control or S02 control. For the 1990 emission estimates, before
obtaining further test reports, the factors were 0.61 for HC1 and 0.56
for HF.
The chloride concentrations were not available for coals from the
following States: Alaska, Illinois, Indiana, Iowa, Missouri, Utah,
3-9
-------
and Washington. Chloride concentrations were assigned, as shown in
Table 3-1, for coals originating from these States.10
3.4.6 Emission Modification Factors for Inorganic HAPs
To address the partitioning of the HAP stream through the
combustion and pollution control process, partitioning factors known
as EMFs were developed from inorganic HAP testing data. The EMFs are
fractions of the amount of a HAP compound exiting a device (boiler or
air pollution control device [APCD]) divided by the amount of the same
HAP compound entering that device.11 These EMFs are averaged by taking
the geometric mean of similar devices (e.g., all oil-fired tangential
boilers, all cold-side ESPs). Geometric means are used because of the
presence of outlying data points, the small amount of data, and the
general fit of the data to a log-normal curve. These geometric means
are then applied to the kg/yr feed rates entering the boiler, the
effect of which either reduces or leaves unchanged the emissions that
pass through them. Those EMFs calculated as being greater than 1.0
(i.e., more material exiting a device than entering it) were set to
equal 1.0.
Nearly all EMFs were computed from three data samples before and
three data samples after the particular device. When all six data
samples for a particular EMF computation were nondetects, the EPA
decided to disregard the EMF. As such, EMFs were computed when there
was at least one detected sample among the six measured samples. The
EMFs developed for 1990 emission estimates were revised to include
additional test report data for 1994 emission estimates. Appendix D
discusses in more detail the methodology used to develop emission
totals.
The EMFs were computed with data from different test reports but
for similar devices (i.e., cold-side ESPs, front-fired boilers in oil-
fired units). The data from coal-fired units were not segregated by
State of coal origin. The EMFs from devices are generally segregated
into only coal-, oil-, or gas-fired bins.
The EFP itself uses EMFs to partition the emissions as they
proceed from the fuel through the unit to the stack exit as follows.
The average concentrations of metallic HAPs in an individual fuel by
State (based on USGS data) were multiplied by the amount of fuel that
the unit burned in 1990 or 1994. After accounting for coal cleaning
(bituminous coal only), the emission concentration of an inorganic HAP
was converted to an emission rate in kg/yr entering the boiler. The
emission rate entering the boiler was then modified by EMFs for the
boiler, particulate control device (when applicable), and the
S02 control device (when applicable).
As stated above, these geometric mean EMFs were then applied to
the fuel HAP concentration estimates and the kg/yr fuel feed rates
entering the boiler, the effect of which either reduced or left
unchanged the emissions that passed through it, depending on the value
of the EMF.
3-10
-------
Table 3-1. Assigned Chloride and HC1 Concentrations in Coal, by
State of Coal Origin 10
State
Alaska
Illinois
Indiana
Iowa
Missouri
Utah
Washington
Conversion of assigned ppmw chloride to
assigned HCI ppmw
54x0.63 =
1,136x0.63 =
1,033x0.63 =
1,498x0.63 =
1,701 x0.63 =
220x0.63 =
104x0.63 =
Assigned ppmw HCI in coal
34.0
715.7
650.8
943.7
1,071.6
138.6
65.5
Appendix C contains all of the EMFs used to develop the 1990 and
1994 unit emission estimates for inorganic HAPs.
3.4.7 Acid Gas HAPs
The method used with HCI or HF emissions allowed direct
conversion from coal chlorine or fluorine content into boiler
emissions, as described earlier, that could be further modified for
systems with PM control or S02 control.
Hydrochloric acid and HF EMFs for PM and S02 control devices were
developed with data from test reports in which contractors conducted
tests individually for HCI, chlorine, HF, and fluorine before and
after each control device. These tests were in contrast to the
remaining tests for which HCI and HF values were estimated or omitted
rather than measured.
The next steps after obtaining amounts of HCI or HF leaving the
boiler were to construct EMFs for the PM control device, and then for
the S02 control device. Using chlorine as an example, the measured
amount of HCI entering the PM control device (in kg/yr with suitable
conversion factors) was compared with the measured amount of HCI
leaving the PM control device. Using these two quantities, an EMF was
formed as described in section 3.4.6.
In the final step, EMFs were formed for HCI and HF through the
S02 control device based on the measured mass of HCI or HF entering
that device (leaving the PM control device) and the mass measured at
the exit of the S02 control device. However, a modification was
required to account for flue-gas bypass around the S02 control device.
A portion of the flue gas is bypassed to maintain S02 removal at the
minimum permitted amount. This action is used as a means of reducing
energy required to reheat the flue gas for effective plume rise from
the stack. In developing the HCI and HF EMFs for wet FGDs and dry
scrubbers, the effect of flue gas bypass was treated by analyzing
3-11
-------
utility test data from the four plants (of eight tested) that used
bypasses, reviewing municipal waste incinerator results that showed a
typical HC1 or HF removal efficiency of 95 percent, and having
discussions with industry representatives. Based on the 95 percent
removal efficiency coupled with the measured values for quantity of
flue gas bypassed, an industry average effective value for flue gas
bypass in 1994 was estimated. The value was assumed to be 15 percent
(17 percent for 1990 data) for wet FGDs and 14 percent (for 1990 and
1994 data) for dry scrubber systems. These assumptions were used only
in the development of HC1 and HF EMFs.12
3.4.8 Organic HAPs
Because organic HAPs were not always tested at the entrance and
exit of each control device in the emissions testing, all organic HAP
emissions were addressed by examining the test data and determining
the concentration of a particular HAP exiting the stack. Organic HAP
concentrations were obtained from emission test reports.
Organic stack emissions from coal-fired boilers were first
determined on an emission factor basis (Ib/trillion Btu) to account
for different coal heating values, then converted to a rate basis
(kg/yr of individual HAP). This procedure was necessary because
different coal ranks had different heating values. For example, it
would require burning more lignite to achieve the same heat input to
the boiler as burning bituminous coal. These values were determined
as averages for each type of coal (see Table 3-2).13
If stack emission or APCD exit emission data were reported as
nondetected, and, if at least one-third of the data samples at the
inlet of the APCD were detected concentrations with values comparable
to the nondetected outlet values, EPA used the inlet data directly as
a measure of outlet concentration at the stack. If the outlet
nondetected values were significantly different from the inlet
detected values, the data were not used. For each individual organic
HAP observed in testing, a median concentration was obtained. This
fuel-specific median concentration was then individually multiplied by
each utility unit's fuel consumption. The result was a fuel-specific
emission rate for all organic HAPs that were observed at least once
during testing.
3.4.9 Model Estimates for the Year 2010
Emission estimates for 2010 were derived from the same basic 1990
model described above. However, changes to input files were made to
accommodate expected changes in fuel usage by fuel type, generating
capacity, and responses to Phases I and II of the 1990 amendments
under Title IV. The details of these expected changes are described
in section 2.7.
In summary, the input files for the 2010 analyses were modified
to account for the expected increases in nationwide coal and natural
gas, and expected decrease in oil, use in the utility industry
resulting from the industry growth described in section 2.7.1. Units
3-12
-------
Table 3-2. Average Higher Heating Values of Coal1
Class and group9
Agglomerating
character
Fixed carbon limits,
% (dry, mineral-
matter-free basis)
Equal or
greater
than
Less
than
Volatile matter limits,
% (dry, mineral-
matter-free basis)
Equal or
greater
than
Less
than
Calorific value limits,
Btu/lb (moist," mineral-
matter-free basis)
Equal or
greater than
Less
than
Average
1. Bituminous
1 . Low-volatile bituminous coal
2. Medium-volatile bituminous coal
3. High-volatile A bituminous coal
4. High-volatile B bituminous coal
5. High-volatile C bituminous coal
High-volatile C bituminous coal
commonly
agglomerating'
«
«
«
«
agglomerating
78
69
...
—
...
—
86
78
69
—
...
—
14
22
31
—
...
—
22
31
...
—
...
—
...
—
14,000"
13,000"
1 1 ,500
10,500
...
—
...
14,000
13,000
11,500
Average of Averages (Value used in EFP for bituminous coal)
14,000
13,500
12,250
11,000
12,688
II. Subbituminous
1. Subbituminous A coal
2. Subbituminous B coal
3. Subbituminous C coal
nonagglomerating
«
«
...
—
...
...
—
...
...
...
...
—
...
10,500
9,500
8,300
1 1 ,500
10,500
9,500
Average of Averages (Value used in EFP for Subbituminous coal)
1 1 ,000
10,000
8,900
9,967
III. Lignitic
1. Lignite A
2. Lignite B
nonagglomerating
«
—
...
—
...
—
...
—
...
6,300
...
8,300
6,300
Average of Averages (Value used in EFP for lignite coal)
7,300
6,300
6,800
a This classification does not include a few coals, principally nonbanded varieties, which have unusual physical and chemical
properties and which come within the limits of fixed carbon or calorific value for high-volatile and Subbituminous ranks. All of
these coals either contain less than 48 percent dry, mineral-matter-free fixed carbon or have more than 15,500 moist, mineral-
matter-free Btu per pound.
b Moist refers to coal containing its natural inherent moisture but not including visible water on the surface of the coal.
0 It is recognized that there may be nonagglomerating varieties in these groups of the bituminous class, and there are notable
exceptions in high-volatile C bituminous group.
d Coals having 69 percent or more fixed carbon on the dry, mineral-matter-free basis shall be classified by fixed carbon, regardless
of calorific value.
known to be retired during the period 1990-2010 were removed from the
input files. Announced options (described in section 2.7.2) taken for
compliance with the Acid Rain Program (e.g., coal switching, FGD
installation) were accounted for in the input files. Thus, emission
changes resulting from these activities were included in the 2010
analyses.
3-13
-------
However, as noted in section 2.7.3, control strategies taken in
response to such activities as industry restructuring and global
warming abatement initiatives were not included in the 2010 analyses.
To the extent that such strategies include increased use of FGD units
or natural gas, the projected 2010 HAP emissions could be over-
estimated in this report. For example, analyses performed to assess
compliance with the revised national ambient air quality standards
(NAAQS) for ozone and PM indicate that mercury emissions in 2010 may
be reduced by approximately 16 percent (11 tons per year) over those
projected in this report, primarily due to increased FGD use and fuel
switching.15 To the extent that such strategies include the increased
use of coal, as may be happening at least in the short term under
industry restructuring, the projected 2010 HAP emissions could be
underestimated in this report.
3.5 SELECTED ESTIMATED NATIONWIDE HAP EMISSIONS
Based on the screening assessment presented in chapter 5, a
subset of HAPs was determined to be of potential concern to public
health and was, therefore, given a priority label for further
analysis. Table 3-3 presents estimated 1990, 1994, and 2010
nationwide emissions of this subset of HAPs from utility units (see
also Table A-l, Appendix A). With expected changes in input data and
new test data for EMFs, the opportunity was taken to improve the model
for more effective use in estimating 1994 emissions. Depending on
reported and projected fuel switching and fuel use, units brought on
line or taken off line, the revisions to EMFs between 1990 and 1994,
and improvements in the model's handling of fuel combustion, estimated
HAP emissions across the 1990 to 2010 span may increase uniformly,
decrease uniformly, peak in 1994, or show a minimum in 1994. These
varied responses among the HAPs are thought to occur primarily because
of industry changes rather than changes to the model.
3.6 COMPARISON OF EFP ESTIMATES WITH TEST DATA
Comparisons were made between test data from 19 utility boiler
stacks and 1994 predicted emissions for the same plants using the
EFP.16 Results suggest that the EFP performs as expected (i.e., across
a range of boilers and constituents the average of the predicted
values agrees well with the average of the reported values). This
close prediction occurs even with large differences between predicted
and reported values for individual boilers and constituents. It
should be stressed that the EFP is designed to provide a reasonable
estimate of nationwide emissions based on summing a large number of
imperfect boiler estimates that are expected to cluster about a
reasonable estimate of the true value. The EFP program also provides
reasonable estimates for the boiler-specific emissions used for the
exposure modeling analyses (see chapter 6). However, there are
uncertainties and, based on an uncertainty analysis, it is estimated
that emission values may be over- or underestimated by as much as
roughly a factor of three for any specific boiler.
3-14
-------
Table 3-3. Selected Nationwide HAP Emissions
Coal-fired electric utility plants
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Hydrogen chloride
Hydrogen fluoride
2,3,7, 8-tetrachlorodibenzo-p-dioxin (TEQ)
n-nitrosodimethylamine
Oil-fired electric utility plants
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Nickel
Hydrogen chloride
2,3,7, 8-tetrachlorodibenzo-p-dioxin (TEQ)
Natural-gas-fired electric utility plants
Arsenic
Nickel
Formaldehyde
Selected nationwide HAP
emissions (estimated) in
tons/year for 1990
60.93
7.13
3.33
73.27
75.47
163.97
45.80
143,000
19,500
0.000097
5.84
5.02
0.46
1.71
4.74
10.58
9.28
0.25
392.8
2,860
0.000016
0.15
2.19
35.62
Selected nationwide HAP
emissions (estimated) in tons/year
for 1994
55.81
7.93
3.15
61.60
61.77
167.72
51.34
134,000
23,100
0.000121
6.09
3.51
0.40
1.09
3.91
8.92
7.30
0.19
322.0
2,100
0.000009
0.18
2.42
39.23
Selected nationwide HAP
emissions (estimated) in tons/year
for 2010
70.61
8.20
3.82
87.43
86.89
219.02
59.74
1 55,000
25,700
0.000108
7.73
2.54
0.23
0.86
2.40
5.35
4.70
0.13
198.2
1,450
0.000003
0.25
3.49
56.58
w
I
I-1
Ul
-------
For three elements and 19 boilers, averages for estimates of
three individual elements (arsenic, chromium, and nickel) were
different from the test values by +60, -32, and -3 percent,
respectively. The highest individual difference between predicted and
reported values was represented by a factor of 2,600.
Table 3-4 presents comparisons for the individual plants for
arsenic, chromium, and nickel. Averages for each element and for the
combination of all three elements are also given. The reported values
are ratios of EFP estimates to measured values in terms of pounds of
element emitted per trillion Btu heat input. Plants 1 through 17 fire
coal, whereas plants 18 and 19 fire oil. At least one of the plants
fires a combination of coal and petroleum coke.
Possible reasons were examined for large differences between
projected and actual emissions. In the 1990 EFP, only one fuel (and
if coal, from only one State) was assumed to be burned. Although the
1994 version of the EFP was designed to accommodate multiple types of
fuel, the one-State-of-origin restriction for coal was still used.
However, at least one of the plants burned combinations of coal and
petroleum coke, but the EFP recognized only coal from one State. The
petroleum coke used by the plant had nickel concentrations that may
have been more than 100 times higher than that found in the Montana
coal assigned to that plant by the EFP, and concentrations of nickel
in the plant's ash on the order of 1,000 times higher than that found
in coal ash. In this case, the EFP underpredicted actual nickel
emissions by factors of up to 2,600 as mentioned above. The EFP was
not sufficiently detailed to recognize mixtures of coal and petroleum
coke (i.e., the EFP did not include the nickel contributed by the
petroleum coke). Plants burning such mixtures will likely have their
nickel emissions underestimated by the EFP.
3.7 CHARACTERISTIC PLANT EMISSIONS
To give the reader a better grasp of the potential emissions of
selected HAPs from an individual utility unit, a set of characteristic
units was chosen (one for each fuel type). The EFP and organic HAP
stack emission factors were then used to determine the units'
projected HAP emissions of concern (according to the health hazard
assessment).
In presenting the characteristic coal-fired unit, the EPA looked
for an existing utility unit that had the characteristics of a typical
coal-fired unit in the United States. Once the specific plant was
chosen, its 1994 HAP emissions of concern (projected by the EFP and
organic HAP stack emission factors) were listed (see Table 3-5).
The most important parameter of the characteristic oil and gas-
fired plants (see Tables 3-6 and 3-7) is their fuel consumption, as
there are usually no control devices to reduce emissions. The fuel
consumptions chosen are the averages of each fuel type (oil or gas).
3-16
-------
Table 3-4. Comparison of Utility Boiler Emissions from EFP
Estimates and from Tests3
Plant number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Average: EFP/test
Arsenic EFP/test
0.12
0.24
1.84
1.66
9.83
0.92
2.00
0.18
0.0010
0.65
0.20
0.09
0.04
4.13
0.01
6.40
2.01
0.05
0.03
1.60
Chromium EFP/test
0.14
0.10
0.48
1.30
1.63
0.26
0.51
1.17
0.0036
0.27
0.19
0.0041
2.40
1.15
0.04
0.72
1.23
—
—
0.68
Nickel EFP/ test
0.0035
0.30
0.13
1.97
1.08
0.36
0.81
10.00
0.0004
0.20
0.16
0.0730
1.37
0.17
0.14
0.93
0.65
0.02
0.03
0.97
EFP = emission factor program
a Values presented are the ratio of emission factor program estimates to test data in terms of lb/1CP Btu.
The characteristic unit emissions in these three tables are
actually projected emission outputs from the EFP for three existing
units. They are chosen for having the most prevalent fuel,
boiler/furnace, and control device type in their fuel class (coal,
oil, or gas). They are also chosen for having megawatt capacities
that are the average for their fuel class.
It should be noted, however, that characteristic emissions are
based on 1994 fuel consumption values, and the emission testing (on
which the EFP is based) was performed under essentially steady-state
conditions (with little or no variation from the baseline operating
condition). Therefore, the characteristic emissions from testing are
a snapshot in time. In reality, emissions of HAPs are not constant,
steady-state values, but fluctuate with operating conditions as well
as changes in fuel. That is why the fuel heat content was used as an
indicator of emissions rather than the plant capacity or utilization.
3-17
-------
Table 3-5. Emissions from a Characteristic Coal-Fired Electric Utility Plant (1994;
Megawatts
Coal source
Fuel consumption
Particulate matter control device
Sulfur dioxide control device
1994
approximately 325 MWe
Kentucky (bituminous, sulfur content of 1 .5 - 3.5 %)
approximately 600,000 tons/year
Electrostatic precipitator (cold-side)
Compliance fuel/State implementation plan (SIP)
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Hydrogen chloride
Hydrogen fluoride
2,3,7,8-tetrachlorodibenzo-p-dioxin (TEQ)
n-nitrosodimethylamine
Selected HAP emissions (estimated) in tons/year for 1994
0.050
0.0081
0.0023
0.110
0.021
0.092
0.045
191.8
14.31
0.000000103
0.0052
w
I
-------
Table 3-6. Emissions from a Characteristic Oil-Fired Electric
Utility Plant (1994)
Megawatts
Fuel
Fuel consumption
Particulate matter control device
Sulfur dioxide control device
1994
approximately 160 MWe
Residual oil (sulfur content < 1 .5 %)
approximately 640,000 barrels / year
None
Compliance fuel/State implementation plan (SIP)
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Nickel
Hydrogen chloride
2,3,7,8-tetrachlorodibenzo-p-dioxin (TEQ)
Selected HAP emissions (estimated) in tons/year for 1994
0.0062
0.0002
0.0140
0.0062
0.014
0.019
0.0012
1.69
9.4
2.27E-08
Table 3-7. Emissions from a Characteristic Natural Gas-Fired
Electric Utility Plant (1994)
Megawatts
Fuel consumption
Particulate matter control device
Sulfur dioxide control device
1994
approximately 240 MWe
approximately 4,600,000,000 cubic feet / year
N/A
N/A
Arsenic
Nickel
Formaldehyde
Selected HAP emissions (estimated) in tons/year for 1994
0.0003
0.0041
0.067
3.8 UNCERTAINTY ANALYSIS OF EMISSION FACTOR PROGRAM
In order to quantify the uncertainty in the EFP output, the EFP
was subjected to an uncertainty analysis using the statistical method
known as a Monte Carlo analysis. The results of this analysis are
presented in section 6-14.
3-19
-------
3.9 REFERENCES
1. U.S. Environmental Protection Agency. Evaluation of Method 29
for the Measurement of Mercury Emissions in Exhaust Gases from a
Coal Fired Electric Utility. EPA-454/d-95-001. Office of Air
Quality Planning and Standards, Source Characterization Group B.
Research Triangle Park, NC. May 1995.
2. U.S. Environmental Protection Agency. Fourier Transform Infrared
(FTIR) Method Validation at a Coal-fired Boiler. EPA-454/R-95-
004. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. July 1993.
3. Utility Data Institute. EEI Power Statistics Data Base.
Washington, DC. 1992. (excluding non utility units)
4. Ref. 3.
5. Memorandum from Cole, J. D., Research Triangle Institute to
Maxwell, W. H., Environmental Protection Agency. February 3,
1993. Addressing fuel consumption gaps in the EEI power
statistics data base data.
6. Memorandum from Heath, E., RTI to Maxwell, W. H., EPA. July 14,
1993. State of coal origin used in the computer emission
program.
7. Akers, D., C. Raleigh, G. Shirley, and R. Dospoy. The Effect of
Coal Cleaning on Trace Elements, Draft Report, Application of
Algorithms, prepared for EPRI by CQ Inc. February 11, 1994.
8. Memorandum from Heath, E., RTI to Maxwell, W. H., EPA. April 5,
1994. Proposed coal cleaning factors.
9. Memorandum from Turner, J. H., RTI, to Cole, J. D., RTI.
December 7, 1997. Methodology for determining 1994 HC1 and HF
concentrations from utility boilers.
10. Memorandum from Heath, E., RTI to Maxwell, W. H., EPA. May 27,
1994. USGS data gaps in chloride concentrations for seven
states.
11. Memorandum from Cole, J. D., RTI, to Maxwell, W. H., EPA. March
31, 1994. Emission factor memorandum.
12. Memorandum from Cole, J. D., RTI to Maxwell, W. H., EPA. May 9,
1994. Emission modification factors for HC1 and HF including FGD
system bypass.
13. Singer, J. G., ed. Combustion Fossil Power, 4th ed. Combustion
Engineering, Incorporated, Windsor, CT. 1991. pp. 2-3, modified
table.
3-20
-------
14. Ref. 12, p. 2-3.
15. U.S. Environmental Protection Agency. Regulatory Impact Analysis
for the Particulate Matter and Ozone National Ambient Air Quality
Standards and Proposed Regional Haze Rule. Appendix A:
Emissions and Air Quality and Appendix H: Economic Impact
Analysis and Supporting Information. Office of Air Quality
Planning and Standards. Research Triangle Park, NC. July 16,
1997.
16. Memorandum from Turner, J. H., RTI, to Cole, J. D., RTI.
December 8, 1997. Comparison of RTI 1994 emission model
projections with test data.
3-21
-------
4.0 INTRODUCTION FOR THE HEALTH HAZARD RISK ANALYSIS
4.1 INTRODUCTION AND BACKGROUND
The EPA partially evaluated the potential hazards and risks for
the year 1990 and for the year 2010. A significant portion of the
analyses focuses on inhalation risks due to utility emissions within a
50 km radius of each facility (i.e., local analysis). The analyses of
long-range transport or regional analysis (i.e., emissions dispersion
and exposure outside of 50 km) and multipathway assessment (e.g., risk
due to ingestion and dermal exposure) were limited, mostly
qualitative, and considered only a few pollutants. This situation
does not necessarily mean that inhalation exposure within 50 km is the
most important route of exposure. For some of the HAPs emitted from
utilities (e.g., mercury and dioxins), noninhalation exposure through
ingestion is likely to be the dominant route of human exposure.1'2
The estimates of risks due to inhalation exposure presented in
this report are the incremental increased risks due to utility
emissions only. For the most part, this assessment does not consider
exposure to emissions from other sources and does not consider
background levels of the HAPs in the environment. However, background
concentrations were evaluated to a limited extent and are discussed
briefly in later sections of this report.
This chapter begins with a summary of risk assessment principles
and guidelines as used by the EPA and discussions of pertinent reports
such as the National Research Council report Science and Judgement in
Risk Assessment3 and the EPA Science Policy Council's (SPC's) Guidance
for Risk Characterization.4 Section 4.2 presents the general approach
and methods for this health hazard risk assessment. Section 4.3
discusses health effects data. Section 4.4 describes the methodology
used in the inhalation exposure assessment, and section 4.5 describes
the methodology for estimating inhalation risks.
4.1.1 Principles of Risk Assessment
Risk assessment is a multidisciplinary evaluation of factual
information as a basis for estimating and evaluating the potential
health effects that individuals or populations may experience as a
result of exposure to hazardous substances. Risk assessments
typically involve both qualitative and quantitative information.
Risk estimates describe the nature and likelihood of adverse
effects and the probabilities that these health effects will occur in
an exposed population. Numerical risk estimates can be calculated for
two categories of adverse health effects:
• Risk of developing cancer
• The likelihood of developing adverse health effects other
than cancer (e.g., asthma).
4-1
-------
To derive statements of risk or the likelihood of adverse health
effects, quantitative information on exposure is combined with
information on toxicity. This process is different for carcinogens
and noncarcinogens due to the underlying assumptions that cancer is a
nonthreshold phenomenon and that thresholds exist for adverse health
effects other than cancer (i.e., noncancer effects).
In 1983, the National Academy of Sciences (NAS) established a
framework to guide risk assessments by Federal agencies.5 As defined
by the NAS, risk assessment consists of four steps:
• Hazard assessment, or hazard identification
• Dose-response assessment
• Exposure assessment
• Risk characterization.
Hazard identification is the review of relevant toxicologic,
biological, and chemical information to determine whether or not a
pollutant may cause adverse health effects. It is a qualitative
assessment of the potential of a pollutant to increase the incidence
of an adverse health effect if exposure to the pollutant occurs.6'7
Dose-response assessment defines the relationship between the
degree of exposure (or amount of dose) observed in animal or human
studies and the magnitude of the observed adverse health effects.
This usually includes a quantitative measure of adverse health effects
for a range of doses. For carcinogens, dose-response data are used to
calculate quantitative estimates of the increased risk of developing
cancer per unit of exposure (e.g., inhalation unit risk estimates
[lUREs]). For noncarcinogens, dose-response data are used to
calculate "safe" levels (e.g., inhalation reference concentrations
[RfCs]) .
Exposure assessment estimates the extent of pollutant exposure
via various routes (e.g., oral, inhalation, dermal) to individuals or
populations. For air pollutants, this often involves the application
of exposure models.
Risk characterization is the integration of the hazard
identification, dose-response, and exposure assessments to describe
the nature, and often to estimate the magnitude, of the health risk in
a given population.6 The risk characterization also includes a
presentation of the qualitative and quantitative uncertainties in the
assessment.
Risk assessment should not be confused with risk management.
Risk management is the process of developing and weighing policy
options and selecting appropriate actions. Risk management integrates
the results of the risk assessment with other information such as
economic, engineering, political and social factors and uses this
integrated information to make policy and regulatory decisions.
4-2
-------
4.1.2 U.S. EPA Risk Assessment Guidelines
Several publications were used to establish the methods for this
risk assessment. The methods generally follow the risk assessment
guidelines published by the U.S. EPA in 1986.7 Other sources consulted
for preparation of this assessment are discussed briefly below.
4.1.3 Risk Assessment Council (RAG) Guidance
The RAG of the EPA evaluated EPA risk assessment practices in
1992 and recommended guidance on risk assessment focusing on the risk
assessment-risk management interface and risk characterization.8 Major
elements relevant to this study are summarized below:
• Complete presentation of risk is needed including
discussions of uncertainty and statements of confidence
about data and methods used. The assessment should clearly
identify all assumptions, their rationale, and the effect of
reasonable alternative assumptions on the conclusions and
estimates.
• Assessors should use consistent and comparable risk
descriptors. For example, assessments should include
descriptions of risk to individuals and to populations, and
presentations of central tendency and worst-case portions of
the range of risk; if feasible, highly exposed or highly
susceptible groups should be identified.
4.1.4 NAS Report Science and Judgement in Risk Assessment
In 1994, the National Research Council (NRC) of the NAS released
a report Science and Judgement in Risk Assessment, which contains a
critique of existing EPA methods and several recommendations for
improvements.3 A few of the recommendations important for the utility
assessment are described briefly here.
The NRC stated that default options are a reasonable way to deal
with uncertainty about underlying mechanisms in selecting methods and
models. However, default options should be explicitly identified, and
the basis explained fully.
The NRC believes the EPA should undertake an iterative approach
to risk assessment. An iterative approach starts with relatively
inexpensive screening techniques to estimate chemicals without health
concerns followed by more resource-intensive levels of data gathering
and model application.
It is appropriate to use "bounding" estimates for screening
assessments to determine whether further levels of analysis are
necessary. For example, if there are no health impacts even in a
worst-case assumption scenario, then it may not be necessary or
desirable to proceed with further analysis.
4-3
-------
4.1.5 SPC's Guidance for Risk Characterization
In 1995, the SPC of the EPA provided guidance for characterizing
risk.4 A few points are briefly summarized here.
Risk assessors should be sensitive to distinctions between risk
assessment and risk management. Risk assessors are charged with
(1) generating a credible, objective, realistic, and scientifically
balanced analysis; (2) presenting information on hazard, dose
response, exposure, and risk; and (3) clearly describing confidence,
strengths, uncertainties, and assumptions.
The risk characterization should include qualitative and
quantitative descriptions of risk. Both high-end and central tendency
descriptors should be used to convey the variability in risk levels
experienced by different individuals in the population. The
assessment should identify and discuss important strengths,
limitations and uncertainties, and degree of confidence in the
estimates and conclusions. The assessment should also include
discussions of data quality and variability.
4.2 GENERAL APPROACH AND METHODS FOR THE UTILITY HEALTH HAZARD RISK
ASSESSMENT
Emissions of HAPs can be a threat to public health if sufficient
exposure occurs. For many HAPs, exposure through inhalation is the
major concern. However, humans can also be exposed to HAPs via
indirect pathways (multipathway) such as through ingestion or dermal
exposure to HAPs through other media such as food, water, or soil that
has been contaminated by the deposition of the HAPs. Indirect
exposure is primarily a concern for HAPs that are persistent and
bioaccumulate.
To assess the public health concerns due to emissions of HAPs
from utilities, the EPA conducted inhalation and multipathway exposure
and risk analyses. First, a screening assessment was conducted on 67
HAPs potentially emitted from utilities to determine priority HAPs.
After the screening assessment was completed, further analyses were
conducted for the priority HAPs. In addition to the inhalation risk
assessment, the EPA conducted multipathway analyses of radionuclides,
mercury, arsenic, and dioxins; a long-range transport modeling
analysis for mercury, arsenic, chromium, nickel, cadmium, and lead;
and a limited qualitative assessment of the potential hazards due to
multipathway exposure to a few other persistent, bioaccumulative HAPs.
Chapter 5 presents the screening assessment. Chapter 6 presents
the inhalation risk assessment for 14 priority HAPs. Chapter 7
presents an assessment of mercury. Chapter 8 presents a qualitative
discussion of lead and cadmium. Chapter 9 presents the assessment for
radionuclides. Chapter 10 and 11 present screening level multipathway
assessments for arsenic and dioxins, respectively; and Chapter 12
discusses potential impacts of hydrogen chloride and hydrogen fluoride
emissions.
4-4
-------
4.3 HEALTH EFFECTS DATA: HAZARD IDENTIFICATION AND DOSE RESPONSE
Health effects data include qualitative and quantitative data on
hazard identification and dose response. These data are closely
related and evaluated concurrently in toxicologic studies. Therefore,
this section of the report includes summary discussions of both. For
detailed information on health effects data for seven of the priority
HAPs emitted from utilities, the reader is referred to Appendix E.
Most of the health effects data used were obtained from EPA's
Integrated Risk Information System (IRIS). IRIS is an online database
maintained by the EPA, which contains chemical- specific health risk
information. The data provided in IRIS have been reviewed by EPA work
groups and represent Agency consensus.9 Primarily, EPA-verified risk
values were used in this study. However, for HAPs without IRIS data,
health data from other toxicologic data sources were used. If other
data sources were used, they are indicated by footnotes in tables, end
notes, or discussed in the text.
4.3.1 Hazard Identification for Carcinogens
Animal and human cancer studies are evaluated to determine the
likelihood that a chemical causes cancer in humans. The evidence for
each chemical is determined to be sufficient, inadequate, or limited.
Other types of experimental evidence (e.g., in vitro genotoxicity
studies) may be used to support the epidemiological or animal bioassay
results.7'10 The EPA uses a weight-of-evidence, three-step procedure to
classify the likelihood that the chemical causes cancer in humans. In
the first step, the evidence is characterized separately for human
studies and for animal studies. The human studies are examined
considering the validity and representativeness of the populations
studied, any possible confounding factors, and the statistical
significance of the results. The animal studies are evaluated to
decide whether biologically significant responses have occurred and
whether the responses are statistically significant. Second, the
human and animal evidence is combined into an overall classification.
In the third step, the classification is adjusted upward or downward,
based on an analysis of other supporting evidence. Supporting
evidence includes structure-activity relationships (i.e., the
structural similarity of a chemical to another chemical with known
carcinogenic potential), studies on the metabolism and
pharmacokinetics of a chemical, and short-term genetic toxicity
tests.6'7 The result is that each chemical is placed into one of the
five categories listed in Table 4-1.
4.3.2 General Discussion of Dose Response
The NAS5 defined dose-response assessment as:
"...the process of characterizing the relation
between dose of a chemical administered or
received and the incidence of adverse health
effects in exposed populations and estimating the
incidence of the effect as a function of human
4-5
-------
Table 4-1. Weight-of-Evidence (WOE) Classification
Group
A
B1
B2
C
D
E
Description
Known Human Carcinogen
Probable Human Carcinogen
Probable Human Carcinogen
Evidence in Humans
Limited Human Data Are Available
Sufficient Evidence in Animals and Inadequate or No
Possible Human Carcinogen
Not Classifiable as to Human
Carcinogenicity
Evidence of Noncarcinogenicity for Humans
exposure to the agent. It takes account of
intensity of exposure, age pattern of exposure,
and possibly other variables that may affect
response, such as sex, lifestyle, and other
modifying factors."
In general, as dose increases, so does the probability that an
adverse effect will occur. Critical to a dose-response assessment is
the basic assumption that thresholds exist for particular compounds
and particular health effects, and thus doses below the threshold
would not result in adverse effects. Thresholds may exist if the body
has the ability to detoxify or compensate for exposures to pollutants
or if multiple numbers of cells perform the same function. When doses
increase to the point that the body can no longer accommodate or
compensate for the exposure to pollutants, adverse health effects can
be observed and the likelihood of effect increases with increased
dose. For "nonthreshold" toxicants, it is assumed that there is no
threshold concentration or dose below which health effects do not
appear and that any exposure means an increase in risk.
The EPA assumes that cancer is a nonthreshold disease; that is,
any exposure to a chemical carcinogen, no matter how low, contributes
to an increased lifetime probability (i.e., risk) of developing
cancer. In contrast, chemicals causing health effects other than
cancer are typically defined as having a threshold exposure
concentration or dose below which adverse health effects are not
expected to occur. The threshold concept influences the way in which
dose-response modeling or dose-response assessment is done.
Assessments of carcinogens and noncarcinogens are conducted separately
and are based on different assumptions and methods. Information for
carcinogens and noncarcinogens is discussed separately in this
section.
4-6
-------
4.3.3 Dose-Response Evaluation for Carcinogens
For chemicals that have been classified as carcinogens (WOE = A,
B, or C), the dose-response data are evaluated; and, if data are
adequate, then the EPA calculates quantitative estimates of the
increased risk of developing cancer per unit of exposure. For
example, for air pollutants, an IURE is calculated. The IURE for a
pollutant is the estimated increased risk (upper limit probability) of
a person developing cancer from breathing air containing a
concentration of 1 microgram of the pollutant per cubic meter (//g/m3)
of air for a lifetime (70 years). The EPA also calculates oral unit
risk estimates for assessing cancer risks from ingestion exposure.6'7'9
Since risks at low exposure levels cannot be measured directly, a
number of mathematical models have been developed to extrapolate from
high to low dose to calculate the unit risk estimates. The linearized
multistage model, which is the default model generally used by the
EPA, leads to a plausible upper limit to the risk that is consistent
with some proposed mechanisms of carcinogenesis. The true risk is
unlikely to exceed the value predicted by the linear multistage model
and may be lower; as low as zero is a possibility.7 For most HAPs
included in this assessment, the EPA has used the linear multistage
low-dose extrapolation model. However, there are a few important HAPs
with WOE ratings of "A" (e.g., chromium VI, arsenic) for which the EPA
used other linear extrapolation models. The lUREs for these HAPs are
also considered upper limit estimates of the risks at low
concentration because of the use of linear high- to low-dose
extrapolation and other factors. Table 5-1 presents a summary of the
EPA-verified cancer health effects data for HAPs emitted from
utilities. Table 5-1 also contains some health effects data that are
not EPA-verified.
The EPA assumes that, for carcinogenesis, no threshold for dose-
response relationships exists or that, if one does exist, it is very
low and cannot be reliably identified. As a result, any increase in
dose is associated with an increase in risk of developing cancer.
Although a number of theories exist to explain the process of
carcinogenesis, the multistage process is the most widely accepted.
The multistage process consists of three distinct stages: initiation,
promotion, and progression.11 One reason the multistage process is so
well accepted is that it has been demonstrated experimentally for a
number of carcinogens and has been shown to adequately describe
carcinogenesis in the cells of some animal tissues, including the
skin, lung, liver, and bladder.12 Individual carcinogens can affect
one or more of these stages.
The method for deriving lUREs based on animal data is different
than the method used for deriving lUREs based on human data. When
animal data are used, EPA typically determines the 95th percentile
confidence limit of the mean of the dose-response curve, then
extrapolates linearly down to zero. When human data are used, EPA
typically determines the "maximum likelihood" estimate of the dose-
response curve, then extrapolates linearly down to zero.
4-7
-------
There are factors involved with the human occupational data that
may result in high- or low-biasing effects, including uncertainties in
the estimation of individual exposures and the assumption that the
susceptibility of the exposed workers in the epidemiology studies is
equivalent to the susceptibility of the general population.
4.3.4 Long-Term Noncancer Health Effects Data
Pollutants can cause a variety of noncancer effects including
neurological, reproductive, developmental, and immunological toxicity.
Noncancer effects can be reversible or irreversible and can occur
following acute (short-term) exposure or chronic (long-term) exposure.6
Subchronic and chronic animal and human studies are evaluated to
determine potential adverse noncancer effects and the estimated doses
or exposure concentrations that cause those effects. If data are
sufficient, the EPA calculates an inhalation RfC, which is an estimate
(with uncertainty spanning perhaps an order of magnitude) of the daily
inhalation exposure of the human population (including sensitive
subgroups) that is likely to be without appreciable risk of
deleterious effects during a lifetime. The RfC is derived based on
the assumption that thresholds exist for certain toxic effects such as
cellular necrosis but may not exist for other toxic effects such as
carcinogenicity. The RfC is calculated as follows: EPA reviews many
human and/or animal studies to determine the highest dose level tested
at which the critical adverse effect does not occur—i.e., the no-
observed-adverse-effect level (NOAEL)—or the lowest dose level at
which the critical adverse effect is observed, the lowest-observed-
adverse-effect level (LOAEL). The NOAEL from an animal study is
adjusted for exposure duration and respiratory tract differences
between animals and humans. EPA then applies uncertainty factors to
adjust for the uncertainties in extrapolating from animal data to
humans (10), and for protecting sensitive subpopulations (10). Also,
a modifying factor is applied to reflect professional judgment of the
entire database. The inhalation RfC considers toxic effects for both
the respiratory system (portal-of-entry) and for effects peripheral to
the respiratory system (extrarespiratory effects). Exposures below
the RfC are not likely to be associated with adverse noncancer health
effects including respiratory, neurologic, reproductive, developmental,
and other effects. In this report the RfC is expressed in micrograms
of pollutant per cubic meter of air (ug/m3) . The EPA also calculates a
similar value, called the reference dose for assessing ingestion
exposure and noncancer hazards. The RfD is expressed in units of
mg/kg/d. Doses below the RfD are not expected to result in adverse
noncancer health effects. The EPA considers reproductive and
developmental effects when establishing RfCs and RfDs. If data are
absent, an uncertainty factor is often added to adjust the RfC or RfD
downward. Doses or concentrations above the RfD or RfC do not
necessarily indicate that adverse health effects will occur.9'13'14 As
the amount and frequency of exposures exceeding the RfC or RfD
increases, the possible occurrence of adverse effects in the human
population also increases. When exceedances of the RfD or RfC are
predicted, the data on exposure and health effects should be evaluated
4-8
-------
further to determine the data quality, uncertainties, degree of
exceedance, and the likelihood, frequency, and severity of potential
adverse effects. Evaluating this information helps to characterize
the public health concerns.
The EPA ranks each RfC as either low, medium, or high in three
areas: (1) confidence in the study on which the RfC was based, (2)
confidence in the database, and (3) overall confidence in the RfC.13'14
Table 5-2 presents the EPA-verified RfCs and other health effects
information for HAPs identified in the emissions data.
4.3.5 Short-Term Noncancer Health Effects Data
Short-term exposure to HAPs can also cause adverse noncancer
health effects. There are no EPA-verified acute health effects
benchmarks available for the priority HAPs. However, reference
exposure levels (RELs) for acute exposures were obtained from the
California Air Resources Board's Risk Assessment Guidelines for the
Hot Spots Program.15
4.3.6 Summary of Health Effects Data Sources
As mentioned, IRIS was the primary source of information on
health effects. However, other sources were also consulted such as
the Toxicological Profiles published by the Agency for Toxic
Substances and Disease Registry (ATSDR), the monographs published by
the International Agency for Research on Cancer (IARC), and various
EPA and non-EPA documents.
4.4 METHODOLOGY FOR ESTIMATING INHALATION EXPOSURE FOR LOCAL ANALYSIS
Exposure assessment is the determination or estimation
(qualitative or quantitative) of the magnitude, frequency, duration,
and route of exposure. An exposure assessment for air pollutants
typically has four major components:
• Emissions characterization
• Environmental fate and transport
• Characterization of the study population
• Exposure calculation.
This section summarizes the local inhalation exposure assessment
approach including discussions of the Human Exposure Model (HEM),
data, default options, and limitations. The long-range transport
exposure analysis is explained in section 6.6.
4.4.1 Emissions Characterization
The emissions data gathered from 52 utility units (described in
chapter 3) were used as the basis for estimating emissions of HAPs
from 684 utility plants in the U.S. As described in chapter 3, a
computer program was developed to estimate emissions from each utility
unit based on boiler type, electric output, fuel type, and APCDs.
This resulted in average annual emissions estimates for each HAP from
all 684 utility plants. The emissions estimates are believed to be
4-9
-------
reasonable estimates of the emissions from the utility plants.
However, there are uncertainties in the emissions estimates. The
emissions estimates, which were calculated using the geometric mean of
the test data, are believed to be central tendency estimates. That
is, the true emissions could be higher or lower than predicted with
the emissions model. Based on an uncertainty analysis, the EPA
predicts that the emissions estimates for any individual plant are
likely to be within a factor of plus or minus three of actual
emissions (see chapters 3 and 6 for further discussion of emissions
data and emissions estimates).
4.4.2 Atmospheric Fate and Transport
To arrive at long-term (annual) average ambient air
concentrations within 50 km of the facility, air dispersion modeling
was conducted using the HEM, which utilizes the Industrial Source
Complex Long-Term, version 2 (ISCLT2) dispersion model (see Appendix F
for details). The ISCLT2 was used to estimate atmospheric fate and
transport of HAPs from the point of emission to the location of
exposure. The ISCLT2 uses emissions source characterizations and
meteorological data to estimate the transport and dispersion of HAPs
in the atmosphere and to estimate the ambient HAP concentrations
within 50 km of each source (i.e., local analysis). Plant-specific
parameters needed for modeling (e.g., stack heights, stack
temperature, stack exit velocity, stack diameter, latitude, and
longitude) were obtained from the UDI/EEI database. Emissions
estimates, also needed as input to the model, were obtained from the
analysis described in chapter 3. Long-range transport (beyond 50 km)
was also addressed and is described in chapters 6 and 7.
The ISCLT2 uses meteorological data in the form of STability
ARray (STAR) data summaries. The STAR summaries contain joint
frequencies of occurrence of windspeed, wind direction, and
atmospheric stability. These factors are combined into an overall
frequency distribution. The meteorological database is based on
hourly surface observations obtained mostly from the Office of Air
Quality Planning and Standard's (OAQPS's) Technology Transfer Network
(TTN). The TTN contains data files of surface observations from
National Weather Service locations (primarily airports) across the
United States and its Territories. The STAR summaries combine several
available years (typically 6 years) into one long-term estimate of the
location's dispersion characteristics. In all cases, the
meteorological data from the site (out of a possible 349 sites)
nearest each plant were used in modeling each utility plant's
emissions. In addition, there are two smaller databases that provide
average mixing height and temperature by atmospheric stability class.
Every STAR site has a matching temperature database. However, the
mixing height database contains information for only 74 sites; as with
the STAR summaries, the nearest site is always selected.
4.4.3 Characterization of Study Population
Census data from 1990, which are the most current and
comprehensive data available, were used in estimating population
4-10
-------
exposures. The data were available on the "block" level, containing
6.9 million records. For each plant, all census blocks within 50 km
were identified and used to estimate local exposure.
4.4.4 Exposure Calculations
Exposure is calculated by multiplying the population (i.e.,
number of people) by the estimated air concentration to which that
population is exposed. The HEM exposure algorithms pair the air
concentration estimates produced by ISCLT2 with the census information
contained in the population database.
Within the HEM, the ISCLT2 calculates air concentrations at
numerous grid points within 50 km of each source. For this study,
grid points were placed around the source along 16 radials, spaced
every 22.5 degrees, at distances of 0.2, 0.5, 1.0, 2.0, 5.0, 10.0,
20.0, 30.0, 40.0, and 50.0 km from the source, for a total of 160 grid
points (which is the default setting). Except for receptors that are
very close to the stack, HEM calculates the air concentration at the
population centroid (the population center of the census block) by
interpolating between the values at the receptors surrounding the
centroid. Exposures were calculated by multiplying the number of
people living within a census block and the modeled air concentration
at the centroid of the census block. When the population blocks are
within 0.5 km of the plant, the population is distributed to each grid
point within 0.5 km to more realistically account for actual locations
of people. (In this region, the areas associated with census blocks
are larger than the sections in the polar grid and thus it is logical
to spread people out by assigning people to grid points rather than
block centroids). Exposure is calculated by multiplying the grid
point concentrations by the number of people assigned to the grid
point. For a more detailed description of the HEM, see Appendix F.
4.5 METHODOLOGY FOR ESTIMATING QUANTITATIVE INHALATION RISKS
Numerical estimates can be calculated for two categories of
adverse health effects:
• Risk of developing cancer
• Likelihood of noncancer health effects.
To derive statements of risk, quantitative information on exposure was
combined with information on dose response.
4.5.1 Estimating Cancer Inhalation Risks
For this analysis, the HEM calculated the cancer risk from
inhalation using standard EPA risk equations and assumptions. The
risk equation, which is a linear, nonthreshold model, defines the
exposure-response relationship. The estimate of the inhalation
exposure concentration (ug/m3) is multiplied by the IURE to calculate
risks for exposed persons who are assumed to be exposed on average to
the modeled ambient concentration of the carcinogen for a lifetime.
Risks are generally expressed as either individual risk or population
4-11
-------
risk. By the nature of the exposure and risk assessment models, the
estimated risks are expressions of the risks associated only with
exposure to utility emissions.
4.5.2 Individual Cancer Risk
Individual risk is commonly used to express risk and is defined
as the increased probability that an exposed individual would develop
cancer following exposure to a pollutant. Individual cancer risks can
be calculated by multiplying the estimated long-term ambient air
concentration (ug/m3) of a HAP (i.e., exposure estimate) by the IURE.
The IURE generally represents an upper bound estimate of the increased
risk of developing cancer for an individual exposed continuously for a
lifetime (70 years) to a specific concentration (e.g., 1 ug/m3) of a
pollutant in the air. The true cancer risk due to exposure to any
particular HAP is unknown and unlikely to be higher than that
predicted with the IURE and could be lower, possibly as low as zero.
If the highest modeled ambient air concentration occurs in an
area (i.e., census block) where no people are known to reside, it is
assumed that, theoretically, a person could be exposed to this
concentration (e.g., someone could move to this location); therefore,
the EPA calculates an estimated risk based on the assumption that
someone is exposed to the concentration. The risk calculated in this
situation is termed the MEI risk. Hence, the MEI risk is the
estimated risk to a theoretical individual exposed to the highest
estimated long-term ambient concentration associated with an emission
source. If the highest modeled ambient air concentration occurs in an
area where people are known to reside, the EPA again calculates an
estimated risk. The risk calculated in this situation is termed the
Maximum Individual Risk (MIR) risk. Hence, the MIR represents the
increased cancer risk to an individual exposed at the highest
estimated long-term ambient concentration in the area in which people
are assumed to reside. In this report, both the MEI risk and the MIR
are calculated assuming that a person is exposed to the modeled
long-term ambient pollutant concentration for a lifetime (i.e., 70
years). By the nature of the assumption, the MEI estimates must be
equal to or greater than the MIR. For this study, the MEI and MIRs
were either the same or very similar. For oil-fired utilities, the
MEI and MIR were exactly the same because the highest concentration
occurred in an area where people are known to reside. For coal-fired
utilities, the MEI risk was slightly higher than the MIR risk. For
example, the MEI risk for arsenic from coal-fired utilities was 3 x 10 ~6,
and the MIR for arsenic from coal-fired utilities was 2 x 10~6.
4.5.3 Population Cancer Risk
Population risk is an estimate that applies to the entire
population within the given area of analysis. Two population risk
descriptors are:
• The probabilistic number of health effects cases estimated
in the population of interest over a specified time period
(e.g., number of cancer cases per year) or cancer incidence
4-12
-------
• The percentage of the population, or the number of persons,
exposed above a specified level of lifetime risk (e.g., I0~s) .
Each modeled ambient HAP concentration level is multiplied by the
estimated number of people exposed to that level and by the IURE,
providing an estimate of cancer incidence for a 70-year lifetime
exposure. These risk values are summed to give aggregate risks for
the population within the study area (i.e., the total estimated excess
cancer cases in the exposed population). This lifetime risk estimate
is divided by 70 years to calculate annual incidence in units of
cancer cases per year.
4.5.4 Distribution of Individual Risk within a Population
The HEM estimates a distribution of individual risks throughout
the exposed population. The risk distribution presents an estimate of
the number of people exposed to various levels of risk (e.g., the
number of people who are exposed to individual risk levels above 10 ~7,
10-s, ID'5, or ID'4) .
4.5.5 Aggregate Inhalation Cancer Risk
The HEM calculates risk from individual HAPs and does not
calculate total risk for the mixture of pollutants from a single
source. To calculate total risks from the emissions of the mixture of
HAPs, the MIR and cancer incidence attributed to each individual HAP
were identified for each power plant. The MIR and cancer incidence
were then added across HAPs for each plant. This addition is
consistent with the EPA's default procedure for assessing mixtures.7
The highest total MIR across all plants was identified and the total
cancer incidence was summed across all plants. Given the structure of
the HEM output, it is not possible to calculate total risk (summed
across all carcinogens) for the entire exposed population. Therefore,
the population distribution by total risk is not presented in this
report.
4.5.6 Estimating Noncancer Inhalation Risks
The concepts of individual and population risks also apply to
noncancer risks. However, cancer risks and noncancer risks are
estimated differently. The noncancer result is not a measure of risk,
but rather indicates the possibility for an adverse health effect. To
assess potential noncancer health effects, the EPA evaluated exposure
to the individuals predicted to receive the maximum modeled
concentration.
Unlike cancer risk characterization, noncancer risks are not
expressed as a probability of an individual suffering an adverse
effect. Instead, the potential for noncancer effects is evaluated by
comparing an exposure concentration over a specified period of time
(e.g., a lifetime) with a toxicity benchmark (e.g., the inhalation
RfC) .
4-13
-------
4.5.7 Inhalation Hazard Quotient (HQ)
The HQ, a ratio of exposure (E) to the RfC, is commonly
calculated. The HQ indicates whether the concentration or dose to an
individual has the potential to cause an adverse effect. HQ values at
or less than 1 imply that exposures are at or below the RfC and not
likely to cause adverse effects. An HQ value exceeding 1 implies that
the RfC is exceeded, and the likelihood of adverse effects increases
as the amount and frequency of exposures exceeding the RfC increase.
In risk assessments in which RfCs are used and exposures are
approaching or exceeding the RfC, information about its derivation,
data, assumptions, and uncertainties should be evaluated along with
the HQ values to determine the concerns for public health and
likelihood for adverse effects. For example, the critical health
effect associated with the RfC, the type of epidemiologic or
toxicologic studies considered, the degree of exceedance, the
uncertainty and modifying factors used in deriving the RfC, and the
uncertainties and degree of confidence in the RfC should be evaluated
to characterize the potential concerns for public health.
4.5.8 Total Risk for Noncancer Effects
The Hazard Index (HI) is used to address total risks from
multiple chemicals and is the sum of HQ values for individual
pollutants to which an individual is exposed. As an initial screen,
the individual HQ values are added within a power plant and the
highest HI across all plants is identified. Similar to the HQ, hazard
indexes at or less than 1 indicate that adverse noncancer health
effects are not expected to occur. As the HI increases, approaching
or exceeding 1 (unity), concern for the potential hazard of the
chemical mixture increases. If the HI exceeds unity, the mixture has
exceeded the equivalent of the RfC for the mixture. The HI should not
be interpreted as a probability of risk nor as a strict delineation of
safe and unsafe levels.7'16
The HI approach assumes that simultaneous exposures to several
chemicals (even at subthreshold levels) could, in combination, result
in an adverse health effect. Even if no single compound exceeds its
RfC, the HI for the overall mixture may exceed 1. If the HI exceeds
unity, the HI should be reevaluated using HQ values summed only for
noncarcinogens with similar target organs based on EPA risk assessment
guidelines7 and assuming that each target organ has a threshold that
must be exceeded before adverse effects can occur and that toxicity
among target organs is independent. In addition, the mixture of
pollutants should be assessed for potential synergistic or
antagonistic effect if the HI is near or at unity and if sufficient
data are available. The EPA has produced a database called Mixtox17
that contains information about potential effects of mixtures of
pollutants. If the HI is at or near unity, then Mixtox can be used to
evaluate the mixture. For this study, only the maximum HI values
associated with a power plant were calculated.
4-14
-------
4.5.9 Direct Inhalation Exposure and Risk Default Options
The EPA's risk assessment guidelines contain a number of "default
options." These options are used in the absence of convincing
scientific knowledge about which ones of several competing models and
theories are correct. Several of the defaults are generally
conservative (i.e., they represent a choice that, although
scientifically plausible given the existing uncertainty, is more
likely to result in overestimating rather than underestimating risk).
However, some of the default options are not necessarily conservative,
and may actually lead to an underestimate of the risks. A number of
default options and assumptions were incorporated into the HEM
inhalation exposure assessment. These include:
• The HEM only estimates exposure within 50 km of each plant.
Exposure due to long-range transport is not considered in
the HEM analysis. (Long-range transport is analyzed in
section 6.6.)
• Dispersion occurs as predicted by a Gaussian plume model in
flat terrain.
• The closest meteorological station to each utility plant is
assumed to represent the weather patterns at the utility
plant site.
• Exposure is based on centroids of census blocks since
locations of actual residences are not in the database.
• For MEIs and MIRs, people are assumed to reside at the same
location for their entire lifetimes (assumed to be 70 years)
• Indoor concentrations are assumed to be the same as outdoor
concentrations.
• The average lifetime exposure is based on the assumption
that all exposures occurring at home; exposure due to
movement between home, school, work, etc., is assumed to be,
on average, equal to exposure at home.
• Utilities emit HAPs at rates predicted by the emissions
factor program described in chapter 3 at the same level for
a 70-year lifetime of exposure. Only stack emissions were
considered. Fugitive dust from coal piles was not included.
• The HEM only estimates exposure due to direct inhalation.
The HEM does not estimate exposure from indirect pathways
(i.e., multipathway exposure).
• The population database is not adjusted for population
growth.
4-15
-------
• Varying exposures as a result of differences such as age,
sex, health status, and activity are not considered.
The impacts of using some of these default values (e.g.,
emissions estimates, indoor vs. outdoor concentrations, breathing
rates, closest meteorological station, flat terrain and others) are
evaluated in sections 6.12 and 6.13 and Appendix G of the EPA's 1996
Interim Utility Air Toxics Report. Other default parameters are
described and discussed in Appendix F and in various sections of
chapter 6 of this report.
4-16
-------
4.6 REFERENCES
1. U.S. Environmental Protection Agency. Draft Dioxin Reassessment
Report. EPA/600/BP-92/001. Office of Research and Development.
1994.
2. U.S. Environmental Protection Agency. Deposition of Air
Pollutants to the Great Waters, First Report to Congress. EPA-
453/R-93-055. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. 1993.
3. National Research Council. Science and Judgment in Risk
Assessment. National Academy Press, Washington, DC. 1994.
4. U.S. Environmental Protection Agency. Guidance for Risk
Characterization, Science Policy Council. Washington, DC.
February 1995.
5. National Research Council. Risk Assessment in the Federal
Government: Managing the Process. National Academy Press,
Washington, DC. 1983.
6. U.S. Environmental Protection Agency. A Descriptive Guide to
Risk Assessment Methodologies for Toxic Air Pollutants. EPA-
453/R-93-038. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. 1993.
7. U.S. Environmental Protection Agency. Risk Assessment Guidelines
of 1986. EPA-600/8-87/045. (Guidelines for Carcinogen Risk
Assessment, Guidelines for Mutagenicity Risk Assessment,
Guidelines for Health Risk Assessment of Chemical Mixtures,
Guidelines for Health Assessment of Suspect Developmental
Toxicants, Guidelines for Estimating Exposures) Office of Health
and Environmental Assessment, Washington, DC. 1987.
8. U.S. Environmental Protection Agency. Guidance on Risk
Characterization for Risk Managers and Risk Assessors.
Memorandum from F. Henry Habicht, II, Deputy Administration,
Washington, DC. 1992.
9. U.S. Environmental Protection Agency. Integrated Risk
Information System (IRIS) Database. Environmental Criteria and
Assessment Office, Cincinnati, OH. 1994.
10. U.S. Environmental Protection Agency. Risk Assessment Guidance
for Superfund - Human Health Evaluation Manual, Part A.
EPA/540/1-89-002. Office of Solid Waste and Emergency Response,
Washington, DC. July 1989.
4-17
-------
11. Weinstein, I. B. The relevance of tumor promotion and multistage
carcinogensis to risk assessment. In: Banbury Report 19: Risk
Quantitation and Regulatory Policy. Edited by Hoel, D. G., R. A.
Merrill, and F. P. Perera. Cold Spring Harbor Laboratory, Cold
Spring Harbor, NY. 1985.
12. National Research Council, Safe Drinking Water Committee.
Drinking Water and Health. Volume 6. National Academy Press.
Washington, DC. 1986.
13. U.S. Environmental Protection Agency. Interim Methods for
Development of Inhalation Reference Concentrations. EPA 600/8-
90/066A. Office of Health and Environmental Assessment,
Washington, DC. August 1990.
14. U.S. Environmental Protection Agency. General Quantitative Risk
Assessment Guidelines for Noncancer Health Effects. ECAO-CIN-
538. Office of Research and Development, Washington, DC. June
1990.
15. California Air Resources Board, Risk Assessment Guidelines for
the Hot Spots Program. October 1993.
16. U.S. Environmental Protection Agency. General Quantitative Risk
Assessment Guidelines for Noncancer Health Effects. ECAO-CIN-
538. Office of Research and Development, Cincinnati, OH. June
1990.
17. MIXTOX (version 1.5). An Information System on Toxicological
Interactions for MS-DOS Personal Computer, Environmental Criteria
and Assessment Office, Cincinnati, OH. September 1992.
4-18
-------
5.0 SCREENING ASSESSMENT TO DETERMINE PRIORITY HAPS
The EPA conducted a screening assessment for the 67 HAPs
identified in the emissions database to determine priority HAPs for
further analyses. The HAPs were prioritized based on their potential
to pose hazards and risks through inhalation or multipathway exposure.
5.1 MODELING DESCRIPTION
To screen for inhalation risks, the EPA conducted exposure
modeling, using the Human Exposure Model (HEM), to estimate direct
inhalation exposure within 50 km of each utility plant for 66 of the
67 HAPs identified in the emissions database. The emissions estimates
used for all the HEM modeling presented in this study (chapters 5 and
6) were estimates of annual average emissions for all plants (see
chapter 3 for details on how emissions estimates were calculated).
The HEM modeling predicted annual average ambient air concentrations
in all the census blocks located within 50 km of each facility (see
chapters 4, 6, and Appendix F for detailed discussions of the HEM
modeling for this study,and the input data, assumptions and default
parameters). Radionuclides could not be modeled adequately using the
HEM. Therefore, radionuclides were screened based on previous studies
conducted in the 1980s.1 The screening for radionuclides is discussed
in section 5.6. A general description of the HEM, input data, and
default options is presented in chapter 4. The HEM is also described
in detail in Appendix F.
Using the average annual emission estimates (discussed in
chapter 3) for each HAP for all 684 plants along with site-specific
parameters as input (e.g., location, stack height, stack exit
velocity, stack temperature, and population data), the HEM was
utilized to estimate inhalation exposures for the maximally exposed
individuals (MEIs). The exposure estimates were then combined with
health effects data to estimate risks due to inhalation exposure for
the MEIs. Based on these MEI risks, priority HAPs were selected for
inhalation risk assessment. As described below, the screening for
multipathway assessment was based on factors such as persistence of
the HAP, bioaccumulation potential, and toxicity by ingestion
exposure.
5.2 SCREENING CRITERIA
First, HAPs were screened based on cancer risk effects due to
inhalation exposure. The maximum modeled air concentrations for each
HAP were multiplied by the lUREs to estimate upper limit increased
lifetime cancer risks to the MEIs. If the highest MEI risk was
greater than 1 in 10 million (i.e., 1 x 10"7), the HAP was considered a
priority for further analysis.
Second, HAPs were screened for noncancer effects due to long-term
(chronic) inhalation exposure. The maximum modeled air concentrations
were compared to RfCs. Hazard quotients (HQ) were calculated by
dividing the maximum modeled concentrations by the RfCs. If the
5-1
-------
highest HQ was greater than 0.1, then the HAP was considered a
priority for further analysis.
Third, in addition to the inhalation screening assessment, HAPs
were prioritized for potential multipathway exposure and risks. The
nonradionuclide HAPs were prioritized for multipathway assessment
based on persistence of the HAP, tendency to bioaccumulate, toxicity
by ingestion exposure route, and quantity of emissions. This resulted
in five nonradionuclide HAPs being identified as priorities for
multipathway assessment.
After HAPs were prioritized based on the above criteria,
additional HAPs were prioritized because of potential concerns for
noncancer effects due to short-term inhalation exposures. Also,
radionuclides were identified as a priority for multipathway
assessment based on results of previous studies. The following
sections present more details about each of the screening analyses.
5.3 INHALATION SCREENING ASSESSMENT FOR COAL-FIRED UTILITIES
Emissions data were available for 66 nonradionuclide HAPs from
coal-fired utilities. Cancer and noncancer quantitative health
effects data were available from IRIS and various EPA documents for 50
of the 66 nonradionuclide HAPs. Table 5-1 presents the results for
HAPs that are considered carcinogens and for which a quantitative
cancer risk estimate was available. Table 5-2 presents results for
the noncancer screening assessment using EPA-verified RfCs. Table 5-3
presents the HAPs for which no EPA-approved quantitative health data
were available for assessment.
5.4 INHALATION SCREENING ASSESSMENT FOR OIL- AND GAS-FIRED UTILITIES
Emissions data were available for 28 HAPs from oil-fired utilities.
EPA cancer and noncancer quantitative health effects data were available
for 22 of the 28 HAPs. Table 5-4 presents the results for HAPs from oil-
fired utilities that are considered carcinogens and for which
quantitative cancer risk estimates were available. Table 5-5 presents
results for the noncancer screening assessment for HAPs from oil-fired
utilities for which EPA-approved RfCs were available. Table 5-6 presents
HAPs from oil-fired utilities for which no EPA-verified quantitative
health data were available for assessment.
For gas-fired utilities, emissions data were available for
14 HAPs. Table 5-7 presents the screening results for gas utilities.
5-2
-------
Table 5-1. Inhalation Screening Assessment for Carcinogenic HAPs
from Coal-Fired Utilities for Which Quantitative Cancer Risk
Estimates Were Available
Hazardous air pollutant
Arsenic compounds
Chromium (11 percent Vl)e
Beryllium
Cadmium
Nickel compounds'
Dioxins9
PAHsh
Naphthalene
Hexaclorobenzene
Carbon tetrachloride
Quinoline'
Vinylidene chloride
Formaldehyde
n-Nitrosodimethylamine'
1,1,2-Trichloroethane
Acetaldehyde
Benzene
Benzyl chloride
Bis(2-ethylhexyl)phthalate'
Bromoform
Chloroform
Ethylene dichloride
Isophorone'
Methyl chloride'
Highest
MEI
conc.a
(ug/m3)
0.0014
0.0023
0.00025
0.00009
0.0027
2x10'9
0.00002
0.00009
9x10'6
0.00038
0.000006
0.0011
0.00047
0.00008
0.00054
0.00078
0.00029
6x10'7
0.00047
0.00077
0.00037
0.00036
0.003
0.0007
EPA
WOE
A
A
B2
B2
Af
B2
B2
C
B2
B2
C
C
B1
B1
C
B2
A
B2
B2
B2
B2
B2
C
C
IUREb
(ug/m3)
0.0043
0.00166
0.0024
0.0018
0.00048
30.09
0.0021h
4x10'6
0.00046
0.000015
0.0035'
5X10'5
1 x10'5
0.014
2x10'5
2x10'6
8x10'6
5X10'5
4x10'6i
1 x10'6
2X10'5
3X10'5
3x10'7i
2x10'6i
MEI
cancer
risk0
>io-6
>io-6
> 10'7
> 10'7
> 10'7
7x10'8
4x10'8
4x10-10
4x10'9
6x1Q-g
2x10'8
6x10'8
6x10'7
1 x10'6
9x10'9
2x10'9
2x10'9
3x10-11
2x10'9
9x10-10
9x10'9
9x10'9
9x10-10
1 x10'9
Primary type
of cancer
assoc. w/
inhalation1'
Lung
Lung
Lung
Lung
Lung & nasal
Tongue, lung,
nasal, liver
Lung (BAP)
NA
NA
Liver
NA
NA
Nasal, lung
Liver & other
NA
Nose &
larynx
Leukemia
NA
NA
NA
Kidney & liver
NA
NA
Kidney
MEI cancer
risk >10'7
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
Yes'
No
No
No
No
No
No
No
No
No
No
(continued)
5-3
-------
Table 5-1. (Continued)
Hazardous air pollutant
Methylene chloride
Trichloroethylene1
Pentachlorophenol1
Tetrachloroethylene1
Highest
MEI
conc.a
(ug/m3)
0.0015
0.00036
1 x 10'6
0.00036
EPA
WOE
B2
B2/C
B2
B2/C
IUREb
(ug/m3)
5x10'7
2x10'6i
3x1Q-5i
6x10'7i
MEI
cancer
risk0
7x10-10
6x10-10
3x10-11
2x10-10
Primary type
of cancer
assoc. w/
inhalation1'
Liver & lung
Lung, liver, &
testicular
NA
Liver
MEI cancer
risk >10'7
No
No
No
No
IURE = Inhalation Unit Risk Estimate. The IURE is the estimated increased risk of cancer from breathing 1 ug of
pollutant per cubic meter of air for 70 years.
MEI = Maximally Exposed Individual.
NA = Not available.
WOE = Weight of Evidence, for carcinogenicity. See section 4.3.1 and Table 4-1. for explanation of WOE.
a This is the highest estimated ambient concentration (annual average) due to emissions from the one highest risk
coal-fired utility based on HEM modeling of all coal-fired utilities in the U.S.
b lUREs obtained from EPA's Integrated Risk Information System (IRIS),2 unless indicated otherwise by footnotes.
c This is the estimated increased lifetime cancer risk to the highest MEI due to inhalation exposure.
d This column presents the type of cancer observed in experimental animal studies or human studies. For more details see
Appendix E and/or various references.
e For coal-fired utilities it is assumed that 11 percent of chromium is chromium VI and that the remainder is chromium III.
For oil-fired utilities it is assumed that 18 percent of chromium is chromium VI. This is based on limited speciation data
described in Appendix H.3 It is assumed that the carcinogenic effects are caused only by the Cr VI fraction. The IURE
was calculated by multiplying the IURE on IRIS for Cr VI (1.2 x 10'2) by 0.11 (11 percent).
f For this screening assessment all nickel was assumed to be as carcinogenic as nickel subsulfide. This assumption is
considered an "upper bound" conservative assumption. Nickel risk uncertainty issues are discussed more thoroughly in
Chapter 6.
9 The emissions were estimated using the toxic equivalency (TEQ) approach described in the draft EPA Dioxin
Reassessment Report.4 Exposure was estimated by modeling the TEQ emissions with HEM. The IURE is for 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) and was obtained from the draft EPA Dioxin Reassessment Report.
h To estimate the potential risk from polycyclic aromatic hydrocarbon (PAH) emissions, first the EPA summed the
emissions of the 7 PAHs that are classified as B2 carcinogens (WOE = B2).5 (These are listed in Appendix H).1 Second,
exposure was estimated by using the HEM and modeling the sum of the 7 PAHs. Third, the estimated exposure to the 7
B2 PAHs were multiplied by the lUREfor benzo[a]pyrene (BAP) (2.1 x 10"3). However, this IURE has not been verified by
the EPA and has not been peer reviewed. It is an interim value with significant uncertainties and is intended for screening
assessment only. This IURE was calculated by converting the oral unit risk estimate of 2.1 x 10"4per ug/L to inhalation
units. The conversion assumes equal absorption and metabolism and assumes equal risk from the different routes of
exposure, which may not be the case.
' The lUREs for these HAPs are not EPA-verified and are intended for screening assessment only. Readers must exercise
caution interpreting the results using these numbers. These lUREs were obtained from Documentation ofDe Minimis
Emission Rates - Proposed 40 CFR Part 63, Subpart B, Background Document.6 This document was developed to
support the proposed rulemaking pursuant to 112(g) of the Clean Air Act (Federal Register, Volume 59, No. 63, April 1,
1994). There are significant uncertainties associated with these lUREs. They are not EPA-verified. They are interim
screening values intended for the screening assessment only. For further discussion of the health data and uncertainties,
see the de minimis document cited above.
' The risk estimate for n-nitrosodimethylamine is highly suspect and uncertain because the emissions estimates were
based on one measured value and several nondetect values. Based on available information, the MEI risk estimate
presented here is likely to be a significant overestimate of the true risks posed by n-Nitrosodimethylamine.
5-4
-------
Table 5-2. Inhalation Screening Assessment for Noncancer Effects
of HAPs Emitted from Coal-Fired Utilities for Which Inhalation
Reference Concentrations Are Available
Hazardous air
pollutant
2-Chloro-acetophenone
Acrolein
Cumene6
Ethyl benzene
Ethyl chloride
Hexane
Hydrogen chloride
Hydrogen cyanide
Lead''9
Manganese
Mercury9
Methyl bromide
Methyl chloroform6
Methyl ethyl ketone
MTBE
Styrene
Toluene
RfC
(Mg/m3)
0.03
0.02
9.0e
1000
10000
200
20
3.0
1.5'
0.05
—
5.0
1000e
1000
3000
1000
400
Noncancer
health effect on
which RfC is
based3
Hyperplasia of nasal
resp. epith. in rats
Metaplasia and
inflammation rat nasal
epithel.
—
Developmental effects
Delayed fetal
ossification
CMS & nasal epith.
lesions humans
Hyperplasia of nasal
mucosa & larynx in
rats
CMS symptoms and
thyroid effects
CMS & devel. humans
CMS, humans
—
Lesions of olfactory
epithelium
Hepatotoxicity
Decreased fetal birth
weight (mice)
Increased liver &
kidney weight (rat)
CMS in humans
Neurological effects;
degeneration of nasal
epithelium
Confidence
in RfCb
low
med
NA
low
med
med
low
low
NA
med
-
high
NA
low
med
med
med
Highest
MEI
conc.c
(M9/m3)
3X10'5
4x10'4
0.00003
0.00005
0.0003
0.00009
2.3
0.0033
0.007
0.02
0.001
0.0001
0.0004
0.0009
0.0002
0.00036
0.0004
Max. HQ
0.001
0.02
3x10'6
5x10'8
2x10'8
5x10'6
0.115
0.001
0.0057
0.4
-
2x10'5
4x10'7
9x10'7
7x10'8
4x10'7
1 x10'6
Highest
HQd
>0.1
No
No
No
No
No
No
Yes
No
No
Yes
No
No
No
No
No
No
No
(continued)
5-5
-------
Table 5-2. (Continued)
Hazardous air
pollutant
1,3-Dichloropropene
Vinyl acetate
RfC
(Mg/m3)
20
200
Noncancer
health effect on
which RfC is
based"
Hypertrophy/hyperpla
sia
of nasal respiratory
epithelium
Nasal epithelium
lesions
Confidence
in RfCb
high
high
Highest
MEI
conc.c
(M9/m3)
0.00054
0.00005
Max. HQ
3X10'5
3x10'7
Highest
HQd
>0.1
No
No
CNS = Central nervous system.
HQ = Hazard Quotient. The ratio of exposure concentration/RfC. An HQ < 1.0 indicates that no adverse health
effects are expected to occur (see Chapter 4 for discussion of HQ).
MEI = Maximally Exposed Individual.
NA = Not applicable.
RfC = Reference concentration (inhalation).
a This is the critical adverse noncancer health effect that was observed in animal or human studies27
b This is the overall confidence in the RfC as reported on IRIS.
c This is the highest estimated ambient concentration (annual average) due to coal-fired utility emissions based on
HEM modeling within 50 km of all coal-fired utilities in the United States.
d If HQ > 0.1, this means that the highest modeled concentration is greater than 1/10 of the RfC. This value (0.1)
was used as criteria in screening assessment. This is not considered a level of concern, but rather it is a
conservative level to ensure that potentially important HAPs are not missed by screen. See text for explanation.
e The RfC was obtained from the 1992 EPA Health Effects Summary Tables8 It has not been verified by the EPA's
RfC/RfD workgroup.
f No RfC is available for lead compounds. Therefore, as a substitute, the lead National Ambient Air Quality
Standard (1.5 ug/m3) was used in this assessment.9 However, the lead NAAQS is note considered equivalent to
an RfC. The lead NAAQS is based on a quarterly average, but the exposure estimates here represent annual
averages. The reader should exercise caution when interpreting the HQ for lead. Lead has also been classified
as a carcinogen.2*
9 These compounds may also be a health concern from multipathway exposure. The assessment here considers
only inhalation exposure. Considering multipathway exposure may increase the risk estimates for these
pollutants. Multipathway screening assessment is discussed in section 5.5.
5-6
-------
Table 5-3. Inhalation Screening Assessment for HAPS Emitted from
Coal-Fired Utilities for Which No EPA-Verified Health Benchmarks
Are Available (Comparison of Highest Modeled Air Concentration to
Various Non-EPA Health Benchmarks)
Hazardous air
pollutant
Acetophenone
Antimony compounds
Carbon disulfide
Chlorobenzene
Cobalt compounds
o & p-Cresols
Cumene
Dibutyl phthalate
Hydrogen fluoride
Methyl methacrylate
MIBK
Phenol
Phthalic anhydride
Phosphorus
Propion-aldehyde
Selenium compounds
m,o,p-Xylenes
2,4-Dinitrotoluene
Methyl iodide
NIOSH
REL/4203
(Atg/m3)
NA
1.2
7.1
NA
0.12
24
580
11.9
6.0
NA
490
48
14
0.24
NA
0.48
1000
3.6
24
OSHA
PEL/420a
(M9/m3)
NA
1.2
29
833
0.12
52
580
11.9
6.0
980
980
45
14
0.24
NA
0.48
1000
3.6
67
ACGIH
TLV/4203
(ug/m3)
NA
1.2
74
830
0.12
580
11.9
6.0C
980C
490
45C
14
NA
0.48C
1000
0.36
29
Highest
MEI conc.b
(Atg/m3)
0.00008
0.0005
0.0005
0.00037
0.0017
0.0003
0.00003
0.00033
0.365
0.00013
0.00058
7x10'4
6x10-4
0.0036
0.0012
0.0056
0.0005
1 x10-6
0.00005
Maximum
HQ
NA
4x10'4
7x10-5
4x10-7
0.014
1 x 10-5
3x10-6
3x10-5
0.06
1 x 1 0'7
1 x 10-6
2x10-5
4x10-5
0.015
NA
0.012
5x10-7
3x10-6
2x10-6
HQ>0.1
NA
No
No
No
No
No
No
No
No
No
No
No
No
No
NA
No
No
No
No
ACGIH = American Conference of Government Industrial Hygienists, which is a professional society, not a
government agency.
HQ = Hazard Quotient. The ratio of exposure concentration/RfC. An HQ < 0.1 indicates that no adverse
health effects are expected to occur (see Chapter 4 for discussion of HQ).
MEI = Maximally Exposed Individual.
NIOSH = National Institute for Occupational Safety and Health, a U.S. government organization that focuses on
research.
OSHA = Occupational Safety and Health Agency, a U.S. Government Agency
5-7
-------
Footnotes for Table 5-3 (continued)
PEL = Permissible Exposure Levels. These are legal limits established by OSHA.
REL = Recommended Exposure Levels. NIOSH develops these recommended levels to protect workers.
TLV = Threshold Limit Values. The TLV are established by ACGIH and are used by industrial hygienists in
the work place to assess the potential concerns for worker exposure.
a The NIOSH, OSHA, and ACGIH are primarily involved with the safety and health of workers. The RELs, PELs,
and TLVs are similar. Breathing concentrations below these levels are expected to be reasonably protective of
health workers, exposed for 8 hours per day, 5 days per week (-40 hours). However, there are uncertainties and
often the data are less then complete. Also, for some of these values (especially the PELs), measurement
techniques and economic factors are sometimes factored in.10'11'12
Occupational Exposure Limits (OELs) are being used in this study for screening assessment purposes only. For
this screening assessment, the REL, PEL, and TLV were divided by 420 (4.2 x 10 x 10). The 4.2 is the conversion
factor to extrapolate from a 40 hr/weekto a 168 hr/week. A factor of 10 is used to adjust for sensitive
subpopulations. Another factor of 10 is used to account for additional uncertainties associated with these values.
A similar method was used by the California Air Resources Board (CARS) in the Air Toxics "Hot Spots" Program. 13
CARS also divides the TLV by 420 to calculate some of their noncancer reference exposure levels (4.2 to account
for exposure time adjustment, 10 to account for sensitive individuals, and another 10 because health effects are
sometimes observed at the TLV level).
b This is the highest estimated ambient concentration due to coal-fired utility emissions based on HEM modeling
within 50 km of all coal-fired utilities in the United States.
c These values are the same as the CARS Noncancer Reference Exposure Levels used in the "Hot Spots
Program.""
5-8
-------
Table 5-4. Inhalation Screening Assessment for Carcinogenic HAPS
from Oil-Fired Utilities for Which Quantitative Cancer Risk
Estimates Were Available
Hazardous air pollutant
Arsenic
Chromium (18 percent Cr
Vl)e
Beryllium
Cadmium
Nickel compounds'
Dioxins9
PAHsh
Formaldehyde
Acetaldehyde
Benzene
Methylene chloride
Naphthalene
Tetrachloroethylene'
Highest
MEI conc.a
(^g/m3)
0.0032
0.0025
0.0003
0.0009
0.21
4x10-g
0.00003
0.007
0.0019
0.0003
0.008
0.00008
0.00013
EPA WOE
A
A
B2
B2
A2
B2
B2
B1
B2
A
B2
C
B2/C
IUREb per
M9/rn3
0.0043
0.0028
0.0024
0.0018
0.00048
30.0
0.0021
1.3x10-5
2.2 x10-6
8.3 x10-6
4.7 x10-7
4.2 x10-6
5.8 x10-7
Cancer
MEI Risk0
>io-6
>io-6
>io-7
>io-6
>io-6
1 x ID'7
6x10-8
9x10-8
4x10-g
3x10-g
4x10-g
3x10-10
8x10-11
Primary type of
cancer
associated w/
Inhalation1'
Lung
Lung
Lung
Lung
Lung & nasal
Tongue, lung,
nasal, liver,
thyroid
Lung (BAP)
Nasal, lung
Nasal & laryngeal
Leukemia
Liver & lung
-
Liver
MEI cancer
risk>10-7
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
IURE = Inhalation Unit Risk Estimate. The IURE is the estimated increased risk of cancer from breathing 1 ug of
pollutant per cubic meter of air for 70 years.
MEI = Maximally Exposed Individual.
WOE = Weight of Evidence, for carcinogenicity. See section 4.3.1 and Table 4-1. for explanation of WOE.
a This is the highest estimated ambient concentration (annual average) due to emissions from the one highest risk
coal-fired utility based on HEM modeling of all coal-fired utilities in the U.S.
b lUREs obtained from EPA's Integrated Risk Information System (IRIS)2, unless indicated otherwise by footnotes.
c This is the estimated increased lifetime cancer risk to the highest MEI due to inhalation exposure.
d This column presents the type of cancer observed in experimental animal studies or human studies. For more details see
Appendix E and/or various references.
e For coal-fired utilities it is assumed that 11 percent of chromium is chromium VI and that the remainder is chromium III.
For oil-fired utilities it is assumed that 18 percent of chromium is chromium VI. This is based on the limited speciation
data described in Appendix H.3 It is assumed that the carcinogenic effects are caused only by the Cr VI fraction. The
IURE was calculated by multiplying the IURE on IRIS for Cr VI (1.2x 10'2) by 0.11 (11 percent).
f For this screening assessment all nickel was assumed to be as carcinogenic as nickel subsulfide. This assumption is
considered an "upper bound" conservative assumption. Nickel risk uncertainty issues are discussed more thoroughly in
Chapter 6.
9 The emissions were estimated using the toxic equivalency (TEQ) approach described in the draft EPA Dioxin
Reassessment Report.1 Exposure was estimated by modeling the TEQ emissions with HEM. The IURE is for 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) and was obtained from the draft EPA Dioxin Reassessment Report.
5-9
-------
Footnotes for Table 5-4. (Continued)
h To estimate the potential risk from polycyclic aromatic hydrocarbon (PAH) emissions, first the EPA summed the
emissions of the 7 PAHs that are classified as B2 carcinogens (WOE = B2).5 (These are listed in Appendix H3). Second,
exposure was estimated by using the HEM and modeling the sum of the 7 PAHs. Third, the estimated exposure to the 7
B2 PAHs were multiplied by the ILJREfor benzo[a]pyrene (BAP) (2.1 x 10"3). However, this IURE has not been verified by
the EPA and has not been peer reviewed. It is an interim value with significant uncertainties and is intended for screening
assessment only. This IURE was calculated by converting the oral unit risk estimate of 2.1 x 10"4per //g/L to inhalation
units. The conversion assumes equal absorption and metabolism and assumes equal risk from the different routes of
exposure, which may not be the case.
' The ILJREs for these HAPs are not EPA-verified and are intended for screening assessment only. Readers must exercise
caution interpreting the results using these numbers. These ILJREs were obtained from Documentation ofDe Minimis
Emission Rates - Proposed 40 CFR Part 63, Subpart B, Background Document. - This document was developed to
support the proposed rulemaking pursuant to 112(g) of the Clean Air Act (Federal Register, Volume 59, No. 63, April 1,
1994). There are significant uncertainties associated with these ILJREs. They are not EPA-verified. They are interim
screening values intended for the screening assessment only. For further discussion of the health data and uncertainties,
see the de minimis document cited above.
5-10
-------
Table 5-5. Inhalation Screening Assessment for Noncancer Effects
of HAPS Emitted from Oil-Fired Utilities for Which EPA-Verified
Inhalation Reference Concentrations Are Available
Hazardous air
pollutant
Ethyl benzene
Hydrogen
chloride
Lead6
Manganese
Mercury
Methyl
chloroform'
Toluene
Vinyl acetate
RfC
(ug/m3)
1000
20
1.5
0.05
-
1000f
400
200
Critical noncancer health
effect that RfC is based ona
Developmental effects
Hyperplasia of nasal
mucosa,
larynx, and trachea in rats
Neurotoxicity and
developmental in humans
Neurobehavioral effects in
humans
-
Hepatotoxicity'
Neurological effects
Nasal lesions
Overall
confidence
in RfCb
Low
Low
NA
Medium
-
NA
Medium
High
Highest
MEIC
cone.
(ug/m3)
1 x ID'4
1.1
0.005
0.002
0.00014
0.0018
0.002
0.0012
Highest
HQ
1 x10-7
0.16
0.003
0.04
-
2x10-6
5x10-6
6x10-6
HQ>0.1d
No
Yes
No
No
No
No
No
No
HQ = Hazard Quotient. The ratio of exposure concentration/RfC. An HQ < 1.0 indicates that no adverse
health effects are expected to occur (see Chapter 4 for discussion of HQ).
MEI = Maximally Exposed Individual.
NA = Not applicable.
RfC = Reference concentration (inhalation).
a This is the critical adverse noncancer health effect that was observed in animal or human studies21
b This is the overall confidence in the RfC as reported on IRIS.
c This is the highest estimated ambient concentration (annual average) due to coal-fired utility emissions based on
HEM modeling within 50 km of all coal-fired utilities in the United States.
d If HQ > 0.1, this means that the highest modeled concentration is greater than 1/10 of the RfC. This value (0.1)
was used as criteria in screening assessment. This is not considered a level of concern, but rather it is a
conservative level to ensure that potentially important HAPs are not missed by screen. See text for explanation.
e No RfC is available for lead compounds. Therefore, as a substitute, the lead National Ambient Air Quality
Standard (1.5 ug/m3) was used in this assessment.2 However, the lead NAAQS is note considered equivalent to
an RfC. The lead NAAQS is based on a quarterly average, but the exposure estimates here represent annual
averages. The reader should exercise caution when interpreting the HQ for lead. Lead has also been classified
as a carcinogen.16
f These compounds may also be a health concern from multipathway exposure. The assessment here considers
only inhalation exposure. Considering multipathway exposure may increase the risk estimates for these
pollutants. Multipathway screening assessment is discussed in section 5.5.
5-11
-------
Table 5-6. Inhalation Screening Assessment for HAPS Emitted from
Oil-Fired Utilities for Which No EPA-Verified Health Benchmarks
Are Available (Comparison of Highest Modeled Concentration to
Various Non-EPA Health Benchmarks)
Pollutant
Cobalt compounds
Hydrogen fluoride
Phenol
Phosphorus
Selenium compounds
m,o,p-Xylenes
NIOSH
REL/4203
(Mg/m3)
0.12
6.0
48
0.24
0.48
1040
OSHA
PEL/420a
(Mg/m3)
0.12
6.0
45
0.24
0.48
1040
ACGIH
TLV/4203
(Mg/m3)
0.12
NA
45
0.24
NA
1040
Highest
MEI cone.
(Hg/m3)b
0.0096
0.03
0.006
0.026
0.001
0.0005
MaxHQ
0.08
0.005
0.0001
0.1
0.002
5x10-7
MaxHQ
>0.1
No
No
No
No
No
No
ACGIH = American Conference of Government Industrial Hygienists, which is a professional society, not a
government agency.
HQ = Hazard Quotient, the ratio of exposure concentration/RfC. An HQ < 0.1 indicates that no adverse
health effects are expected to occur (see Chapter 4 for discussion of HQ).
MEI = Maximally Exposed Individual.
NIOSH = National Institute for Occupational Safety and Health, a U.S. government organization that focuses on
research.
OSHA = Occupational Safety and Health Agency, a U.S. Government Agency
PEL = Permissible Exposure Levels. These are legal limits established by OSHA.
TLV = Threshold Limit Values. The TLV are established by ACGIH and are used by industrial hygienists in
the work place to assess the potential concerns for worker exposure.
a The NIOSH, OSHA, and ACGIH are primarily involved with the safety and health of workers. The RELs, PELs,
and TLVs are similar. Breathing concentrations below these levels are expected to be reasonably protective of
health workers, exposed for 8 hours per day, 5 days per week (-40 hours). However, there are uncertainties and
often the data are less then complete. Also, for some of these values (especially the PELs), measurement
techniques and economic factors are sometimes factored in.mils
Occupational Exposure Limits (OELs) are being used in this study for screening assessment purposes only. For
this screening assessment, the REL, PEL, and TLV were divided by 420 (4.2 x 10 x 10). The 4.2 is the conversion
factor to extrapolate from a 40 hr/weekto a 168 hr/week. A factor of 10 is used to adjust for sensitive
subpopulations. Another factor of 10 is used to account for additional uncertainties associated with these values.
A similar method was used by the California Air Resources Board (CARS) in the Air Toxics "Hot Spots" Program. —
CARS also divides the TLV by 420 to calculate some of their noncancer reference exposure levels (4.2 to account
for exposure time adjustment, 10 to account for sensitive individuals, and another 10 because health effects are
sometimes observed at the TLV level).
b This is the highest estimated ambient concentration due to coal-fired utility emissions based on HEM modeling
within 50 km of all coal-fired utilities in the United States.
5-12
-------
Table 5-7. Inhalation Screening Assessment for HAPS Emitted from
Gas-Fired Utilities
Hazardous air
pollutant
Arsenic
Nickel compounds6
Naphthalene
Toluene
Lead
Formaldehyde
Mercury
Benzene
Phosphorus
Cobalt
Highest MEI
conc.a (ug/m3)
2x1Q-5
0.0003
0.0001
0.0018
0.00006
0.008
0.0000002
0.0003
0.0002
0.00002
IUREb
(ug/m3)
0.0043
0.00048
4x10-6
NA
NA
1.3x10-5
NA
8.3 x10-6
NA
NA
HEM Cancer
MEI Risk0
1 x10-7
2x10-7
4x10-10
NA
NA
1 x10-7
NA
2x10-g
NA
NA
RfC
(ug/m3)
NA
NA
NA
400
1.5
NA
-
NA
0.24f
0.1 2f
Highest HQd
NA
NA
NA
4.5 x10-6
4x10-5
NA
-
NA
0.0008
0.0002
HEM = Human Exposure Model
HQ = Hazardous Quotient. The ratio of exposure concentration/RfC. An HQ < 1.0 indicates that no adverse
health effects are expected to occur (see Chapter 4 for discussion of HQ).
IURE = Inhalation Unit Risk Estimates
MEI = Maximally Exposed Individual.
a This is the highest estimated ambient concentration (annual average) due to emissions from the one highest risk
coal-fired utility based on HEM modeling of all coal-fired utilities in the U.S.
b lUREs obtained from EPA's Integrated Risk Information System (IRIS),2 unless indicated otherwise by footnotes.
c This is the estimated increased lifetime cancer risk to the highest MEI due to inhalation exposure.
d If HQ > 0.1, this means that the highest modeled concentration is greater than 1/10 of the RfC. This value (0.1)
was used as criteria in screening assessment. This is not considered a level of concern, but rather it is a
conservative level to ensure that potentially important HAPs are not missed by screen. See text for explanation.
e For this screening assessment all nickel was assumed to be as carcinogenic as nickel subsulfide. This
assumption is considered an "upper bound" conservative assumption. Nickel risk uncertainty issues are
discussed more thoroughly in Chapter 6.
f These values are not RfCs. They are TLV/420. See Tables 5-3 and 5-6.
5.5 MULTIPATHWAY SCREENING ANALYSIS FOR NONRADIONUCLIDE HAPS
5.5.1 Overview
In past years, most analyses of human health risk associated with
atmospheric emissions of nonradionuclide HAPs from combustion sources
have focused only on exposures occurring by inhalation. The
inhalation exposure pathway is generally the significant pathway for
human exposure to air pollutants. In the past decade, though, studies
have linked elevated levels of pollutants in soils, lake sediments,
and cow's milk to atmospheric transport and deposition of pollutants
from combustion sources.14 Scientists have collected convincing
5-13
-------
evidence showing that toxic chemicals released to air can travel long
distances and be deposited to land and water at locations both near
and far from their original emission sources.15 Many studies indicate
that deposition of atmospherically emitted pollutants can result in
indirect avenues of exposure for humans.16 For some HAPs, these
noninhalation routes of exposure can be as significant, or more
significant, than inhalation.
Certain HAPs have been associated with significant adverse
effects on human health and wildlife from noninhalation exposure
pathways." HAPs that pose a concern for noninhalation exposure
generally have common characteristics. They are persistent in the
environment, have the potential to bioaccumulate, and exhibit toxicity
via ingestion. For lipophilic contaminants such as dioxins, furans,
polychlorinated biphenyls, and certain pesticides and for metals such
as lead and mercury, exposures through food consumption have been
demonstrated to be dominant contributors to total dose for
nonoccupationally exposed populations.17 It is also likely that
atmospheric pollution from combustors and other thermal processes
significantly contributes to the ubiquitous presence of some of the
highly persistent lipophilic compounds."
Multipathway exposure to HAPs can potentially occur through the
following exposure routes:
- Soil ingestion - Fruit ingestion
- Soil dermal contact - Vegetable ingestion
- Water ingestion - Ingestion of animal fats
- Inhalation - Milk ingestion
- Fish and meat ingestion - Ingestion of other food products.
The following section presents the screening assessment to
prioritize the nonradionuclide HAPs for further analysis of
multipathway exposures and risks. Chapters 7, 8, 10, and 11 present
the multipathway assessments for the selected priority nonradionuclide
HAPs.
5.5.2 Prioritization of HAPs for Multipathway Exposure Assessment
The 66 nonradionuclide HAPs potentially emitted by utilities were
evaluated for their potential to cause health effects through
noninhalation exposure pathways. To select the highest priority HAPs
for multipathway exposure assessment, a four-step process was
followed. This process involved assessing the HAPs for their
potential to be of concern for exposure through noninhalation
pathways, evaluating their toxicity, and considering the emission
levels from utilities. First, a subset of HAPs was selected from the
list of 66 nonradionuclide HAPs by using the HAP ranking presented in
Attachment A (draft Focus Chemicals Report) of the EPA document,
Schedule for Standards: Methodology and Results for Ranking Source
Categories Based on Environmental Effects Data.18 The four criteria
evaluated and used in this ranking were human toxicity, aquatic
toxicity, bioconcentration potential, and environmental persistence.
Environmental partitioning was not used as a ranking criterion but was
5-14
-------
used as a "qualifying" criterion. The HAP ranking method is a
modified version of the Inerts Ranking Program (IRP) methodology
developed by EPA's Office of Research and Development, Environmental
Research Laboratory - Duluth, for evaluating pesticide ingredients.
The IRP scoring method was modified for scoring the environmental
criteria and for determining overall scores for the HAPs. For the
environmental criteria modification, acute aquatic toxicity and
chronic aquatic toxicity were combined into a single aquatic toxicity
criterion that is based strictly on chronic toxicity data when such
data are available. Each criterion, except environmental persistence,
allowed a possible score of 0, 1, 2, or 3. A score of 0 indicates
that no data are available, and scores of 1, 2, and 3 indicate low,
medium, and high concern, respectively. For environmental
persistence, substances were assigned a score of 1 or 3 only, because
data did not support finer discrimination. The method for deriving
the overall score was similar to that for deriving the original IRP
score. For each substance, the overall score was derived by adding
the scores for the four criteria, dividing by the number of criteria
for which there were data, and then multiplying by 10 to produce an
overall score on a scale of 10 to 30.
The HAP ranking in the Focus Chemicals report is a ranking of all
of the HAPs based on the overall score for each HAP. Of the 66 HAPs
potentially emitted to air by utilities, those that ranked the
highest, with overall ranking scores of greater than 23, were selected
for further evaluation. The cutoff score of 23 was selected because,
at this level and below, a HAP would have scores of 1 or 2, indicating
low and medium concern, respectively, for at least two of the four
criteria. Thirteen HAPs were selected on these criteria. The high
ranking reflects that these 13 HAPs are more likely to be highly
persistent in the environment and/or to bioaccumulate, as well as to
potentially be toxic to humans. The 13 HAPs selected and their
ranking scores are listed in Table 5-8. This approach to ranking the
HAPs is a screening-level, hazard-based ranking of chemicals. This
approach yielded a subset of 13 HAPs from which five HAPs were chosen
for further evaluation.
In the second step of the process, additional information was
gathered for each of the 13 selected HAPs to determine the HAPs that
are most important for multipathway assessment for the utilities.
Where available and applicable, the RfD, the oral unit risk estimate
(CURE), the EPA WOE classification, and the emissions estimate were
obtained for each of the 13 HAPs. This information is presented in
Table 5-9.
Several criteria were used to further prioritize HAPs for
multipathway exposure assessment. The six HAPs with the highest
noncancer toxicity (i.e., lowest RfDs [less than 1 x 10 ~3] ) , as well as
the HAPs with EPA-verified OUREs and a WOE classifica-tion of A or B,
were selected. Also, due to their extremely high toxicity and the
concern that they are "no threshold" or extremely low threshold
chemicals, 2,3,7,8-TCDD and lead compounds were also included.
Mercury was also selected because mercury is persistent, tends to
5-15
-------
Table 5-8. Thirteen HAPs Selected from the Hazard-Based
Multipathway Ranking (shown in order of ranking), and the
Overall and Individual Criterion Scores Assigned to Each
HAP
2,3,7,8-TCDD (dioxins)
Cadmium compounds
Mercury compounds
Hexachlorobenzene
Selenium compounds
Lead compounds
Cobalt compounds
Pentachlorophenol
Arsenic compounds
POM (PAH)
Beryllium compounds
Cyanide compounds
Manganese compounds
Overall
score
30
30
30
28.75
28.75
27.5
26.67
25
25
25
23.75
23.75
23.75
Human
toxicity
score
3
3
3
2.5
2.5
3
3
2
3
3
2.5
2.5
1.5
Aquatic
toxicity
score
3
3
3
3
3
3
2
3
2
3
2
3
2
Bioconcentration
potential score
3
3
3
3
3
2
0
2
2
3
3
3
3
Environmental
persistence
score
3
3
3
3
3
3
3
3
3
1
3
1
3
HAP = Hazardous air pollutant
TCDD = Tetrachlorodibenzo-p-dioxins
PAH = Polycyclic aromatic hydrocarbons
POM = Polycyclic organic matter
bioaccumulate, and is relatively toxic by ingestion exposure. This
second step in the prioritization process resulted in eight HAPs being
selected: 2,3,7,8-TCDD, lead, mercury, arsenic, cadmium,
hexachlorobenzene, beryllium, and polycyclic organic matter (POM). In
the next step in the selection process (the third step), emission
estimates from utilities were examined for each of the eight HAPs.
For two of the HAPs, POM and hexachlorobenzene, the emissions data for
utilities did not support their inclusion in further assessments. For
hexachlorobenzene, emissions were not considered high (0.7 ton/yr)
nationwide. Also, this estimate was highly uncertain because of the
very limited emissions data for hexachlorobenzene. For POM, the
emission levels of 1.9 ton/yr from coal-fired utilities and less than
1 ton/yr for oil-fired utilities are low relative to other
anthropogenic sources of POM. For the final step, the two lowest-
ranking of the six remaining HAPs (cadmium and beryllium) were
compared with each other in terms of toxicity, emissions, and the
original ranking scores they were assigned. Cadmium was selected for
further assessment, rather than beryllium, because of its higher
5-16
-------
Table 5-9. Comparison of Cancer and Noncancer Effects Benchmarks
and Emissions Estimates for 13 Selected HAPs
HAP
2,3,7,8-TCDD (dioxins)(TEQ)
Lead compounds
Mercury compounds
Arsenic compounds"
Cadmium compounds
Hexachlorobenzene
Selenium compounds
Beryllium compounds
Cyanide compounds
Manganese compounds
Pentachlorophenol
Cobalt compounds
POM (PAHf
RfD
(mg/kg/day)
NA
No
threshold c
1 x 1Q-4
3x10'4
5x10'4
8x10'4
5x1Q-3
5x1Q-3
5x1Q-3
5x1Q-3
NA
NA
NA
CURE
(per pg/L)
3x10+ob
-
-
5.0 x1Q-5
-
4.6 x1Q-5
-
1.2 x10'4
-
-
-
-
2.1 x10'4
WOE
B2
B2
C
A
B2
B2
-
B2
-
-
B2
-
B2
Coal-fired
emissions
estimates3
(ton/yr)
9.7 x1Q-5
7.5x10+1
4.6x10+1
6.1 x10+1
3.3
0.7
1.5x10+2
7.1
2.2x10+2
1.6x10+2
6.4x10-2
2.1 x10+1
1.9
Oil-fired
emissions
estimates3
(ton/yr)
0.69 x10-5
11
0.3
5
2
NA
2
0.5
NA
10
NA
20
< 1
HAP = Hazardous air pollutant
NA = Not available
OURE = Oral unit risk estimate
PAH = Polycyclic aromatic hydrocarbons
POM = Polycyclic organic matter
RfD = Reference dose
TEQ = Toxicity equivalent
TCDD = Tetrachlorodibenzo-p-dioxin
WOE = weight of evidence
a This is an estimate of total nationwide emissions from the source category.
b This is an unverified oral unit risk estimate*
c The Agency has determined that some of the effects of lead, particularly changes in the levels of certain blood
enzymes and in aspects of children's neurobehavioral development, may occur at blood lead levels so low as to
be essentially without a threshold.
d RfD is for inorganic arsenic. There was not a clear consensus for developing this value. See Appendix E and/or
the IRIS database for details.
e POM emissions were estimated by summing the emissions estimates for each individual PAH listed in Appendix H
of the Interim Final Utility HAP Report.3
5-17
-------
ranking scores for human and aquatic toxicity and its lower Rfd,
representing higher noncancer toxicity. This resulted in the
selection of five highest priority HAPs for multipathway exposure
assessment. The five HAPs selected to be highest priority for further
analysis were 2,3,7,8-TCDD, lead compounds, mercury compounds, arsenic
compounds, and cadmium compounds. This prioritization method for HAPs
for multipathway analysis has limitations and uncertainties, was based
on limited data, was not quantitative, and was based largely on
subjective decisions; therefore, the selection of among the most
important to assess for multipathway exposures and are considered a
reasonable starting point for further multipathway analyses.
5.6 SELECTION OF HAPS FOR FURTHER ANALYSIS
In the initial phase of the screening assessment 12 HAPs were
selected as priority. Pollutants were selected as priority in the
initial phase if they met one of the following three criteria:
(1) the MEI inhalation cancer risk was estimated to be greater than 1
in 10 million (i.e., 1 x 10"7) ; (2) maximum inhalation exposure
concentration was greater than one-tenth the RfC (i.e., if the HQ was
greater than 0.1); or (3) the emitted HAP is persistent in the
environment, tends to bioaccumulate, and emissions are significant
enough that there are potential concerns for human health from
multipathway exposure. The risk levels chosen for the first
two criteria (i.e., 1 x 10"7 and 1/10 the RfC) are lower than levels
that have been considered historically as levels for regulatory and
policy decisions (e.g., 1 x 10~s for cancer and RfC for noncancer).
These lower values were chosen for screening purposes so that it would
be unlikely that potentially important HAPs would be missed by screen.
That is, these conservative levels were chosen to ensure that all
potentially important HAPs would be identified by the screen. The
third criterion was primarily chosen to identify HAPs that are
considered a potential concern from multipathway exposure. Based on
these three criteria, 12 HAPs (arsenic, beryllium, cadmium, chromium,
dioxin/furans, nickel, n-nitrosodimethylamine, hydrogen chloride,
manganese, lead, mercury, and formaldehyde) were chosen to be
priorities for further assessment.
Radionuclides were also chosen as a priority for multipathway
assessment because previous risk assessments indicate that
radionuclides from utilities could potentially cause cancer risks
greater than 1 x 10~s for MEIs.-
In addition, three HAPs (HC1, HF, and acrolein) were chosen as
priority for assessment of potential noncancer effects due to short-
term (acute) exposures. The prioritization of HAPs for short-term
exposure analysis was based on review of health effects data,
emissions estimates, and recommendations from the peer review
panel.-'"'" Hydrogen chloride, HF, and acrolein were the three HAPs
considered to be of highest potential concern for health effects due
to short-term exposures. Table 5-10 presents the 15 HAPs that were
selected as priority based on the screening assessment.
5-18
-------
Table 5-10. Pollutants Considered Priority for Further Analysis
Based on Results of Screening Assessment
Pollutant
Acrolein3
Arsenic
Beryllium
Cadmium
Chromium
Dioxins/furans
Nickel
Radionuclides"
n-Nitroso-
dimethylamine0
Hydrogen chloride
Hydrogen flouride3
Manganese
Lead
Mercury
Formaldehyde
Priority
for coal
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
No
Priority
for oil
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
No
Priority
for gas
No
Yes
No
No
No
No
Yes
Yes
No
No
No
No
No
No
Yes
Inhalation
MEI cancer
risk>107
No
Yes
Yes
Yes
Yes
Yes (oil)
Yes
NA
Yes
No
No
No
No
No
Yes (gas)
Noncancer
Inhalation
HQ>0.1
No
NA
NA
NA
NA
NA
NA
NA
NA
Yes
No
Yes
No
No
NA
Priority for
multipathway
assessment
No
Yes
No
Yes
No
Yes
No
Yesb
No
No
No
No
Yes
Yes
No
HQ = Hazard quotient
MEI = Maximally exposed individual
NA = Not applicable
a Acrolein and hydrogen fluoride did not pass screen based on RfC analysis. However, these two HAPs were
identified as priority because of potential concern for acute exposure.
b Radionuclides were considered priority based on previous risk assessments conducted in the 1980s on
radionuclides from utilities.1
c The MEI risk estimate for n-Nitrosodimethylamine is highly uncertain (and likely to be a high estimate) because
emission estimates were based on one measured value and several measurements below the maximum detection
limit. Therefore, n-Nitrosodimethylamine was not selected as priority for further assessment.
5.7 LIMITATIONS OF SCREENING ASSESSMENT
The HAPs that were not chosen for further analysis were below the
screening level and not considered priority for this report. These
HAPs are considered lower priority and are not likely to present
significant risks to public health. Based on available data and the
screening analyses, the 53 HAPs that were not chosen for further study
are not likely to be a concern for public health. However, due to
uncertainties and limitations in the data, it is not possible to fully
and conclusively determine that they do not pose a threat to public
health. It is possible (although unlikely) that future data, such as
revised emissions data or new toxicologic data, could possibly warrant
further evaluation of some of these 53 HAPs in the future.
5-19
-------
5.8 REFERENCES
1. U.S. Environmental Protection Agency. Risk Assessments.
Environmental Impact Statement. NESHAPS for Radionuclides.
Background Information Document-Volume 2. EPA/520/1-89-006-1.
Office of Radiation Programs, Washington, DC. 1989.
2. U.S. Environmental Protection Agency. Integrated Risk
Information System (IRIS) Database, Environmental Criteria and
Assessment Office, Cincinnati, OH. 1994.
3. U.S. Environmental Protection Agency. Study of Hazardous Air
Pollutant Emissions from Electric Utility Steam Generating
Units—Interim Final Report. Volume III. Appendices H-M. EPA-
453/R-96-013C. October 1996.
4. U.S. Environmental Protection Agency. Estimating Exposure to
Dioxin-Like Compounds. Volume III: Site Specific Assessment
Procedures. Review Draft. EPA/600/6-88/005C. Office of
Research and Development. 1994.
5. U.S. Environmental Protection Agency. Provisional Guidance for
Qualitative Risk Assessment for Polycyclic Aromatic Hydrocarbons.
EPA/600/R-93/089. Washington, DC. 1993.
6. U.S. Environmental Protection Agency. Documentation of De
Minimis Emission Rates - Proposed 40 CFR Part 63, Subpart B,
Background Document. Research Triangle Park, NC. 1994.
7. U.S. Environmental Protection Agency. Interim Methods for
Development of Inhalation Reference Concentrations. EPA 600/8-
90/066A. Office of Health and Environmental Assessment,
Washington, DC. August 1990.
8. U.S. Environmental Protection Agency. Health Effects Assessment
Summary Tables. Office of Research and Development, Office of
Emergency and Remedial Response. Washington, DC. 1992.
9. Code of Federal Regulations. Title 40-part 50, Section 50.12.
Office of the Federal Register, National Archive and Recoreds
Administration. U.S. Government Printing Office. Washington,
DC.
10. U.S. Department of Health Services. NIOSH Pocket Guide to
Chemical Hazards. Public Health Service, Centers for Disease
Control, National Institute for Occupational Safety and Health,
Cincinnati, OH. 1994.
11. American Conference of Governmental Industrial Hygienists.
Threshold Limit Values. Cincinnati, Ohio. 1993-94.
5-20
-------
12. Calabrese and Kenyon. Air Toxics and Risk Assessment. Lewis
Publishers, Inc., Chelsea, MI. 1991.
13. California Air Pollution Control Officers Association. Air
Toxics "Hot Spots" Program. Revised 1992. Risk Assessment
Guidelines. California. October 1993.
14. Fradkin, L., R. J. F. Bruins, D. H. Cleverly, et al. Assessing
the risk of incinerating municipal solid waste: The development
and application of a methodology. In: Municipal Waste
Combustion and Human health. CRC Press, Palm Beach, FL. 1988.
15. U.S. Environmental Protection Agency. Deposition of Air
Pollutants to the Great Waters: First Report to Congress. EPA-
453/R-93-055. Office of Air Quality Planning and Standards,
Research Triangle park, NC. 1994.
16. U.S. Environmental Protection Agency. Methodology for Assessing
Health Risks Associated with Indirect Exposure to Combustor
Emissions, Interim Final. EPA/600/6-90/003. Office of Health
and Environmental Assessment, Washington, DC. 1990.
17. U.S. Environmental Protection Agency. Memo to Carol M. Browner,
July 29, 1994. "Draft 'Addendum to the Methodology for Assessing
Health risks Associated with Indirect Exposure to Combustor
Emissions'." EPA-SAB-IAQC-94-009b. Science Advisory Board.
Washington, DC. 1994.
18. U.S. Environmental Protection Agency. Schedule for Standards:
Methodology and Results for Ranking Source Categories Based on
Environmental Effects Data. Attachment A. EPA-453/R-93-053.
Office of Air Quality Planning and Standards, Research Triangle
Park, NC. 1993.
5-21
-------
6.0 INHALATION RISK ASSESSMENT
6.1 BASELINE ASSESSMENT OF INHALATION EXPOSURES AND RISKS FOR
13 PRIORITY POLLUTANTS
This chapter presents estimates of risks due to inhalation
exposure to 13 of the 14 priority HAPs identified in the screening
assessment (chapter 5). The assessment of risks presented in this
chapter is more refined and complete than the screening level
assessment presented in chapter 5. The risk estimates presented in
this chapter are believed to be more accurate and more comprehensive
than those presented in chapter 5. However, it is important for the
reader to understand that the risk estimates presented in this chapter
are still generally considered high-end estimates, and there are still
substantial uncertainties and data gaps in the risk assessment
presented in this chapter. Further assessment would be needed in
several areas to gain a better understanding of the actual risks posed
by electric utilities.
Radionuclides were the one priority HAP not included in this
chapter because the analysis for radionuclides requires an air
dispersion model that predicts the impacts of the radioactive decay
process. The radionuclide analysis is presented in chapter 9. In
this section, for the 13 priority HAPs, risks have been calculated
using the HEM for HAP emissions from all 684 utilities, and using the
standard HEM default options and assumptions described in chapter 4.
The HEM estimates ambient air concentrations within 50 km of each
utility. Therefore, the baseline risk estimates reflect only
inhalation exposure within 50 km of each utility (i.e., local
analysis). In addition, the baseline risks presented in section 6.1
do not account for background levels, long-range transport, complex
terrain, indirect exposures, or overlapping plumes. These issues are
analyzed and discussed in later sections of the report.
Not incorporating the above factors may lead to underestimating
risks. However, there are several important assumptions that were
incorporated into the baseline assessment that are generally
conservative (i.e., more likely to overestimate rather than
underestimate risks). For example, the baseline assessment assumes
that MEIs are exposed to the modeled concentrations for 70 years.
Also, the cancer potency values (i.e., lUREs) that were used in this
assessment are considered "upper limit" estimates.1 The lUREs
represent upper bound estimates of the cancer risks posed by these
HAPs. The true risks are not known and could be as low as zero.
These are just a few of the assumptions and uncertainties associated
with the baseline assessment. Later sections describe many of the
data inputs and default assumptions and discuss various issues and
uncertainties.
The HEM exposure modeling conducted for the screening assessment
(chapter 5) was very similar to the HEM exposure modeling conducted
for the baseline assessment (this section). The same default options
6-1
-------
described in chapter 4 and same input data were used. However, there
is one important difference. For the baseline assessment, a
distinction was made between urban and rural locations. If a utility
plant is located in an urban area, it was modeled using the urban mode
(i.e., dispersion is assumed to be characteristic of emissions emitted
by a facility in an urban location where there are buildings nearby).
If a utility plant is located in a rural location, it was modeled
using the rural mode (i.e., dispersion is assumed to be characteristic
of a facility located in a rural location). Dispersion of the
pollutant plume in an urban area is expected to exhibit greater
turbulence because of heat transfer and obstacles (i.e., large
buildings). Therefore, using the urban default setting typically
results in higher predicted air concentrations as compared to the
concentrations predicted using the rural default setting. In the
screening assessment, all plants were modeled using the urban default
because using the urban default typically leads to more conservative
estimates of human exposures and risks. However, using the urban and
rural distinction is believed to reflect more realistic conditions;
therefore, it was considered appropriate to use the urban versus rural
distinction in the baseline assessment, and in subsequent HEM modeling
analyses presented throughout chapter 6. The urban and rural options
(which differ in the assumptions for surface roughness) and their
impact on the risk estimates are discussed in detail in section 6.2.
The uncertainty analysis (presented in later sections of
chapter 6) suggests that the baseline risk estimates are generally
conservative, but within the range of reasonable estimates.
Therefore, the results presented in this section (baseline risk
estimates) are generally considered reasonably high-end estimates of
the risks due to inhalation exposure of utility HAP emissions within
50 km of the utility plants. This conservatism is considered
appropriate given EPA's mandate of public health protection. Central
tendency estimates of risks as well as background exposures and risks
are discussed, and analyzed to a limited degree, in later sections of
this chapter.
6.1.1 Baseline Inhalation Risks for Coal-Fired Utilities
for Priority HAPs
A total of 426 coal-fired units were modeled with the HEM using
1990 emissions and population data. Two of the plants resulted in
individual risks less than 1 x 10 "9 and were excluded from the
presentation of results, thus reducing the total number of plants to
424. Table 6-1 summarizes the following: the predicted high-end MEI
risks; high-end MIRs; the high-end estimate of the number of persons
predicted to be exposed above individual cancer risk levels of 1
chance in 1 million (i.e., 1 x 10"s); the number of plants whose
emissions result in those risk levels; and the maximum HQs.
6.1.1.1 Individual Cancer Risk. Table 6-1 presents the
estimated risks due to inhalation exposure within 50 km for each HAP
across all coal-fired plants. As stated previously, the MEI is
6-2
-------
Table 6-1. Summary of High-End Risk Estimates from Chronic
Inhalation Exposure by HAP for 424 U.S. Coal-Fired Utilities
Based on the Baseline Inhalation Risk Assessment
Pollutant
Arsenic
Beryllium
Cadmium
Chromiumd
Dioxin/furans
Hydrogen chloride
Lead
Manganese
Mercury
Nickel6
Total
Carcinogens
Highest
MEI cancer
risk3
3x10'6
3x10'7
2x10'7
2x10'6
5x10'8
NA
NA
NA
NA
4x10'7
4x10'6
Highest
cancer
MIRb
2x10'6
2x10'7
1 x 1Q-7
1 x 1Q-6
3x10'8
NA
NA
NA
NA
2x10'7
3x10'6
Population
with risk
>10-6
850
0
0
110
0
NA
NA
NA
0
0
NA
# Plants
with MIR
>10-6
2
0
0
1
0
NA
NA
NA
0
0
2
Cancer
incidence
(cases/yr)c
0.05
0.004
0.0006
0.02
0.001
NA
NA
NA
NA
0.003
0.1
Noncarcinogen
Max. HQ
NA
NA
NA
NA
NA
0.1
0.001
0.05
-
NA
0.1 to 0.2
HQ = Hazard quotient, the ratio of exposure concentration to the reference concentration (RfC). HQ values below 1 are not
expected to result in adverse effects.
MEI = Maximally exposed individual, which is calculated using the highest annual average concentration predicted with the
HEM. An individual may or may not be exposed at that point. This value may be greater than the MIR.
MIR = Maximum individual risk is the highest risk identified at the centroid of a census tract to which a population is assigned.
See chapter 4 for description of MEI and MIR.
NA = Not available.
Total = Total MEI are the sum of MEI for individual HAPs within a plant. The total HQ (=HI) is the sum of the HQs within a
plant.
a Of all 424 coal-fired plants modeled with the HEM, this is the estimated increased inhalation cancer risk for a theoretical person
assumed to be exposed for 70 years to the highest modeled HAP ambient air concentration around any of the 424 plants.
b Of all 424 coal-fired plants modeled with the HEM, this is the highest MIR.
0 The cancer incidence could be up to roughly 7 times greater when considering the impacts of long-range transport (i.e., exposure
outside of 50 km) from all coal-fired utilities combined. See section 6.6 for discussion of long-range transport.
d Assumes that 11 percent of total chromium emitted is hexavalent chromium, the species of chromium responsible for
carcinogenic potential. Trivalent chromium, which would also be present, is assumed not to have carcinogenic effects.
e The nickel emitted is a mixture of various nickel compounds such as soluble nickel. This analysis assumes that the mix of nickel
compounds emitted is 50% as carcinogenic as nickel subsulfide.
calculated based on the maximum modeled ambient concentration even
though a person may or may not reside in the vicinity of the maximum
concentration. The MEI risk was highest for arsenic (a Class A, human
carcinogen) at 3 x I0~6 for the "highest-risk" coal-fired plant. The
highest estimated MIR at a single plant was 2 x I0~s for arsenic.
Table 6-1 shows that arsenic emissions from two plants resulted in
6-3
-------
MIRs greater than or equal to 1Q~6. The MIRs for the remaining 424
coal-fired plants were lower than 1 x I0~6. Figures 6-la and 6-lb show
that most inhalation risks were considerably lower than 1 x 10 ~6. The
risk for chromium assumes that 11 percent of total chromium is
hexavalent chromium, (a Class A, human carcinogen). The limited
emissions speciation data (described in Appendix H of the EPA Interim
Final Utility Report, Volume III)2 found hexavalent chromium between
0.3 and 34 percent of total chromium. The average percentage of
chromium VI based on limited speciation data was 11 percent. The
other HAPs do not appear to make a significant contribution to the MIR
from coal-fired plants. Figures 6-la and 6-lb present the
distribution of plants at different MIR levels for the major
carcinogens. Arsenic and chromium are the major contributors of
inhalation cancer risks from coal-fired utilities. Of the 424 coal-
fired plants, the median MIR is 2 x 10 "8 for arsenic and 2 x 10"9 for
chromium. The 90th percentile MIR of all 424 plants modeled is
1 x 10"7 for arsenic and 4 x 10"8 for chromium (that is, 10 percent of
the plants have MIR levels above, and 90 percent of the plants have
MIR levels below, these risk levels).
The total MIR due to inhalation exposure to the aggregate of HAPs
for each plant was calculated by summing the MIR for each HAP for each
plant. There are two coal-fired plants with total MIRs greater than
1 x 10"s. The highest total MIR for a single coal-fired plant is
3 x 10"s. Of the 424 coal-fired plants modeled, the median total MIR
is 5 x 10"8, and the 90th percentile is 2 x 10"7 (that is, 10 percent of
plants pose an MIR greater than 2 x 10 ~7) .
6.1.1.2 Population Cancer Risk. The population distribution at
various risk levels is shown in Table 6-2 for each of the five major
carcinogenic HAPs. As with the MIR, arsenic and chromium are the
major contributors. The high-end estimate of number of people exposed
to risks of 1 x 10"s or greater from exposure to arsenic is 850 and
from exposure to chromium is about 107. That is, based on this
assessment, it is unlikely that more than 850 people are exposed to
inhalation risks greater than 1 x 10"s due to coal-fired utility
nonradionuclide HAP emissions, and most likely fewer people are
exposed (possibly as few as zero) to this level of inhalation risk.
The HEM also calculated the annual incidence of cancer expected
for each of the HAPs due to inhalation exposure within 50 km. As
shown in Table 6.1, the high-end estimate for total cancer incidence
from the nonradionuclide carcinogenic HAPs was estimated, using the
HEM, to be as high as 0.1 cancer case per year for coal-fired plants
due to emissions within 50 km of each plant. However, the estimate
for incidence increases by about a factor of seven when considering
emissions dispersion beyond 50 km (see section 6.6). Arsenic and
chromium are again the major contributors and account for almost 90
percent of the estimated cancer incidences.
6-4
-------
Arsenic
250 -,
5; 200 -
T3
il 150 -
"ro
o
^ 100 -
o
| 50 -
^
z
n
148
235
40
2
<1E-£
1E-8to1E-7 1E-7to1E-6
Maximum Individual Risk (MIR)
1E-6 to 1E-5
Chromium
£-\J\J
1
rT 200-
n i- «-'«-'
T3
iZ 150 -
i
ro
0
^ 100 -
o
te
-g 50 -
3
-z.
n -
218
iao
10 .
1
<1E-8
1E-8 to 1E-7 1E-7 to 1 E-6
Maximum Individual Risk (MIR)
1 E-6 to 1 E-5
Total: All Carcinogenic HAPs
350 -
ro 30° -
D.
-D 250 -
11 200 -
ro
0 150 -
"5
£ 100 -
1 50 -
n -
44
289
91
2
1 E-8
1 E-8 to 1 E-7 1 E-7 to 1 E-6
Maximum Individual Risk (MIR)
1 E-6 to 1 E-5
Figure 6-la. Number of Coal-Fired Utilities Posing Various Levels
of Maximum Individual Risk (by levels of MIR)
6-5
-------
450
400
350 —
300
a>
il 250
To
o
^ 200
150 —
100
50 —
415
1 0
Cadmium
D<1E-8 H1E-8to1E-7 E1E-7 to 1E-6 B1E-6to1E-5
218
196
148
10
1
40
• • • ' h
Chromium Arsenic
Maximum Individual Risk (MIR)
44
91
Total: All Carcinogenic HAPs
Figure 6-lb. Number of Coal-Fired Utilities Posing Various Levels
of Maximum Individual Risk (by levels of MIR)
-------
Table 6-2. Summary of High-End Estimates of Population Exposed
at Various Levels of Inhalation Risk or Greater by HAP: Coal-
Fired Utilities
Risk level
5x10'6
2.5 x10'6
1 x10'6
5x10'7
2.5 x10'7
1 x10'7
Arsenic
0
0
852
5,990
88,800
1,710,000
Chromium
0
0
107
2,160
8,630
80,500
Beryllium
0
0
0
0
0
1,280
Cadmium
0
0
0
0
0
107
Dioxins/furans
0
0
0
0
0
0
6.1.1.3 Noncancer Risk. The maximum HQ estimated for
noncarcinogenic HAPs emitted from coal-fired power plants was 0.1 for
HC1. The next highest was 0.05 for manganese. HQ values for all
other HAPs were at least an order of magnitude lower. This assessment
does not include background concentrations due to other sources.
6.1.2 Baseline Inhalation Risks for Oil-Fired Utilities
A total of 137 oil-fired plants were modeled using 1990 HAP
emissions and population data. The HEM estimated the high-end
individual and population risks for each of the HAPs evaluated. Eight
plants had risks less than 1 x 10 "9 and were excluded from the
presentation. Table 6-3 presents the results.
6.1.2.1 Individual Cancer Risk. For oil-fired utilities, the
HEM predicts that people live in the location of highest modeled
ambient air concentration; therefore, the MEI and the MIR are equal.
The maximum MEI/MIR estimated for a single carcinogenic HAP across all
plants was 5 x 10"5 from inhalation of nickel compounds.
There are numerous uncertainties that are discussed and analyzed in
later sections, but the EPA believes that the uncertainties associated
with nickel speciation are worth discussing here. There are substantial
uncertainties associated with nickel speciation. In this analysis, as a
conservative assumption, the mix of nickel compounds emitted by oil-fired
utilities was assumed to be 50 percent as carcinogenic as nickel
subsulfide, which is a Class A human carcinogen and has the highest
cancer potency of all nickel compounds evaluated by the EPA. The limited
speciation data indicate that 3 to 26 percent of nickel emissions (from
oil-fired utilities) are sulfidic nickel.2'3 It is not known how much of
the sulfidic nickel is in form of nickel subsulfide. The remainder of
the nickel is a combination of various nickel compounds for which the EPA
has not yet determined the carcinogenic potential. Many nickel compounds
are thought to have some carcinogenic potential via inhalation exposure
although the potency is not known. This issue is discussed further in
section 6.10.
6-7
-------
Table 6-3. Summary of the High-End Risk Estimates from
Inhalation Exposure for Priority HAPs for 137 U.S. Oil-Fired
Utilities Based on the Baseline Risk Assessment
Pollutant
Arsenic
Beryllium
Cadmium
Chromium"
Dioxin/furans
Hydrogen chloride
Lead
Manganese
Mercury
Nickel0
Total
Carcinogens
Highest
Cancer MIR
1 x1Q-5
7x10'7
2x10'6
5x10'6
1 x 1Q-7
NA
NA
NA
NA
5X10'5
6x1Q-5
Population
with risk
>10-6
2,400
0
45
2,300
0
0
0
0
0
110,000
NA
# Plants with
MIR
^o-6
2
0
1
1
0
0
0
0
0
11
11
Cancer
incidence3
(cases/yr)
0.04
0.002
0.005
0.02
0.0007
NA
NA
NA
NA
0.2
0.3
Noncarcinogen
MAXHQ
NA
NA
NA
NA
NA
0.06
0.004
0.04
-
NA
NA
HQ = Hazard quotient, which is the ratio of exposure concentration to the reference concentration (RfC). HQ values below 1
are not expected to result in adverse effects.
MIR = Maximum individual risk is the highest risk identified at the centroid of a census tract to which a population is assigned.
See chapter 4 for description of MIR and MEI.
NA = Not available.
Total = Total MIR is the sum of the MIR for individual HAPs within a plant. The total HQ (=HI) is the sum of the HQs within a
plant.
a The cancer incidence could be up to roughly 7 times greater when considering the impacts of long-range transport (i.e., exposure
outside of 50 km) from all utilities combined. See section 6.6 for discussion of long-range transport.
b Assumes that 18 percent of total chromium emitted is hexavalent chromium, the species of chromium responsible for
carcinogenic potential. Trivalent chromium, which would also be present, is assumed to be noncarcinogenic.
0 This analysis conservatively assumes that all nickel emitted from utilities is 50 percent as carcinogenic as nickel subsulfide (the
highest potency of nickel compounds tested). However, the nickel emitted is a mixture of various nickel compounds such as
soluble nickel, nickel oxides, and sulfidic nickel. Emissions tests indicate that 3 to 26 percent of the nickel emissions are sulfidic
nickel. Nickel subsulfide is one of the possible forms of sulfidic nickel. It is not known how much of the sulfidic nickel is in the
form of nickel subsulfide. Many nickel compounds are thought to have carcinogenic potential although the potency is not known.
See section 6.10 for further discussion and analysis of nickel speciation uncertainty.
6-8
-------
Figures 6-2a and 6-2b show the distribution of plants at
different MIR levels for the major carcinogenic HAPs. The median MIR
across all plants is 1 x 10~7 for nickel and 1 x 10"8 for arsenic. The
90th percentile for MIR among plants is 5 x 10 "7 for nickel (that is,
90 percent of plants are estimated to pose risks less than 5 x 10"7 due
to nickel emissions) and 1 x 10"7 for arsenic.
The total MIR was calculated for each facility by summing the
MIRs for individual HAPs. The highest total high-end MIR from the sum
of high-end risks for each carcinogen is 6 in 100,000 (6 x 10 ~5) at
only one plant. The second and third highest-risk oil-fired plants
pose MEI inhalation risks of 2 x 10"5 and 1 x 10"5, respectively. The
total high-end MIR exceeded 1 x 10"s as a result of HAP emissions from
11 oil-fired plants. The median total MIR for all plants is
approximately 4 x 10"7, and the 90th percentile is approximately
2 x 10"s. However, these estimates are considered conservative, high-
end estimates because they are mainly due to nickel emissions and the
assumption that the mix of nickel compounds is 50 percent as
carcinogenic as nickel subsulfide (see section 6.10 for discussion).
6.1.2.2 Population Cancer Risk. The population distribution at
various risk levels is shown in Table 6-4 for each of six carcinogens.
As with the MIR, nickel, arsenic, and chromium are the major
contributors to the total population exposed to risk levels of 1 in
1,000,000 (1 x 10"s) or more. The high-end estimate for number of
people exposed to risks of 1 x 10~s or greater is 110,000 for nickel,
and about 2,400 for arsenic and chromium. That is, based on this
assessment, it is unlikely that more than 110,000 people are exposed
to inhalation risks greater than 1 x 10"s due to oil-fired utility
nonradionuclide HAP emissions, and most likely fewer people are
exposed to this level of inhalation risk.
Incidences from each nonradionuclide HAP were summed to estimate
total cancer incidence, which was estimated to be as high as 0.3 case
per year from these 137 oil-fired plants. Nickel accounts for over 60
percent of the total annual incidence and arsenic contributes roughly
about 15 percent.
As with individual risk estimates for oil-fired plants, there are
significant uncertainties associated with these population risk
estimates because of the uncertainties associated with nickel
speciation and other uncertainties as discussed in later sections of
this report.
6.1.2.3 Noncancer Risks Due to Chronic Exposures. The highest
HQ resulting from oil-fired utility emissions was 0.04 for manganese.
6-9
-------
Arsenic
J2
§ 80 -,
0.
"g 60 -
LJ_
b
15 9n
i- zu
0)
_Q
p
58
40
20
1 1 0
I I I I I
< 1E-8 1E-8 to 1E-7 1E-7 to 1E-6 1E-6 to 1E-5
Maximum Individual Risk (MIR)
1 E-5 to 1E-4
1 E-4 to 1E-3
80 -i
72
Chromium
"S 60 -
iZ
o
° 20 -
0)
p n -
47
1 0 0
< 1 E-8 1 E-8 to 1 E-7 1 E-7 to 1 E-6 1 E-6 to 1 E-5
Maximum Individual Risk (MIR)
1 E-5 to 1 E-4
1 E-4 to 1 E-3
CO
I 60
°- 50
| 40
0= 30
"B 20
_g 10
I 0
51
Nickel
47
77>e nickel risks presented here
are based on an assumption that
the mix of nickel compounds
emitted by utilities are 50 percent
28
as carcinogenic as nickel subsulfide.
9
I 1 2 °
< 1 E-8 1 E-8 to 1 E-7 1 E-7 to 1 E-6 1 E-6 to 1 E-5
Maximum Individual Risk (MIR)
1 E-5 to 1 E-4
1 E-4 to 1 E-3
Total: All Carcinogenic HAPs
48
52
T, ^U-
2 40-
LJ-
6 3°-
•B 20 -
8 10-
E n _
26
9
I 1 2 o
I I i i
< 1 E-8 1 E-8 to 1 E-7 1 E-7 to 1 E-6 1 E-6 to 1 E-5
Maximum Individual Risk (MIR)
1 E-5 to 1 E-4
1 E-4 to 1 E-3
Figure 6-2a. Number of Oil-Fired Utilities Posing Various Levels
of Maximum Individual Risk (by levels of MIR)
-------
120
100
re 80
5 60
0)
£1
E 40
3
20
98
D<1 E-8 B1E-8 to 1E-7 D1E-7 to 1 E-6 D] 1E-6 to 1E-5 • 1E-5 to 1E-4
29
72
47
58
49
20
* The nickel risks presented here are based on an assumption
that the mix of nickel compounds emitted by utilities are 50
percent as carcinogenic as nickel subsulfide.
51
28
1 1
47
Cadmium
Chromium Arsenic
Maximum Individual Risk (MIR)
Nickel*
48
52
26
Total: All Carcinogenic HAPs
Figure 6-2b. Number of Oil-Fired Utilities Posing Various Levels
of Maximum Individual Risk (by levels of MIR)
-------
Table 6-4. Summary of High-End Estimates of Population Exposed
Through Inhalation at Various Levels of Risk or Greater from Oil-
Fired Utilities
Risk level
5x1Q-5
2.5 x1Q-5
1 x1Q-5
5x10'6
2.5 x10'6
1 x10'6
5x10'7
2.5 x10'7
Nickel
45
89
2,200
2,300
9,900
110,000
1,600,000
7,000,000
Arsenic
0
0
45
89
2,280
2,370
32,600
287,000
Chromium
0
0
0
45
89
2,280
2,280
9,490
Cadmium
0
0
0
0
0
45
89
2,280
Beryllium
0
0
0
0
0
0
45
89
Dioxins/
furans
0
0
0
0
0
0
0
0
Note: Double counting of population around facilities within 50 km of each other may occur. Exposed individuals
are included in the statistics for each plant within 50 km, presumably at different risk levels. See Section 6.5 and
Appendix F for further discussion of double counting and related issues.
6.1.3 Baseline Risks from Gas-Fired Utilities
Risks were estimated from 267 gas-fired facilities. Table 6-5
summarizes the results. The HAP emissions from only one plant
resulted in high-end risks greater than 1 in 10 million (10~7) with 23
persons exposed above that level. For noncarcinogens, the maximum HQ
was 1 x 10"7. The estimated risks due to HAP emissions from gas-fired
utilities are low.
6.2 DISTINGUISHING BETWEEN URBAN AND RURAL LOCATIONS
The HEM has two distinct modeling options (urban or rural)
intended to simulate atmospheric dispersion behavior of gases via
different surface roughness.4 The urban option assumes that there are
buildings near the emission source and that turbulence results because
of these surfaces and other urban effects such as heat transfer from
buildings and roadways. The rural option assumes that there are not
any major buildings nearby, and therefore emissions dispersion plumes
are not as turbulent. Typically, for tall stacks, the urban option
predicts higher exposure concentrations and, therefore, higher risks
to nearby populations.
In the screening assessment(presented in chapter 5), the urban
option was used in all modeling runs. However, to assess the impact
of this default option on the risk assessment results, all of the
priority HAPs were modeled distinguishing between urban and rural
locations. As an option provided by the U.S. EPA Guidelines on Air
6-12
-------
Table 6-5. Summary of High-End Inhalation Risk Estimates for
Gas-Fired Utilities
Pollutant
Arsenic
Lead
Mercury
Nickel3
Carcinogens
MEI risk
2x10'7
NA
NA
1 x10'7
Population MIR
>10-6
0
NA
NA
0
# Plants MIR
>10-6
0
NA
NA
0
Noncarcinogen
HQmax
NA
1 x10'7
NA
NA
HQ = Hazard quotient, which is the ratio of exposure concentration to the reference concentration (RfC). HQ
values below 1 are not expected to result in adverse effects.
MEI = Maximum exposed individual, which is calculated using the highest annual average concentration. An
individual may or may not be exposed at that point. This value may be greater than the MIR, which is
calculated at the centroid of a census block.
MIR = Maximum individual risk is the highest risk identified at the centroid of a census tract to which a population
is assigned.
NA = Not available
a The nickel emitted is a mixture of various nickel compounds. This analysis assumes that all nickel emitted is 50
percent as carcinogenic as nickel subsulfide.
Quality Models (40 CFR, Appendix W to Part 51) , 4 it was assumed that if
21,000 people lived within a 3-km radius of the plant (i.e., density =
750 people/km2) , then the area was urban and was modeled using the
urban modeling option. If less than 21,000 people lived within a 3-km
radius, then the area was considered rural and the rural modeling
option was chosen.4 Tables 6-6 and 6-7 present the results. There
were some differences in site-by-site estimates. As Tables 6-6 and
6-7 show, choosing the urban default option versus a more refined
selection of surface roughness options has some impact on the overall
results. There were only slight changes in the results for oil-fired
utilities. Many oil-fired facilities are located in urban areas. The
differences in the risk estimates from coal-fired utilities were
greater. Generally, risk estimates are lower when urban and rural
modeling distinctions are used.
The use of the refined analysis, whereby surface roughness
distinction was made for urban and rural locations, was considered
appropriate for all the inhalation exposure modeling analyses
presented in chapter 6 since it is believed to more realistically
reflect the location of utilities and the impacts of rural and urban
conditions on the dispersion of pollutants. Although the EPA believes
using this distinction is appropriate, there are still uncertainties
and limitations to this approach, which are discussed in later
sections of this report.
6-13
-------
Table 6-6. Comparison of High-End Inhalation Cancer Risk
Estimates Based on (1) HEM Modeling Using Urban Default
Assumption and (2) HEM Modeling Using Urban vs. Rural Distinction
Pollutant and fuel
As, (Coal)
Cr, Coal
(assuming 11%
CrVI)
Be, Coal
Cr, Oil (assuming
18% CrVI)
Be, Oil
Cd, Oil
Ni, Oil
As, Oil
Urban default
MEI risk
6x1Q-6
3x10'6
6x1Q-7
5x10'6
7x10'7
1.6 x10'6
5x1Q-5
1 x1Q-5
Cancer
incidence
(cases/year)
0.08
0.03
0.006
0.02
0.002
0.007
0.3
0.05
Population3
with cancer
risk >10'6
21,000
890
0.0
2,300
0.0
45
155,000
4,600
Rural vs. urban
MEI risk
3x10-6
2x10'6
3x10-7
5x10'6
7x10-7
1.6x10-6
5x10-5
1 x10-5
Cancer
incidence
(cases/year)
0.05
0.02
0.004
0.02
0.002
0.005
0.2
0.04
Population3
with cancer
risk>10-6
850
110
0.0
2,300
0
45
110,000
2,400
MEI = Maximally exposed individual
a The number of people estimated to be exposed to ambient air concentrations causing a high-end estimated
increased risk of cancer of 1 in 1 million or greater.
Table 6-7. Comparison of High-End Inhalation Noncancer Risk
Estimates Based on (1) HEM Modeling Using Urban Default
Assumption and (2) HEM Modeling Using Urban vs. Rural Distinction
Pollutant and fuel
Hcl, from Coal
Mn, from Coal
Urban default
MEIHQ
2.3/20 = 0.12
0.02/0.05 = 0.4
Selection of appropriate setting
(rural vs. urban)
MEIHQ
2.3/20 = 0.12
0.002/0.05 = 0.04
HQ = Hazard quotient
MEI = Maximally exposed individual.
6-14
-------
6.3 INHALATION RISK ESTIMATES FOR THE YEAR 2010
The EPA analyzed potential inhalation risks from utility
emissions for the year 2010. This analysis was conducted to estimate
hazards and risks to public health after imposition of the
requirements of the Clean Air Act Amendments of 1990. The primary
differences between the 1990 and 2010 scenario are increased emissions
from coal-fired utilities and decreased emissions from oil-fired
utilities. Other predicted changes include the installation of
scrubbers for a small number of facilities, the closing of a few
facilities, and an increase in production of other facilities. The
details of the expected changes are explained in chapters 2 and 3.
Similar to any analyses that predict future events, significant
uncertainties are associated with the method used for projecting risks
of HAP emissions to the year 2010. Moreover, there are several other
potential future actions or programs (e.g., PM and ozone NAAQS
implementation, climate change programs, electricity restructuring),
which could have an impact on HAP emissions, that were not considered
in the projections made for this 2010 analysis because of the
uncertainties and unknowns about how these programs will affect HAP
emissions. However, even with these limitations, the method used by
EPA is considered reasonable given the available data.
The exposures and risks for the year 2010 were estimated using
the HEM, utilizing the same modeling assumptions, defaults, and inputs
used in the 1990 risk estimates, except that the emissions inputs were
changed to 2010 estimates. Instead of modeling all 15 priority HAPs a
second time, the EPA modeled a subset of HAPs that appear to present
the majority of the inhalation risks from utility emissions. The
analysis of this subset of priority HAPs provides information
regarding the anticipated potential public health risks due to
inhalation for the year 2010.
The results (Tables 6-8 and 6-9) indicate that, based on the
expected changes between 1990 and the year 2010, the inhalation risks
from coal-fired utilities will not change substantially, and the risks
from oil-fired utilities will decrease by roughly a factor of 2.
6.4 ASSESSMENT OF POTENTIAL RISKS DUE TO SHORT-TERM EXPOSURE
The potential for exceeding short-term reference exposure levels
(RELs)5 was evaluated for compounds emitted from coal- and oil-fired
utilities. The RELs (1-hour averages) are set to prevent adverse
acute responses in the exposed population. The pollutants of highest
concern were acrolein, HC1, and HF because these pollutants are
potentially emitted in significant quantities and are toxic due to
short-term (acute) exposures. Although the Agency has not determined
RELs for these compounds, REL values were obtained from the California
Air Pollution Control Officers Association (CAPCOA) Air Toxics 'Hot
Spots' Program Risk Assessment Guidelines, October 1993.5 The CAPCOA
RELs are listed in Table 6-10.
6-15
-------
Table 6-8. Estimated High-End Inhalation Cancer Risks for the
Year 2010 Compared to 1990 for Coal- and Oil-Fired Utilities
Pollutant and fuel
As from Coal
Be from Coal
Cd from Coal
Cr from Coal (11%
CrVI)
Dioxins from Coal
Ni from Coal
CrfromOil(18%Cr
VI)
Be from Oil
Cd from Oil
Ni from Oil
Dioxins from Oil
As from Oil
Cancer risk 2010
MEI risk3
3x10'6
3x10'7
3x10'8
1.4 x10'6
6x10'8
2x10'7
3x10'6
4x10'7
8x10'7
3x1Q-5
7x10'8
7x10'6
Cancer
incidence
(cases/year)
0.051
0.004
0.0007
0.021
0.0012
0.003
0.009
0.0008
0.0026
0.1
0.0004
0.026
Population w/
MIR>10-6
590
0.0
0.0
399
0.0
0.0
89
0.0
0.0
11,000
0.0
2,300
Cancer risk 1990
MEI risk3
3x10-6
3x10-7
2x10-7
2x10-6
5x10-8
4x10-7
5x10-6
7x10-7
2x10-6
5x10-5
1 x 1 0'7
1 x 10-5
Cancer
incidenceb
(cases/year)
0.045
0.0035
0.0006
0.02
0.001
0.003
0.02
0.0017
0.0053
0.2
0.0007
0.042
Population w/
MIR >10-6
852
0.0
0.0
107
0.0
0.0
2,300
0.0
45
110,000
0.0
2,400
Note: The EPA used urban vs. rural modeling data distinction in this analysis.
MEI = Maximally exposed individual.
MIR = Maximum individual risk is the highest risk identified at the centroid of a census tract to which a population is assigned.
a These MEI risk estimates are for the "highest risk" plant.
b This is the estimated cases of cancer predicted to occur in the United States due to emissions of this HAP from all utilities of that
fuel type based on the HEM analysis.
Table 6-9. Estimated High-End Inhalation Noncancer Risks for
Coal-Fired Utilities for the Year 2010 Compared to the Year 1990
Pollutant
HCI
Manganese
RfC (M9/m3)
20
0.05
Highest MEI
Cone, for 2010
2.6 A^g/m3
0.003 A^g/m3
Maximum HQfor
2010
0.1
0.06
Highest MEI Cone.
for 1990
2.3 A^g/m3
0.002 A^g/m3
Maximum HQfor
1990
0.1
0.05
HQ = Hazard quotient.
MEI = Maximally exposed individual.
RfC = Reference concentration.
6-16
-------
Table 6-10. Noncancer Reference Exposure Levels (Acute) from
CAPCOA5
Pollutant
Acrolein
Hydrochloric acid
Hydrogen fluoride
REL - Hourly average concentration (/zg/m3)
2.5
3,000
580
CAPCOA = California Air Pollution Control Officers Association.
REL = Reference exposure level.
The utilities modeled included the coal-fired and the oil-fired
utility that presented the highest predicted long-term concentrations
as determined from the earlier HEM screening analysis. In addition,
the largest emitter of each compound from a coal-fired and oil-fired
utility was modeled. Note that acrolein was not detected in the
emissions tests for oil-fired utilities.
6.4.1 Methodology. The EPA used a short-term air dispersion
model (called TSCREEN) that considers the potential range in
meteorological conditions at the utility plant site to estimate the
maximum 1 hour concentration of the three compounds in the vicinity of
selected coal- and oil-fired utilities. TSCREEN provides estimates of
1-hour concentrations at various distances from the stack being
analyzed. The user specifies the minimum distance to the stack at
which concentrations will be predicted. For all utilities modeled,
100 meters from the stack was selected.
The reported concentrations are the maximum predicted from a
range of atmospheric stability classes and windspeeds. The modeler
must also specify whether urban or rural meteorological conditions
exist at the utility site. Urban was selected to maximize the
predicted concentrations.
Each of the selected plants emitted the HAPs from several stacks
at the site. Because the TSCREEN model can evaluate only one emission
point at a time, some adjustments were required for each utility's
emission parameters. The concept was to select one stack and one
emission rate with one set of stack parameters that would represent
the multiple stacks and their corresponding emissions and stack
parameters. If the stacks at each utility varied in height or other
release characteristics (e.g., stack temperature, stack gas exit
velocity), emissions were assumed to be emitted under conditions to
maximize downwind concentrations: from the shortest stack present,
the lowest temperature among the stack characteristics, and the lowest
exit velocity (see Table 6-11). The emissions rate was calculated by
summing the emissions from each stack.
6-17
-------
Table 6-11. Sample Stack Parameters for Typical Utility Plant
Stack
1
2
3
4
Stack height (m)
75
75
70
75
Exit velocity (mis)
15
15
22
12
Stack temperature (K)
400
400
390
410
To illustrate this methodology, a sample utility is presented in
Table 6-11. The resulting inputs to the dispersion model for this
sample utility would have been one stack with a stack height of 70 m,
exit velocity of 12 m/s, and temperature of 390 K. An average of the
inside stack diameters for the four stacks would be used (see
Table 6-12).
6.4.2 Results
As shown in Table 6-13, for all scenarios and all pollutants
modeled, the predicted maximum concentrations were more than 100 times
lower than the RELs. The emission rate used for each compound
represents an average. The analysis does not address peak short-term
emissions that may result from upsets or other atypical operations.
Peak emission episodes would reduce the gap between predicted maximum
concentrations and REL, but the peak hourly emission rates are not
expected to be 100-fold higher than the average.
The TSCREEN can also incorporate terrain characteristics.
Terrain was not considered an important factor in the analysis since
the utilities that caused the highest individual risk in the HEM
analysis were located in relatively flat terrain. (The effects of
terrain are analyzed in Appendix G of the EPA Interim Final Utility
Report, Volume II.)6 Although hilly terrain can cause an estimated
15-fold higher predicted long-term concentration than flat terrain,
this increase would still not result in exceedances of RELs for the
three compounds.
6.5 OVERLAPPING PLUMES/DOUBLE COUNTING
In general, the default standard mode of operation for the HEM is
to evaluate exposure to each source, one at a time, out to 50 km from
the plant. Each source's exposure is independently estimated, and
detailed exposure estimates are not saved for the next source's
exposure analysis. Summary information, such as the total numbers of
people who are exposed, is saved. Thus, if two plants are located
very close together, the HEM would independently estimate the total
number of people exposed to each plant's emission and sum the two
totals even though the same people are being exposed to both plant's
6-18
-------
Table 6-12. Stack and Emission Values Input to TSCREEN
Pollutant
Stack height (m)
Stack gas exit
velocity (mis)
Stack diameter
(m)
Stack gas
temperature (K)
Emission rate
(9/s)
COAL
HF
HCI
Acrolein
49
49
49
45.7
47.5
45.7
2.5
2.5
2.5
395
395
395
0.42
9.07
0.01
OIL
HF
HCI
42
42
12.3
12.3
3.0
3.0
396
396
0.06
1.24
Table 6-13. Results of the TSCREEN Model
Pollutant
Acrolein
HCI
HF
Reference
exposure
levels
(hourly avg
^g/m3)
2.5
3,000
580
Coal-fired maximum
predicted concentration
(hourly avg A*g/m3)
0.016
21.5
1.0
Oil-fired maximum
predicted concentration
(hourly avg ^g/m3)
not emitted
5.5
0.3
How much lower?
AAC/Pred
coal
150
140
580
oil
1,200
2,100
Note: Since the largest emissions are generally associated with taller stacks, other analyses indicated that the
estimated concentrations were generally a factor of 2 lower than that presented.
emissions. In this mode, the HEM will most likely overestimate the
number of people who are exposed when two or more plants are within 50
km of each other. This effect has been called "double-counting."
Although not intuitive, experience has shown that this effect is not
of great concern when estimating the risks to the MEI and to the
population as a whole. Because of the linear nature of the exposure
and risk models, the population risks (cancer cases per year) are the
same whether one calculates the exposure one plant at a time or
calculates the exposure from nearby plants together; only the number
of people who are estimated to be in the exposed group will differ.
In the case of the risk to the MEI, nearby plants can only
significantly change the estimated maximum concentration when plants
of equal emission rates are located very close to each other, perhaps
within several hundred meters. This is very unlikely for the utility
industry.
6-19
-------
There is an option to the HEM, called single-count, which can
provide further insight into this potential problem of double-counting.
This option still evaluates exposure on a source-by-source basis, but
exposure is calculated for each population census block within 50 km and
this detailed information is saved (stored in the computer memory).
As each source is considered, the exposure estimate for each census
block is added to the previous source's exposure estimates at the same
census block. At the end of the computer run, the computer has a
total exposure estimate for each census block in the United States
and, by adding the census block exposure estimates together, provides
a national level estimate of total exposure. For this study, single-
count HEM runs have indicated that individuals may be living within 50
km of up to 12 coal-fired plants or 17 oil-fired plants; thus, a
concern has arisen over multiple exposures to many plants. However,
the single-count analyses conducted for arsenic emissions indicate
that overlapping effects from nearby sources do not significantly
change the estimated risks (see Table 6-14) .
6.6 ASSESSMENT OF EXPOSURE DUE TO LONG-RANGE TRANSPORT
6.6.1 History and Background Information
During the mid-1970s, SRI International developed a Lagrangian
puff air pollution model called the EUROPEAN Regional Model of Air
Pollution (EURMAP) for the Federal Environment Office of the Federal
Republic of Germany.7 This regional model simulated monthly S02 and
sulfate (SO^") concentrations and wet and dry deposition patterns, and
generated matrices of international exchanges of sulfur for 13
countries of western and central Europe. In the late 1970s, the EPA
sponsored SRI International to adapt and apply EURMAP to eastern North
America. The adapted version of this model, called Eastern North
American Model of Air Pollution (ENAMAP), also calculated monthly S02
and SO^" concentrations and wet and dry deposition patterns, and
generated matrices of interregional exchanges of sulfur for a user-
defined configuration of regions.7'8 In the early 1980s, EPA modified
and improved the ENAMAP model to increase its flexibility and
scientific credibility.
By 1985, simple parameterizations of processes involving fine
(diameters < 2.5 //m) and coarse (2.5 //m < diameters < 10.0 //m) PM were
incorporated into the model. This version of the model, renamed the
Regional Lagrangian Model of Air Pollution (RELMAP), is capable of
simulating concentrations and wet and dry deposition patterns of S02,
304", and fine and coarse PM and can also generate source-receptor
matrices for user-defined regions. In addition to the main model
program, the complete RELMAP modeling system includes 19 preprocessing
programs that prepare gridded meteorological and emissions data for
use in the main program. A complete scientific specification of the
RELMAP as used at EPA for sulfur modeling is provided in RELMAP: a
Regional Lagrangian Model of Air Pollution - User's Guide.9 The next
section discusses modifications made to the original sulfur version of
RELMAP to enable the simulation of atmospheric particulate metals
(arsenic, cadmium, chromium, lead and nickel).
6-20
-------
Table 6-14. Comparison of Risk Estimates for Single-Count Versus
Double-Count Runs to Assess the Impact of Overlapping Plumes
HAP, fuel, year
As, Coal, 1990
As, Oil, 1990
As, Coal, 2010
As, Oil,2010
Single-count runs
MEI risk
3x10-6
1 x10-5
3x10-6
7x10-6
Incidence
0.05
0.04
0.05
0.03
Population w/
risk>10-6
850
2,200
590
2,200
Double-count runs
MEI risk
3x10-6
1 x10-5
3x10-6
7x10-6
Incidence
0.05
0.04
0.05
0.03
Population w/
risk>10-6
850
2,400
590
2,300
HAP = Hazardous air pollutants.
MEI = Maximally exposed individual.
6.6.2 RELMAP Modeling Approach for Particulate Metals
6.6.2.1 Introduction. Previous versions of RELMAP have been
described by Eder et al.9 and Clark et al.10 The goal of the current
effort was to model the emission, transport, and fate of airborne
cadmium, arsenic, chromium, nickel, and lead from utilities in the
continental United States for the year 1989. Modifications to the
RELMAP simulation for arsenic, cadmium, chromium, lead, and nickel
were based on the assumption that these emissions are in particulate
form.
The RELMAP may be run in either of two modes. In the field mode,
wet deposition, dry deposition, and air concentrations are computed at
user-defined time intervals. In the source-receptor mode, RELMAP also
computes the contribution of each source cell to the deposition and
concentration at each receptor cell. For this study, only the field
mode of RELMAP operation was used. With over 10,000 model cells in
the high-resolution receptor grid and a significant fraction of these
cells also emitting the five metals, the data accounting task of a
source-receptor run for all utility sources could not be performed
with the computing resources and time available.
Unless specified otherwise in the following sections, the
modeling concepts and parameterizations described by Eder et al.9were
preserved for this RELMAP modeling study.
6.6.2.2 Physical Model Structure. Because of the long
atmospheric residence time of fine PM, significant long-range
transport was expected. For this study, RELMAP simulations were
limited to the area bounded by 25 and 55 degrees north latitude and 60
and 130 degrees west longitude and with a minimum spatial resolution
of one-half degree longitude by one-third degree latitude
6-21
-------
(approximately 40 km2) to provide high-resolution coverage over the
entire continental United States.
Since the descriptive document by Eder et al. 9 was produced, the
original three-layer puff structure of RELMAP was replaced by a four-
layer structure. The following model layer definitions were used for
the RELMAP particulate metal simulations:
Layer 1 top - 30 to 50 m above the surface
(season-dependent)
Layer 2 top - 200 m above the surface
Layer 3 top - 700 m above the surface
Layer 4 top - 700 to 1,500 m above the surface
(month-dependent).
6.6.2.3 Treatment of Emissions. All of the utilities within
each high-resolution RELMAP grid cell were treated as a single
integrated point source located at the center of the grid cell. As
mentioned in earlier sections of this report, the utility database
contained the necessary information to satisfy the RELMAP data needs,
including long-term particulate metals emission rates, stack
parameters, and plant location. All point source emissions (assumed
to be in steady state) were introduced into model layer 2 to account
for the effective stack height of the point source type in question.
Effective stack height is the actual stack height plus the estimated
plume rise. The layer of emission is inconsequential during the
daytime when complete vertical mixing is imposed throughout the
four layers. At night, since there is no vertical mixing, source
emissions to layer 1 are subject to dry deposition while point source
emissions to layer 2 are not. Large industrial emission sources and
sources with very hot stack emissions tend to have a larger plume
rise, and their effective stack heights might actually be larger than
the 700-m top of layer 2. However, since the layers of the pollutant
puffs remain vertically aligned during advection, the only significant
process affected by the layer of emission is nighttime dry deposition.
6.6.2.4 Lagrangian Transport and Deposition. In the model, each
pollutant puff begins with an initial mass equal to the total emission
rate of all sources in the source cell multiplied by the model time-
step length. For particulate metals, as for most other pollutants,
emission rates for each source cell were defined from input data and a
time step of three hours was used. The initial horizontal area of
each puff was set to 1,200 km2, instead of the standard initial size of
2,500 km2, in order to accommodate the finer grid resolution used for
the modeling study; however, the standard horizontal expansion rate of
339 km2/h was not changed. Although each puff was defined with four
separate vertical layers, each layer of an individual puff was
advected through the model cell array by the same wind velocity field.
Thus, the layers of each puff always remained vertically stacked.
6-22
-------
Wind field initialization data for a National Weather Service
prognostic model, the Nested Grid Model (NGM), were obtained from the
National Oceanographic and Atmospheric Administration's (NOAA's)
Atmospheric Research Laboratory for the entire year of 1989. Wind
analyses for the vertical level of approximately 1,000 meters above
ground level of the NGM were used to define translation of puffs
across the model grid, except during the months of January, February,
and December when the 600-m vertical level was used to reflect a more
shallow mixed layer.
Pollutant mass was removed from each puff by the processes of wet
deposition and dry deposition. The model parameterizations for these
processes are discussed in section 6.6.3. Precipitation data for the
entire year of 1989, obtained from the National Climatic Data Center,
were used to estimate the wet removal of all pollutant species
modeled. Wet and dry deposition mass totals are accumulated and
average surface-level concentrations are calculated monthly for each
model cell designated as a receptor. Except for cells in the far
southwest and eastern corners of the model domain where there were no
wind data, all cells were designated as a receptor for the particulate
metal simulations. When the mass of pollutant on a puff declines to a
user-defined minimum value, or when a puff moves out of the model
grid, the puff and its pollutant load are no longer tracked. The
amount of pollutant in the terminated puff is taken into account in
monthly mass balance calculations so that the integrity of the model
simulation is assured. Output data from the model include monthly wet
and dry deposition totals and monthly average air concentrations for
each modeled pollutant in every receptor cell.
6.6.2.5 Vertical Exchange of Mass with the Free Atmosphere. To
accurately simulate the long atmospheric lifetime of some pollutants,
the RELMAP was adapted to allow a treatment of the exchange of mass
between the surface-based mixed layer and the free atmosphere above.
The RELMAP no longer requires that the pollutant mass remain entirely
within the mixed layer. As an intuitive approximation, a pollutant
depletion rate of 5 percent per three-hour time step was chosen to
represent this diffusive mass exchange. This rate of mass exchange
used in the RELMAP was chosen to approximate the average daily
exchange of mass obtained from similar Lagrangian model exercises in
Europe. When compounded over a 24-hour period, the mass exchange rate
of 5 percent every three hours removes 33.6 percent of an inert,
non-depositing pollutant. Since each of the modeled particulate
metals undergo significant wet and dry deposition, their effective
diffusion rate out of the top of the model is less than 33.6 percent
per day. The mass lost through this vertical exchange through the top
of the model is accounted for and is reported as a model output for
mass balance checks.
6.6.3 Model Parameterizations
6.6.3.1 Chemical Transformation. The simplest pollutant type to
model with RELMAP is the inert type. To model inert pollutants, one
6-23
-------
can simply omit chemical transformation calculations and not be
concerned with chemical interactions with other chemical species.
Arsenic, cadmium, chromium, lead and nickel were treated as inert
pollutants.
6.6.3.2 Dry Deposition. All five metals were assumed to be
totally in particulate form. Since each of these metals and their
compounds make up only a small fraction of the total particulate mass
loading of the atmosphere, we modeled each as a minor component of the
general population of conglomerate aerosol particles. Lead has been
generally associated with fine particle sizes (
-------
Table 6-15. Windspeeds Used for Each Pasquill Stability
Category in CARB Subroutine Calculations
Stability category
A
B
C
D
E
F
Windspeed (mis)
2.0
3.0
4.0
5.0
3.0
2.0
CARB = California Air Resource Board.
Table 6-16. Roughness Length Used for Each Land-Use Category in
CARB Subroutine Calculations
Land-use category
Urban
Agricultural
Range
Deciduous forest
Coniferous forest
Mixed forest/wetland
Water
Barren land
Nonforested wetland
Mixed agricultural/range
Rocky open areas
Roughness length (ms)
Spring-summer
0.5
0.15
0.12
0.5
0.5
0.4
ID'6
0.1
0.2
0.135
0.1
Autumn-winter
0.5
0.05
0.1
0.5
0.5
0.4
ID'6
0.1
0.2
0.075
0.1
CARB = California Air Resource Board.
6.6.4 Exposure and Risk Estimates
6.6.4.1 Air HAP Concentration Estimates. Table 6-17 presents the
average and maximum annual air particulate metal concentration for the
continental United States as predicted by the RELMAP analysis for four
metal particulates (i.e., arsenic, cadmium, chromium, and nickel). The
results of the RELMAP modeling for lead are presented in chapter 8.
Results are presented as both combined impacts (e.g., emissions from both
coal and oil utility combustion combined) and segregated impacts(e.g.,
emissions from coal and oil utilities separately).
Figures 6-3 through 6-14 graphically present the RELMAP air
concentration results for each of the metals and each combustion fuel
type. In general, air concentrations of the four metal particulates,
as a result of utility oil and coal combustion, are predicted by
RELMAP to be maximum in the eastern part of the United States.
6-25
-------
Table 6-17. RELMAP Predicted Air Concentrations
Pollutant
Arsenic
Cadmium
Chromium
Nickel
Fuel
Coal
Oil
Coal &Oil
Coal
Oil
Coal &Oil
Coal
Oil
Coal &Oil
Coal
Oil
Coal &Oil
Max concentration
(Mg/m3)
2.5 E-04
4.4 E-05
2.5 E-04
8.0 E-06
1.0 E-05
1.1 E-05
2.2 E-04
3.0 E-05
2.2 E-04
1.7 E-04
2.6 E-03
2.6 E-03
Average concentration
(Atg/m3)
8.0 E-06
5.0 E-07
8.5 E-06
3.0 E-07
1.6 E-07
4.6 E-07
1.2 E-05
4.7 E-07
1.2 E-05
7.8 E-06
3.9 E-05
4.7 E-05
RELMAP = Regional Lagrangian Model of Air Pollution.
Air concentrations as predicted with RELMAP for emissions from
coal-fired utilities are predicted to be maximum along the western
slopes of the Appalachians Mountains and the northern Ohio River
Valley. Air concentrations as a result of utility oil combustion are
predicted to be maximum along the coastal northeast and the Florida
Peninsula.
6.6.4.2 Estimating Exposures and Risks. Once the grid cell
concentrations are known (as predicted by RELMAP), public exposure and
risks can be calculated. The population database within the HEM model
contains the centroid coordinates and number of people living within
each census block group. By applying this database to the predicted
RELMAP air concentrations both average and total population exposure
and risk can be estimated by the following six-step approach:
1. For each census block group determine which RELMAP grid cell the
census block group centroid is located within. All the people
living within the census block group are assumed to be exposed to
the predicted RELMAP HAP concentration for that corresponding
grid cell.
where:
C
•(n.pol)
= C
(i, j,po1)
n,pol)
the concentration of pollutant "pol" at census
block group "n" in //g/m3.
the RELMAP concentration of pollutant "pol" at
grid cell (i,j) in //g/m3 where the centroid of
census block group "n" is locate within grid
cell (i,j).
6-26
-------
CTl
I
tsJ
0.00'\ to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 6-3. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Arsenic from Coal Utilities: Predicted Air
Concentration of Arsenic, Units: ng/m3
-------
CTl
I
tSJ
00
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 6-4. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Arsenic from Oil Utilities: Predicted Air
Concentration of Arsenic, Units: ng/m3
-------
CTl
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 6-5. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Arsenic from Coil and Oil Utilities: Predicted Air
Concentration of Arsenic, Units: ng/m3
-------
en
i
OJ
o
0.0001 to 0.0005
0.0005 to 0.001
0.001 to 0.002
0.002 to 0.005
> = 0.005
Figure 6-6. _ Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Cadmium from Coal Utilities: Predicted Air
Concentration of Cadmium, Units: ng/m3
-------
i
u>
0.0001 to 0.0005
0.0005 to 0.001
0.001 to 0.002
0.002 to 0.005
> = 0.005
Figure 6-7. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Cadmium from Oil Utilities: Predicted Air
Concentration of Cadmium, Units: ng/m3
-------
CTl
I
ISJ
0.0001 to 0.0005
0.0005 to 0.001
0.001 to 0.002
0.002 to 0.005
> = 0.005
Figure 6-8. Results of the RELMAP Modeling Analysis from 1990
Emissions Estimates for Cadmium from Coal and Oil Utilities-
Predicted Air Concentration of Cadmium, Units: ng/m3
-------
CTl
I
OJ
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 6-9. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Chromium from Coal Utilities: Predicted Air
Concentration of Chromium, Units: ng/m3
-------
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
Figure 6-10. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Chromium from Oil Utilities: Predicted Air
Concentration of Chromium, Units: ng/m3
-------
CTl
I
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 6-11. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Chromium from Coal and Oil Utilities: Predicted Air
Concentration of Chromium, Units: ng/m3
-------
I
OJ
CTl
0.01 to 0.05
0.05 to 0.1
0.1 to 0.2
0.2 to 0.5
> = 0.5
Figure 6-12. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Nickel from Coal Utilities: Predicted Air
Concentration of Nickel, Units: ng/m3
-------
CTl
I
OJ
~J
0.01 to 0.05
0.05 to 0.1
0.1 to 0.2
0.2 to 0.5
> = 0.5
Figure 6-13. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Nickel from Oil Utilities: Predicted Air
Concentration of Nickel, Units: ng/m3
-------
I
(-0
00
0.01 to 0.05
0.05 to 0.1
0.1 to 0.2
0.2 to 0.5
> = 0.5
Figure 6-14. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Nickel from Coal and Oil Utilities: Predicted Air
Concentration of Nickel, Units: ng/m3
-------
Determine the average individual lifetime cancer risk for people
living in the census block group by multiplying the HAP
concentration for the census block group (determined in step 1)
by the IURE (which is the estimated increase risk of cancer for
an individual exposed to the pollutant concentration of 1 //g/m3
for 70 years). The IURE is explained in detail in chapter 4.
R(n,PoD = C(nipol) * IURE(pol)
where:
R(npol) = the average individual lifetime cancer risk
for pollutant "pol" at census block group "n"
in //g/m3 .
IURE(pol) = the estimated increase risk of cancer for an
individual exposed to a concentration of 1 //g/m3
of pollutant "pol" for 70 years (1 per //g/m3 ) .
Determine the lifetime population cancer incidence for each census
block group by multiplying the average individual risk (from step 2)
by the corresponding population of the census block group.
LCI(n,pol) = R(n,pol) * Pop(n)
where :
LCI(npol) = the Lifetime (70 years) Population Cancer
Incidence for census block group "n" for
pollutant "pol" in cases/lifetime.
Pop(n) = the 1990 Census total population for census
block group "n" .
The lifetime cancer incidence can be estimated by summing the
lifetime population cancer incidence for each census block group
(from step 3 ) .
LCI (total, pol) = Z^CI(nipol) n=l,466,318
where :
LCI (total pol) = the Lifetime (70 years) Population Cancer
Incidence for the continental U.S. (466,318
census block groups in the continental United
States) for pollutant "pol" in cases/lifetime.
The annual population cancer incidence in the continental U.S. is
predicted by dividing the lifetime cancer incidence (from step 4)
by 70 years.
ACI(total,pol) = LCI(total,pol) / 70 years
where :
ACI (totai, poi) = tne Annual Population Cancer Incidence for the
continental United States.
6-39
-------
6. The average individual exposure over the entire continental U.S.
population for a pollutant is estimated by dividing the total
population exposure by the total number of people living in the
continental United States.
AIE(total,pol) = PE(total/pol) / Pop(total)
where:
AIE(total/pol) = the average individual annual exposure for the
continental United States for pollutant "pol"
PE(total/pol) = the sum of the concentration of pollutant
"pol" for all census block groups
(IC n=i,466,3i8) in ug/m3.
Pop(total) = the 1990 Census total population for the
continental United States. (=247,000,000
people)
The results of the exposure and risk estimates are shown in Table 6-18.
To evaluate potential impacts due to long-range transport (LRT),
the coal and oil utility emissions were modeled both together and
separately. By applying the algorithm described above, a combined
(e.g.,oil and coal emissions) value of 0.42 cancer cases/year was
estimated for arsenic emissions over the continental United States.
The contribution of oil emissions is predicted to be about 11 percent
of the total inhalation risks from arsenic emissions.
The LRT population exposure and cancer incidence for the four
metals from RELMAP for coal-fired utilities is about seven times
greater (i.e., 600 percent greater) than the population risks
estimated in the local analysis alone (i.e., within 50 km) using the
HEM. However, the population exposure and cancer incidence for oil-
fired utilities from the RELMAP study are only slightly (about 10
percent) greater than the exposures and incidence predicted with the
HEM analysis (See Table 6-19) .
The total (coal and oil)population risk estimate for chromium
emissions is predicted to be 0.22 cases/year. The contribution of oil
fired chromium emissions is predicted to be about 10 percent of the
total risk from chromium emissions. The LRT population risk estimate
for chromium from RELMAP is about five times greater than the
population risks estimated modeling chromium emissions using the HEM
model.
The high-end estimate for total population cancer incidence due
to nickel emissions is predicted to be up to 0.2 cases/year. The
contribution of coal-fired nickel emissions is predicted to be about
10 percent of the total risk from nickel emissions. The LRT population
risk estimate for nickel is about equal to the population risks
6-40
-------
Table 6-18. Predicted Exposure and High-End Risk Estimates Based
on RELMAP Modeling of Particulate Metal Emissions from All Oil-
and Coal-Fired Utilities in the United States.
Pollutant
Arsenic
Cadmium
Chromium
Nickel
Fuel
Coal
Oil
Coal & Oil
Coal
Oil
Coal & Oil
Coal
Oil
Coal & Oil
Coal
Oil
Coal & Oil
Maximum
Concentration
(//g/m3)
2.5e-04
4.4e-05
2.6e-04
8.0e-06
1.1e-05
1.1e-05
2.2e-04
3.0e-05
2.2e-04
1 .7e-04
2.6e-03
2.6e-03
Average Exposure
Concentration
^g/m3)
2.4e-05
3.1e-06
2.8e-05
7.6e-07
9.4e-07
1.7e-06
3.3e-05
2.7e-06
3.5e-05
2.2e-05
2.4e-04
2.6e-04
Unit Risk
Estimate
(per//g/m3)
4.3e-03
4.3e-03
4.3e-03
1 .8e-03
1.8e-03
1.8e-03
1 .3e-03
2.2e-03
1.7e-03
2.4e-04
2.4e-04
2.4e-04
Maximum Exposed
Individual (MEI)
Risks
1.1e-06
1.9e-07
1.1e-06
1 .4e-08
1.9e-08
1.9e-08
2.9e-07
6.6e-08
3.9e-07
4.0e-08
5.0e-07
5.0e-07
Population
Risks
(cases/yr)
0.4
0.05
0.4
0.005
0.006
0.01
0.2
0.02
0.2
0.02
0.2
0.2
MEI = Maximally exposed individual.
RELMAP = Regional Lagrangian Model of Air Pollution.
estimated by modeling nickel emissions using the HEM model because
most of the nickel exposure is due to oil-fired utilities.
The potential impacts to the MEIs appear to be considerably less
than for population exposures for each metal particulate. The maximum
RELMAP ambient concentrations (Table 6-17) are each less than 20
percent of the highest HEM concentration for coal- and oil-fired
utilities (see Tables 5-1 and 5-4). Also, a comparison of results for
MEI risks in Table 6-18 with MEI risks in Tables 6-1 and 6-3 shows the
differences in MEI results from the local versus long-range transport
analyses.
These metal particulates are also associated with other HAP
particulate matter in the utility emissions and probably act in a
similar manner in the atmosphere. In addition, these other HAPs are
generally emitted in roughly proportional quantities for each fuel
type and are emitted from the same set of plant locations. Therefore,
the factor of 7 can be applied to these other HAPs from coal-fired
utilities to roughly estimate the potential impact of long-range
transport of HAPs on the overall cancer incidence. Therefore,
considering local and LRT, the cancer incidence due to inhalation
exposure to HAP emissions is estimated to be as high as about 0.5
cancer cases per year for oil-fired utilities and as high as about 1.3
cases per year for coal-fired utilities. Gas-fired utilities are
estimated to present far less population cancer risks than oil-, and
6-41
-------
Table 6-19. Summary of the High-End Estimates of the Inhalation
Risk Estimates Due to Local and Long-Range Transport
LOCAL IMPACTS (dispersion within 50 km of each utility plant)3
Pollutant
Radionuclides"
Nickel0
Chromium
Arsenic
Cadmium
All othersM
Total6
Oil-fired plants
Maximally exposed
individual (MEI)
1 x1Q-5
5x1Q-5
5x10'6
1 x1Q-5
2x10'6
8x10'7
6x1Q-5
Annual increased
cancer Incidence
0.2
0.2
0.02
0.04
0.005
0.005
0.5
Coal-fired plants
Maximally exposed
individual (MEI)
2x10'8
4x10-7
2x10-6
3x10-6
2x10-7
8x10-7
4x10-6
Annual increased
cancer incidence
0.1
0.003
0.02
0.05
0.0006
0.004
0.18
LOCAL PLUS LONG-RANGE IMPACTS (dispersion from utility emission points to borders of continental U.S.)
Pollutant
Radionuclidesb
Nickel0
Chromium
Arsenic
Cadmium
All othersM
Total6
Oil-fired plants
Maximally exposed
individual (MEI)
1 x 1Q-5
5x1Q-5
5x10'6
1 x1Q-5
2x10'6
8x10'7
6x1Q-5
Annual increased
cancer incidence
0.2
0.4
0.02
0.05
0.006
0.006
0.5
Coal-fired plants
Maximally exposed
individual (MEI)
Not estimated
4x10-7
3x10-6
4x10-6
3x10-7
1 x10-6
5x10-6
Annual increased
cancer incidence^
0.7
0.02
0.15
0.4
0.005
0.028
1.3
a There are uncertainties associated with these risk estimates. See sections 6.4 for discussion.
b Radionuclides and "all others were not modeled with RELMAP. The cancer incidence for these HAPs in the
Local Plus Long-Range Impacts section were estimated by assuming the results from the RELMAP modeling for
Cr, As, Cd, and Ni could be applied to these other HAPs. Hence, the cancer incidence for radionuclides and "all
others" for oil-fired utilities were assumed to be the same as incidence from the local (HEM) analysis, and the
incidence for radionuclides and "all others" from coal-fired utilities were assumed to increase by a factor of seven.
c Assumes that the nickel mixture is 50 percent as carcinogenic as nickel subsulfide.
d Estimated risks due to exposure to all remaining HAPs analyzed (i.e., excluding nickel, arsenic, chromium,
cadmium, and radionuclides).
6 Aggregate risk (risk due to inhalation exposure to all carcinogenic HAPs, assuming additivity of risks).
f These population risk estimates are predicted directly from RELMAP which includes the local and regional
impacts.
6-42
-------
coal-fired utilities. Therefore, adding these estimates (0.5 + 1.3),
it is predicted that up to about 2 cancer cases per year occur due to
inhalation exposure to HAP emissions from all utility plants (coal-,
oil-, and natural gas-fired) in the continental United States.
However, as stated earlier, the lUREs are upper bound estimates of the
cancer risks posed by HAPs at low exposure levels, and the true risk
is unknown and could be as low as zero. Also, the inhalation high-end
risk estimates are conservative. Therefore, the cancer incidence due
to inhalation exposure to utility HAP emissions is predicted to be no
greater than 2 cases per year in the United States. Most likely fewer
than 2 cases/yr occur in the United States due to inhalation exposure
to utility HAP emissions.
There are numerous uncertainties in the modeling, the
assumptions, the extrapolations, and the resulting cancer incidence
estimates. Since the exposure concentrations for much of the exposed
population are quite low, this analysis relies heavily on the
assumption of cancer being a nonthreshold phenomenon and the
assumption that the dose-response curve for these carcinogens is
linear at very low doses. Also, there are considerable uncertainties
in the risk estimates and incidence estimates for nickel because of
the uncertainties associated with the emissions of different nickel
species and the uncertainties in the health effects for each of those
forms. Further evaluation of the data, models, and methods could be
useful to reduce the uncertainties and to fully evaluate the impacts
of long-range transport.
6.7 DISCUSSION OF BACKGROUND EXPOSURES
6.7.1 Arsenic
Over 250 sites have reported ambient arsenic data to the EPA's
Aerometric Information Retrieval System. Up to the year 1987, arsenic
was measured by performing an analysis of the filter catch from 24-
hour high-volume total-suspended-particulate (TSP) sampling devices.
The Agency compared the results of the dispersion modeling to
available data in 1987 (latest available data) and attempted to
provide insight into typical arsenic concentrations in areas away from
utilities and to provide a check on the credibility of the predicted
concentrations.
A review of 1987 ambient arsenic data indicated that the minimum
concentration that could be detected was about 3 ng/m3. Much of the
reported data were at or below the minimum detectable level (MDL); for
instance, 145 of the 261 total sites reported no values above the MDL.
At sites not located near known, large arsenic emitters, such as
copper smelters, the largest annual concentration reported was about 8
ng/m3. Further analysis indicated the large majority (about
75 percent) of monitors were located within 50 km of at least one
coal- or oil-fired utility plant, and six sites were located within 50
km of at least 10 plants. On the other hand, there were 59 sites that
were not within 50 km of any coal- or oil-fired utility plant.
6-43
-------
Typical arsenic concentrations can be determined by reviewing the
data from the 59 sites not near utility plants. Of the 59 sites, 8
were known to be near large arsenic sources and were not
representative of typical sites. Only 13 of the remaining 51 sites
recorded annual arsenic concentrations above the MDL. The highest
concentration reported was about 8 ng/m3. Thus, based on these data,
typical concentrations are probably not much higher than 8 ng/m3 and
are most likely to be lower (or much lower) than the MDL of 3 ng/m3.
In fact, as seen from review of the data collected near utilities,
this result is typical of all the available arsenic data, when the
monitors are not located near large arsenic-emitting sources. If a
person were exposed to this highest measured concentration of 8 ng/m3
(or 0.008 //g/m3) for a lifetime and the IURE is used to estimate the
cancer risk, this person would have an estimated increased cancer risk
of 3 x 10"5. However, typical background arsenic inhalation exposures,
which are likely to be less than 3 ng/m3 would likely pose risks lower
than 1 x 10-5.
Next, a comparison was made between the predicted arsenic
concentrations and the measured values near the plants. The highest
arsenic long-term concentration estimated for any utility plant using
the HEM was about 3 ng/m3. The estimated maximum concentrations
predicted with the HEM for all the other utility plants were lower or
much lower than 3 ng/m3. The monitor nearest the plant that caused the
maximum arsenic concentration was about 12 km away, and that monitor
did not register any concentrations above the MDL of 3 ng/m3. The air
dispersion analysis using the HEM predicted an arsenic concentration
of 0.05 ng/m3 at that monitoring site, so concentrations due to utility
emissions were not expected to register on this monitor. At the site
where the highest arsenic concentration was reported, the air
dispersion analysis predicted arsenic concentrations well below 0.01
ng/m3.
The information presented above is useful for gaining a general
sense of the potential background air concentrations of arsenic.
However, it is difficult to draw conclusions from the comparison of
the modeled concentrations and the measurement data. Direct
comparisons between estimated and measured values can be misleading.
As suggested by the analysis of sites away from where arsenic
concentrations were detected, there are confounding factors. One
confounding factor occurs because arsenic is a naturally occurring
element in the earth's crust. Some arsenic is expected to be in every
TSP filter catch (i.e., a natural background concentration that would
be present even without nearby anthropogenic sources). There is a
second confounding factor because any other PM-arsenic source in the
area will also have an impact on the monitor. So, for arsenic, the
monitored concentrations are measuring a combination of
concentrations: (1) from natural background, (2) from other arsenic
sources, and (3) from nearby utilities. Thus, the monitored values
are always expected to exceed the impact from the plant's emissions.
6-44
-------
6.7.2 Chromium, Nickel, Manganese, and HC1
Chromium and nickel ambient data were also available. The
results in analyzing these data led to conclusions similar to those
drawn from the arsenic analysis. Much of the data were below
detectable levels and did not provide much insight into the relative
concentration impacts from utility emissions. However, data presented
in a 1994 EPA Report15 indicate that chromium levels in some urban
areas have been measured to be roughly from 0.8 to 16 ng/m3, which
would equate to a high-end increased cancer risk of 1 x 10 ~5 to 2 x 10 "4
if it is assumed that the chromium is hexavalent and that a person is
exposed to those levels for 70 years (i.e., lifetime). In addition,
based on data presented in the above 1994 EPA report, nickel levels in
a few urban areas are roughly between 0.1 to 20 ng/m3. Assuming the
nickel mix is 50 percent as carcinogenic as nickel subsulfide and
assuming people are exposed to these levels for a lifetime, this
concentration range would correspond to high-end risks of roughly
between 3 x 10'8 to 7 x lO'5.
Based on the HEM modeling, manganese and HC1 were the two HAPs
that appear to be of highest potential concern for noncancer effects
due to inhalation exposure. However, in the assessment of noncancer
health effects due to inhalation exposure to HAP emissions from
utilities, the highest HEM-modeled concentrations of manganese and HC1
from the highest-risk plants were estimated to be 10 times lower than
the RfC. All other HEM-modeled concentrations for HC1 and manganese
were even lower. Therefore, regardless of background exposure levels,
the emissions of HC1 and Mn from utilities are not likely to
contribute significantly to an RfC exceedance. For this reason, the
EPA did not conduct an analysis of ambient air background exposures
for these two HAPs for this report.
6.8 CHROMIUM SPECIATION UNCERTAINTY AND IMPACT ON RISK ESTIMATES
Available health effects data indicate there are significant
differences in toxicity between the trivalent chromium (Cr III) and
the hexavalent chromium (Cr VI). Chromium VI is classified as a human
carcinogen (WOE = A) based on human and animal studies that show an
increase in lung cancer. Available data are not sufficient to
determine the carcinogenicity of Cr III (WOE = D). Cr III appears to
be much less toxic than Cr VI.16'17 For more information on chromium
toxicity see Appendix E.
Data on speciation of chromium were available from 11 test sites.
The limited emissions speciation data2 indicate that somewhere between
0.4 percent and 34 percent of the emitted chromium is chromium VI.
The average chromium VI from the coal-fired utilities was 11 percent;
the average from oil-fired utilities was 18 percent.
To assess the potential impact of the range of chromium
speciation on the risk results, the utilities were modeled using the
HEM assuming different speciation percentage assumptions. Tables 6-20
and 6-21 present the results of the assessment.
6-45
-------
Table 6-20. Chromium Speciation Analysis for Coal-Fired
Utilities: Inhalation Risk Estimates due to Chromium Emissions
Based on Various Assumptions of Percent Chromium VI
% Chromium VI assumption3
Assume 1 00% Cr VI
Assume 23% Cr VI
Assume 11%CrVI
Assume 0.4% Cr VI
Lifetime
MEI risk
2x10-5
4x10-6
2x10-6
7x10'8
Lifetime
MIR
1 x10-5
2x10-6
1 x10-6
4x10'8
Population w/>10"6
lifetime cancer risk
69,000
2,300
110
0.0
Cancer incidence
(cases/year)
0.2
0.04
0.02
0.0007
MEI = Maximally exposed individual.
MIR = Maximum individual risk.
a Based on speciation data from emissions tests for four coal-fired test sites, the average percent Cr VI was 11 percent, the
maximum was 23 percent, and the minimum was 0.4 percent. The remaining chromium emissions are assumed to be Cr III. It is
assumed that the cancer risk is due only to Cr VI emissions. Because carcinogenicity data for chromium III are very limited and
uncertain, it was assumed that Cr III does not pose cancer risk. It is not known whether the Cr III emissions contribute to the
cancer risk.
Table 6-21. Chromium Speciation Analysis for Oil-fired
Utilities: Inhalation Risk Estimates due to Chromium Based on
Various Assumptions of Percent Chromium VI
% Chromium VI
assumption3
100%CrVI
34%CrVI
18%CrVI
5% Cr VI
Lifetime MEI risk
3x10-5
1 x 10-5
5x10-6
1.5x10-6
Lifetime MIR
3x10-5
1 x10-5
5x10-6
1.5x10-6
Population w/> 10"6
lifetime cancer risk
40,000
2,300
2,300
45
Cancer incidence
(cases/year)
0.1
0.04
0.02
0.005
MEI = Maximally exposed individual.
MIR = Maximum individual risk.
a Based on limited speciation data from emissions tests for seven oil-fired test sites, the average percent Cr VI was 18 percent,
the maximum was 34 percent, and the minimum was 5 percent Cr VI, it was assumed that chromium III does not pose a cancer
risk. It is assumed that the remainder of the chromium emissions are Cr III. It is assumed that the cancer risk is due only to Cr
VI emissions.
6-46
-------
6.9. ISSUES WITH ARSENIC CANCER UNIT RISK ESTIMATE AND IMPACT ON
INHALATION RISK ESTIMATES
Arsenic is considered a human carcinogen (WOE = A). The EPA
reviewed the dose-response data in 1986 and established an IURE
of 4.29 x 10"3 per //g/m3.17 This IURE is the EPA-verified value
currently available on IRIS. A more in-depth discussion of the cancer
health effects data is provided in Appendix E.
The EPRI submitted a paper on arsenic carcinogenicity to the EPA
IRIS office. This paper suggested that the IURE should be
approximately three times lower than the current EPA-verified value as
a result of reviewing new data.18 The EPRI asked the EPA to review the
new data and consider revising the arsenic unit risk estimate based on
the most current data and analyses.
The EPA has initiated the review process. However, to conduct a
thorough review and analysis of the data and to calculate a new risk
estimate is time consuming. A full review and IRIS update could not
be completed in time for this report. However, the EPA has done a
cursory review of the paper submitted by EPRI along with other
relevant data. Based on this initial review by EPA scientists, it
appears that the EPRI-proposed IURE is within the range of plausible
estimates of cancer potency.19 The Canadians have also reviewed the
available data recently and established an IURE of 6 x 10 ~3. 20 The
Canadian IURE also appears to be within the plausible range of potency
for arsenic.3 Table 6-22 compares EPRI, EPA-verified, and Canadian
inhalation risk estimates.
Since a full review of the unit risk could not be completed in
time for this report, and to help characterize the potential range of
risk due to arsenic exposure, an assessment was conducted that
presents the estimated risks due to inhalation exposure using three
different lUREs (Table 6-22) . It should be noted that this
presentation does not present the full range of uncertainty, but
rather presents the impact on the results due to the three different
estimates of the unit risk.
6.10 NICKEL SPECIATION UNCERTAINTY AND IMPACT ON RISK ESTIMATES
There are significant uncertainties associated with nickel
speciation. Nickel exists in four different valence states and can be
combined with many other elements to form different nickel compounds.
Numerous nickel compounds are known to exist. 21
At the time emissions data were being analyzed for this report
(1992 to 1994), total nickel was measured at nearly all sites, but
only two sites (both oil-fired utilities) provided data on speciated
nickel. The species measured were soluble nickel (water-soluble salts
such as nickel sulfate and nickel chloride), sulfidic nickel (such as
nickel subsulfide, nickel monosulfide, and nickel sulfide), metallic
nickel (including alloys), and oxidic nickel (including nickel oxide,
6-47
-------
Table 6-22. High-End Arsenic Inhalation Risk Estimates:
Comparison of Results Using the EPRI, EPA-Verifled, and Canadian
IUREsa
Arsenic from
oil-fired utilities
Arsenic from
coal-fired utilities
Risk estimates using EPRI
IURE (1.4x10^ per uglm3)
MEI risk
4x10'6
6x10'7
#>10'6
2,200
0.0
Incidence
0.014
0.015
Risk estimates using EPA
IURE(4.3x10^per/ig/m3)
MIR
1 x10'5
3x10'6
#>io-6
2,400
850
Incidence
0.042
0.045
Risk estimate w/Canadian
IURE(6x10^per/ig/m3)
MIR
2x10'5
4x10'6
#>io-6
3,000
850
Incidence
0.05
0.06
EPA = U.S. Environmental Protection Agency
EPRI = Electric Power Research Institute
IURE = Inhalation unit risk estimate
MEI = Maximally exposed individual
MIR = Maximum individual risk
a The EPRI IURE for arsenic (1.4 x 10-3 per /jglm *) is three times lower than the EPA-verified IURE for arsenic (4.3 x 103 per
//g/m3), and, the Canadian value is approximately 35 percent greater than the EPA estimate . 17
complex oxides, and silicates). The average values of the two test
sites were: 58 percent soluble nickel, 3 percent sulfidic nickel, and
39 percent nickel oxides.2 More recently, EPRI submitted a fax to the
EPA summarizing nickel speciation data from 5 test sites.3 Based on
the data presented in the fax from EPRI, 25 to 60 percent of nickel
emissions are soluble nickel compounds, 4 to 26 percent are sulfidic
nickel compounds, 0 to 4 percent are metallic nickel compounds, and 27
to 70 percent are oxidic nickel compounds.
The available health effects data vary significantly from species
to species. Human epidemiologic data indicate that at
least some forms of nickel are carcinogenic to humans by inhalation
exposure.22'23 Nickel refinery dust and nickel subsulfide are
classified as human carcinogens (WOE = A). The IURE for nickel
refinery dust is 2.4 x 10"4. Based on an assumption that nickel
subsulfide constitutes 50 percent of the refinery dust, a potency
estimate (IURE) of 4.8 x 10"4 was assigned to nickel subsulfide.
Nickel carbonyl is classified as a probable human carcinogen (WOE =
B2), but no IURE has been established. These are the only species
currently classified by the EPA as carcinogens. The International
Agency for Research on Cancer (IARC) considers nickel monoxide, nickel
hydroxide, and metallic nickel as having sufficient evidence in
experimental animals for carcinogenicity. 24 The IARC considers nickel
compounds to be carcinogenic to humans and metallic nickel to be
possibly carcinogenic. The State of California concludes that the
class of nickel compounds is potentially carcinogenic by inhalation.24
The American Conference of Governmental Industrial Hygienists (ACGIH)
has stated that all nickel compounds should be considered carcinogenic
6-48
-------
for risk management purposes.25 However, there are still significant
uncertainties regarding the carcinogenicity of many of the nickel
compounds. Available data are insufficient to confirm the
carcinogenicity of many nickel compounds.
Cancer lUREs are only available for nickel subsulfide and nickel
refinery dust. The cancer potency of the other nickel compounds that
may be carcinogenic is not known. Results of animal studies suggest
that nickel subsulfide is the most carcinogenic form.23'24 Based on the
limited speciation data, no more than 10 percent of the nickel
compounds are likely to be nickel subsulfide. Therefore, the nickel
risk estimates presented in previous sections (where it is assumed the
mix of nickel compounds emitted from utilities are 50 percent as
carcinogenic as nickel subsulfide) are considered conservative, high-
end risk estimates.
To assess the potential impact of the speciation uncertainty, the
EPA conducted an assessment for cancer risks using different
assumptions for speciation and cancer potency. The assessment
(summarized in Table 6-23) provides a range of the potential cancer
risks due to nickel emissions.
6.10.1 Alternative Analysis for Estimating Population Risks.
Figure 6-15 summarizes the impact of using alternative IURE values for
nickel (as a percent of the nickel subsulfide IURE) on annual cancer
incidence. The estimated annual cancer incidence due to oil-fired
utilities is 0.3 case per year using the assumption that the potency
(IURE) of the mixture of nickel compounds emitted from oil-fired
utilities is 50 percent the potency of nickel subsulfide, about
0.15 case/yr if the IURE is assumed to be 20 percent as potent as
nickel subsulfide, and about 0.1 case per year if the IURE is assumed
to be 10 percent nickel subsulfide. Likewise, there would be changes
in the number of persons potentially exposed at various risk levels.
If the IURE were 20 percent nickel subsulfide, about 9,930 persons
would be exposed at an MIR > 1 x 10~s. Figure 6-15 does not capture
the full potential range of estimated population risks. It is
possible that the potency of the mix of nickel compounds emitted from
oil-fired utilities is less than 10 percent as carcinogenic as nickel
subsulfide. Therefore, the cancer incidence could possibly be lower
than that shown in Figure 6-15. The cancer incidence due to nickel
emissions could possibly be as low as zero.
In addition to the cancer effects, nickel also causes noncancer
health effects, such as allergenicity and respiratory effects.
Currently, no RfC is available for nickel compounds. However, there
are various health benchmarks in the literature that are useful for
screening purposes to give some idea whether or not the exposure
estimates are likely to cause noncancer health effects. The EPA
conducted such an assessment (see Table 6-24) .
6-49
-------
Table 6-23. Nickel from Oil-Fired Utilities: Inhalation Cancer
Risk Estimates Based on Various Assumptions of Speciation and
Cancer Potency
Nickel Speciation3
100%Nisubsulfide
20% Ni subsulfide
10% Ni subsulfide
1% Ni subsulfide
Cancer potency
(IURE)b
4.8 x10'4
9.6 x10-5
4.8 x10-5
4.8 x10-6
MIR
9.6 x10-5
2x10-5
9.6 x10-6
9.6 x10-7
# People >10'6 risk
1,600,000
9,900
2,300
0.0
Annual incidence
0.4
0.08
0.04
0.004
IURE = Inhalation unit risk estimate
MIR = Maximum individual risk
a The limited nickel speciation data indicate that nickel is a combination of nickel oxide, soluble nickel, sulfidic nickel, and
insoluble nickel. The limited speciation data indicate that less than 10 percent of the nickel is nickel subsulfide.
b The Inhalation Unit Risk Estimate (IURE) of 4.8 x 104 is the IURE for nickel subsulfide found on IRIS. For each of these cases,
it is assumed that either 100 percent, 20 percent, 10 percent, or 1 percent of the nickel is nickel subsulfide, and that only this
fraction is contributing to the cancer risk. The cancer risk due to the other nickel compounds is not known.
6.11 POTENTIAL INCREASED DIOXIN EMISSIONS FROM UTILITIES WITH
ELECTROSTATIC PRECIPITATORS
Emissions data for dioxins and dibenzofurans were available from
only nine test sites. None of these sites have hot-side electrostatic
precipitators (ESPs) installed for controlling emissions. The EPA
discovered that dioxin emissions from municipal waste combustors
(MWCs) with hot-side ESPs could be 5 to 15 times greater than
emissions from a similar source without a hot-side ESP.26 Since this
phenomenon was observed at MWCs, the EPA assumes that it is possible
that the same situation may possibly occur at utilities. However, at
this time, sufficient information is not available to assess the
potential risks due to dioxin emissions from the utility plants with
hot-side ESPs. Currently, the DOE is planning to conduct an emission
test at a facility with a hot-side ESP; however, at this time, no data
are available for dioxins from hot-side ESP units.
6.12 DISCUSSION OF UNCERTAINTY AND ASSUMPTIONS FOR DOSE-RESPONSE
ASSESSMENT FOR CARCINOGENS
Information related to dose-response assessment for the HAPs is
summarized here to identify the assumptions, methods, data
used, and uncertainty associated with the dose-response measures.
This information is useful to place the quantitative risk estimates
into context with respect to their associated uncertainty and
conservatism.
6-50
-------
CTl
I
Ul
S 0.5
S. 0.45
$ 0.4
re
oT 0.35
o
c
I 0.3
'o
^ 0.25
-------
Table 6-24. Comparison of Nickel Exposure to Various Noncancer
Health Benchmarks
Various health benchmarks for nickel compounds
# People exposed0 above the benchmark
Maximum HQd
CARB RELa = 0.24 jug/m3
0.0
0.82
EPRIb value = 2.4 //g/m3
0.0
0.082
CARB = California Air Resources Board
EPRI = Electric Power Research Institute
HQ = Hazard quotient
REL = Reference exposure level
a This value was obtained from the CARB Hot Spots Program5 CARB calculated this number by dividing the Threshold Limit
Value (TLV) of 0.1 mg/m3 by 420. The TLV is a level set by the American Conference of Government Industrial Hygienists
(ACGIH) as a guideline to protect workers. The 420 accounts for extrapolating from a 40-hour work week to a 168-hour week
(4.2x), extrapolating from healthy workers to sensitive subpopulations (10x), and another factor of 10x because adverse health
effects are often seen at the TLV.
b The EPRI benchmark 27 was calculated by dividing the TLV by 42. The 42 accounts for extrapolating from a 40-hour work week
to a 168-hour week, and a 10x is applied to account for sensitive subpopulations.
0 The exposed population is estimated from the results of the Inhalation Human Exposure Modeling.
d The HQ is calculated by dividing the modeled concentration by the health benchmark. It is the ratio of the estimated highest
exposed concentration to the benchmark concentration. A value of 1 or higher indicates that the exposure is above the health
benchmark.
6.12.1 Default Options
The EPA uses default options when dealing with competing
plausible assumptions and uncertainty in estimating cancer unit risks.
The use of these default options is intended to lead to unit risk
estimates that, although plausible, are believed to be more likely to
overestimate than to underestimate the risks. The use of these
defaults has led EPA scientists to conclude that the resulting unit
risk estimates are upper limits. That is, the actual risks are
unlikely to be greater than these estimates, and may be lower; they
could also be zero. Below are several of the major default options
used in cancer dose-response assessment identified by NRC.28 However,
it must be noted that the preliminary HAPs of interest in this study
for cancer risks (i.e., arsenic, chromium VI, and nickel subsulfide)
have lUREs and WOE that are based on human epidemiology studies;
therefore, many of the assumptions listed below are not relevant for
much of this study.
• Laboratory animals are a surrogate for humans in assessing
cancer risks; positive cancer-bioassay results in
laboratory animals are taken as evidence of a chemical's
cancer-causing potential in humans.
• Humans are as sensitive as the most sensitive animal
species, strain, or sex evaluated in a bioassay with
appropriate study-design characteristics.
6-52
-------
• Agents that are positive in long-term animal experiments
and also show evidence of promotion or cocarcinogenic
activity should be considered as complete carcinogens.
• Benign tumors are surrogates for malignant tumors, so
benign and malignant tumors are added in evaluating
whether a chemical is carcinogenic and in assessing its
potency.
• Chemicals act like radiation at low exposures (doses) in
inducing cancer; i.e., intake of even one molecule of a
chemical has an associated probability for cancer
induction that can be calculated, so the appropriate model
for relating exposure-response relationships is the
linearized multistage model.
• Important biological parameters, including the rate of
metabolism of chemicals, in humans and laboratory animals
are related to body surface area. When extrapolating
metabolic data from laboratory animals to humans, one may
use the relationship of surface area in the test species
to that in humans in modifying the laboratory animal data.
• A given unit of intake of a chemical has the same effect,
regardless of the time of its intake; chemical intake is
integrated over time, irrespective of intake rate and
duration.
• Unless there are data to the contrary, individual
chemicals act independently of other chemicals in inducing
cancer when multiple chemicals are taken into the body;
when assessing the risks associated with exposures to
mixtures of chemicals, one treats the risks additively.
6.12.2 Models, Methods, and Data
In a dose-response assessment, the likelihood of developing
cancer is determined quantitatively for any given level of exposure to
a carcinogen.29 The two basic reasons for conducting a cancer
dose-response assessment are (1) to extrapolate from high to low
doses, and (2) to extrapolate from animal to human responses. Both
epidemiologic and toxicologic studies are conducted at doses higher
than those normally encountered in the environment. Therefore, in
order to determine response at lower doses, an extrapolation from high
to low dose must be performed. Many models are available for dose-
response estimation and high- to low-dose extrapolation. The
dose-response assessment must also extrapolate from animals to humans
if only animal data are available. This interspecies extrapolation is
carried out by applying a scaling factor to the experimental data30 or
through the use of physiologically based pharmacokinetic (PBPK) data.
6.12.2.1 Mathematical Dose-Response Extrapolation Models. No
single dose-response model is appropriate in all situations. A
6-53
-------
dose-response model is usually selected on an agent-specific basis.
However, two categories of dose-response models are generally used in
carcinogen risk assessment—mechanistic models and
tolerance-distribution models.
Mechanistic models describe some mechanism by which
carcinogenesis is believed to occur. All of the mechanistic models
assume that a tumor originates from a single cell that has been
altered by either the agent or one of its metabolites.30 Examples of
mechanistic models are the one-hit, multi-hit, and multistage models.
The one-hit model assumes that a single hit at a critical site
can result in malignant transformations. This model is conservative
(i.e., reduces the chance of underestimating risk) because it does not
account for cellular or deoxyribonucleic acid (DNA) repair mechanisms.
The multi-hit mechanistic model, an adaptation of the one-hit model,
assumes more than one chemical exposure or biological event is
required to elicit a carcinogenic response. The linearized multistage
model is the most frequently used of the low-dose extrapolation
models. It corresponds to the most commonly accepted theory of
carcinogenesis (the multistage process) and is the model most
frequently used by EPA in conducting dose-response assessments. This
model assumes that a cell progresses through a number of distinct
stages before becoming malignant. Like the one-hit model, the
multistage model is approximately linear in the low-dose region.
The second type of dose-response model, the tolerance
distribution model, is an empirical model that assumes for each
individual in a population there is a tolerance level below which that
person will not respond to the exposure.30 These models assume a
variability among individual tolerance levels that can be described in
terms of a probability distribution. This concept of individual
tolerance levels differs from the "threshold" concept used in most
noncancer risk assessment, which posits a general level of exposure
that is "safe" for most of the population. Tolerance distribution
models are actually based on the "nonthreshold" concept of
carcinogenesis because they refer to an infinite number of individual
tolerance levels or thresholds distributed along a curve. The
low-dose extrapolation techniques based on the tolerance distribution
theory include the probit (log-probit), logit (log-logistic), and the
Weibull model.
If animal data are used in the dose-response assessment, scaling
factors are commonly used to calculate a human equivalent dose. These
scaling factors are applied to animal data to account for differences
between humans and animals regarding body size, lifespan, route,
metabolism, and duration of exposure.1
Standardized dosage scales such as mg/kg body weight/day, ppm, in
the diet or water, and mg/m2 body surface area/day are commonly used to
allow for comparison of data across species.1 The EPA considers
extrapolation on the basis of surface area most appropriate because
6-54
-------
particular pharmacologic effects commonly correlate to surface area.
Because the body surface area is proportional to the animal's weight
to the two-thirds power, and because weight is more easily determined
than surface area, equivalent dose can be calculated as follows:
da/bwa<2/3' = dh/bwh<2/3'
where
da = experimental animal dose (mg)
dh = equivalent human dose (mg)
bwa = weight of experimental animals (kg)
bwh = weight of average human (kg).
6.12.2.2 Discussion of the Derivation of lUREs. An IURE
represents an upper limit increased cancer risk estimate from a
lifetime (70-year) exposure to a concentration of 1 //g/m3 in the
ambient air. This IURE is typically derived from the slope factor,
which is a plausible upper-bound estimate of the availability of a
response per unit intake or exposure concentration of a chemical over
a lifetime.1 When the slope factor is generated from the linearized
multistage model, it is denoted as qx*. Slope factors are usually
expressed in terms of (mg/kg-day)^ when derived from oral data and
(mg/m3)"1 when derived from inhalation data. The following equation is
used to convert a slope factor to an IURE for air contaminants:
IURE = Slope Factor x 1/70 kg x 20 m3/d x 10"3.
To calculate the IURE, it is assumed that a 70-kg individual with a
breathing rate of 20 m3/d is exposed to the carcinogen over a 70-year
lifespan. The factor of 10~3 in the IURE equation is required to
convert from milligrams to micrograms. The IURE is based on the
assumption of low-dose linearity. If a nonlinear low-dose-response
extrapolation model were used, the unit risk would differ at different
dose levels, and the dose-response assessment output could be
expressed as a dose corresponding to a given level of risk, analogous
to the risk-specific dose, rather than as a single IURE.
If the IURE is derived from animal data, it usually represents
the upper 95th percent confidence limit of the slope factor as
suggested by the variation within the animal data. Using the upper
95th percent confidence limit reduces the probability of
underestimating the unit risk.
For four priority HAPs (arsenic, chromium, radionuclide, and
nickel), human epidemiologic data are available and were used to
derive a maximum likelihood estimate (MLE) of the IURE. The MLE is
defined as a statistical best estimate of the value of a parameter
from a given data set.30 Therefore, the difference between the
upper-bound estimate and the MLE is that the upper bound is a
conservative measure of risk while the MLE is a statistically best
estimate.
6-55
-------
6.12.3 Discussion of Uncertainty in lUREs
Uncertainty is associated with the IURE because many assumptions
have been made in the process of deriving it. Uncertainty arises from
several areas in a dose-response assessment including intra- and
interspecies variability, high- to low-dose extrapolation, route-to-
route extrapolation, and the development of equivalent doses. One
type of potential uncertainty is often called the "healthy worker
effect." This results because the lUREs for some HAPs (e.g., arsenic,
chromium VI) are based on studies of workers exposed during their
working careers. The sensitivity of the workers to developing cancer
may not be the same as the sensitivity of the general population.
Therefore, there is uncertainty in the representativeness of the
worker population for calculating an IURE for the general population.
There may also be uncertainties because of truncation of observation
periods in most epidemiology studies. In addition, there are
uncertainties in the estimates of individual dose or exposure in the
epidemiology studies.
When using animal studies to estimate dose-response, the
assumption that administered dose is proportional to delivered dose is
typically used when estimating human equivalent doses. However,
physiological and pharmacokinetic differences between experimental
animal species and humans may result in differences in delivered
target organ dose. Not accounting for these may introduce uncertainty
in the estimation of human equivalent dose.
Low-dose extrapolation models can result in estimates of risks
that differ by several orders of magnitude. Therefore, selection of
model is critical. Some uncertainties may result in high biases,
others may result in low biases.
The IURE is based on the assumption that exposure to a particular
agent occurs over a 70-yr lifetime under constant conditions and
assumes that risk is independent of dose rate. Actually, the exposed
population is not exposed either continuously or at a constant level.
It is unknown how the detoxification and repair mechanisms may act at
higher or lower dose rates or with intermittent exposures, thereby
introducing uncertainty in the risk estimate. Variable exposure
concentrations introduce uncertainty. If detoxification and repair
mechanisms are more efficient with intermittent exposures (allowing
for recuperation or repair), the IURE would over-estimate risks when
compared to the total dose received. By contrast, if these mechanisms
were less efficient at an intermittently higher dose rate, the IURE
may underestimate risk when compared to total dose.
Risks from multiple carcinogens are typically estimated assuming
dose additivity. However, uncertainties are associated with this
approach. The risk summation technique assumes exposures are in the
low-dose range where responses are linear; however, at higher risk
levels, nonlinearity may need to be considered. The additivity
approach also assumes that each chemical acts independently (i.e.,
that there are no synergistic or antagonistic chemical interactions
6-56
-------
and that all chemicals produce the same effect). If these assumptions
are incorrect, over- or underestimation of the actual multiple-
substance risk could occur.31 Several other limitations to this
approach must be acknowledged. Because the IURE is typically an upper
95th percentile estimate of potency and upper 95th percentiles of
probability distributions are not strictly additive, the total cancer
risk estimate might become artificially more conservative as risks
from a number of different carcinogens are summed. However, the
human-derived potency estimates, which are the most important for this
analysis (e.g., radionuclides, arsenic, chromium VI, and nickel
subsulfide), are not based on the upper 95th percentile. These lUREs
are based on a maximum likelihood estimate. Therefore the potential
for artificially conservative estimates resulting from summing risks
of individual HAPs may not be an issue for this risk assessment.
Uncertainty in the breathing rate relates to the level of
activity. The breathing rates in epidemiological studies on which the
cancer slope factors (CSFs) are based are typically higher than the
standard 20 m3/d for the general population. Uncertainty in the
deposition fractions varies between individuals due to variation in
breathing rates, particle sizes, and the sizes of lung passages.
Retention half-times typically are distributed lognormally though
there is little information on how they differ between the
(epidemiological) study population and the general population. Life-
time averaged retention half-times should be slightly lower in the
general population due to the inclusion of young ages for which the
retention half-times are usually lower than adult values.
6.12.4 Variability in Cancer Dose-Response Assessment
Human beings vary substantially in their inherent susceptibility
to carcinogenesis. Person-to-person differences in behavior, genetic
makeup, and life history can influence susceptibilities. Such
interindividual differences can be inherited or acquired. Acquired
differences that can significantly affect an individual's
susceptibility to carcinogenesis include the presence of concurrent
viral or other infectious diseases, nutritional factors such as
alcohol and fiber intake, and temporal factors such as stress and
aging. Evidence regarding the individual mediators of susceptibility
supports the plausibility of a continuous distribution of
susceptibility in the human population.
Some researchers have attempted to determine the range of
susceptibility due to the general variability in physiological
parameters that may affect target organ dose. Their results indicate
that the difference in susceptibility between the most sensitive 1
percent of the population and the least sensitive 1 percent might be
as small as a factor of 36 (if the logarithmic standard deviation was
0.9) or as large as a factor of 50,000 (if the logarithmic standard
deviation were 2.7) .28
Certain groups of individuals within the population are
inherently more sensitive to carcinogen exposure than others. Factors
6-57
-------
that influence susceptibility include age, race, sex, and genetic
predisposition. An example of a sensitive subpopulation is children.
This subpopulation can be more sensitive to certain chemicals and more
susceptible to cancer for a variety of reasons, including:
• Children have faster breathing rates than adults and,
thus, inhale larger quantities of a pollutant, relative to
their body weights.
• Organs in children are still growing and developing and
are, therefore, more prone to disruption by an
environmental agent.
• Young organisms appear to be inherently susceptible to
many carcinogens. Young experimental animals have been
shown repeatedly to acquire more tumors in a shorter time
with a smaller dose than adult animals.
In most circumstances, as with this study, there are not enough data
available to perform separate quantitative dose-response assessments
for these sensitive subpopulations. Obviously, children are not
included in the work force at plants where much of the epidemiology
data are collected.
As stated above, the IURE is based on the assumption that
exposure to a particular agent occurs over a 70-year lifetime under
the same conditions to which the study group was exposed. For
animals, it is essentially steady and constant exposure over a
lifetime; for humans, it is varying exposure over their working career
at a particular plant. In effect, this assumes that risk is
independent of dose rate. Recent research suggests that cellular
repair mechanisms exist that can reverse the damage caused by a
carcinogen, and it is likely that these mechanisms operate most
effectively after low doses or in the absence of repeated doses.
Therefore, variability in exposure would also influence or create a
variability in how effective the IURE predicts risk.
6.13 PRELIMINARY QUANTITATIVE UNCERTAINTY AND VARIABILITY
ANALYSIS FOR INHALATION EXPOSURE AND RISK ASSESSMENT
6.13.1 Introduction
Risk assessment is a complex process, and uncertainty will be
introduced at every step in the analysis. Even using the most
accurate data with the most sophisticated models, uncertainty is
inherent in the process. There are a number of uncertainties
associated with the exposure assessment of emissions from utilities.
These include parameter estimation (test results), model choice, and
the use of simplifying assumptions.
Uncertainty in emissions and exposure estimates can result from
uncertainty (i.e., doubt or ignorance of the true value) or from
variability (i.e., known range of values over time, space, or within a
6-58
-------
population). A quantitative uncertainty analysis was conducted for
the direct inhalation exposure part of this risk assessment. The
evaluation of uncertainty in the estimation of emissions, dispersion
and exposure is summarized here. See Appendix G of the EPA Interim
Final Utility Report for details on the uncertainty analysis.6 This
uncertainty evaluation does not include consideration of the impacts
(and associated uncertainties) due to long-range transport and
multipathway exposures. The focus of this particular analysis is the
uncertainties and variability of the inhalation exposure within 50 km
of the plants.
The need for formal uncertainty analysis as a part of any risk
assessment and its aid in conveying results of the risk assessment are
widely accepted, having been proposed in both the EPA Risk
Characterization Guidance and the NRC Committee Report: Science and
Judgement in Risk Assessment.28 Furthermore, any procedure that relies
on a combination of point values (some conservative and some not
conservative) yields a point estimate of exposure and risk that falls
at an unknown percentile of the full distributions of exposure and
risk.
The risk estimates presented in previous sections were derived by
utilizing various input data and assumptions. The results were
presented as point estimates of risks. The following uncertainty
analysis was conducted to determine the degree of conservatism
associated with these point estimates.
The uncertainty analysis focused on the three HAPs (nickel,
arsenic, and chromium) that accounted for over 95 percent of the high-
end estimate of cancer incidence. An analysis of uncertainty on these
three HAPs accounts for much of the uncertainty in the overall risk
estimates.
6.13.2 Approach to Quantitative Uncertainty Analysis
Uncertainty has been classified into four types (parameter
uncertainty, model uncertainty, decision-rule uncertainty, and
variability). The first two, parameter uncertainty and model
uncertainty, are generally recognized by risk assessors as major
sources of uncertainty. Parameter uncertainty occurs through
measurement errors, random errors, or systematic errors when variables
cannot be measured precisely either because of equipment limitations
or because the quantity being measured varies spatially or temporally.
Model uncertainty can result from surrogate variables, excluded
variables, abnormal conditions, and/or incorrect model form.
Decision-rule uncertainty arises out of the need to balance different
social concerns when determining an acceptable level of risk, which
can affect the choice of model, data, or assumptions. Variability is
often used interchangeably with the term "uncertainty," but this is
not strictly correct. Variability is the unchanging and underlying
distribution of a parameter based on physical, chemical, and/or
biological processes (e.g., body weight within a population). Even if
6-59
-------
variability is known (therefore, not in itself uncertain), it still
contributes to overall uncertainty of the risk assessment.
This uncertainty analysis focused on parameter uncertainty within
the models and data available. Other uncertainties which were not
amenable to quantitative evaluation are discussed qualitatively in
section 6.12. Table 6-25 briefly summarizes information regarding the
parameters used in the risk estimation process. Model uncertainties
are not addressed in the quantitative uncertainty analysis, but are
described qualitatively. Variability has been evaluated separately
for exposure-response, but is included in the overall estimate of
uncertainty related to emissions and exposure. The goal of this
uncertainty analysis is to estimate the range of possible risk
estimates considering the parameter uncertainty and variability. It
should also be noted that there are other sources of uncertainty, some
of which may be significant, which could not be evaluated
quantitatively. These uncertainties are qualitatively discussed.
The approach used in this analysis was to identify the
uncertainty with each of the parameters used in the risk estimation
process. First, the uncertainty associated with each of these
variables was described using an appropriate statistic (e.g., mean and
standard error of means) or as a probability density function (the
relative probability for discrete parameter values). The standard
error of the mean (SE) for each parameter was the estimate of
uncertainty and variability used rather than the standard deviation
for each parameter. Since the available dose-response data are based
on lifetime exposure, and the cancer risk assessment is concerned with
long-term average exposures, the SE is a more appropriate statistic.
However, it should be noted that using the SE from a sample may be an
overconfident estimate (i.e., too narrow a range) of uncertainty (see
Appendix G of the EPA Interim Final Utility Report for explanation).6
In general, numerical methods (e.g., Monte Carlo simulation) were
then used to develop a composite uncertainty distribution by combining
the individual distributions. In Monte Carlo simulations, the risk
and/or model equations are repeatedly solved using randomly sampled
values from the specified distributions to calculate a distribution of
risk values. These risk distributions were derived for estimates of
MIR and population risks. Because variability was not specifically
differentiated in the analysis of emissions and dispersion modeling,
uncertainty and variability were simulated together in a one-
dimensional Monte Carlo simulation.
6-60
-------
Table 6-25. SUMMARY OF BASIC PARAMETERS USED IN THE INHALATION RISK ASSESSMENT FOR UTILITIES
[Default Option/Assumptions
(and departure from default and reason why)
Conservatism
Data 1 Uncertainty/Variability
(e.g., source, quality) | (quantitative and qualitative)
Distribution
Judgement
Strategy
EMISSIONS CHARACTERIZATION
Fuel consumption
Coal State of Origin
Trace element
concentration
Coal cleaning factor
(CCF)
EMF: Boiler and
APCD
Coal:
• 1990 (or 1989, geo mean 1980-8) UDI/EEI
data self-reported by the utilities to DOE.
•Adjustments made for heating value of
different coals
• Based on total tonnage
Oil: assumed to be residual oil, quantity
consumed in gallons is converted to mass
based on an assumption of uniform density.
•Assume all from the state where majority of
fuel consumed based on total tonnage
•Arithmetic average of coal type for state of
origin which is used most at the facility.
•Oil: average HAP concentration in test data
of residual fuel oil No. 6 (about 80% of all oil
burned).
Process of preparing coal for shipment may
reduce some mineral matter. Since about
77% of eastern and midwestern bituminous
shipments are cleaned a CCF was applied to
all bit. coal
Geometric mean of test data, measured in
gas stream, ash stream was ignored
Low, average
value used
Unknown
Low, average
value used
Low, average
value used
Low, average
value used
• UDI Database: self-reported,
with no QC or validation
•Average heating values used for
coal type (lign = 6800 BTU/lb, bit.
= 12688 BTU/lb, subbit. = 9967
BTU/lb)
• UDI database identifying
majority use
• USGS core/channel sampling
(extraction) of economically
feasible coal seams (n= 3331)
Testing of coal shipments from
Wyoming, Colorado and Illinois
Coal: 19facilitestested of varying
configurations, combinations of
boiler type and APCD.
Oil: teesting at 2 facilities.
Each test point was at least a
triplicate sample
•Accuracy of self-reported values
• Fuel consumption overtime due to demand,
sulfur content, etc...
•Coal from several states may be used at
one plant, mix of states coal actually used
• Relative composition of state coals due to
availability, cost, sulfur content etc.. may
change over time
Coal:
•Cone, measured in extracted coal, not in
coal shipments, reductions in trace element
cone, may occur during processing
•Coal seams measured may not actually
being used for shipment
•Coal from other states used at plant may
contribute significantly
•Variability within a coal seam, between coal
seams within a state.
OIL:
• Density will vary among No. 6 fuel oils
which means that the volume and mass
consumed will vary. Concentration of HAPs
within oil will vary.
•Coal cleaning data may not apply to other
types of coal
• Effectiveness of coal cleaning may vary
according to variability in the sulfur and ash
content within a coal seam and the variability
in processing
•Were units tested representative of units in
operation
• Unit performance likely to vary over time
due to fuel and operating parameters.
Normal
NA
Log-normal
Normal
Beta or
Triangular (if n=1)
Based on engineering
judgement.
Relative contribution
between states held
constant
prevents the possibility
of negative
concentration with no
upper limit
engineering process
Beta: constrained
within 0 and 1 and
distribution defined by
data..
Tri.: no distribution can
be estimated from
single point, value
used as the apex
bounded by 0 and 1 .
CTl
I
CTl
-------
Table 6-25. Continued
[Default Option/Assumptions
(and departure from default and reason why)
Conservatism
Data 1 Uncertainty/Variability
(e.g., source, quality) | (quantitative and qualitative)
Distribution
Judgement
Strategy
DISPERSION AND EXPOSURE MODELING
Dispersion
Roughness (rural v.
urban)
Terrain
Meteorology
Effective Stack
Height
Location of Exposed
Population
Gaussian plume
Population density (within 3 km of plant) is
assumed to be an indicator or proxy for
setting of the plant, and that urban and rural
are representative of surface roughness.
Binary choice of urban v. rural
Assumed to be flat terrain.
Flat terrain used in gaussian plume
dispersion.
The meteorological data from the nearest
STAR location are used to represent the
meteorology near the plant.
Stability classes are assumed to be
represented in the STAR data and implicitly
addressed in the HEM model.
Effective stack height is calculated using the
stack height, exit velocity and exit gas
temperature as reported in the UDF
database.
Population assigned to the centroid of the
census block or, if within 0.5 km, to receptor
grid location for which concentratio nis
estimated. Assumed to represent a persons
average exposure ( they may also spend
time in areas of higher or lower
concentrations)
Unknown
Unknown, may
not account for
values beyond
model defaults
Low, as shown by
complex terrain
analysis
Unknown,
assumed to
represent site met
conditions
Medium, actual
data with
conservative
model
Low/Medium
assignment at 0.5
km accounts for
variability
Limited data on other models
Census data on population within
3 km of facility to indicate urban.
Extensive data on terrain
surrounding each facility.
However, analysis is very
exhaustive and cannot be carried
out for all facilities.
STAR data are typically five-year
averages at 350 airports.
Data from UDI database are as
reported by the facility.
Little or no data on variability in
exit gas temperature orexit
velocity
1990 Census Block data.
• Roughness not binary and not always
attributable to population density (e.g.,
buildings) but other land features as well.
• Urban and rural model default settings may
not represent the entire range of surface
roughness leading to possible extremes not
addressed.
•Terrain effects can be significant leading to
minimal dispersion and high exposure
•Terrain is not a binary parameter and the
degree of terrian differences will vary
between plants.
• Meteorology at plant may be significantly
different than the nearest plant. Different
meteorology may not affect the maximum
concentration but may significantly affect the
number of persons exposed and at risk.
Short-term meteorological conditions (e.g.,
inversion) may affect short-term exposure
levels.
• Meteorological conditions will vary at a site
overtime.
•The effective stack height may vary
significantly from the calculated value due to
variation in exit gas temperature and velocity.
which would greatly impact the resulting
exposure concentration.
•Stack height would not vary as it is a
physical parameter.
• Location of the centroid is estimated, is it
accurately defined. Uncertainty as to where
people are actually located with respect to
the centroid.
•The location of individuals will vary with
respect to the centroid, some being in higher
concentrations some in lower concentrations.
Not analyzed
Each plant run in
both modes
Analyzed
separately
(see Section 3.2)
Three closest met
locations used
High(1.1)Med(1)
Low (0.9)
Analyzed
separately
Beyond scope of
project. Model is EPA
default.
Population may not be
an indicator of actual
surface roughness.
the urban and rural
defaults approximate
the range of
uncertainty.
Subset of plants
analyzed were
representative.
The actual site met
conditions would be
approximated by at
least one of three
closest stations.
Based on subjective
judgement.
Changing grid or
centroid assignment
distances showed little
influence, resolution of
0.5 km minimizes
spatial effects.
CTl
I
CTl
-------
Table 6-25. Continued
[Default Option/Assumptions
(and departure from default and reason why)
Conservatism
Data 1 Uncertainty/Variability
(e.g., source, quality) | (quantitative and qualitative)
Distribution
Judgement
Strategy
EXPOSURE-RESPONSE ASSESSMENT
Exposure Duration
(Population mobility)
Exposure Frequency
Breathing Rate
Lung Deposition
Retention Half-life
Slope factors
Assumes persons spend 70-year lifetime at
the location to which they are assigned. The
concentration at the centroid represents their
average to which they are exposed.
(Indoor/outdoor concentration)
Assumes exposure at 100% of outdoor
concentration.
Assumes that workers breathing rate is
equal to national average.
No adjustment between worker and general
populations.
No adjustment between worker and general
populations.
Used EPA-verified slope factors, best
estimate linearized function.
High (MEI/MIR)
tends to maximize
exposure
High, maximizes
exposure,
Low to Med,
workers may have
higher BR value
Low, average
value used
Low, average
value used
Unknown, but
believed to be high
Very little data on mobility which
are applicable to the range of
populations affected. Most have
been focused on small subsets
(e.g., residency in apartments).
Measurements of indoor/outdoor
concentration ratio, time-activity
patterns, and exposure. No data
specific to locations of electric
utility plants.
Measurements of minute volumes
for different population subgroups.
Measurement of lung deposition
fractions.
Measurement of half-lives, but not
for the specific HAPs evaluated.
Human epidemiological data.
Individuals mobility will affect exposure.
Uncertainty about defining a representative
mobility pattern or distribution which also
accounts for movement to alternatively
polluted areas.
• Population mobility varies dramatically
within a population and for an individual over
time.
•Alternatively can consider time-activity
patterns (e.g., indoor/outdoor, movement
within area) and residence time (average = 9
years, 90th = 30 years).
The relationship between indoor and outdoor
concentration is complex because infiltration
is affected by climate, building type,
ventilation etc..
Infiltration will vary over time due to climate
variability (e.g., open windows).
Breathing rates differ greatly by age and
activity.
Lung depostion can vary by age and activity
level.
Retention half-lives will vary by age, activity
level, and particle size.
Slope factors will differ dramatically based on
model choice, statistical uncertainty in data.
Not analyzed
Log normal,
variability
Normal,
uncertainty
Log normal,
variability
Normal,
uncertainty
Log normal,
variability
Normal,
uncertainty
Log normal,
variability
Normal,
uncertainty
Log normal,
uncertainty
Given the proportion of
the population who are
exposed to emissions
from utility emissions, it
is likely that people
who move will still be
exposed (though at
lower levels).
Based on limited data
and accepted EPA
defaults.
Variability measured,
prevents negative
values.
Uncertainty from
standard error of
mean.
Variability measured,
prevents negative
values.
Uncertainty from
standard error of
mean.
Variability measured,
prevents negative
values.
Uncertainty from
standard error of
mean.
Variability treated
qualitatively.
Uncertainty from the
SEM from existing
data. Model
uncertainty:
qualitatively.
CTl
I
CTl
W
NOTE: Quantitative values for all parameters and their distributions are presented in the body of the text.
Key: High = Most likely to overestimate than underestimate. Can represent an upper bound estimate.
Medium = May either under- or over-estimate. With the use of conservative models usually more likely to over- than under-estimate risk.
Low = Usually an unbiased estimator using the average value. Equally likely to over- or under-estimate risk.
UDI/EEI: Utility Data Institute, Edison Electric Institute
EMF: Emissions modification Factor
APCD: Air pollution control device
Beta Distribution: constrained between two distinct values (e.g., 0 and 1), defined by the mean and standard error of mean (SEM) of the originial data. This distribution maintains the mean and standard error of mean
(SEM). The use of a truncated distribution (either normal or log-normal) can lead to a drift in the mean and/or SEM from the original data.
-------
The uncertainty analysis was conducted on the three major
components of the risk assessment process: emissions
characterization, dispersion and exposure modeling, and
exposure-response assessment. Each of these is summarized briefly
below. Figure 6-16 provides an example of how the uncertainty from
each of these components is combined into an overall distribution. A
detailed uncertainty analysis could not be conducted on all of the
utility plants. Therefore, a total of four plants (two oil-fired and
two coal-fired plants) were selected which contribute most to risk,
the highest estimated incidence and the highest maximum individual
risk. Each of these plants was analyzed for arsenic, nickel, and
chromium. The highest incidence oil-fired plant (Plant No. 29)
accounted for about 7 percent of the annual cancer incidence and,
therefore, was chosen for illustration purposes.
6.13.2.1 Emissions Characterization Uncertainty. An emissions
factor program was developed by EPA to estimate plant-specific
emissions rates based on fuel type, fuel origin, plant configurations,
and emissions testing results. The emission factor program (including
principles and rationale) and the data used are described in chapter 3
and appendix D. This program is based on a mass-balance concept,
reducing concentrations in the fuel due to the impact of the boiler
and control devices.
The parameters used in the emissions characterization were: fuel
consumption (coal: ton/yr, oil: barrel/yr), HAP (trace element)
concentration in fuel, coal cleaning factor (if needed), emissions
modification factors for the boiler (EMFb, boiler-specific factor to
account for amount of HAP entering boiler to that exiting boiler), and
the air pollution control device (APCD), if present (EMFa, APCD-
specific factor to account for amount of HAP entering the APCD to that
exiting APCD).
It should be noted that two different trace metal concentrations
in oil were used. The original data were from the EPRI's Field
Chemistry Emissions Monitoring (FCEM). A subsequent data set was
provided by UARG and their contracting lab (SGS Environmental
Laboratories). An analysis of these data indicated that the samples
were discretely different. It is unknown which "sample population" is
most representative of the oil burned. Therefore, the two data sets
are treated as two distinct cases and are assumed to be representative
of the range of oil being burned by utilities.
6.13.2.2 Plant-Specific Emission Rates. Monte Carlo simulation
was used to develop a distribution of possible plant-specific
emissions rates. Simulations were carried out randomly sampling
values for fuel consumption, HAP concentration, and EMFs. For
illustration purposes, Table 6-26 and Figure 6-17 present the summary
statistics and graphical representation, respectively, of the
emissions predicted for Plant No. 29. This distribution gives some
indication of the degree of uncertainty and the possible range of
emissions estimates that may be experienced. The emissions estimates
6-64
-------
I
"g
w r
£
Q.
1
§
10
C
o
(A
Q
Frequency
i S s
§ C S $
AlHiqiqoJd
§
•H
n
n
«
o,
m
•H
a »
n •
-H 01
3
§
2
W
z
Q
W
CO
(A
C
s
1
S
S
z
c
o
,
o f
Q- S
w -e
UJ (D
DC 9
W c
O .2
Q. 3
X fi
Frequency
(0
tfi
A
S 2
•"8.
1
i
VO
VO
S
•H
h
6-65
-------
oo
-
r-
CO
IO
* '
CD
S
E
I
o"
CO,
CO
"co
-.2 Z> 3*
i= a. -a
c ^ CD
CD o c
1°1
o ^ hr
"co "5 £
C O) 5
'^rV
"i_ LJ- C
O .!= .2
CO ^
2 I
CO CO
"O CD
rn CO
W CO
2
"co
o
a.
'
co o ±:
^ X" co °
W m ^ °>
P S2 -a .g
O)
CD
CD
P c >< ^ ® b
Q.^ 8 d f s s g
-
CD
C
!Q
E
S3
0
'55
CO
o
CO
§ dii^SSoO^^
CO
CD
co"
-
CO
"co
111
CO ^
'^= CD
"E 0.
co
_CD ^_^
'•§ x
S E
O Q.
CD °-
CL -2-
E"°.^*^ Q- 3 co^
CTco^mST—o'a
^COCco'iocDCD0-
w "CD .o .g co -jg > CD
•i cu 5 •§ "I I ¥ S
iiiiffii
0 g. » g (3 = ^ i-
S CD Q. .^ 0 co
.92 E > ii ^ = ^
(0
C7)
O =
6-66
-------
Figure 6-17. Summary of Results of Monte Carlo Simulation of
HAP Emissions from Oil-Fired Plant No. 29
FCEM Concentration Data
SGS Concentration Data
CellE11
.336-r-
Forecast: Plant 29 As Emissions
Frequency Chart 2,942 Trials Shown
.252 ._
2 .168
a
.084 .
.000
lllllu.
494
100.00
200.00
kg/yr
300.00
400.00
Cell E12
Forecast: PI 29 As Emission (SGS)
Frequency Chart 2,941 Trials Shown
.027
.020 ._ .
.013
.007
.000
0.00
125.00
250.00
kg/yr
375.00
Illllilllil
78
58.5
Tl
39 2
19.5
CellH11
.036 +
Forecast: Plant 29 Cr Emissions
Frequency Chart 2,937 Trials Shown
105
52.5
262
CellH12
.087
.065 __
Forecast: PI 29 Cr Emissions (SGS)
Frequency Chart 2,444 Trials Shown
.043._
212
106
£
022 _
.000
0.00
200.00
0.00
100.00
Forecast: Platn 29 Ni Emissions
Cell K11 Frequency Chart 2,930 Trials Shown
.036 H : 1- 106
.027
.018
009
.000
Cell K12
.030f
Forecast: PI 29 Ni Emissions (SGS)
Frequency Chart 2,940 Trials Shown
79.5
53
265
0
0.00 5,000.00 10,000.00 15,000.00 20,000.00
kg/yr
44.5
22.2
16,875.00 22,500.00
FCEM = Field Chemical Emissions Monitoring from EPRI program. Original oil concentration data.
SGS = Subsequent data, trace metal analysis conducted by Utility Air Regulatory Group (UARG) HAP con
from samples collected for radionuclide analysis. The concentration was determined
by SGS Environmental Laboratories, a contractor to UARG.
Combined = Combined forecasts assuming equal probability of the FCEM and SGS data sets.
Initial Point Estimate = The estimate of emissions used in the baseline exposure assessment.
This value was based on the average concentration in the FCEM data.
(Percentile) = The percentile of the predicted distribution corresponding to the initial point estimate.
6-67
-------
used in the baseline risk assessment were generally central tendency
estimates (i.e., geometric means). In general, the 95 percent
confidence range for long-term average emissions estimates are within
a factor of 2 to 3 of the emissions estimates used in the risk
assessment. For example, the 95th percentile of the overall range of
predicted emissions ranged from about 0.9 times the original emissions
estimate from plant No. 343 for nickel to about 2.5 times the original
estimate for arsenic. As shown in Table 6-26, the original estimate
of emissions from the baseline risk assessment ranged from the 22nd to
the 95th percentile of the range of emissions predicted under the
uncertainty analysis.
A preliminary evaluation of the EFP was conducted. Comparisons
were made of test data from 19 utility boiler stacks (17 coal-fired, 2
oil-fired) against predicted emissions for the Table 6-26. Summary of
Results for Monte Carlo Simulation of HAP Emissions (kg/year) from
Oil-fired Plant #29 same plants using the EFP. For each facility, the
emission estimate from the EFP was divided by the reported value from
the corresponding test report. A value of 1 meant that the EFP
exactly predicted the test results, values lower than 1 indicated
the EFP underpredicted emissions, while values higher than 1 indicated
the EFP overestimated emissions. In general, the results suggested
that the EFP performs reasonably well for predicting emissions on a
national basis. The average of the ratios across all stacks and
constituents was 1.08, while averages for arsenic, chromium, and
nickel were 1.6, 0.68, and 0.97, respectively.
However, while the model did well in predicting overall or
average emissions across a range of utility boilers, large differences
between predicted and reported values were found for a few individual
boilers and constituents. The largest difference was for an
individual boiler for which estimated emissions were about 2,600 times
lower than reported test results. However, it was determined that
this facility was a low-risk plant in the overall analysis; therefore,
increasing emissions of this plant by 2,600 would not change the
overall results. The EFP tended to underestimate rather than
overestimate emissions about 70 percent of the time within this sample
of boilers. A preliminary evaluation of facilities with large
differences between projected and actual emissions found that these
facilities were likely to burn multiple fuel types. In addition,
variability in fuel composition might also lead to large differences
between measured and calculated emissions. Since most of the higher
risk plants do not fall into this category, the differences here are
not expected to impact significantly on the overall risk estimates.
See Appendix G of the EPA Interim Final Utility Report for further
discussion.s
6.13.2.3 Dispersion and Exposure. Air dispersion modeling is
complex and nonlinear, cannot be carried out with the use of
spreadsheets, and requires significant time to conduct the modeling
and process the data for each run. To better estimate percentile
values above 90 percent, a stochastic (Monte Carlo) approach requires
6-68
-------
large numbers (thousands) of repetitive runs (3,000 was used for the
emissions estimates) needed to generate a distribution. Given the
time and resources required for single runs, the Monte Carlo approach
was not feasible and an alternative approach was needed to evaluate
the uncertainty in dispersion and exposure modeling.
The degree of dispersion and resulting exposure is affected by
three major parameters: plant stack parameters (e.g., stack height,
stack gas temperature, and exit velocity), meteorologic conditions,
and surface roughness (urban vs. rural). The uncertainty analysis,
therefore, focused on these three parameters. The three factors being
evaluated are nonlinear with respect to each other and require a
separate HEM run for each parameter value. Therefore, a test matrix
approach was used to evaluate uncertainty in the exposure modeling
component of the exposure assessment. A limited number of options
were developed to represent the expected range of uncertainty for each
of these three categories of parameters as follows:
Surface roughness: urban or rural mode
Stack parameters: represented as high (1.1 x UDI values),
medium (UDI values), and low (0.9 x UDI
values) estimates for stack gas temperature
and flue gas exit velocity
Meteorology: three closest meteorology locations in the
STAR database.
As a result, for each plant, a total of 18 different HEM runs were
made covering each combination of dispersion parameters. For the
purposes of this uncertainty analysis, it was assumed that there is
insufficient information to determine the relative correctness of each
combination and, therefore, each was considered equally likely to
represent the possible range of values. The coefficients for
estimating maximum concentration and total exposure (per 1 kg/yr
emission) resulting from each of these 18 HEM runs were summarized for
each plant.
6.13.2.4 Exposure-Response Assessment. The variability of the
quantitative relationship between exposure and the excess probability
of cancer for different humans and the uncertainty in the mean (taken
here also to be the "best estimate" or "maximum likelihood estimate")
quantitative relationship between exposure and the excess probability
of cancer were both addressed. As with the uncertainty analysis for
emissions and dispersion, efforts were limited to arsenic, chromium,
and nickel. Specific parameters, for which uncertainty about the mean
value (or best estimate for a given parameter within the exposed
population) was addressed, include exposure frequency, exposure
duration, breathing rate, deposition fractions, and retention half-
times. Uncertainty related to the IURE focused on data and the use of
epidemiologic data (typically from workers) extrapolated to the
general population.
6-69
-------
The software program Crystal Ball8 (Decisioneering, Inc., Denver,
CO) conducted stochastic (Monte Carlo) simulations of the risk
estimates, incorporating the uncertainty for each parameter. A
probability distribution that best represents the variable, its
average value, and a measure of uncertainty about the average value
was developed for each parameter. The simulation consists of
conducting repeated calculations (thousands) of risk using values for
each parameter sampled from the distribution of values for that
parameter.
The study of variability focused on how parameter values would be
expected to vary among individuals within the general population and
how that would affect the estimation of risk and incidence. The
parameters for which some measure of variability among individuals
within the general population was addressed include exposure duration,
exposure frequency, breathing rate, deposition rate, and retention
times in the lung. No specific measures of variability were available
for how the IURE for these three HAPs may differ among individuals.
However, limited data indicate that the IURE differs between smokers
and nonsmokers and this difference was incorporated in the analysis.
6.13.3 Discussion of Results of the Quantitative Uncertainty Analysis
The risk estimation process used in the baseline assessment
utilized a combination of parameters, each with varying degrees of
conservatism (the degree of overestimation, or underestimation). In
general, the estimates of maximum individual risk and annual cancer
incidence derived in the baseline risk assessment were conservative,
generally around the 95th percentile on the distribution. The 95th
percentile is roughly 10 times the median and about 5 times the mean.
The distribution of estimates of MIR for Plant No. 29 are presented in
Table 6-27. The sensitivity analysis indicated that the dispersion
coefficient (surface roughness) was the most significant parameter for
estimating uncertainty MIR and incidence, followed by the EMFs. The
deposition fraction, retention time, and exposure frequency also
contributed significantly in the variability of these estimates.
The EPA risk assessments are generally conservative (more likely
to overestimate than underestimate risks). Often there is a concern
that the use of several conservative assumptions results in risk
estimates that are unrealistic and beyond the range of possible risks
(i.e., overly conservative). The results of the uncertainty analysis
indicate that the baseline inhalation risk estimates are reasonably
conservative (predicted to be roughly around the 90th or 95th
percentile). The uncertainty analysis supports the general conclusion
that the baseline risk estimates are likely to be reasonable high-end
estimates.
6-70
-------
Table 6-27. Distribution of MIR: Plant No. 29:
Comparison of FCEM and SGS Concentration Data
MIR, Plant No. 29
Uncertainty
Arsenic Chromium
Nickel
Mean
Initial Point Estimate
(percentile)
Percentiles:
0.0%
2.5%
5.0%
10%
25%
50%
75%
90%
95.0%
97.5%
Ratio
95th baseline
95th median
95th mean
FCEM
1E-07
(96)
2E-12
3E-09
6E-09
1E-08
3E-08
6E-08
8E-08
2E-07
5E-07
1E-06
0.8
8.7
3.5
SGS
6E-07
6E-07
(71)
1E-09
8E-09
2E-08
3E-08
7E-08
2E-07
7E-07
1E-06
2E-06
4E-06
4.1
10.6
3.8
FCEM
1E-07
(87)
2E-10
2E-09
3E-09
5E-09
1E-08
4E-08
1E-07
3E-07
5E-07
7E-07
2.3
12.8
4.2
SGS
4E-08
2E-07
(98)
1E-11
7E-10
1E-09
3E-09
7E-09
1E-08
2E-08
8E-08
1E-07
2E-07
0.7
9.6
3.9
FCEM
2E-06
(90)
2E-09
2E-08
5E-08
9E-08
2E-07
6E-07
2E-06
4E-06
7E-06
1E-05
1.7
11.1
4.0
SGS
3E-06
4E-06
(85)
6E-09
4E-08
6E-08
1E-07
3E-07
9E-07
3E-06
6E-06
1E-05
2E-05
2.5
10.4
3.8
Variability
Arsenic Chromium
Nickel
Mean
Initial Point
(percentile)
Percentiles
Estimate
0.0%
2.5%
5.0%
10%
25%
50%
75%
90%
95.0%
97.5%
FCEM
1E-07
(95)
5E-12
1E-09
3E-09
6E-09
1E-08
3E-08
6E-08
2E-07
5E-07
1E-06
SGS
6E-07
6E-07
(68)
4E-09
1E-08
2E-08
4E-08
1E-07
3E-07
7E-07
2E-06
2E-06
3E-06
FCEM
1E-07
(90)
4E-10
2E-09
3E-09
6E-09
1E-08
4E-08
1E-07
3E-07
4E-07
6E-07
SGS
3E-08
2E-07
(97)
1E-11
3E-10
7E-10
1E-09
3E-09
7E-09
3E-08
7E-08
1E-07
2E-07
FCEM
2E-06
(90)
5E-09
3E-08
6E-08
1E-07
3E-07
7E-07
2E-06
4E-06
6E-06
9E-06
SGS
2E-06
4E-06
(90)
2E-08
6E-08
1E-07
2E-07
4E-07
1E-06
3E-06
6E-06
9E-06
1E-05
FCEM = Field Chemical Emissions Monitoring from EPRI program. Original oil concentration data.
SGS = Subsequent data, trace metal analysis conducted by Utility Air Regulatory Group (UARG) HAP committee from samples collectec
The concentration was determined by SGS Environmental Laboratories, a contractor to UARG.
Combined = Combined forecasts assuming equal probability of the FCEM and SGS data sets.
Initial Point Estimate = The estimate of emissions used in the baseline exposure assessment.
This value was based on the average concentration in the FCEM data.
(Percentile) = The percentile of the predicted distribution corresponding to the initial point estimate.
6-71
-------
The uncertainty analysis suggests that the most likely inhalation
MIRs (i.e., central tendency MIRs) and most likely cancer incidence
values (i.e., central tendency cancer incidence estimates) may be
roughly 2 to 10 times lower than the high-end MIRs and incidence
estimates presented above. In addition, based on results of the HEM
modeling and the uncertainty analysis, it is predicted that the
average individual risks due to inhalation exposure to utility HAP
emissions for the total exposed U.S. population (roughly 200,000,000
people) are roughly 100 to 1,000 times lower than the high-end MIRs.
However, it should be noted that this analysis has focused only
on parameter uncertainty. Also, not all parameters were included.
For example, residence time and activity patterns were not assessed
quantitatively in the uncertainty analysis. As a result, the
uncertainty presented here may underestimate the overall uncertainty.
6.14 QUALITATIVE DISCUSSION OF ADDITIONAL UNCERTAINTIES
There are several areas of uncertainty that were not covered in
the quantitative analysis. Several of these were discussed in
previous sections of this report. Further discussion of two areas of
uncertainty is provided below.
6.14.1 Uncertainty Using lUREs
As discussed in section 6.12, there are uncertainties associated
with the lUREs. Many of these uncertainties were not included in the
quantitative uncertainty analysis because adequate data were not
available.
6.14.2 Residence Time and Activity Patterns
In the baseline assessment for the MEI risks, it was assumed that
people are exposed to the modeled concentration at their residence for
70 years. This approach assumes that people spend most of their time
at home and that the average concentration at their residence
represents the average concentration to which they are exposed.
Electric utility plants typically have high stacks compared with many
other air pollutant point sources. As a result, ground-level
concentrations (and concomitant exposures) would tend to vary less
with distance than other sources. Therefore, movement by individuals
within the grid would have minimal impact on exposures. The EPA
realizes that the average person does not live in the same house for
70 years. However, adjusting for exposure due to changes in residence
is no easy task, especially for utilities since plants are located
nationwide and roughly 80 percent of the United States population live
within 50 km of at least one plant.
This uncertainty was not quantified for several reasons. First,
a person who moves out of one residence may move into another
residence still in the high-concentration area (e.g., person moves
next door). Second, a person may move away from an area for a period
of time, then move back to the same location. Third, since there is
typically more than one person located in the high exposure area, if
6-72
-------
all except one move away (e.g., one person in the census block stays
for 70 years), then the assumption of 70-year residence time holds for
the MEI. And, fourth, a person may move from the area of exposure of
one utility into an exposure area of another. This person's exposure
may change, but may not become zero. Therefore, 70-year exposure is
considered a conservative, but reasonable, assumption for the MEI.
However, it is still quite uncertain how much residence time and
activity patterns would affect the risk estimates.
6-73
-------
6.15 REFERENCES
1. U.S. Environmental Protection Agency. Risk Assessment Guidelines
of 1986. EPA-600/8-87/045. (Guidelines for Carcinogen Risk
Assessment, Guidelines for Mutagenicity Risk Assessment,
Guidelines for Health Risk Assessment of Chemical Mixtures,
Guidelines for Health Assessment of Suspect Developmental
Toxicants, Guidelines for Estimating Exposures) Office of Health
and Environmental Assessment, Washington, DC. 1987.
2. U.S. Environmental Protection Agency. Air Pollutant Emissions
from Electric Utility Steam Generating Units—Interim Final.
Volume III. Appendices H-M. EPA-453/R-96-013C. 1996.
3. EPRI. Personal communication by fax from Paul Chu to Chuck
French and Bill Maxwell, EPA. November 27, 1995.
4. U.S. Environmental Protection Agency. Guidelines on Air Quality
Models. Code of Federal Regulations, 40, Appendix W to Part 51,
July 1, 1994.
5. California Air Pollution Control Officers Association. Air
Toxics "Hot Spots" Program, Risk Assessment Guidelines. October
1993.
6. U.S. Environmental Protection Agency. Air Pollutant Emissions
from Electric Utility Steam Generating Units—Interim Final.
Volume II. Appendices A-G. EPA-453/R-96-013b. October 1996.
7. Johnson, W. B., Wolf, D. E. and Mancuso, R. L. Long Term
Regional Patterns and Transfrentier Exchanges of Airborne Sulfur
Pollution in Europe. Atmospheric Environment. Volume 12. 1978.
Pp. 511-527.
8. U.S. Environmental Protection Agency. ENAP-1 Long-Term SO2 and
Sulfate Pollution Model - Adaptation and Application to Eastern
North America. EPA 600/4-80-039. Environmental Sciences
Research Laboratory, Research Triangle Park, NC. May 1980.
9. U.S. Environmental Protection Agency. RELMAP: A Regional
Lagrangian Model of Air Pollution - User's Guide. (Eder et al.
1986) Atmospheric Sciences Research Laboratory, Research Triangle
Park, NC. March 1986.
10. Clark, T. L., Blakley, P., Mapp, G. Model Calculations of the
Annual Atmospheric Deposition of Toxic Metals to Lake Michigan.
Paper presented at the 85th Annual Meeting of the Air and Waste
Management Association. June 1992.
11. Sehmel, G. A. Particle and Gas Dry Deposition: A Review.
Atmospheric Environment. Volume 14. Pp. 983-1011. 1980.
6-74
-------
12. California Air Resource Board. Deposition Rate Calculations for
Air Toxics Source Assessments. Air Quality Modeling Section,
Technical Support Division, California. September 16, 1987.
13. Alcamo, J., Bartnicki, J., Olendrzynski, K. and Pacyna, J.
Computing Heavy Metals in Europe's Atmosphere - I. Model
Development and Testing. Atmospheric Environment. volume 26A,
No. 18. Pp. 3355-3369. 1992.
14. Chan, W. H., Tang, A. J. S., Chung, D. H. S., Lusis, M. A.
Concentration and Deposition of Trace Metals in Ontario - 1982.
Ontario Ministry of the Environment, Air Resources Branch.
Toronto, Ontario, Canada. Revised February 5, 1986.
15. U.S. Environmental Protection Agency. A Screening Analysis of
Ambient Monitoring Data for the Urban Area Source Program. EPA-
453/R-94-075. October 1994.
16. U.S. Environmental Protection Agency. Health Effects Notebook
for Hazardous Air Pollutants. EPA-456-d-94-1003. Air Risk
Information Support Center. Research Triangle Park, NC.
December 1994.
17. U.S. Environmental Protection Agency. Integrated Risk
Information System (IRIS) Database, Environmental Criteria and
Assessment Office, Cincinnati, OH. 1994.
18. Viren Jr., Silvers A. Unit risk estimates for airborne arsenic
exposure: an updated view based on recent data from two copper
smelters. Regulatory Toxicology and Pharmacology Volume 20.
1994 pp. 125-138.
19. Memorandum from Hugh McKinnon, M. D., to Seitz, John, EPA.
April 26, 1995. Advice on unit risk estimate for airborne
arsenic.
20. Government of Canada. Canadian Environmental Protection Act
Priority Substances List Assessment Report. Arsenic and Its
Compounds. 1993.
21. Handbook of Chemistry and Physics, 60th ed., Chemical Rubber
Company Press, Cleveland, OH. 1980.
22. U.S. Environmental Protection Agency. Health Effects Document
for Nickel and Mercury Compounds, Final Report. EPA/600/8-
83/012ff. September 1986.
23. International Agency for Research on Cancer. IARC Monographs on
the Evolution of Carcinogenic Risks to Humans: Chromium, Nickel
and Welding. Volume 49. Lyon, France. 1990.
6-75
-------
24. California Air Resources Board. Initial Statement of Reasons for
Rulemaking. Proposed Identification of Nickel as a Toxic Air
Contaminant. 1991.
25. American Conference of Governmental Hygienists. Threshold Limit
Values for Chemical Substances and Physical Agents and Biological
Exposure Indices. 1995.
26. Memorandum from Maxwell, William, EPA to the EPA Air Docket No.
A-92-55. March 8, 1996. Dioxin from Hot-side ESP units.
27. Letter and enclosure from Peck, Stephen C., Electric Power
Research Institute, to Maxwell, William H., EPA:ESD.
September 15, 1995. Transmittal of unlicensed electric utility
trace substances synthesis report.
28. National Research Council. Science and Judgment in Risk
Assessment. National Academy of Sciences, Washington, DC. 1994.
29. Office of Science and Technology Policy, Executive Office of the
President. Chemical Carcinogens: A Review of the Science and Its
Associated Principles. Washington, DC. 50 Federal Register
10372. March 14, 1985.
30. U.S. Environmental Protection Agency. Guidelines for
Developmental Toxicity Risk Assessment: Office of Health and
Environmental Assessment, 54 Federal Register 6398-63826, 1991.
31. U.S. Environmental Protection Agency. A Descriptive Guide to
Risk Assessment Methodologies for Toxic Air Pollutants. EPA-
453/R-93-038. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. 1993.
6-76
-------
7.0 MERCURY ASSESSMENT
7.1 OVERVIEW
Mercury is a highly persistent, naturally occurring metal in the
environment. Mercury is typically found in the environment in the
elemental state Hg(0). When it bonds to other chemical elements, it
is commonly found as a cation. The mercuric ion may bind to a number
of inorganic anions; these are generally referred to as species of
divalent mercury (Hg[II]). The mercuric ion may also form one or two
bonds with a methyl group forming either monomethyl- or
dimethylmercury.
The tendency of this metal to bioaccumulate in aquatic food webs
has been well documented.1"3 Mercury is toxic to humans from both the
inhalation4 and oral exposure routes.4"6 Mercury is also toxic to other
mammals7"10arid to birds.11"18 Questions remain regarding both the quantity
of mercury and the duration of the exposure required to elicit
responses in humans and animals, but it is widely accepted that
exposures to mercury produce neurotoxicity. Mercury contamination of
freshwater fish is a potential concern in the United States as
indicated by numerous fish advisories19 and mercury-related water
quality standards issued by State Agencies. The 1997 EPA Mercury
Study Report to Congress20 presents a more complete assessment of the
health effects, exposures, risks, ecological effects, sources, and
control technologies. This chapter presents an abbreviated assessment
of mercury as it is relevant to utilities, which is largely based on
information presented in EPA's Mercury Study Report to Congress.20
Fish consumption dominates the pathway for human and wildlife
exposure to methylmercury. The EPA's 1997 Mercury Study Report to
Congress supports a plausible link between anthropogenic releases of
mercury from industrial and combustion sources in the United States
and methylmercury in fish. However, these fish methylmercury
concentrations also result from existing background concentrations of
mercury (which may consist of mercury from natural sources, as well as
mercury which has been re-emitted from the oceans or soils) and
deposition from the global reservoir (which includes mercury emitted
by other countries). Given the current scientific understanding of
the environmental fate and transport of this element, it is not
possible to quantify how much of the methylmercury in fish consumed by
the U.S. population is contributed by U.S. emissions relative to other
sources of mercury (such as natural sources and re-emissions from the
global pool). As a result, it cannot be assumed that a change in
total mercury emissions will be linearly related to any resulting
change in methylmercury in fish, nor over what time period these
changes would occur. This is an area of ongoing study.
7.1.1 The Mercury Cycle
Environmental mercury passes through various environmental
compartments and may change physical form and chemical species during
7-1
-------
this process; these movements are conceptualized as a cycle. The
mercury cycle has been studied and described in several recent reports
and its understanding continues to undergo refinement.3'21"24
Given the present understanding of the mercury cycle, the flux of
mercury from the atmosphere to land or water at any one location is
comprised of contributions from: the natural global cycle; the global
cycle perturbed by human activities; regional sources; and local
sources. Recent advances allow for a general understanding of the
global mercury cycle and the impact of anthropogenic sources. It is
more difficult to make accurate generalizations of the fluxes on a
regional or local scale due to the site-specific nature of emission
and deposition processes.
7.1.1.1 The Global Mercury Cycle Past and Present. As a
naturally occurring element, mercury is present throughout the
environment in both environmental media and biota.25 In a 1979 report
edited by Nriagu, various authors estimated the global distribution of
mercury and concluded that by far the largest repository is ocean
sediments. Ocean sediments contain an estimated 1017 g of mercury,
mainly as HgS. According to estimates in the report edited by Nriagu,
ocean waters contain around 1013 g, soils and freshwater sediments
1013 g, the biosphere 1011 g (mostly in land biota) , the atmosphere
108 g, and freshwater contains on the order of 107 g. This budget
excludes "unavailable" mercury in mines and other subterranean
repositories. A more recent estimate of the global atmospheric
repository by Fitzgerald3 is 25 Mmol or approximately 5xl09 g; this is
50 times the previous estimate of Nriagu.25
Recent estimates of annual total global mercury emissions from
all sources (natural and anthropogenic) are about 5,000 to 5,500 tpy.26
Of this total, about 1,000 tpy are estimated to be natural emissions
and about 2,000 tpy are estimated to be contributions through the
natural global cycle of re-emissions of mercury associated with past
anthropogenic activity. Current anthropogenic emissions account for
the remaining 2,000 tpy. Point sources such as fuel combustion; waste
incineration; industrial processes (e.g., chlor-alkali plants); and
metal ore roasting, refining, and processing are the largest point
source categories on a world-wide basis. Given the global estimates
of 5,000 to 5,500 tpy (which are highly uncertain), U. S.
anthropogenic mercury emissions are estimated to account for roughly 3
percent of the global total, and U. S. utilities are estimated to
account for roughly 1 percent of total global emissions.
A number of different techniques have been used to estimate the
pre-industrial mercury concentrations in environmental media before
anthropogenic emissions contributed significantly to the global
mercury cycle. Figure 7-1 shows estimated current and preindustrial
budgets and fluxes. It is difficult to separate current mercury
concentrations by origin (i.e., anthropogenic or natural) because of
the continuous cycling of the element in the environment. For
7-2
-------
Estimated Current
Mercury Budgets
and Fluxes
Estimated Pre-Industrial
Mercury Budgets
and Fluxes
Anthropogenic
4,000
Hg(p) 98% Hg(gaseous) Hg(0)
2% Hg(particulate)
Local & Regional
Deposition
2,000
Global Terrestrial
Deposition
3,000
Global Marine
Deposition
2.000
Paniculate
Removal
200
All Fluxes in 101 Kg/y
All Pools in 10J Kg
Air
1,600
p) 98% Hg(gas«ous) Hg(0)
2% Hg(particulate)
Global Terrestrial
Deposition Global Marine
1,000 Deposition
600
Parti culate
Removal
60
All Fluxes in 10' Kg/y
All Pools in 101 Kg
Source: Adapted from Mason et al., 1994.
Figure 7-1. Comparison of Estimated Current and Pre-Industrial Mercury Budgets and Fluxes
7-3
-------
example, anthropogenic releases of elemental mercury may be oxidized
and deposit as divalent mercury far from the source; the deposited
mercury may be reduced and re-emitted as elemental mercury only to be
deposited again continents away. Not surprisingly, there is a broad
range of estimates and a great deal of uncertainty with each. When
the estimates are combined, they indicate that between 40 and 75
percent of the current atmospheric mercury concentrations are the
result of anthropogenic releases. The Expert Panel on Mercury
Atmospheric Processes27 concluded that pre-industrial atmospheric
concentrations constitute approximately one-third of the current
atmospheric concentrations. The panel estimated that anthropogenic
emissions may currently account for 50-75 percent of the total annual
input to the global atmosphere.27 The estimates of the panel are
corroborated by Lindqvist et al. ,28 who estimated that 60 percent of
the current atmospheric concentrations are the result of anthropogenic
emissions and Porcella,29 who estimated that this fraction was 50
percent. Horvat et al.30 assessed the anthropogenic fraction as
constituting 40 to 50 percent of the current total. This overall
range appears to be in agreement with the several-fold increase noted
in inferred deposition rates.31'32'33 The percentage of current total
atmospheric mercury which is of anthropogenic origin may be much
higher near mercury emissions sources.
A better understanding of the relative contribution of mercury
from anthropogenic sources is limited by substantial remaining
uncertainties regarding the level of natural emissions as well as the
amount and original source of mercury that is re-emitted to the
atmosphere from existing reservoirs. Recent estimates indicate that
of the approximately 200,000 tons of mercury emitted to the atmosphere
since 1890, about 95 percent resides in terrestrial soils, about 3
percent in the ocean surface waters, and 2 percent in the atmosphere.27
More study is needed before it is possible to accurately differentiate
natural fluxes from these reservoirs from re-emissions of mercury
originally released from anthropogenic sources. For instance,
approximately one-third of total current global mercury emissions are
thought to cycle from the oceans to the atmosphere and back again to
the oceans, but a major fraction of the emissions from oceans consists
of recycled anthropogenic mercury. It is believed that as little as
20 to 30 percent of the current oceanic emissions are from mercury
originally mobilized by natural sources.34 Similarly, a potentially
large fraction of terrestrial and vegetative emissions consists of
recycled mercury from previously deposited anthropogenic and natural
ions .27
Comparisons of contemporary (within the last 15-20 years)
measurements and historical records indicate that the total global
atmospheric mercury burden has increased since the beginning of the
industrialized period by a factor of between two and five.
Contamination from some anthropogenic processes that are no longer in
use produces continuing significant releases to surface water,
groundwater, and the atmosphere. It is estimated that the mercury
content of typical lakes and rivers has been increased by a factor of
7-4
-------
two to four since the onset of the industrial age.25 For example,
analysis of sediments from Swedish lakes shows mercury concentrations
in the upper layers that are two to five times higher than those
associated with pre-industrialized times. More recently, researchers
in Sweden estimated that mercury concentrations in soil, water and
lake sediments have increased by a factor of four to seven in southern
Sweden and two to three in northern Sweden in the 20th century.23 In
Minnesota and Wisconsin, an investigation of whole-lake mercury
accumulation indicates that the annual deposition of atmospheric
mercury has increased by a factor of three to four since pre-
industrial times. Similar increases have been noted in other studies
of lake and peat cores from this region; results from remote lakes in
southeast Alaska also show an increase, though somewhat lower than
found in the upper midwest United States.27
Although it is accepted that atmospheric mercury burdens have
increased substantially since the preindustrial period, it is
uncertain whether overall atmospheric mercury levels are currently
increasing, decreasing, or remaining stable. Measurements over remote
areas of the Atlantic Ocean show increasing levels up until 1990 and a
decrease for the period 1990-1994.35 At some locations in the upper
Midwest of the United States, measurements of deposition rates suggest
decreased deposition. However, other measurements at remote sites in
northern Canada and Alaska show deposition rates that continue to
increase.36'37 Since these sites are subject to global long-range
sources and few regional sources, these measurements may indicate a
still increasing global atmospheric burden. More research is
necessary; a multi-year, world-wide atmospheric mercury measurement
program may help to better determine current global trends.38
7.1.1.2 Regional and Local Mercury Cycles. According to one
estimate, roughly one half of the total anthropogenic mercury
emissions eventually enter the global atmospheric cycle,-39 the
remainder is removed through local or regional cycles. Mercury
emissions from utilities are believed to exist primarily in two forms,
divalent or elemental mercury. Divalent mercury, or Hg(II), is a
positive ion (missing two electrons) with a electric charge of plus 2
(i.e., Hg++, or oxidized mercury). Elemental mercury, or Hg(0), has a
neutral charge (i.e., Hg°) . An estimated 5 to 10 percent of primary
Hg(II) emissions are deposited within 100 km of the point of emission
and a larger fraction on a regional scale. Hg(0) that is emitted may
be removed on a local and regional scale to the extent that it is
oxidized to Hg(II). Some Hg(0) may also be taken up directly by
foliage; most Hg(0) that is not oxidized will undergo long-range
transport due to the insolubility of Hg(0) in water. In general,
primary Hg(II) emissions will be deposited on a local and regional
scale to the degree that wet deposition processes remove the soluble
Hg(II). Dry deposition may also account for some removal of
atmospheric Hg(II). Assuming constant emission rates, the quantity of
mercury deposited on a regional and local scale can vary depending on
source characteristics (especially the species of mercury emitted),
meteorological and topographical attributes, and other factors.27 For
7-5
-------
example, deposition rates at some locations have been correlated with
wind trajectories and precipitation amounts.40'41 Although these
variations prohibit generalizations of local and regional cycles, such
cycles may be established for specific locations. For example, unique
mercury cycles have been defined for Siberia on a regional scale42 and
for the area downwind of a German chlor-alkali plant on a local
scale.43 Mercury cycles dependent on local and regional sources have
also been established for the Upper Great Lakes region44'45 and the
Nordic countries.40
While the overall trend in the global mercury burden since pre-
industrial times appears to be increasing, there is some evidence that
mercury concentrations in the environment in certain locations have
been stable or decreasing over the past few decades. For example,
preliminary results for eastern red cedar growing near industrial
sources (chlor-alkali, nuclear weapons production) show peak mercury
concentrations in wood formed in the 1950s and 1960s, with stable or
decreasing concentrations in the past decade.27 Some results from peat
cores and lake sediment cores also suggest that peak mercury
deposition in some regions occurred prior to 1970 and may now be
decreasing.31'32'33'37 Data collected over 25 years from many locations in
the United Kingdom on liver mercury concentrations in two raptor
species and a fish-eating grey heron indicate that peak concentrations
occurred prior to 1970. The sharp decline in liver mercury
concentrations in the early 1970s suggests that local sources, such as
agricultural uses of fungicides, may have led to elevated mercury
levels two to three decades ago.46 Similar trends have been noted for
mercury levels in eggs of the common loon collected from New York and
New Hampshire.47 The downward trend in mercury concentrations observed
in the environment in some geographic locations over the last few
decades generally corresponds to regional mercury use and consumption
patterns over the same time frame (consumption patterns are discussed
in Volume II of the Mercury Study Report to Congress). 20
7.1.2 Atmospheric Processes
Basic processes involved in the atmospheric fate and transport of
mercury include: (1) emissions to the atmosphere; (2) transformation
and transport in the atmosphere; (3) deposition from the air; and then
(4) re-emission to the atmosphere. Each of these processes is briefly
described below.
7.1.2.1 Emissions of Mercury. As discussed fully in Volume II
of the Mercury Study Report to Congress, 20 mercury is emitted to the
atmosphere through both naturally occurring and anthropogenic
processes. Natural processes include volatilization of mercury in
marine and aquatic environments, volatilization from vegetation,
degassing of geologic materials (e.g., soils), and volcanic emissions.
The natural emissions are thought to be primarily in the elemental
mercury form. Conceptually, the current natural emissions can arise
from two components: mercury present as part of the pre-industrial
equilibrium and mercury mobilized from deep geologic deposits and
added to the global cycle by human activity.
7-6
-------
Anthropogenic mercury releases are thought to be dominated on the
national scale by industrial processes and combustion sources that
release mercury into the atmosphere. Available information indicates
that stack emissions include both gaseous and particulate forms of
mercury. Gaseous mercury, Hg(g), emissions are believed to include
both elemental and oxidized chemical forms, while particulate mercury,
Hg(p), emissions are thought to be composed primarily of oxidized
compounds due to the relatively high vapor pressure of Hg(0). The
analytic methods for mercury speciation of exit gasses and emission
plumes are being refined, and there is still controversy in this
field. Chemical reactions occurring in the emission plume are also
possible. Available information suggests that the speciation of
mercury emissions depend on the fuel used (e.g., coal, oil), flue gas
cleaning and operating temperature, and possibly other factors. The
exit stream is thought to range from almost all divalent mercury to
nearly all elemental mercury. Most of the mercury emitted at the
stack outlet is found in the gas phase although exit streams
containing soot can bind up some fraction of the mercury. The
divalent fraction is split between gaseous and particle bound phases.28
Much of this Hg(II) is believed to be mercuric chloride (HgCl2) . 48
An emission factor-based approach was used to develop the
nationwide emission estimates for the fossil fuel combustion
categories presented in Table 7-1. The emission factors presented are
estimates based on ratios of mass mercury emissions to measures of
source activities and nationwide source activity levels. The reader
should note that the data presented in this table are estimates;
uncertainties include the precision of measurement techniques and the
calculation of emission factors, estimates of pollutant control
efficiency, and nationwide source class activity levels. The
estimates may also be based on limited information for a particular
source class, thereby increasing the uncertainty in the estimate
further. Due to these and other uncertainties, other sources have
calculated different total emissions estimates using similar methods.49
7.1.2.2 Transformation and Transport of Atmospheric Mercury.
Hg(0) has an atmospheric residence time of about one year and will
thus be distributed fairly evenly in the troposphere. Oxidized
mercury may be deposited relatively quickly if it is precipitated out,
leading to a residence time of hours to months. Longer residence
times are possible as well; the atmospheric residence time for some
Hg(II) associated with fine particles may approach one year.49
The transformation of Hg(0)(g) to Hg(II)(aqueous) and Hg(II)(p)
in cloud water demonstrates a possible mechanism by which natural and
anthropogenic sources of Hg(0) to air can result in mercury deposition
to land and water. This deposition can occur far from the source due
to the slow rate of Hg(0)(g) uptake in cloud water. It has been
suggested that this mechanism is important in a global sense for Hg
pollution, while direct wet deposition of anthropogenic Hg(II) is the
most important locally.3'28 Gaseous Hg(II) is expected to deposit at a
faster rate after release than particulate Hg(II) assuming that most
7-7
-------
Table 7-1. Best Point Estimates of National Mercury Emission
Rates by Category
Sources of mercury"
Area sources
Lamp breakage
General laboratory use
Dental preparations
Landfills
Mobile sources
Paint use
Agricultural burning
Point Sources
Combustion sources
Utility boilers
Coal
Oil
Natural gas
MWCs"
Commercial/industrial boilers
Coal
Oil
MWIs"
Hazardous waste combustors"
Residential boilers
Oil
Coal
SSIs
Wood-fired boilers'
Crematories
Manufacturing sources
Chlor-alkali
Portland cement*
Pulp and paper manufacturing
Instruments manufacturing
Secondary Hg production
Electrical apparatus
Carbon black
Lime manufacturing
Primary lead
Primary copper
Fluorescent lamp recycling
Batteries
Primary Hg production
Mercury compounds
Byproduct coke
Refineries
Miscellaneous sources
Geothermal power
Turf products
Pigments, oil, etc.
TOTAL
1994-1 995 Mg/yrb
3.1
1.4
1.0
0.6
<0.1
c
c
c
140.9
125.2
46.8
(46.7)d
(0.2)
(<0.1)
26.9
25.8
(18.8)
(7.0)
14.6
6.4
3.3
(2.9)
(0.4)
0.9
0.2
<0.1
14.4
6.5
4.4
1.7
0.5
0.4
0.3
0.3
0.1
0.1
<0.1
<0.1
<0.1
c
c
c
c
1.3
1.3
g
g
144
1994-1 995 tons/yrb
3.4
1.5
1.1
0.7
<0.1
c
c
c
155.7
137.9
51.5
51.3
(0.2)
(<0.1)
29.6
28.4
(20.7)
(7.7)
16.0
7.1
3.6
(3.2)
(0.5)
1.0
0.2
<0.1
15.8
7.1
4.8
1.9
0.5
0.4
0.3
0.3
0.1
0.1
<0.1
<0.1
<0.1
c
c
c
c
1.4
1.4
g
g
158
% of Total inventory"
2.2
1.0
0.7
0.4
0.0
c
c
c
97.8
86.9
32.6
32.5
(0.1)
(0.0)
18.7
17.9
(13.1)
(4.9)
10.1
4.4
2.3
(2.0)
(0.3)
0.6
0.1
0.0
10.0
4.5
3.1
1.2
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
0.0
c
c
c
c
0.9
0.9
g
g
100
MWC = Municipal waste combustor; MWI = medical waste incinerator; SSI = sewage sludge incinerator.
Numbers do not add exactly because of rounding.
Insufficient information to estimate 1994-1995 emissions.
Parentheses denote subtotal within larger point source category.
For the purpose of this inventory, cement kilns that burn hazardous waste for fuel are counted as hazardous waste combustors.
Includes boilers only; does not include residential wood combustion (wood stoves).
Mercury has been phased out of use.
U.S. EPA has finalized emission guidelines for these source categories which will reduce mercury emissions by at least an additional
90 percent over 1995 levels.
7-8
-------
of the particulate matter is less than 1 urn in diameter. An
atmospheric residence time of. % - 2 years for elemental mercury
compared to as little as hours for some Hg(II) species is expected.50
This behavior is observed in the modeling results presented in this
effort as well. It is possible that dry deposition of Hg(0) can occur
from ozone mediated oxidation of elemental mercury taking place on wet
surfaces, but this is not expected to be comparable in magnitude to
the cloud droplet mediated processes.
This great disparity in atmospheric residence time between Hg(0)
and the other mercury species leads to very much larger scales of
transport and deposition for Hg(0). Generally, air emissions of Hg(0)
from anthropogenic sources, fluxes of Hg(0) from contaminated soils
and water bodies, and natural fluxes of Hg(0) all contribute to a
global atmospheric mercury reservoir with a holding time of % to 2
years. Global atmospheric circulation systems can take Hg(0)
emissions from their point of origin and carry them anywhere on the
globe before transformation and deposition occur. Emissions of all
other forms of mercury are likely to be deposited to the earth's
surface before they thoroughly dilute into the global atmosphere.
Continental-scale atmospheric modeling, such as that performed for
this study using the Regional Lagrangian Model of Air Pollution
(RELMAP), can explicitly simulate the atmospheric lifetime of gaseous
and particulate mercury species, but not Hg(0). Although Hg(0) is
included as a modeled species in the RELMAP analysis, the vast
majority of Hg(0) emitted in the simulation transports outside the
spatial model domain without depositing, and the same is generally
thought to happen in the real atmosphere. Natural Hg(0) emissions and
anthropogenic Hg(0) emissions from outside the model domain are
simulated in the form of a constant background Hg(0) concentration of
1.6 ng m~3, approximating conditions observed in remote oceanic
regions.3 This background Hg(0) concentration is subject to simulated
wet deposition by the same process as explicitly modeled anthropogenic
sources of Hg(0) within the model domain.
Explicit numerical models of global-scale atmospheric mercury
transport and deposition have not yet been developed. As the general
understanding of the global nature of atmospheric mercury pollution
develops, numerical global-scale atmospheric models will surely
follow.
7.1.2.3. Deposition of Mercury. The divalent species emitted,
either in the vapor or particulate phase, are thought to be subject to
much faster atmospheric removal than elemental mercury.51'52 Both
particulate and gaseous divalent mercury is assumed to dry deposit
(this is defined as deposition in the absence of precipitation) at
significant rates when and where measurable concentrations of these
mercury species exist. The deposition velocity of particulate mercury
is dependent on atmospheric conditions and particle size. Particulate
mercury is also assumed to be subject to wet deposition due to
scavenging by cloud microphysics and precipitation. The gaseous
divalent mercury emitted is also expected to be scavenged readily by
7-9
-------
precipitation. Divalent mercury species have much lower Henry's law
constants than elemental mercury, and thus are assumed to partition
strongly to the water phase. Dry deposition of gas phase divalent
mercury is thought to be significant due to its reactivity with
surface material. Overall, gas phase divalent mercury is more rapidly
and effectively removed by both dry and wet deposition than
particulate divalent mercury,51'52'53 a result of the reactivity and water
solubility of gaseous divalent mercury.
In contrast, elemental mercury vapor is not thought to be
susceptible to any major process of direct deposition to the earth's
surface due to its relatively high vapor pressure and low water
solubility. On non-assimilating surfaces elemental mercury deposition
appears negligible,51 and though elemental mercury can be formed in
soil and water due to the reduction of divalent mercury species by
various mechanisms, this elemental mercury is expected to volatilize
into the atmosphere.27 In fact, it has been suggested that in-situ
production and afflux of elemental mercury could provide a buffering
role in aqueous systems, as this would limit the amount of divalent
mercury available for methylation.3 Water does contain an amount of
dissolved gaseous elemental mercury, 54 but it is minor in comparison to
the dissolved-oxidized and particulate mercury content.
There appears to be a potential for deposition of elemental
mercury via plant-leaf uptake. Lindberg et al.51 indicated that forest
canopies could accumulate elemental mercury vapor via gas exchange at
the leaf surface followed by mercury assimilation in the leaf interior
during the daylight hours. This process causes a downward flux of
elemental mercury from the atmosphere, resulting in a deposition
velocity. Recent evidence55 indicates that this does occur but only
when air concentrations of elemental mercury are above an equilibrium
level for the local forest ecosystem. At lower air concentration
levels, the forest appears to act as a source of elemental mercury to
the atmosphere, with the measured mercury flux in the upward
direction. Lindberg et. al.5S noted this may be explained by the
volatilization of elemental mercury from the canopy/soil system, most
likely the soil. Hanson et al.55 stated that "dry foliar surfaces in
terrestrial forest landscapes may not be a net sink for atmospheric
elemental mercury, but rather as a dynamic exchange surface that can
function as a source or sink dependent on current mercury vapor
concentrations, leaf temperatures, surface condition (wet versus dry)
and level of atmospheric oxidants." Similarly, Mosbaek et al.57
convincingly showed that most of the mercury in leafy plants is due to
air-leaf transfer, but that for a given period of time the amount of
elemental mercury released from the plant-soil system greatly exceeds
the amount collected from the air by the plants. It is also likely
that many plant/soil systems accumulate airborne elemental mercury
when air concentrations are higher than the long-term average for the
particular location, and release elemental mercury when air
concentrations fall below the local long-term average. On regional
and global scales, dry deposition of Hg(0) does not appear to be a
7-10
-------
significant pathway for removal of atmospheric mercury, although
approximately 95 percent or more of atmospheric mercury is Hg(0).3
There is an indirect pathway, however, by which elemental mercury
vapor released into the atmosphere may be removed and deposited to the
earth's surface. Chemical reactions occur in the aqueous phase (cloud
droplets) that both oxidize elemental mercury to divalent mercury and
reduce the divalent mercury to elemental mercury. The most important
reactions in this aqueous reduction-oxidation balance are thought to
be oxidation of elemental mercury with ozone, reduction of divalent
mercury by sulfite (S03~2) ions, or complexation of divalent mercury
with soot to form particulate divalent mercury:
Hg(0)(g) -> Hg(0)(aq)
Hg(0) (aq) + 03(aq) -> Hg(II) (aq)
Hg(II)(aq) + soot/possible evaporation -> Hg(II)(p)
Hg(II)(aq) + SCV2(aq) -> Hg(0)(aq)
(g) = gas phase molecule
(aq) = aqueous phase molecule
(p) = particulate phase molecule
The Hg(II) produced from oxidation of Hg(0) by ozone can be reduced
back to Hg(0) by sulfite; however, the oxidation of Hg(0) by ozone is
a much faster reaction than the reduction of Hg(II) by sulfite. Thus,
a steady state concentration of Hg(II)(aq) is built up in the
atmosphere and can be expressed as a function of the concentrations of
Hg(0)(g), 03(g), H+ (representing acids) and S02 (g) .28 Note that H+ and
S02 (g), although not apparent in the listed atmospheric reactions,
control the formation of sulfite.
The Hg(II)(aq) produced would then be susceptible to atmospheric
removal via wet deposition. The third reaction, however, may
transform most of the Hg(II)(aq) into the particulate form, due to the
much greater amounts of soot than mercury in the atmosphere. The soot
concentration will not be limiting compared to the concentration of
Hg(II)(aq), and S atoms in the soot matrix will bond readily to the
Hg(II)(aq). The resulting Hg(II)(p) can then be removed from the
atmosphere by wet deposition (if the particle is still associated with
the cloud droplet) or dry deposition (following cloud droplet
evaporation). It is possible that dry deposition of Hg(0) can occur
from ozone mediated oxidation of elemental mercury taking place on wet
surfaces, but this is not expected to be comparable in magnitude to
the cloud droplet mediated processes.58
Mercury released into the atmosphere from natural and
anthropogenic sources deposits mainly as Hg(II), from either direct
deposition of emitted Hg(II) or from conversion of emitted elemental
Hg(0) to Hg(II) through ozone-mediated reduction. The former process
may result in elevated deposition rates around atmospheric emission
sources and the latter process results in regional/global transport
followed by deposition.
7-11
-------
There is still a great deal of uncertainty with respect to the
amount of dry deposition of mercury. Once deposited, mercury appears
to bind tightly to certain soil components. The deposited Hg(II) may
revolatilize through reduction and be released back to the atmosphere
as Hg(0). Soil Hg(II) may also be methylated to form methylmercury;
these two forms may remain in the soil or be transported through the
watershed to a water body via runoff and leaching. Mercury enters the
water body through direct deposition on the watershed, and mercury in
water bodies has been measured in both the water column and the
sediments. Hg(II) in the water body may also be methylated to form
methylmercury; both Hg(II) and methylmercury may be reduced to form
Hg(0) which is reintroduced to the atmosphere.
7.1.2.4 Re-emissions of Mercury into the Atmosphere. Re-
emission of deposited mercury results most significantly from the
evasion of elemental mercury from the oceans. In this process,
anthropogenically emitted mercury is deposited to the oceans as Hg(II)
and then reduced to volatile Hg(0) and re-emitted. According to one
estimate, this process accounts for approximately 30 percent (10
Mmol/year) of the total mercury flux to the atmosphere.39 Overall, 70
to 80 percent of total current mercury emissions may be related to
anthropogenic activities.34 By considering the current global mercury
budget and estimates of the preindustrial mercury fluxes, Mason et
al.39 estimate that total emissions have increased by a factor of 4.5
since preindustrial times, which has subsequently increased the
atmospheric and oceanic reservoirs by a factor of 3. The difference
is attributed to local deposition near anthropogenic sources.
Although the estimated residence time of elemental mercury in the
atmosphere is about 1 year, the equilibrium between the atmosphere and
ocean waters results in a longer time period needed for overall change
to take place for reservoir amounts. Therefore, by substantially
increasing the size of the oceanic mercury pool, anthropogenic sources
have introduced long- term perturbations into the global mercury
cycle. Fitzgerald and Mason34 estimate that if all anthropogenic
emissions were stopped, it would take about 15 years for mercury pools
in the oceans and the atmosphere to return to pre-industrial
conditions. The Science Advisory Board, in its review of the EPA's
Mercury Study, concluded that it could take significantly longer.
There is scientific agreement, however, that the slow release of
mercury from terrestrial sinks to freshwater and coastal waters will
persist for a long time, probably decades, which effectively increases
the length of time anthropogenic emissions would impact the
environment. This is particularly significant given that the surface
soils contain most of the pollution-derived mercury of the industrial
period. As a result, it is uncertain at this time how long it would
take after reductions in anthropogenic emissions for mercury levels in
the global environment, including fish levels, to return to true
background levels. The slow release of mercury from terrestrial sinks
to freshwater and coastal waters will likely persist for much longer,
possibly decades, effectively increasing the lifetime of anthropogenic
mercury further.34 This may be particularly significant considering
that surface soils currently contain most of the pollution-derived
7-12
-------
mercury of the industrial period. Thus, re-emissions of past
anthropogenic mercury emissions will contribute to long-term
influences on the global biogeochemical cycle for mercury.
7.1.3 Terrestrial and Aquatic Fate of Mercury
7.1.3.1 Mercury in Soil. Once deposited, the Hg(II) species are
subject to a wide array of chemical and biological reactions. Soil
conditions (e.g., pH, temperature and soil humic content) are
typically favorable for the formation of inorganic Hg(II) compounds
such as HgCl2, Hg(OH)2 and inorganic Hg(II) compounds complexed with
organic anions.59 Although inorganic Hg(II) compounds are quite
soluble (and, thus, theoretically mobile) they form complexes with
soil organic matter (mainly fulvic and humic acids) and mineral
colloids; the former is the dominating process. This is due largely
to the affinity of Hg(II) and its inorganic compounds for sulfur-
containing functional groups. This complexing behavior greatly limits
the mobility of mercury in soil. Much of the mercury in soil is bound
to bulk organic matter and is susceptible to elution in runoff only by
being attached to suspended soil or humus. Some Hg(II), however, will
be absorbed onto dissolvable organic ligands and other forms of
dissolved organic carbon (DOC) and may then partition to runoff in the
dissolved phase. Currently, the atmospheric input of mercury to soil
is thought to exceed greatly the amount leached from soil, and the
amount of mercury partitioning to runoff is considered to be a small
fraction of the amount of mercury stored in soil. The affinity of
mercury species for soil results in soil acting as a large reservoir
for anthropogenic mercury emissions.60'23 For example, note the mercury
budget proposed by Meili et al.so Even if anthropogenic emissions were
to stop entirely, leaching of mercury from soil would not be expected
to diminish for many years.23 Hg(0) can be formed in soil by reduction
of Hg(II) compounds/complexes mediated by humic substances.25 This
Hg(0) will vaporize eventually and re-enter the atmosphere.
Methylmercury can be formed by various microbial processes acting on
Hg(II) substances. Approximately 1-3 percent of the total mercury in
surface soil is methylmercury, and as is the case for Hg(II) species,
it will be bound largely to organic matter. The other 97-99 percent
of total soil mercury can be considered largely Hg(II) complexes,
although a small fraction of Hg in typical soil will be Hg(0).S1
7.1.3.2 Plant and Animal Uptake of Mercury. While there is a
great deal of uncertainty surrounding air-to-plant transfer of
mercury, some evidence indicates that this pathway may be an important
source of mercury to soils via defoliation. Overall, mercury
concentrations in plants, even those whose main uptake appears to be
from the air, are expected from modeling results to be low. This
prediction is corroborated by low reported mercury concentrations in
most green plants, although the data set of these values is not
complete and there are some exceptions. The bulk of the mercury in
plants appears to be inorganic.50 Livestock typically accumulate
little mercury from foraging or silage/grain consumption, and the
mercury content of meat is low. Due to these factors, the terrestrial
7-13
-------
pathway is not expected to be significant, particularly when compared
to the consumption of fish by humans. Since this is not an exposure
pathway of concern for mercury, it was not included in the modeling
that follows.
7.1.3.3 Mercury in the Freshwater Ecosystem. There are a number
of pathways by which mercury can enter the freshwater environment:
Hg(II) and methylmercury from atmospheric deposition (wet and dry) can
enter water bodies directly; Hg(II) and methylmercury can be
transported to water bodies in runoff (bound to suspended soil/humus
or attached to dissolved organic carbon); and Hg(II) and methylmercury
can leach into the water body from groundwater flow in the upper soil
layers. Once in the freshwater system, the same complexation and
transformation processes that occur to mercury species in soil will
occur in aquatic media along with additional processes due to the
aqueous environment. Mercury concentrations are typically reported
for particular segments of the water environment; the most common of
these are the water column (further partitioned as dissolved or
attached to suspended material), the underlying sediment (further
divided into surface sediments and deep sediments), and biota
(particularly fish).
Most of the mercury in the water column, Hg(II) and
methylmercury, will be bound to organic matter, either to dissolved
organic carbon62'28 or to suspended particulate matter. In most cases,
studies that refer to the dissolved mercury in water include mercury
complexes with DOC. Studies indicate that about 25-60 percent of
Hg(II) and methylmercury organic complexes are particle-bound in the
water column. The rest is in the dissolved, bound-to-DOC phase.25'63
Hg(0) is produced in fresh water by humic acid reduction of Hg(II) or
demethylation of methylmercury. Some will remain in the dissolved
gaseous state, but most will volatilize. As noted previously, Hg(0)
constitutes very little of the total mercury in the water column but
may provide a significant pathway for the evolution of mercury out of
the water body via Hg(II) or methylmercury -> Hg(0) -> volatilization.
For many lakes, however, sedimentation of the Hg(II) and methylmercury
bound to particulate matter is expected to be the dominant process for
removal of mercury from the water column.24
Generally, no more than 25 percent of the total mercury in a
water column exists as a methylmercury complex; typically, less than
10 percent is observed. This is a result of methylation of Hg(II)
which is thought to occur in the bottom sediment and the water column
by microbial action and abiotic processes. An equilibrium is soon
established between Hg(II) and methylmercury in freshwater systems; in
a number of sediment-water systems, it has been found that
methylmercury concentrations in waters were independent of water
column residence time or time in contact with sediments.64
Methylmercury in the water column which is lost through demethylation,
exported downstream, or taken up by biota is thought to be replaced by
additional methylation of Hg(II) compounds to sustain equilibrium.
7-14
-------
Once entering a water body, mercury can remain in the water
column, be lost from the lake through drainage water, revolatilize
into the atmosphere, settle into the sediment or be taken up by
aquatic biota. After entry, the movements of mercury through any
specific water body may be unique. Only mercury in the water column,
the sediment, and other aquatic biota appears to be available to
aquatic organisms for uptake.
Methylation appears to be a key step in the entrance of mercury
into the food chain.24 The biotransformation of inorganic mercury
species to methylated organic species in water bodies can occur in the
sediment and the water column. Abiotic processes (e.g., humic and
fulvic acids in solution) also appear to methylate the mercuric ion.
All mercury compounds entering an aquatic ecosystem are not
methylated, and demethylation reactions as well as volatilization of
dimethylmercury decrease the amount of methylmercury available in the
aquatic environment. It is clear that there is a large degree of
scientific uncertainty and variability among water bodies concerning
the processes that methylate mercury.24
Methylmercury is very bioavailable and accumulates in fish
through the aquatic food web; nearly 100 percent of the mercury found
in fish muscle tissue is methylated.24 Methylmercury appears to be
primarily passed to planktivorous and piscivorous fish via their
diets. Larger, longer-lived fish species at the upper end of the food
web typically have the highest concentrations of methylmercury in a
given water body. Most of the total methylmercury production ends up
in biota, particularly fish. Overall, methylmercury production and
accumulation in the freshwater ecosystem places this pollutant into a
position to be ingested by fish-eating organisms.
Methylmercury appears to be efficiently passed through the
aquatic food web to the highest trophic level consumers in the
community (e.g., piscivorous fish). At this point it can be contacted
by fish-consuming humans through ingestion. Methylmercury appears to
pass from the human gastrointestinal tract into the bloodstream more
efficiently than the divalent species.
7.1.3.4 Fate of Mercury in Marine Environments. As noted
earlier, mercury is an atmophillic element and, as such, its global
transport occurs primarily through the atmosphere. Elemental mercury,
the principle species found in the atmosphere, has a high vapor
pressure and a low solubility in water. As a result of these
properties, the half-life of atmospheric mercury is thought to be a
year or longer. Elemental mercury appears to be deposited to ocean
waters primarily through wet deposition. Oxidizing reactions in the
atmosphere may also play a role in the conversion of elemental mercury
to more reactive atmospheric species which are subsequently deposited.
Mercury found in ocean waters and sediments comprises a large
reservoir of the total mercury on the planet. The conceptualization
of oceans as reservoirs of mercury is fitting for they serve both as
7-15
-------
sources of mercury to the atmosphere and as environmental mercury
sinks. S5'ss'67 The forms and species of mercury present in the ocean
waters and sediments may be transformed as a result of both biotic and
abiotic factors within the ocean. The most significant species of
mercury from a human health perspective is monomethylmercury (MHg).
MHg shows strong evidence of bioaccumulation and biomagnification in
the marine food web, potentially posing risks to consumer species
(particularly apex marine predators and piscivores).24
7.2 MERCURY HEALTH EFFECTS
A brief summary of the health effects of methylmercury is
presented here. The 1997 EPA Mercury Study Report to Congress68
contains more information on the health effects of mercury and mercury
compounds.
Most of the population of the earth have some exposure to mercury
as a result of normal daily activities. The general population may be
exposed to mercury through inhalation of ambient air; consumption of
contaminated food, water, or soil; and/or dermal exposure to
substances containing mercury. In addition, some quantity of mercury
is released from dental amalgam.
The health effects literature contains many investigations of
populations with potentially high exposure to mercury, including
industrial workers, people living near point sources of mercury
emissions, people who consume large amounts of fish, and dental
professionals. There also are numerous studies of populations exposed
to high levels of mercury, such as the Minamata poisoning episode in
Japan. Volume IV of the EPA's Mercury Study Report to Congress69
presents measured and predicted mercury exposure for various U.S.
populations.
The form of mercury which is emphasized here is methylmercury
because methylmercury is the form of primary interest for human
exposures for this report. It is acknowledged that humans can be
exposed to elemental and inorganic mercury and that certain
populations can be exposed to many types of organic mercurials, such
as antiseptics and pesticides, which are not discussed here.
7.2.1 Toxicokinetics
The toxicokinetics (i.e., absorption, distribution, metabolism,
and excretion) of mercury is highly dependent on the form of mercury
to which a receptor has been exposed. Below is a brief summary of the
toxicokinetics information for methylmercury.
Methylmercury is rapidly and extensively absorbed through the
gastrointestinal tract. Absorption information following inhalation
exposures is limited. This form of mercury is distributed throughout
the body and easily penetrates the blood-brain and placental barriers
in humans and animals. Methylmercury transport into tissues appears
to be mediated by the formation of a methylmercury-cysteine complex.
7-16
-------
This complex is structurally similar to methionine and is transported
into cells via a widely distributed neutral amino acid carrier
protein. Methylmercury in the body is considered to be relatively
stable and is only slowly demethylated to form mercuric mercury in
rats. It is hypothesized that methylmercury metabolism may be related
to a latent or silent period observed in epidemiological studies
observed as a delay in the onset of specific adverse effects.
Methylmercury has a relatively long biological half-life in humans;
estimates range from 44 to 80 days. Excretion occurs via the feces,
breast milk, and urine.
7.2.2 Biological Effects
The primary targets for toxicity of mercury and mercury compounds
are the nervous system, kidney, and developing fetus. Other systems
that may be affected include the respiratory, cardiovascular,
gastrointestinal, hematologic, immune, and reproductive systems. A
brief summary of the biological effects of methylmercury is presented
here.
Three human studies that examined the relationship between
methylmercury and cancer incidence were considered extremely limited
because of study design inappropriate for risk assessment or
incomplete data reporting. Evidence from animal studies provides
limited evidence of carcinogenicity. Male ICR and B6C3F1 mice exposed
orally to methylmercuric chloride were observed to have an increased
incidence of renal adenomas, adenocarcinomas, and carcinomas. Renal
epithelial cell hyperplasia and tumors, however, were observed only in
the presence of profound nephrotoxicity suggesting that the tumors may
be a consequence of reparative changes to the damaged kidneys. Tumors
were observed at a single site, in a single species and sex.
Methylmercury appears to be clastogenic but not a potent mutagen.
Studies have also shown evidence that methylmercury may induce
mammalian germ cell chromosome aberrations. There are a number of
studies in both humans and experimental animals that show
methylmercury to be a developmental toxicant. Neurotoxicity in
offspring is the most commonly observed effect and the effect seen at
lowest exposures.
A significant body of human studies exists for evaluating the
potential systemic toxicity of methylmercury. This data base is the
result of studying two large scale poisoning episodes in Japan and
Iraq as well as several epidemiological studies assessing populations
that consume significant quantities of fish. In addition, much
research on the toxicity of methylmercury has been conducted in
animals including non-human primates.
The critical target for methylmercury toxicity is the nervous
system. The developing fetus may be at particular risk from
methylmercury exposure. Offspring born of women exposed to high doses
of methylmercury during pregnancy have exhibited a variety of
developmental neurological abnormalities, including the following:
7-17
-------
delayed onset of walking, delayed onset of talking, cerebral palsy,
altered muscle tone and deep tendon reflexes, and reduced neurological
test scores. Maternal toxicity may or may not have been present
during pregnancy for those offspring exhibiting adverse effects. For
the general population, the critical effects observed following
methylmercury exposure are multiple central nervous system effects
including ataxia and paresthesia.
A latent or silent period has been observed in some
epidemiological and animal studies indicating a delay in the onset of
adverse effects. It is hypothesized this delay may be related to
methylmercury metabolism.
7.2.3 Sensitive Subpopulations
A susceptible population is a group that may experience more
severe adverse effects at comparable exposure levels or adverse
effects at lower exposure levels than the general population. The
greater response of these sensitive subpopulations may be a result of
a variety of intrinsic or extrinsic factors. For mercury, the most
sensitive subpopulations may be developing organisms. Data are also
available indicating that other factors may be associated with the
identification of sensitive subpopulations including the following:
age; gender; dietary insufficiencies of zinc, glutathione, or
antioxidants; predisposition for autoimmune glomerulonephritis; and
predisposition for acrodynia.
7.2.4 Interactions
There are data demonstrating that a number of substances affect
the pharmacokinetics and/or toxicity of mercury compounds. Of most
interest is the potential interaction of selenium and mercury.
Selenium is known to bioaccumulate in fish, so exposure to
methylmercury from fish consumption may be associated with exposure to
increased levels of selenium. There are data indicating that selenium
co-administered with methylmercury can form selenium-methylmercury
complexes. The formation of these complexes may temporarily prevent
methylmercury-induced tissue damage but also may delay excretion of
the methylmercury. Thus, formation of selenium-methylmercury
complexes may not reduce methylmercury toxicity but rather may delay
onset of symptoms. More information is needed to understand the
possible interaction of selenium with methylmercury. There is also
potential for interaction between various forms of mercury and
ethanol, thiol compounds, tellurium, potassium dichromate, zinc,
atrazine, and vitamins C and E.
7.2.5 Hazard Identification/Dose-Response Assessment
The available toxicological and epidemiological evidence was
evaluated, and U.S. EPA risk assessment guidelines and methodologies
were applied to hazard identification for various endpoints; namely,
carcinogenicity, germ cell mutagenicity, developmental toxicity, and
7-18
-------
general systemic toxicity. Data supported quantitative assessments of
systemic toxicity. An oral reference dose (RfDa) was calculated for
methylmercury. U.S. EPA derived the RfD for methylmercury by
extrapolating from the high-dose exposures that occurred in the Iraq
incident. Data for carcinogenicity of inorganic and methylmercury
were judged to be inadequate in humans and limited from animal
bioassays. The carcinogenicity data for methylmercury were not
sufficient to support a quantitative assessment. Table 7-2 summarizes
the hazard identification and dose-response information for organic
mercury.
7.2.6 Ongoing Research
While much data has been collected on the potential toxicity of
mercury and mercury compounds, much is still unknown. Two ongoing
epidemiological studies are now providing critical information on the
developmental toxicity of methylmercury. One study, being conducted
in the Seychelles Islands, is evaluating dose-response relationships
in a human population with dietary exposures (fish) at levels believed
to be in the range of the threshold for developmental toxicity. The
second study, conducted in the Faroe Islands, is assessing mercury
exposure in a population that consumes a relatively large quantity of
marine fish and marine mammals. Children exposed to methylmercury in
utero and followed through 6 years of age have been assessed for
mercury exposure and neurological developmental. Because of various
limitations and uncertainties in all of the available data, the U.S.
EPA and other Federal agencies intend to participate in an interagency
review of the human data on methylmercury, including the most recent
studies from the Seychelle Islands and the Faroe Islands. The
purposes of this review are to refine the estimates of the level of
exposure to mercury associated with subtle neurological endpoints and
to further consensus among all of the Federal agencies. After this
process, the U.S. EPA will determine if a change in the RfD for
methylmercury is warranted.
7.2.7 Research Needs
Specifically, information is needed to reduce the uncertainties
associated with the current oral RfD for methylmercury. More work
with respect to both dose and duration of exposure would also allow
for potentially assessing effects above the RfD. Limited evidence
suggests that methylmercury is a possible human carcinogen. Research
on mode of action in induction of tumors at high doses will be of
particular use in defining the nature of the dose response
relationship for carcinogenicity.
The oral RfD is an estimate (with uncertainty spanning perhaps an order
of magnitude) of a daily exposure to the human population (including
sensitive subpopulations) that is likely to be without an appreciable
risk of deleterious health effects during a lifetime.
7-19
-------
Table 7-2. Summary of U.S. EPA Hazard Identification/Dose-
Response Assessment for Methylmercury
Form
of
mercury
Organic
Oral RfD
(mg/kg-day)
0.000-T
(methyl-
mercury)
Inhalation
RfC
(mg/m3)
n/a
Cancer
weight-of-
evidence
rating
C, possible
human
carcinogen
Cancer
slope
factor
n/a
Germ cell
mutagenicity
High weight of
evidence
Developmental
toxicity
data base
characterization
Sufficient human
and animal data
1 Critical effect is neurological toxicity in progeny of exposed women, RfD calculated using a benchmark dose (10%).
There are many uncertainties associated with the health effects
data analysis, due to an incomplete understanding of the toxicity
of methylmercury. The sources of uncertainty include the
following:
• The data serving as the basis for the methylmercury RfD were
from a population ingesting contaminated seed grain. The
nutritional status of this group may not be similar to that
of U.S. populations. The exposure was for a short, albeit
critical, period of time. It is likely that there is a
range of response among individuals to methylmercury
exposure. The selenium status of the exposed Iraqi
population is not certain, nor is it established the extent
to which selenium has an effect on mercury toxicity.
• There was no NOAEL (no-observable-adverse-effect level) for
estimation of a threshold for all developmental endpoints.
A benchmark was estimated using a Weibull model on grouped
data. Use of an estimate other than the 95 percent lower
limit on 10 percent response provides alternate estimates.
Other modeling approaches using data which have not been
grouped provide similar estimates. Benchmark doses, NOAELs,
and LOAELs from other human studies provide support for the
benchmark used in the RfD.
• Ingestion levels of methylmercury associated with measured
mercury in hair were estimated based on pharmacokinetic
parameters derived from evaluation of the extant literature.
Use of other plausible values for these parameters results
in (relatively small) changes in the exposure estimate.
• While there are data to show that the developing fetus is
more susceptible to methylmercury toxicity than adults,
there are not sufficient data to support calculation of a
separate RfD for children (vs. adults).
7-20
-------
To improve the risk assessment for methylmercury, U.S. EPA would
need the following:
• Results from ongoing studies in human populations with
measurable exposure to methylmercury, and new research on
actual consumption patterns and estimated methylmercury
exposure of the subpopulations of concern, with validation
by analysis of hair samples from a representative sample of
members of this subpopulation.
• Reproductive studies and analysis.
• Data on mode of action of methylmercury tumor induction.
• Validated physiologically-based pharmacokinetic models for
mercury which include a fetal component.
Based on the extant data and knowledge of developing studies, the
following outcome can be expected:
• Human populations exposed to sufficiently high levels of
methylmercury either in utero or post partum will have
increased incidence of neurotoxic effects.
7.3 MERCURY CONCENTRATIONS IN BIOTA
The Mercury Study Report to Congress documents many
concentrations in animals and plants.24 Concentrations in abiotic
environmental components consist primarily of inorganic species.
While these concentrations may be elevated in specific areas, fish
concentrations are generally of highest concern when assessing risks
posed by emitted mercury. The concern stems from the consumption of
fish by humans and the form of mercury, methylmercury, which fish
bioaccumulate. Methylmercury, which is the primary form of mercury
found in fish tissue, is a human neurotoxin and is readily absorbed
into the human body through the gastrointestinal tract. Fish
methylmercury concentrations result from existing environmental
concentrations of mercury (which may consist of mercury from
anthropogenic and natural sources, as well as mercury which has been
re-emitted from the oceans or soils) and deposition from the global
reservoir (which includes mercury emitted by other countries).
Given the current scientific understanding of the environmental
fate and transport of this element, it is not possible to quantify how
much of the methylmercury in fish consumed by the U.S. population is
contributed by U.S. emissions relative to other sources of mercury
(such as natural sources and re-emissions from the global pool). As a
result, it cannot be assumed that a change in total mercury emissions
will be linearly related to any resulting change in methylmercury in
fish, nor over what time period these changes would occur. This is an
area of ongoing study.
7-21
-------
7.4 MEASUREMENT DATA NEAR UTILITIES
The measured mercury concentrations in environmental media around
utilities are briefly summarized in this section. These data are not
derived from a comprehensive study of mercury around utilities.
Despite the need for this effort, such a study does not appear to
exist. The quality of the following studies has not been assessed by
the U.S. EPA. The data do not appear to be directly comparable among
themselves because of differences in analytic techniques and
collection methods used. Some of these studies are from older
literature and may not reflect current mercury emissions from the
sources described.
Anderson and Smith70 measured mercury levels in environmental media
and biota around a 200-MW coal-fired utility in Illinois. The
facility had two 152-m-high smokestacks and was equipped with an
electrostatic precipitator. Commercial operations at the facility had
been ongoing for 6 years when sampling was conducted (from 1973
through 1974). Levels of mercury detected in atmospheric particulate
samples collected 4.8 and 9.6 km downwind of the facility were not
statistically significantly elevated when compared with samples
collected 4.8 km upwind of the site. Mercury levels detected in
samples from the upper 2 cm of downwind agricultural soils (sample
mean 0.022 ppm mercury) were statistically significantly elevated when
compared with upwind samples (0.015 ppm mercury). Core sediment
sampling from a nearby lake bed showed statistically significant
elevations in sediment mercury concentrations after plant operations
began (sample mean 0.049 ppm mercury) when compared with sediment
deposits prior to operation (0.037 ppm mercury). No increases were
observed in mercury levels in fish from the nearby lake when compared
with fish from remote lakes.
Crockett and Kinnison71 sampled the arid soils around a 2,150-MW
coal-fired utility in New Mexico in 1974. The four-stack (two stacks
76 m high and two 91 m high) facility had been operational since 1963
and had an estimated mercury release rate of 850 kg/yr. The rainfall
in the area averaged 15 to 20 cm/yr. Although a mercury distribution
pattern was noted, soil mercury levels near the facility did not
differ significantly from background. Given the high amounts of
mercury released by the facility and the insignificant amounts
detected, the authors speculated that much of the mercury emitted was
transported over a larger area, rather than deposited locally.
Measurement data near other types of anthropogenic sources are
discussed in the 1997 EPA Mercury Study Report to Congress. 24
7.5 MODEL FRAMEWORK
This section describes the models and modeling scenarios used to
predict the environmental fate of mercury. Measured mercury
concentrations in environmental media were used when available to
parameterize these models. Human exposures to mercury were predicted
based on modeling results.
7-22
-------
7.5.1 Models Used
The extant measured mercury data alone were judged insufficient
for a national assessment of mercury exposure for humans from utility
units. Thus, the decision was made to model the mercury emissions.
In this study, there were three major types of modeling efforts:
(1) modeling of mercury atmospheric transport on a regional basis;
(2) modeling of mercury atmospheric transport on a local scale (within
50 km of source); and (3) modeling of mercury fate in soils and water
bodies into biota, as well as the resulting exposures to human fish
consumers. The models used are described in Table 7-3.
7.5.2 Modeling of Long-Range Fate and Transport of Mercury
The goal of this analysis was to model the emission, transport,
and fate of airborne mercury over the continental United States using
the meteorologic data for the year of 1989 and the most current
utility mercury emissions data. The results of the simulation were
intended to be used to answer a number of fundamental questions.
Probably the most general question was "How much mercury emitted by
utility boilers is deposited back to United States soils and water
bodies over a typical year?" It is known that year-to-year variations
in accumulated precipitation and wind flow patterns affect the
observed quantity of mercury deposited to the surface at any given
location. Meteorological data for the year 1989 was used since most
of the continental United States experienced near average weather
conditions during that year. To estimate the quantity of mercury
emitted by utilities that deposits in the United States, and
specifically which geographic regions may be more highly impacted,
information on chemical and physical forms of the mercury emissions
was needed since these characteristics determine the rate and location
of the wet and dry deposition processes for mercury.
The RELMAP model was used to predict the average annual
atmospheric mercury concentration and the wet and dry deposition flux
for each % degree longitude by Vz degree latitude grid cell
(approximately 40 km square) in the continental United States. The
emission, transport, and fate of airborne mercury over the continental
United States was modeled using meteorological data for the year 1989.
The utility emission data used were those presented in the Mercury
Study Report to Congress.20 Emission data are shown in Table 7-4.
The RELMAP model was originally developed to estimate
concentrations of sulfur and sulfur compounds in the atmosphere and
rainwater in the eastern United States. The primary modification of
RELMAP was the handling of three species of mercury (elemental,
divalent, and particulate) and carbon soot (or total carbon aerosol).24
A complete description of the RELMAP mercury model is presented in the
Mercury Study Report to Congress.2* The results of the RELMAP modeling
are shown in Figures 7-2 through 7-4.
7-23
-------
Table 7-3. Models Used to Predict Mercury Air Concentrations,
Deposition Fluxes, and Environmental Concentrations
Model
RELMAP
ISC3
IEM-2M
Description
Predicts average annual atmospheric mercury concentration and wet and dry
deposition flux for each 40 km2 grid in the U.S. due to all anthropocentric sources
mercury in the U.S.
Predicts annual average atmospheric concentrations and deposition fluxes within
km of mercury emission source
of
50
Predicts environmental mercury concentrations based on air concentrations and
deposition rates to watershed and water body. Predicts human exposure based on
these predicted concentrations and human activity patterns.
RELMAP = Regional Lagrangian Model of Air Pollution
IEM = Indirect exposure methodology
ISC = Industrial Source Complex
Table 7-4. Mercury Emissions Inventory Used in the RELMAP
Modeling (Based on the 1994-95 Estimates)
Mercury emission source type
Electric utility boilers (coal, oil and gas)
Emissions
(kg/yr)
46,183
Assumed speciation percentages
Hg(0) a
50
Hg2+b
30
HgPc
20
a Hg(0) represents elemental mercury gas
b Hg2+ represents divalent mercury gas
c Hgp represents particulate mercury
7-24
-------
0.05 to 1
1to2
2 to 5
5 to 10
> = 10
Figure 7-2. Total Modeled Mercury Deposits from Wet and Dry Deposition
from Coal Utilities Based on 1994 Emissions Estimates
as Modeled with RELMAP, Units: ug/m2/yr
-------
-J
ISJ
cr\
0.05 to 1
1to2
2 to 5
5 to 10
> = 10
Figure 7-3. Total Modeled Mercury Deposits from Wet and Dry Deposition
from Oil Utilities Based on 1994 Emissions Estimates
as Modeled with RELMAP, Units: ug/m2/yr
-------
-J
I
0.05 to 1
1to2
2 to 5
5 to 10
> = 10
Figure 7-4. Total Modeled Mercury Deposits from Wet and Dry Deposition
from Coal and Oil Utilities Based on 1994 Emissions Estimates
as Modeled with RELMAP, Units: ug/m2/yr
-------
A computer simulation of long-range transport of mercury
emissions from all United States sources conducted for the EPA's 1997
Mercury Study Report to Congress suggests that about one-third (~ 52
tons) of the 158 tpy of United States anthropogenic emissions are
deposited, through wet and dry deposition, within the lower 48 States.
The remaining two-thirds (~ 107 tons) is transported outside of United
States borders where it diffuses into the global reservoir. In
addition, the computer simulation suggests that another 35 tons of
mercury from the global reservoir is deposited for a total deposition
of roughly 87 tons. Although this type of modeling is uncertain, the
simulation suggests that about three times as much mercury is being
added to the global reservoir from United States sources as is being
deposited from it. What is not uncertain is that additional emissions
to air will contribute to levels in the global reservoir, and
concomitant deposition to water bodies.
Long-range transport modeling conducted as part of this Utility
Study predicts that approximately 30 percent (i.e., 15 tpy) of the
utility mercury emissions deposit in the continental United States.
The estimated annual deposition rates resulting from utility mercury
emissions range from 0.5 to greater than 10 //g per square meter.
7.5.3 Modeling the Local Transport of Mercury in the Atmosphere
The program used to model the transport of the anthropogenic
mercury within 50km of an emissions source was the Industrial Source
Complex Version 3 (ISC3) gas deposition model obtained from the United
States EPA's Support Center for Regulatory Air Models (SCRAM) website
(the program is called GDISCDFT). This model has a gas dry deposition
model that was applied in this study. The issues related to using
this program to model emitted mercury in the local atmosphere are
detailed in Volume III of the Mercury Study Report to Congress. 24
The phase and oxidation state of emitted mercury is thought to
be of critical importance in determining atmospheric fate. Only Hg(0)
and Hg+2 were considered in the air dispersion modeling. At the point
of stack emission and during atmospheric transport, the contaminant is
partitioned between two physical phases: vapor and particle-bound.
It was assumed that 25 percent of the divalent emissions from an
individual source would attach to particles in the plume; particle
sizes were assumed to reflect ambient particle data.
7.5.3.1 Development and Description of Model Plants. Model
plants representing four utility boilers were developed to represent a
range of mercury emissions from this source. Parameters for each
model plant were selected after evaluation of the characteristics of a
given source category and current knowledge of mercury emissions from
that source category. Important variables for the mercury risk
assessment included mercury emission rates, mercury speciation, and
mercury transport/deposition rates. Important model plant parameters
included stack height, stack diameter, stack volumetric flow rate,
stack gas temperature, plant capacity factor (relative average
operating hours per year), stack mercury concentration, and mercury
7-28
-------
speciation (see Table 7-5). Emission estimates were assumed to
represent typical emission levels emitted from existing sources.20'24
7.5.3.2. Hypothetical Locations of Model Plants. There are a
variety of geographic aspects that can influence the dispersion of
mercury emissions from a utility boiler. These aspects include
factors that affect the environmental chemistry of a pollutant and the
physics of plume dispersion. Environmental chemistry can include
factors such as the amount of wet deposition in a given area. Factors
affecting plume dispersion include terrain, wind direction and average
wind speed.
Because wet deposition may be an important factor leading to
mercury exposures, especially for the more soluble species emitted,
the meteorology of a location was used as a selection criterion. Two
different types of meteorology were deemed necessary to characterize
the environmental fate and transport of mercury: an arid/semi-arid
site and a humid site. The criterion specifically utilized was total
yearly rainfall. (See Volume III of the Mercury Study Report to
Congress for details . ) 24
Terrain features refer to the variability of the receptor height
with respect to a local source. Two main types of terrain were used
in the modeling: simple, and complex. Simple terrain is defined as a
study area that is relatively level and well below stack top (rather,
the effective stack height). Complex terrain refers to terrain that
is not simple, such as source located in a valley or a source located
near a hill. This included receptors that are above or below the top
of the stack of the source. Complex terrain can affect
concentrations, plume trajectory, and deposition. Due to the
complicated nature of plume flow in complex terrain, it is probably
not possible to predict impacts in complex terrain as accurately as
for simple terrain. In view of the wide range of uncertainty inherent
in accurately modeling the deposition of the mercury species
considered, the impacts posed by complex terrain were not incorporated
in the local scale analysis.
Two generic sites are considered: a humid site east of 90
degrees west longitude, and a more arid site west of 90 degrees west
longitude (these are described in Volume III of the Mercury Study
Report to Congress) .24 The primary differences between the two sites as
parameterized were the assumed erosion characteristics for the watershed
and the amount of dilution flow from the water body. The eastern site
had generally steeper terrain in the watershed than the other site. A
circular drainage lake with a diameter of 1.78 km and average depth of
5 m, with a 2 cm benthic sediment depth, was modeled at both sites.
The watershed area was 37.3 km2.
7.5.4 Modeling Mercury in a Watershed
Atmospheric mercury concentrations and deposition rates estimated
from RELMAP and ISC3 drive the calculations of mercury in watershed
soils and surface waters. The soil and water concentrations, in turn,
7-29
-------
Table 7-5. Process Parameters for Model Plants
Model plant
Large coal-fired
utility boiler
Medium coal-fired
utility boiler
Small coal-fired
utility boiler
Medium oil-fired
utility boiler
Plant size
975 Megawatts
375 Megawatts
100 Megawatts
285 Megawatts
Capacity
(% of year)
65%
65%
65%
65%
Stack
height (ft)
732
465
266
290
Stack
diameter (ft)
27
18
12
14
Hg emission
rate
(kg/yr)
230
90
10
2
Speciation
percent
(Hg(0)/Hg27Hgp)
50/30/20
50/30/20
50/30/20
50/30/20
Exit
velocity
(m/sec)
31.1
26.7
6.6
20.7
Exit
temperature
(°F)
273
275
295
322
Hg(0)
Hg2*
Hgp
elemental mercury;
divalent vapor phase mercury;
particle-bound mercury
drive calculations of concentrations in the associated biota and fish,
which humans are assumed to consume. The watershed model used for
this report, Indirect Exposure Methodology Version 2M (IEM-2M), was
adapted from the more general IEM-2 methodology30'31 to handle mercury
fate in soils and water bodies. It is described completely in the EPA
Mercury Study Report to Congress, Volume III.24
7.5.4.1 Overview of the Watershed Model. IEM-2M simulates three
chemical components: elemental mercury, Hg(0); divalent mercury,
Hgll; and methylmercury, MHg. In the previous version of IEM-2, these
components were assumed to be in a fixed ratio with each other as
specified by the fraction elemental (fJ and fraction methyl (f3) .
This version calculates the fractions in each component based on
specified or calculated rate constants.
IEM-2M is composed of two integrated modules that simulate
mercury fate using mass balance equations describing watershed soils
and a shallow lake, as illustrated in Figures 7-5 and 7-6. The mass
balances are performed for each mercury component, with internal
transformation rates linking Hg(0), Hgll, and MHg. Sources include
wetfall and dryfall loadings of each component to watershed soils and
to the water body. An additional source is diffusion of atmospheric
Hg(0) vapor to watershed soils and the water body. Sinks include
leaching of each component from watershed soils, burial of each
component from lake sediments, volatilization of Hg(0) and MeHg from
the soil and water column, and advection of each component out of the
lake.
At the core of IEM-2M are nine differential equations describing
the mass balance of each mercury component in the surficial soil
layer, in the water column, and in the surficial benthic sediments.
The equations are solved for a specified interval of time, and
predicted concentrations output at fixed intervals. For each
7-30
-------
Burial
Advection
Water column
benthic
transformation
Csoll
cw
^atm
Dyds
yds
yws
total mercury concentration in upper soil
total mercury concentration in water body
vapor phase mercury concentration in air
average dry deposition to watershed
average wet deposition to watershed
ng/g
ng/L
ng/m3
,ug/m2-yr
Mg/m2-yr
Figure 7-5. Overview of the IEM-2M Watershed Modules
7-31
-------
Local
Source
Center of lake ar /
2.5 km,
10km, or
25km
Prevailing Downwind Direction
Watershed
Figure 7-6. Configuration of Hypothetical Water Body and Watershed
Relative to Local Source
7-32
-------
calculational time step, IEM-2M first performs a terrestrial mass
balance to obtain mercury concentrations in watershed soils. Soil
concentrations are used along with vapor concentrations and deposition
rates to calculate concentrations in various food plants. These are
used, in turn, to calculate concentrations in animals. IEM-2M next
performs an aquatic mass balance driven by direct atmospheric
deposition along with runoff and erosion loads from watershed soils.
The nature of this methodology is basically steady with respect
to time and homogeneous with respect to space. While it tracks the
buildup of soil and water concentrations over the years given a steady
depositional load and long-term average hydrological behavior, it does
not respond to unsteady loading or meteorological events. There are
limitations on the analysis and interpretations imposed by these
simplifications. The model's calculations of average water body
concentrations are less reliable for unsteady environments, such as
streams, than for more steady environments, such as lakes.24 The
description includes a "benchmarking" exercise with an independent
model, the Mercury Cycling Model.
Mhg concentrations in fish are derived from dissolved MHg water
concentrations using bioaccumulation factors (BAF) . 24 Methylmercury
concentrations in fish were derived from predicted water column
concentrations of dissolved methylmercury by using BAFs for trophic
level 4 fish (Table 7-6). The BAFs selected for these calculations
were estimated from existing field data. The BAF (dissolved
methylmercury basis) for trophic level 4 fish is 1.6 x 10s.
Methylmercury was estimated to constitute 7.8 percent of the total
dissolved mercury in the water column, and 65 percent of this was
assumed to be freely dissolved. The potential variability around
these predicted fish residue values is highlighted in Table 7-6, which
shows percentile information for the BAF estimates.
There are several limitations to the modeling analyses. First,
there is a lack of adequate mercury measurement data near the
anthropogenic atmospheric mercury sources considered in this report.
To assess how well the modeled data predict actual mercury
concentrations in different environmental media at a variety of
geographic locations requires a database against which to make these
comparisons. The lack of such measured data preclude a comparison of
the modeling results with measured data around these sources. These
data include measured mercury deposition rates as well as measured
concentrations in the atmosphere, soils, water bodies and biota.
Substantial additional monitoring data would facilitate such
comparison. Second, the IEM-2M has not been validated with site-
specific data. The model was benchmarked against the independently-
derived Regional Mercury Cycling Model (R-MCM), which itself has been
calibrated to several Wisconsin lakes. When driven by the same
atmospheric loading and solids concentrations, IEM-2M predictions of
mercury concentrations compare well with those calculated by R-MCM for
a set of Wisconsin lakes. Additional limitations are discussed in
later sections below.
7-33
-------
Table 7-6. Percentlies of the Methylmercury Bioaccumulation
Factor
Parameter
Trophic 4 BAF
Percentile of distribution
5th
3.3x106
25th
5.0x1 06
50th
6.8x1 06
75th
9.2x1 06
95th
1.4x107
BAF = bioaccumulation factor
7.5.5 Exposure Modeling
7.5.5.1 Description of Hypothetical Human Exposure Scenarios.
Human exposure to environmental mercury is the result of mercury
concentrations at specific human exposure points (e.g., ingested
fish). For each location, mercury exposure was estimated only for
individuals representing several specific subpopulations that consumed
the freshwater fish that inhabited one of the three local lakes. The
individuals representing the subpopulations were defined to model both
average and high-end exposures.
The fish ingestion pathway was the only source of methylmercury
intake assessed. For this assessment, four human fish consumption
scenarios were considered for the hypothetical lakes: (1) an adult
subsistence fish consumer who was assumed to ingest large amounts of
locally-caught fish; (2) a child of a subsistence local fish consumer;
(3) a high-end recreational angler; and 4) an average local fish
consumer. These consumption scenarios were thought to represent
identified fish-consuming subpopulations in the United States.
Fish for human consumption from local water bodies can be derived
from many sources including self-caught, gifts, and grocery and
restaurant purchases. For the purposes of this study, all fish
consumed were assumed to originate from the hypothetical lakes, which
were considered to represent several small lakes that might be present
in the type of hypothetical locations considered. No commercial
distribution of locally caught fish was assumed; exposure to locally-
caught fish was modeled for the fish-consuming subpopulations
described above.
Fish consumption rates for the three fish-consuming
subpopulations were derived from the Columbia River Inter-Tribal Fish
Commission report72 and the draft EPA Exposure Factors Handbook.13
Other estimates of human fish consumption rates are reported in the
Exposure Factors Handbook.13 The estimates presented highlight the
broad variability in consumption rates. The Columbia River Inter-
Tribal Fish Commission report72 estimated fish consumption rates for
members of four tribes inhabiting the Columbia River Basin. The
estimated fish consumption rates were based on interviews with 513
adult tribe members who lived on or near the reservation. The
7-34
-------
participants had been selected from patient registrations lists
provided by the Indian Health Service. Adults interviewed provided
information on fish consumption for themselves and for 204 children
under five years of age.
Fish consumption rates for tribal members are shown in Tables 7-7
and 7-8. The values used in this study are shown in Table 7-9. The
values listed below reflect an annual average, but monthly variations
were also reported. For example, the average daily consumption rate
during the two highest intake months was 107.8 g/day, and the daily
consumption rate during the two lowest consumption months was
30.7 g/day. Fish were consumed by over 90 percent of the surveyed
population, with only 9 percent of the respondents reporting no fish
consumption. The maximum daily consumption rate for fish reported by
one member of this group was 972 g/day. Since most of the population
consisted of fish consumers, utilization of per capita estimates was
considered appropriate.
The Exposure Factors Handbook13 describes many freshwater fish
consumption studies. The mean daily consumption rates derived for
recreational freshwater anglers from the compiled studies range from
5-17 g/day; the derived 95th percentile range was 8-25 g/day. The
value of 30 g/day clearly exceeds the 95th percentile; this individual
is a high-end consumer. The recommended mean intake for subsistence
populations was 70 g/day, and the 95th percentile was 170 g/day. The
value of 60 g/day which is used for the subsistence adult is lower
than the recommended mean. The body weights used for the adult and
child were 70 and 17 kg.74
7.6 RESULTS
Tables 7-10 through 7-13 present the results of the multipathway
modeling analysis. The results are based on a model plant analysis
and are for hypothetical scenarios. Therefore, the results do not
apply to any specific utility plant and contain significant
uncertainties.
In all cases, the average air concentrations are predicted to be
dominated by the regional contribution of utilities rather than the
single local source modeled. This is largely due to the high
effective stack heights exhibited by the sources. The largest
contribution of 35 percent is for the medium coal-fired utility boiler
(MCUB) in the western site. The western site is predicted to have
lower concentrations of mercury as a result of regional transport.
At both the eastern and western sites using the 50th percentile
RELMAP results, the deposition rates, soil concentrations, and fish
concentrations are usually dominated by the local coal-fired utility
source within 10 km of the source. The small coal-fired utility
boiler (SCUB) at the eastern site is the exception due to the higher
deposition rate from regional sources. In the eastern site regional
sources dominate the deposition rates, soil concentrations, and fish
7-35
-------
Table 7-7. Fish Consumption Rates for Columbia River Tribes"
Subpopulation
Total adult population, aged 18 years and older
Children, aged 5 years and younger
Adult females
Adult males
Mean daily fish consumption (g/day)
59
20
56
63
Table 7-8. Daily Fish Consumption Rates Among Adults in the
Columbia River Tribes72
Percentile
50th
90th
95th
99th
g/day
29-32
97-130
170
389
Table 7-9. Fish Consumption Rates Used in This Study
Subpopulation
Subsistence adult
High-end child
Recreational angler
High-end recreational angler
Fish consumption
rate (g/day)a
60 a
20 a
8
30 a
1 Columbia River Inter-Tribal Commission, 1994.
7-36
-------
Table 7-10. Model Results for Eastern Site, RELMAP 50th Percentile (utilities only)
Facility
Large coal-fired utility
boiler
Medium coal-fired utility
boiler
Small coal-fired utility
boiler
Medium oil-fired utility
boiler
Distance
2.5km
10km
25km
2.5km
10km
25km
2.5km
10km
25km
2.5km
10km
25km
Air concentration
(ng/m3) % RELMAP %ISC3
0.026 93% 7%
0.026 91% 9%
0.026 92% 8%
0.027 89% 11%
0.028 88% 12%
0.027 90% 10%
0.028 87% 13%
0.027 91% 9%
0.025 96% 4%
0.024 99% 1%
0.024 99% 1%
0.024 99% 1%
Deposition (ug/m2/yr) % RELMAP %ISC3
17.9 13% 87%
5.27 44% 56%
3.4 69% 31%
9.12 26% 74%
4.17 56% 44%
3.19 73% 27%
3.94 59% 41%
2.93 80% 20%
2.54 92% 8%
2.53 93% 7%
2.41 97% 3%
2.37 99% 1%
Soil concentration
(ng/g) % RELMAP %ISC3
31 13% 87%
9.1 44% 56%
5.8 69% 31%
16 26% 74%
7.2 56% 44%
5.5 73% 27%
6.8 59% 41%
5 80% 20%
4.4 92% 8%
4.3 93% 7%
4.1 97% 3%
4.1 99% 1%
Tier 4 fish concentration
(ug/g) % RELMAP %ISC3
0.43 10% 90%
0.11 42% 58%
0.064 68% 32%
0.21 21% 79%
0.081 54% 46%
0.06 73% 27%
0.08 55% 45%
0.056 78% 22%
0.048 92% 8%
0.048 91% 9%
0.045 97% 3%
0.044 99% 1%
Table 7-11. Model Results for Western Site, RELMAP 50th Percentile (utilities only)
Facility
Large coal-fired utility
boiler
Medium coal-fired utility
boiler
Small coal-fired utility boiler
Medium oil-fired utility
boiler
Distance
2.5km
10 km
25km
2.5km
10 km
25km
2.5km
10km
25km
2.5km
10km
25km
Air concentration
(ng/m3) % RELMAP %ISC3
0.0061 87% 13%
0.0066 80% 20%
0.0073 73% 27%
0.007 76% 24%
0.0081 65% 35%
0.0076 69% 31%
0.0077 69% 31%
0.0067 79% 21%
0.006 89% 11%
0.0054 97% 3%
0.0054 97% 3%
0.0054 98% 2%
Deposition (ug/m2/yr) % RELMAP %ISC3
3.9 10% 90%
1.51 25% 75%
1.4 27% 73%
2.41 16% 84%
1.75 22% 78%
1.26 30% 70%
1.44 26% 74%
0.836 45% 55%
0.535 71% 29%
0.471 80% 20%
0.439 86% 14%
0.405 93% 7%
Soil concentration
(ng/g) % RelMap % ISC3
6.4 10% 90%
2.5 25% 75%
2.3 27% 73%
3.9 16% 84%
2.9 22% 78%
2.1 30% 70%
2.4 26% 74%
1.4 45% 55%
0.88 71% 29%
0.77 80% 20%
0.72 86% 14%
0.66 93% 7%
Tier 4 fish concentration
(ug/g) % RELMAP %ISC3
0.12 8% 92%
0.04 22% 78%
0.035 25% 75%
0.066 14% 86%
0.047 19% 81%
0.032 28% 72%
0.04 22% 78%
0.023 39% 61%
0.013 68% 32%
0.011 79% 21%
0.011 83% 17%
0.0097 93% 7%
-------
Table 7-12. Predicted Exposure Results for Eastern Site, RELMAP 50th Percentile (utilities only)
Facility
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Medium oil-fired utility boiler
Distance
2.5km
10km
25km
2.5km
10km
25km
2.5km
10km
25km
2.5km
10km
25km
MHg Exposure (mg/kg/day)
Child of Subsistence Average recreational
Subsistence Fisher Fisher Recreational Angler angler
3.7E-04 5.1E-04 1.8E-04 4.9E-05
9.0E-05 1.2E-04 4.5E-05 1.2E-05
5.5E-05 7.6E-05 2.8E-05 7.4E-06
1.8E-04 2.4E-04 8.9E-05 2.4E-05
7.0E-05 9.6E-05 3.5E-05 9.3E-06
5.2E-05 7.1E-05 2.6E-05 6.9E-06
6.9E-05 9.4E-05 3.4E-05 9.2E-06
4.8E-05 6.6E-05 2.4E-05 6.4E-06
4.1E-05 5.6E-05 2.0E-05 5.5E-06
4.1E-05 5.7E-05 2.1E-05 5.5E-06
3.9E-05 5.3E-05 1.9E-05 5.2E-06
3.8E-05 5.2E-05 1.9E-05 5.1E-06
% RELMAP %ISC3
10% 90%
42% 58%
68% 32%
21% 79%
54% 46%
73% 27%
55% 45%
78% 22%
92% 8%
91% 9%
97% 3%
99% 1%
-J
I
w
00
Table 7-13. Predicted Exposure Results for Western Site, RELMAP 50th Percentile (utilities only)
Facility
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Medium oil-fired utility boiler
Distance
2.5km
10km
25km
2.5km
10km
25km
2.5km
10km
25km
2.5km
10km
25km
MHg Exposure (mg/kg/day)
Child of subsistence Average recreational
Subsistence fisher fisher Recreational angler angler
1.0E-04 1.4E-04 5.1E-05 1.4E-05
3.5E-05 4.8E-05 1.7E-05 4.6E-06
3.0E-05 4.2E-05 1.5E-05 4.0E-06
5.7E-05 7.8E-05 2.8E-05 7.6E-06
4.0E-05 5.5E-05 2.0E-05 5.4E-06
2.8E-05 3.8E-05 1.4E-05 3.7E-06
3.5E-05 4.7E-05 1.7E-05 4.6E-06
2.0E-05 2.7E-05 9.9E-06 2.6E-06
1.1E-05 1.6E-05 5.7E-06 1.5E-06
9.7E-06 1.3E-05 4.9E-06 1.3E-06
9.3E-06 1.3E-05 4.6E-06 1.2E-06
8.3E-06 1.1E-05 4.2E-06 1.1E-06
% RELMAP % ISC3
8% 92%
22% 78%
25% 75%
14% 86%
19% 81%
28% 72%
22% 78%
39% 61%
68% 32%
79% 21%
83% 17%
93% 7%
-------
concentrations at 25 km from all four sources considered. In the
western site at 25 km the local coal combustion source emissions still
dominate at 25 km except for the SCUB. The deposition rates at both
sites are dominated by the regional sources when compared to the
medium oil-fired utility model plant.
The contribution of the local source is identical for the
deposition and soil concentrations, but this is not true for the fish
concentration. This is because the surface water receives input of
mercury from both direct deposition and from erosion/runoff from the
watershed. The water body is assumed to lie at the end of the
watershed closest to the facility, and so the contribution of the
local source to the deposition rate to the water body is generally
larger than that for the watershed. This results in a slightly higher
contribution of the local source to the predicted fish concentrations.
The multipathway exposure modeling analysis presented in this
chapter contains substantial uncertainties and is based on model
plants and hypothetical scenarios. Therefore, the results do not
apply to any existing utility plant. The analysis and results are
useful for gaining a better qualitative understanding of the potential
environmental fate of mercury emissions from a model utility plant.
However, the quantitative results are uncertain. Further research and
analyses are needed to gain a more complete understanding of the
mercury exposures due to utility emissions.
Based on the model plant, multipathway exposure modeling analysis
of hypothetical scenarios, the daily average methylmercury exposure of
the average hypothetical recreational angler (8 g fish consumed per
day) is not predicted to exceed the RfD of 1 x 10 ~4 mg/kg/day under any
combination of source, site, and distance. The daily average
methylmercury exposure of the high-end hypothetical recreational
angler (30 g fish consumed per day) is predicted to exceed the RfD in
the eastern site at 2.5 km from the large coal-fired utility boiler
(LCUB). The daily average methylmercury exposure of the hypothetical
subsistence angler (60 g fish consumed per day) is predicted to exceed
the RfD in the eastern site at 2.5 km from the LCUB and at 2.5 km from
the MCUB. The daily average methylmercury exposure of the
hypothetical subsistence angler is predicted to exceed the RfD in the
western site at 2.5 km from the LCUB. Fish consumption by children is
predicted to exceed the RfD for several hypothetical general cases: at
2.5 km from the LCUB at either site and at 2.5 km from the MCUB at the
eastern site. Background exposures were not considered in this
analysis. If background exposures due to other anthropogenic and
natural sources were considered, this would obviously result in higher
predicted exposures. Total exposures and background exposures are
discussed and analyzed in the 1997 EPA Mercury Study Report to
Congress.211'26'69
7-39
-------
7.7 CONCLUSIONS
Long-Range Transport Analysis
Based on modeling analysis of the wet and dry deposition of
utility air emissions of all forms of mercury within the continental
United States, the Agency finds that the following geographic areas
have the highest annual rate of total deposition of mercury in all
forms (above the 90th percentile level):
• Southeastern Great Lakes and Ohio River Valley
• Most of central and western Pennsylvania
• The urban corridor from Washington, DC, to New York City.
• In the vicinity of many of the larger cities in the eastern
United States and in numerous isolated locations where
relatively large coal-fired utilities are located.
Areas Predicted to be Least Impacted by Atmospheric Deposition of
Mercury from Utilities
Based on modeling analysis of the wet and dry deposition of
utility emissions of all forms of mercury within the continental
United States, the Agency predicts that the following geographic areas
have the lowest annual rate of total deposition of mercury in all
forms (below the 10th percentile level):
• Most of the Pacific Coast and Great Basin regions
• Parts of the northern Rocky Mountain region.
The three principal factors that contribute to these modeled
deposition patterns are:
• the emission source locations,
• the amount of divalent and particulate mercury emitted or
formed in the atmosphere, and
• climate and meteorology.
A facility located in a humid climate is predicted to have a higher
annual rate of mercury deposition than a facility located in an arid
climate. The critical variables within the model are:
• the estimated washout ratios of elemental and divalent
mercury, and
• the annual amount of precipitation.
7-40
-------
Precipitation is important because it removes various forms of mercury
from the atmosphere and deposits them to the surface of the earth.
The chemical form of emitted mercury is a critical factor in its
fate, transport, and toxicity in the environment. The form
distributions, or speciation factors, define the estimated fraction of
mercury emitted as elemental mercury (HgO) , divalent mercury (Hg2+) , or
mercury associated with particulates (Hgp). A wide variety of
alternate speciation scenarios have been investigated to measure the
sensitivity of the RELMAP results to this uncertainty. 74'75 The results
show that the total simulated wet and dry deposition of mercury to the
continental United States is strongly and positively correlated to the
fraction of mercury emitted as Hg2+ and Hgp for all major source types.
The differences between the results for the eastern and western
sites are due primarily to the differences in the frequency and
intensity of precipitation. At the eastern site, precipitation occurs
about 12 percent of the year, with about 5 percent of this
precipitation of moderate intensity (0.11 to 0.30 inches per hour).
By comparison, at the western site, precipitation occurs about 3
percent of the year, with about 2 percent of the precipitation of
moderate intensity.
Assessment of Watershed Fate
The atmospheric mercury concentrations and deposition rates
estimated using the RELMAP and ISC3 were then used as inputs in the
watershed model, IEM-2M, to derive calculations of mercury in
watershed soils and surface waters. The soil and water
concentrations, in turn, drive calculations of concentrations in the
associated biota and fish, which humans are assumed to consume.
IEM-2M Model Sensitivity
For a specific atmospheric deposition rate, mercury
concentrations in watersheds and water bodies can vary significantly.
Several intrinsic and extrinsic watershed and water body
characteristics influence the mercury concentrations in soil, water,
and fish. These should cause significant variability in mercury
concentrations between regions and among individual lakes within a
region.
Mercury concentrations in watershed soils are strongly influenced
by atmospheric loading and soil loss processes. The influence of
plant canopy and roots in mediating both the loading to the soil and
the loss from the soil is not well characterized at present, although
published studies indicate its potential importance. Reduction of
Hg(II) in the upper soil layer appears to control the volatile loss of
mercury, and variations in this reaction can cause significant
variations in soil mercury levels. The factors controlling mercury
reduction are not well characterized at present. Soil erosion from a
watershed can vary more than 3 orders of magnitude depending on
7-41
-------
rainfall patterns, soil type, topography, and plant cover. High
levels of soil erosion should significantly diminish soil mercury
concentrations. Runoff and leaching are not expected to affect soil
mercury concentrations significantly.
Total mercury concentrations in a water body are strongly
influenced by atmospheric loading and, for drainage lakes, by
watershed loading. Variations in watershed size and erosion rates can
cause significant variability in lake mercury levels. Hydraulic
residence time, the water body volume divided by total flow, affects
the maximum possible level of total water column mercury for a given
loading rate. Parameters controlling mercury loss through
volatilization and net settling can also cause significant variations
among lakes. Mercury loss through settling is affected by in-situ
productivity, by the supply of solids from the watershed, and by the
solids-water partition coefficient. DOC concentrations can
significantly affect partitioning, and thus overall mercury levels.
Mercury loss through volatilization is controlled by the reduction
rate, which is a function of sunlight and water clarity. Reduction
may also be controlled by pH, with lower values inhibiting this
reaction and leading to higher total mercury levels.
Fish mercury levels are strongly influenced by the same factors
that control total mercury levels. In addition, fish concentrations
are sensitive to methylation and demethylation in the water column and
sediments. A set of water body characteristics appear to affect these
reactions, including DOC, sediment TOC, sunlight, and water clarity.
Variations in these properties can cause significant variations in
fish concentrations among lakes. Other factors not examined here,
such as anoxia and sulfate concentrations, can stimulate methylation
and lead to elevated fish concentrations. Fish mercury levels are
sensitive to factors that promote methylmercury mobility from the
sediments to the water column; these factors include sediment DOC and
sediment-pore water partition coefficients.
Limitations of the Local Scale and Watershed Analyses
There are limitations associated with the fate and transport
analyses. These have to do to a large degree with the current state-
of-the-science concerning mercury fate and transport in the
terrestrial and aquatic environments and variability between
waterbodies. A few important limitations were discussed previously.
Additional important limitations are discussed below.
• There is a lack of information characterizing the movement
of mercury from watershed soils to water bodies and the
rates at which mercury converts from one chemical species to
another. There appears to be a great deal of variability in
these factors among watersheds.
7-42
-------
• There are not conclusive data on the amount of and rates of
mercury methylation in different types of water bodies. In
addition, there is a lack of data on the transfer of mercury
between environmental compartments and biologic
compartments; for example, the link between the amount of
mercury in the water body and the levels in fish appears to
vary from water body to water body.
Conclusions Regarding Mercury Fate and Transport in the Environment
The uncertainty inherent in the modeled estimates in this study
arises from many individual assumptions present within the three
models. Because of these uncertainties, EPA interpreted the model
results qualitatively rather than quantitatively as follows.
The analyses conducted for this study as well as for the EPA's
Mercury Study and available scientific knowledge indicate that human
activities today are adding to the mercury reservoirs that already
exist in land, water, and air, both naturally and as a result of
previous human activities.
The analysis of mercury fate and transport conducted for this
study, in conjunction with available scientific knowledge, supports a
plausible link between mercury emissions from utility combustion
sources and mercury concentrations in air, soil, water, and sediments.
The critical variables contributing to this linkage are:
• the species of mercury that are emitted from the sources,
with HgO mostly contributing to concentrations in ambient
air and Hg2+ mostly contributing to concentrations in soil,
water and sediments;
• the overall amount of mercury emitted from a combustion
source;
• the watershed soil loss rates, including reduction and
erosion;
• the water body loss rates, including outflow, reduction, and
settling; and
• the climate conditions.
In addition, this study also supports a plausible link between
mercury emissions from utility combustion sources and methylmercury
concentrations in freshwater fish. The critical variables
contributing to this linkage are:
• the species of mercury that are emitted, with emitted
divalent mercury mostly depositing into local watershed
areas and, to a lesser extent the atmospheric conversion of
7-43
-------
elemental mercury to divalent species which are deposited
over greater distances;
• the overall amount of mercury emitted from a source;
• the watershed soil loss rates, including reduction and
erosion;
• the water body loss rates, including outflow, reduction, and
settling;
• the extent of mercury methylation in the water body;
• the extent of food web bioaccumulation in the water body;
and
• the climate conditions.
From the analysis of deposition and on a comparative basis, the
deposition of Hg2+ close to an emission source is greater for receptors
in elevated terrain (i.e., terrain above the elevation of the stack
base) than from receptors located in flat terrain (i.e., terrain below
the elevation of the stack base). The critical variables are
parameters that influence the plume height, primarily the stack height
and stack exit gas velocity.
On a national scale, an apportionment between sources of mercury
and mercury in environmental media and biota cannot be described in
quantitative terms with the current scientific understanding of the
environmental fate and transport of this pollutant.
Human Exposure
The only exposure route considered was the consumption of
freshwater fish. Consumption of fish is the dominant pathway of
exposure to methylmercury for fish-consuming humans. There is a great
deal of variability among individuals in these populations with
respect to fish consumption rates. As a result, there is a great deal
of variability in exposure to methylmercury in these populations.
While EPA interprets these models qualitatively, some freshwater fish-
consuming individuals are predicted to exceed the RfD as a result of
mercury emissions from the sources considered. Measuring
methylmercury concentrations in fish from these waters and more direct
measures of exposure (e.g., hair or blood data) to humans consuming
fish around these sources should be a research priority.
It is important to note that the utility contribution is only one
component of the total amount of methylmercury in fish. Other
anthropogenic sources, natural sources and the existing background are
expected to influence fish methylmercury levels.
7-44
-------
7.8 DISCUSSION OF POTENTIAL CONCERNS OF MERCURY EMISSIONS FROM
UTILITIES
Mercury is considered the highest priority for multipathway
analyses because it is an environmentally persistent, toxic element.
Mercury is deposited to soil and terrestrial vegetation through
terrestrial exposure pathways, but at levels that do not result in
human exposures likely to be detrimental to health. However, in its
methylated form mercury bioaccumulates in the food web (especially the
aquatic food web). Modeling results suggest that most of the mercury
emitted to the atmosphere is deposited more than 50 km away from the
source, especially sources that have tall stacks. As stated above,
the modeling assessment from the Mercury Study, in conjunction with
available scientific knowledge, supports a plausible link between
anthropogenic mercury emissions and mercury found in freshwater fish.
Additional emissions to air will contribute to levels in the global
reservoir and deposition to water bodies. As a result, mercury
emissions from utility units may add to the existing environmental
burden.
At this time, the available information, on balance, indicates
that utility mercury emissions are of sufficient potential concern for
public health to merit further research and monitoring. The EPA
recognizes that there are substantial uncertainties that make it
difficult to quantify the magnitude of the risks due to utility
mercury emissions, and that further research and/or evaluation would
be needed to reduce these uncertainties. Remaining questions include
the following: (1) what is the quantitative relationship between a
change in United States mercury emissions and the resulting change in
methylmercury levels in fish; (2) what are the actual consumption
patterns and estimated methylmercury exposures of the subpopulations
of concern; (3) what are the actual mercury levels in a statistically
valid and representative sample of the U.S. population and susceptible
subpopulations; (4) what exposure levels are likely to result in
adverse health effects; (5) what affects the formation of
methylmercury in waterbodies and its bioaccumulation in fish; (6) how
much mercury is emitted from natural sources and past anthropogenic
sources; and (7) how much mercury is removed during coal cleaning and
other ongoing practices for pollution control. New data that could
reduce some of the uncertainties are likely to become available in the
next several years, and EPA plans to review and consider these data,
as appropriate, in future decisions.
7-45
-------
7.9 REFERENCES
1. Phillips, G. R., T. E. Lenhart, and R. W. Gregory. Relation
between trophic position and mercury accumulation among fishes
from the Tongue River, Montana. Environ. Res. Volume 22. 1980.
pp. 73-80.
2. Wren, C. D., H. R. MacCrimmon, and B. R. Loescher. Examination
of bioaccumulation and biomagnification of metals in a
precambrian shield lake. Water, Air, Soil Pollut. Volume 19.
1983. pp. 277-291.
3. Fitzgerald, W. F. Global biogeochemical cycling of mercury.
Presented at the DOE/FDA/EPA Workshop on Methylmercury and Human
Health, Bethesda, MD. March 22-23, 1994.
4. U.S. Environmental Protection Agency. Integrated Risk
Information System (IRIS) Database. Environmental Criteria and
Assessment Office, Cincinnati, OH. 1994.
5. Tsubaki, T. and K. Irukayama. Minamata Disease. Methylmercury
Poisoning in Minamata and Niigata, Japan. Kodansha, Ltd and
Elsevier Scientific Publishing Company, Amsterdam. 1977.
6. Bakir, F., S. F. Damluji, L. Amin-Zaki, M. Murtadha, A. Khalidi,
N.Y. Al-Rawi, S. Tikriti, H. I. Dhahir, T. W. Clarkson, J. C.
Smith and R. A. Doherty. Methylmercury poisoning in Iraq.
Science. Volume 181. 1973. pp. 230-241.
7. O'Connor, D. J. and S. W. Nielsen. Environmental survey of
methylmercury levels in wild mink (Mustela vison) and otter
(Lutra canadensis) from the northeastern United States and
experimental pathology of methylmercurialism in the otter.
Worldwide Furbearer Conference Proceedings, 1981.
pp. 1728-1745.
8. Borst, H. A. and C. G. Lieshout. Phenylmercuric acetate
intoxication in mink. Tijdschr. Diergeneesk. Volume 102. 1977.
pp. 495-503.
9. Wobeser, G., N. D. Nielsen and B. Schiefer. Mercury and mink I:
The use of mercury contaminated fish as a food for ranch mink.
Can. J. Comp. Med. Volume 40. 1976. pp. 30-33.
10. Wobeser, G., N. D. Nielsen and B. Schiefer. Mercury and mink II:
Experimental methylmercury intoxication. Can. J. Comp. Med.
Volume 40. 1976. pp. 34-45.
11. Fimreite, N. Effects of methylmercury treated feed on the
mortality and growth of leghorn cockerels. Can. J. Anim. Sci.
Volume 50. 1970. pp. 387-389.
7-46
-------
12. Fimreite, N. Effects of methylmercury on ring-necked pheasants.
Canadian Wildlife Service Occasional Paper Number 9. Department
of the Environment. 1971. p. 39.
13. Fimreite, N. Accumulation and effects of mercury on birds. In:
The Biogeochemistry of Mercury in the Environment. Edited by
J. 0. Nriagu. Elsevier, Amsterdam, The Netherlands. 1979.
pp. 601-628.
14. Heinz, G. H. Effects of low dietary levels of methylmercury on
mallard reproduction. Bull. Environ. Contam. Toxicol.
Volume 11. 1974. pp. 386-392.
15. Heinz, G. H. Effects of methylmercury on approach and avoidance
behavior of mallard ducklings. Bull. Environ. Contam. Toxicol.
Volume 13. 1975. pp. 554-564.
16. Heinz, G. H. Methylmercury: Second-year feeding effects on
mallard reproduction and duckling behavior. J. Wildl. Manage.
Volume 40. 1976. pp. 82-90.
17. Heinz, G. H. Methylmercury: Second-generation reproductive and
behavioral effects on mallard ducks. J. Wildl. Manage. Volume 40.
1976. pp. 710-715.
18. Heinz, G. H. Methylmercury: Reproductive and behavioral effects
on three generations of mallard ducks. J. Wildl. Mgmt. 1979.
pp. 394-401.
19. U.S. Environmental Protection Agency. U.S. Fish Advisory
Bulletin Board. Fish Contamination Section, Office of Science
and Technology, Office of Water. 1994.
20. U.S. Environmental Protection Agency. Mercury Study Report to
Congress. Volume II. An Inventory of Anthropogenic Mercury
Emissions in the United States. Office of Air Quality Planning
and Standards and Office of Research and Development. 1997.
21. Mitra, S. Mercury in the Ecosystem. Trans Tech Publications
Ltd. Switzerland. 1986.
22. Fitzgerald, W. F. and T. W. Clarkson. Mercury and
monomethylmercury: Present and future concerns. Envir. Health
Perspec. Volume 96. 1991. pp. 159-166.
23. Swedish Environmental Protection Agency. Mercury in the
Environment: Problems and Remedial Measures in Sweden.
ISBN 91-620-1105-7. 1991.
7-47
-------
24. U.S. Environmental Protection Agency. Mercury Study Report to
Congress. Volume III Fate and Transport of Mercury in the
Environment. Office of Air Quality Planning and Standards and
Office of Research and Development. 1997.
25. Nriagu, J. 0. The Biogeochemistry of Mercury in the Environment.
Elsevier/North Holland. Biomedical Press: New York. 1979.
26. U.S. Environmental Protection Agency. Mercury Study Report to
Congress. Volume I. Office of Air Quality Planning and
Standards and Office of Research and Development. 1997.
27. Expert Panel on Mercury Atmospheric Processes. Mercury
Atmospheric Processes: A Synthesis Report. Report No. TR-
104214. 1994.
28. Lindqvist, 0., K. Johansson, M. Aastrup, A. Andersson, L.
Bringmark, G. Hovsenius, L. Hakanson, A. Iverfeldt, M. Meili, and
B. Timm. Mercury in the Swedish environment - recent research on
causes, consequences and corrective methods. Water, Air and Soil
Poll. Volume 55. (all chapters).
29. Porcella, D. B. Mercury in the environment: Biogeochemistry.
In: Mercury Pollution Integration and Synthesis. Edited by
Watras, C.J. and J.W. Huckabee. 1994.
30. Hovart, M., L. Liang, N. S. Bloom. Comparison of distillation
with other current isolation methods for the determination of
methylmercury compounds in low level environmental samples.
Part II: Water. Analytica Chimica Acta. Volume 282. 1993.
pp. 153-168.
31. Swain, E. B., D. A. Engstrom, M. E. Brigham, T. A. Henning, and
P. L. Brezonik. Increasing rates of atmospheric mercury
deposition in midcontinental North America. Science. Volume 257.
1992. pp. 784-787.
32. Engstrom D. R., E. B. Swain, T. A. Henning, M. E. Brigham and
P. L. Brezonick. Atmospheric mercury deposition to lakes and
watersheds: A quantitative reconstruction from multiple sediment
cores. In: Environmental Chemistry of Lakes and Reservoirs.
Edited by L. A. Baker. American Chemical Society. 1994.
pp. 33-66.
33. Benoit, J. M., W. F. Fitzgerald, and A. W. H. Damman. Historical
atmospheric mercury deposition in the mid-continental U.S. as
recorded in an ombrotrophic peat bog. In: Mercury Pollution
Integration and Synthesis. Edited by Watras, C. J. and J. W.
Huckabee 1994. pp. 187-202.
7-48
-------
34. Fitzgerald, W. F., and R. P. Mason. The global mercury cycle:
oceanic and anthropogenic aspects. In Baeyens, W., R. Ebinghaus,
and 0. Vasiliev, eds., Global and Regional Mercury Cycles:
Sources, Fluxes and Mass Balances. 1996. Pp. 85-108.
35. Slemr, F. Trends in atmospheric mercury concentrations over the
Atlantic Ocean and at the Wank Summit, and the resulting
constraints on the budget of atmospheric mercury. In Baeyens,
W., R. Ebinghaus, and 0. Vasiliev, eds., Global and Regional
Mercury Cycles: Sources, Fluxes and Mass Balances. 1996.
Pp. 33-84.
36. Lucotte, M., A. Mucci, C. Hillaire-Marcel, P. Pichet, and A.
Grondin. Anthropogenic mercury enrichment in remote lakes of
Northern Quebec (Canada). Water, Air, and Soil Pollution
Volume 80. 1995. pp. 467-476.
37. Engstrom, D. R., and E. B. Swain. Recent Declines in Atmospheric
Mercury Deposition in the Upper Midwest. Environmental Science
and Technology. Volume 31. 1997. pp. 960-967.
38. Fitzgerald, W. F. Is mercury increasing in the atmosphere? The
need for an atmospheric mercury network (AMNET). Water, Air, and
Soil Pollution. Volume 80. 1995. pp. 245-254.
39. Mason, R. P., W. F. Fitzgerald and F.M.M. Morel. The
biogeochemical cycling of elemental mercury: Anthropogenic
influences. Geochimica et Cosmochimicia Acta. Volume 58,
No. 15. 1994. pp. 3191-3198.
40. Jensen, A., and A. Iverfeldt. Atmospheric bulk deposition of
mercury to the southern Baltic Sea area. In Watras, C. J., and
J. W. Huckabee eds. Mercury Pollution: Integration and
Synthesis. Pp. 221-229.
41. Dvonch, J. T., A. F. Vette, G. J. Keeler, G. Evans and
R. Stevens. An intensive multi-site pilot study investigating
atmospheric mercury in Broward County, Florida. Water, Air, and
Soil Pollution. Volume 80. 1995. pp. 169-178.
42. Sukhenko, S. A., and 0. F. Vasiliev. A regional mercury budget
for Siberia and the role of the region in global cycling of the
metal. In Baeyens, W., R. Ebinghaus, and 0. Vasiliev, eds.,
Global and Regional Mercury Cycles: Sources, Fluxes and Mass
Balances. Pp. 123-133.
43. Ebinghaus, R., and 0. Kruger. Emission and local deposition
estimates of atmospheric mercury in North-Western and Central
Europe. In Baeyens, W., R. Ebinghaus, and 0. Vasiliev, eds.,
Global and Regional Mercury Cycles: Sources, Fluxes and Mass
Balances. Pp. 135-159.
7-49
-------
44. Glass, G., J. Sorensen, K. Schmidt, G. Rapp, D. Yap, and
D. Eraser. Mercury deposition and sources for the upper Great
Lakes region. Water, Air and Soil Pollution. Volume 56. 1991.
pp. 235-249.
45. Lamborg, C. H., M. E. Hoyer, G. J. Keeler, I. Olmez, and X.
Huang. Particulate-phase mercury in the atmosphere:
collection/analysis method development and applications. In:
Mercury Pollution Integration and Synthesis. Edited by Watras,
C. J. and J. W. Huckabee. 1994. pp. 251-259.
46. Newton, I., I. Wyllie, and A. Asher. Long-term trends in
organochlorine and mercury residues in some predatory birds in
Britain. Environ. Pollut. Volume 79. 1993. pp. 143-151.
47. Mclntyre, J. W., J. T. Hickey, K. Karwowski, C. L. Holmquist, and
K. Carr. In: Proceedings of the 1992 Conference on the Loon and
its Ecosystem Status. Edited by L. Morse and M. Pokras. U.S.
Fish and Wildlife Service, Concord, NH. 1993. pp. 73-91.
48. Michigan Environmental Science Board. Mercury in Michigan's
Environment: Environmental and Human Health Concerns. Report to
Gov. John Engler. 1993.
49. Porcella, D. B., P. Chu, and M. A. Allan. Inventory of North
American Hg emissions to the atmosphere: Relationship to the
global mercury cycle. In Baeyens, W., R. Ebinghaus, and 0.
Vasiliev, eds., Global and Regional Mercury Cycles: Sources,
Fluxes and Mass Balances. Pp. 179-190.
50. Lindqvist, 0. and H. Rode. Atmospheric mercury-a review. Tellus.
Volume 37B. 1985. pp. 136-159.
51. Lindberg, S. E., T. P. Meyers, G. E. Taylor, R. R. Turner, and
W. H. Schroeder. Atmosphere-surface exchange of mercury to a
forest: Results of modeling and gradient approaches. J. of
Geophy. Res. Volume 97, No. D2. 1992. Pp. 2519-2528.
52. Shannon, J. D., and E. C. Voldner. Modeling; atmospheric
concentrations and deposition of mercury to the Great Lakes.
Presented at the DOE/FDA/EPA Workshop on Methylmercury and Human
Health, Bethesda, MD. March 22-23, 1994.
53. Petersen, G., Iverfeldt, A. and Munthe, J. Atmospheric mercury
species over Central and Northern Europe. Model calculations and
comparison with observations from the Nordic Air and
Precipitation Network for 1987 and 1988. Atmos. Env. Volume 29.
1995. pp. 47-68.
7-50
-------
54. Fitzgerald, W. F., R. P. Mason and G. M. Vandal. Atmospheric
cycling and air-water exchange of mercury over mid-continental
lacustrine regions. Water, Air and Soil Poll. Volume 56. 1991.
pp. 745-767.
55. Hanson, P. J., S. E. Lindberg, K. H. Kim, J. G. Owens and T. A.
Tabberer. Air/surface exchange of mercury vapor in the forest
canopy, I. Laboratory studies of foliar Hg vapor exchange.
Presented at the 3rd International Conference on Mercury as a
Global Pollutant. Whistler, BC, Canada. July 10-14, 1994.
56. Lindberg, S. E., R. R. Turner, T. P. Meyers, G. E. Taylor, and
W. H. Schroeder. Atmospheric concentrations and deposition of Hg
to a deciduous forest at Walker Branch Watershed, Tennessee, USA.
Water, Air and Soil Poll. Volume 56. 1991. pp. 577-594.
57. Mosbaek, H., J. C. Tjell, and T. Sevel. Plant uptake of mercury
in background areas. Chemosphere. Volume 17, No. 6. 1988.
pp. 1227-1236.
58. Lindberg, S. E. Forests and the global biogeochemical cycle of
mercury: the importance of understanding air/vegetation exchange
processes. W. Baeyans et al. (Eds.), Global and Regional Mercury
Cycles: Sources, Fluxes, and Mass Balances, pp. 359-380.
59. Schuster, E. The behavior of mercury in the soil with special
emphasis on complexation and adsorption processes- a review of
the literature. Water, Air and Soil Poll. Volume 56. 1991.
pp. 667-680.
60. Meili, M., A. Iverfeldt and L. Hakanson. Mercury in the surface
water of Swedish forest lakes - Concentrations, speciation and
controlling factors. Water, Air, and Soil Pollution. Volume 56.
1991. pp. 439-453.
61. Revis, N. W., T. R. Osborne, G. Holdsworth, and C. Hadden.
Mercury in soil: A method for assessing acceptable limits. Arch.
Environ. Contam. Toxicol. Volume 19. 1990. pp. 221-226.
62. Lindqvist, 0., K. Johansson, M. Aastrup, A. Andersson,
L. Bringmark, G. Hovsenius, L. Hakanson, A. Iverfeldt, M. Meili,
and B. Timm. Mercury in the Swedish environment - recent
research on causes, consequences and corrective methods. Water,
Air and Soil Poll. Volume 55. (Chapter 4). 1991.
63. Bloom, N. S., C. J. Watras, and J. P. Hurley. Impact of
acidification on the methylmercury cycle of remote seepage lakes.
Water, Air and Soil Poll. Volume 56. 1991. pp. 477-491.
7-51
-------
64. Parks, J. W., A. Lutz, and J. A. Sutton. Water column
methylmercury in the Wabigoon/English River-Lake system: Factors
controlling concentrations, speciation, and net production. Can.
J. Fisher. Ag. Sci. Volume 46. 1989. pp. 2184-2202.
65. Mason, R., and W. Fitzgerald. The distribution and biogeochemical
cycling of mercury in the equatorial Pacific Ocean. Deep Sea
Research. Volume 40, No. 9. 1993. pp. 1897-1924.
66. Mason, R., and W. Fitzgerald. Sources, sinks and biogeochemical
cycling of mercury in the ocean. In: Global and Regional Mercury
Cycles: Sources, Fluxes and Mass Balances. Edited by: W. Baeyens,
R. Ebinghaus, and O.Vasiliev. Kluwer Academic Publishers,
(Netherlands). pp. 249-272.
67. Cossa, D. M. Coquery, C. Gobeil, and J. Martin. Mercury fluxes
at the ocean margins. In: Global and Regional Mercury Cycles:
Sources, Fluxes and Mass Balances. Edited by W. Baeyens,
R. Ebinghaus, and 0. Vasiliev. Kluwer Academic Publishers,
Netherlands. 1996. pp. 229-248.
68. U.S. Environmental Protection Agency. Mercury Study Report to
Congress. Volume V. Health Effects of Mercury and Mercury
Compounds. EPA-452/R-97-007. Office of Air Quality Planning and
Standards and Office of Research and Development.
69. U.S. Environmental Protection Agency. Mercury Study Report to
Congress. Volume IV. An Assessment of Exposure to Mercury in
the United States. EPA-452/R-97-006. Office of Air Quality
Planning and Standards and Office of Research and Development.
1997.
70. Anderson, W. L. and K. E. Smith. Dynamic of mercury at
coal-fired utility power plant and adjacent cooling lake.
Environ. Sci and Technol. Volume 11. 1977. p. 75.
71. Crockett, A. and R. Kinnison. Mercury residues in soil around a
coal-fired power-plant. Envir. Sci. Technol. Volume 13. 1979.
pp. 712-715.
72. Columbia River Inter-Tribal Fish Commission. A Fish Consumption
Survey of the Umatilla, Nez Perce, Yakima and Warm Springs Tribes
of the Columbia River Basin. Technical Report 94-3. October
1994.
73. U.S. Environmental Protection Agency. Exposure Factors Handbook,
Volume 2 of 3. Food Ingestion Factors. SAB Review draft. 1996.
7-52
-------
74. Bullock, Jr., 0. R., W. G. Benjey and M. H. Keating. Modeling of
regional scale atmospheric mercury transport and deposition using
RELMAP. In: Atmospheric Deposition of Contaminants to the Great
Lakes and Coastal Waters. Edited by J. E. Baker. 1997a.
pp. 323-247.
75. Bullock, Jr., 0. R., K. A. Brehme and G. R. Mapp. Lagrangian
modeling of mercury air emission, transport and deposition: An
analysis of model sensitivity to emissions uncertainty. Special
Issue on Mercury as a Global Pollutant: Science of the Total
Environment, in press. 1997b.
7-53
-------
8.0 QUALITATIVE MULTIPATHWAY ASSESSMENT FOR LEAD AND CADMIUM
8.1 BACKGROUND
A multipathway exposure analysis, which is an assessment of
multiple routes of exposure of humans and/or biota to pollutants, is
the appropriate approach for a complete human health risk assessment.
Though it would have been preferable to perform a quantitative
multipathway exposure assessment of utility emissions of all six
priority HAPs (radionuclides, mercury, arsenic, dioxins, cadmium, and
lead), at the time of this study not enough data had been collected or
were available to do so. However, multipathway assessments were
performed for radionuclides, mercury, arsenic, and dioxins. For the
other two HAPs, lead and cadmium, only qualitative assessments of the
potential concerns to human health from noninhalation exposure were
performed.
The completion of quantitative assessments of inhalation
exposures for all HAPs and of multipathway exposures for only
radionuclides, mercury, arsenic, and dioxins does not reflect the
significance of noninhalation exposure and risks from utility HAP
emissions. Rather, it is a reflection of the complexity of assessing
multipathway exposure to a HAP. Due to the intensive data
requirements of a quantitative multipathway exposure assessment of a
HAP, and the limited chemical-specific data available (e.g., chemical-
specific air to plant biotransfer factor, chemical-specific plant
uptake rates) for input into the exposure model, quantitative analyses
were not completed for lead or cadmium. Though it is important to
address the noninhalation exposure pathways for select HAPs, there are
complex issues associated with the analysis of all exposure pathways.
The EPA recognizes that, for some of the utility HAPs
(e.g., mercury, dioxins), noninhalation exposure is a potentially more
significant route of exposure than inhalation exposure. The mercury
assessment suggests that there is a need for further analysis of
noninhalation exposures.
Efforts are underway to collect the chemical-specific data needed
for quantitative multipathway exposure assessment, and further
analyses may be undertaken for additional HAPs in the future, in
recognition of the need to understand the contribution of air
pollutants to risk from noninhalation exposure pathways. For this
report, a qualitative discussion of the potential concerns about
noninhalation exposure to lead and cadmium is presented in the
following sections.
8.2 LEAD COMPOUNDS
Lead compounds are persistent in the environment and have the
tendency to bioaccumulate (see Table 5-8). Lead is known to be toxic
by both the oral and inhalation routes. For these reasons,
5-1
-------
noninhalation exposure to lead emissions from utilities are a
potential concern.
For 1990, the estimated lead emissions from all coal-, oil-, and
gas-fired units were 72 ton/yr, 10 ton/yr, and 0.44 ton/yr, respectively.
To put these emissions estimates into perspective, it was estimated that
a total of 7.2 x 103 metric tons per year of lead were emitted into the
atmosphere from anthropogenic point and nonpoint sources during 1989. 1
When this estimate is compared to the 1990 lead emissions estimate for
utilities, it appears that utilities are responsible for emitting
approximately 1 percent of the total amount of lead emitted annually. In
1989, the major contributors of atmospheric lead included industrial
processes (2.3 x 103 metric tons/year), solid waste management (2.3 x 103
metric tons/year), transportation (2.2 x 103 metric tons/year), and fuel
combustion (0.5 x 103 metric tons/year).3
Lead is a naturally occurring metal that can be found in small
amounts in the earth's crust. However, the primary source of lead
found in the environment is anthropogenic activities. Once emitted to
the atmosphere from stack sources, such as utilities, lead can be
deposited onto environmental media such as soil, water, and
vegetation. In the atmosphere, lead exists primarily as PM in the
form of lead sulfate (PbS04) and lead carbonate (PbC03) . However, it
is not clear how the chemical composition of lead changes during
dispersion. Because a typical residence time of particulate lead is
10 days, long-range transport up to thousands of miles can occur. For
example, lead has been found in sediment cores of lakes in Canada that
were not located near any point sources of lead, indicating that
long-range atmospheric transport may have been occurring. The primary
removal mechanism from the atmosphere is wet or dry deposition onto
soil, water, and plants. Atmospheric deposition is the primary source
of lead found in soils.2 There is evidence that lead enters soil as
lead sulfate or quickly converts to lead sulfate at the soil surface.
Once deposited onto soil, lead tends to sorb strongly to soils and
becomes extremely persistent. Because lead is strongly sorbed to
soils, leaching of significant amounts of lead to groundwater or
surface water is not likely to occur. With the exception of highly
acidic environmental conditions, leaching of lead into groundwater and
surface water occurs very slowly. The major contributors to surface
water lead are atmospheric deposition and urban runoff. Typically in
the aquatic environment, lead is associated with suspended solids and
sediments. The concentration of dissolved lead in water is low
because lead tends to form compounds with low water solubilities that
precipitate out of the water column. The ratio of suspended lead to
dissolved lead is found to vary from 4:1 in rural streams to 27:1 in
urban streams. Many terrestrial plants accumulate lead by root uptake
from soil or by absorption of airborne lead deposited onto plants.
Industrial processes include nonferrous smelters, battery plants, and
chemical plants.
-------
However, the bioavailability of lead to plants from soil is limited
due to the strong absorption of lead to soil.
The highest background levels of lead are found in soils (<10 to
30 Atg/g) and in sediments (the average concentration of lead in river
sediments is 20,000 A«j/g).- In 1988, the average ambient air
concentration for 139 sites monitored by the National Air Surveillance
Network (NASN) was determined to be 0.085 /ug/m3.i This value is well
below the NAAQS of 1 . 5
Concentrations of lead found in foods are given in Table 8-1.
These concentrations range from a low of about 0.002 /ug/g of food to a
high of more than 0.8 //g/g (found in milk) . Background levels of lead
in milk can range from 23 to 79 //g/kg.3 The ATSDR states that, for the
general population, the highest levels of exposure to lead are most
likely to occur through the ingest ion of contaminated food and
drinking water and by the inhalation of lead particles in ambient air.
Furthermore, fruits, vegetables, and grains may contain levels of lead
in excess of background levels as a result of deposition of lead
onto plant surfaces and by plant uptake from soil . As seen in
Table 8-1, the average adult dietary intake of lead for the years
1980-82 was estimated to be 56.5 /ug/day.i However, recent data (1992)
indicate that average dietary intakes have decreased significantly
over the past decade to approximately 2-4 /ug/day. 4 Presumably, this is
at least partially due to the phasing out of lead in gasoline over the
past two decades. In general, human exposure to lead is most likely
to occur through the ingestion of contaminated food and drinking water
and by inhalation of lead particulates emitted from an emission source
or reentrainment of lead-contaminated soil. Lead-based paints, dust,
and chips are also potential avenues of significant exposures for
subpopulations (e.g., children) that may ingest these items.
Lead emissions from utilities do not contribute substantially to
the total amount of lead released annually from anthropogenic
activities (i.e., approximately 1 percent). However, lead emissions
from utilities were not modeled for noninhalation exposures.
Therefore, it is unclear whether the impact of the lead emissions from
utilities is significant. Air concentrations and deposition rates for
lead emissions were estimated using the RELMAP model. The methods and
model are described in detail in chapter 6 (section 6.6) . Figures 8-1
through 8-6 show the results of the lead RELMAP modeling.
8 . 3 CADMIUM COMPOUNDS
As shown in Table 5-8, cadmium is persistent in the environment
and has the potential to bioaccumulate . Health effects data indicate
that cadmium is relatively toxic by both oral and inhalation routes.
Cadmium is a probable (Bl) human carcinogen by the inhalation route
and is relatively potent (IURE = 1.8 x 10 "3 per //g/m3) . However, there
are insufficient data to assess the carcinogenicity from oral
exposure. Regarding noncancer effects, cadmium exposure has been
linked to kidney effects, primarily proteinuria. The RfD for cadmium
8-3
-------
Table 8-1. Concentration of Lead in Various Food Products-'^
Food group
Dairy products
Meat, fish, and poultry
Grain and cereal products
Vegetables
Fruit and fruit juices
Oils, fats, and shortenings
Sugar and adjuncts
Beverages
Concentration (pg/g)
0.003-0.83
0.002-0.159
0.002-0.136
0.005-0.649
0.005-0.223
0.002-0.028
0.006-0.073
0.002-0.041
is 5 x 10"4 mg/kg/d, and is associated with a high confidence level
since it is based on human data. For these reasons, cadmium emissions
from anthropogenic sources have the potential to be a concern for
noninhalation exposure.
For 1990, the estimated cadmium emissions from all coal-, oil-,
and gas-fired units were 1.9 ton/yr, 1.7 ton/yr, and 0.054 ton/yr,
respectively. To put these estimates into perspective, a recent study
by the EPA estimates that about 233 tpy were emitted in the United
States by anthropogenic sources in 1990. 5 Therefore, utilities are
estimated to emit slightly more than 1 percent of anthropogenic
cadmium emissions in the United States.
Cadmium is a naturally occurring metal found in small amounts
throughout the earth's crust. However, the primary source of cadmium
found in the environment is anthropogenic activities. Once released
from stack sources, such as utilities, cadmium can be deposited to
various environmental media. Cadmium emitted from combustion
processes typically exists in the atmosphere as small PM. Cadmium
oxide is the predominant form of cadmium in the atmosphere. However,
the toxicology of cadmium appears not to be dependent on the
chemical's form. Because a typical residence time of particulate
cadmium is between 1 and 10 days, long-range transport (up to
thousands of kilometers) can occur. The primary removal mechanism
from the atmosphere is wet or dry deposition onto soil, water, and
plants. Atmospheric deposition can contribute significantly to the
concentration of cadmium in soil in areas surrounding emission
sources, such as incinerators and areas of heavy vehicular traffic.
Once deposited onto soil, cadmium can be leached into water,
especially under acidic conditions. In the aquatic environment,
cadmium exists primarily as a soluble hydrated ion; as a result, it is
more mobile than other heavy metals, such as lead. However, under
certain environmental conditions, cadmium concentrations have been
found to be at least 1 order of magnitude higher in sediment than in
5-4
-------
00
I
Ul
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 8-1. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Lead from Coal Utilities: Predicted Air
Concentration of Lead, Units: ng/m3
-------
00
I
CTl
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 8-2. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Lead from Oil Utilities: Predicted Air
Concentration of Lead, Units: ng/m3
-------
00
I
0.001 to 0.01
0.01 to 0.02
0.02 to 0.05
0.05 to 0.1
> = 0.1
Figure 8-3. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Lead from Coal and Oil Utilities: Predicted Air
Concentration of Lead, Units: ng/m3
-------
\
0.05 to 2
2 to 5
5 to 10
10 to 20
> = 20
Figure 8-4. Predicted Lead Wet and Dry Deposition from Coal Utilities Based on
1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
00
0.05 to 2
2 to 5
5 to 10
10 to 20
> = 20
Figure 8-5. Predicted Lead Wet and Dry Deposition from Oil Utilities Based on
1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
0.05 to 2
2 to 5
5 to 10
10 to 20
> = 20
Figure 8-6. Predicted Lead Wet and Dry Deposition from Coal and Oil Utilities
Based on 1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
the water column. Accumulation of cadmium in terrestrial and aquatic
plants can occur by root uptake from soil. In addition, terrestrial
plants can be contaminated by absorption of airborne cadmium deposited
onto plants. Accumulation of cadmium in feed crops has the potential
to result in high levels of cadmium in the liver and kidneys of
animals that eat the contaminated feed. Data indicate that cadmium
bioaccumulates in all levels of the food chain.6 Table 8-2 presents
concentrations of cadmium in various foods.
The highest background levels of cadmium are found in soils.-
The mean concentration of cadmium in uncontaminated soil is 0.25 ppm.
Topsoil concentrations can be as much as two times higher than
subsurface concentrations due to atmospheric exposition and
contamination (e.g., landfarming of municipal sludge). Average
ambient air concentrations can range from 1 x 10~6 mg/m3 in remote areas
to 4 x 10"5 in urban areas. Concentrations of cadmium in surface
water, groundwater, and drinking water are typically less than 1 //g/L.
In a study conducted in 27 U.S. cities, 12 food groups were tested and
cadmium was detected in nearly all samples. As shown in Table 8-2,
the highest levels of cadmium were found in leafy vegetables and
potatoes and the lowest levels were found in beverages. Liver and
kidney meats and shellfish were associated with higher concentrations
than other types of meats. Cadmium can accumulate in freshwater and
marine animals at levels up to thousands of times higher than the
levels of cadmium found in the water. In the United States, the adult
intake of cadmium attributable to diet is estimated to be
approximately 30 //g/d. Assuming a gastrointestinal absorption of 5 to
10 percent, the amount of cadmium absorbed from diet is approximately
1 to 3 yug/d. In addition, human exposure can occur at the same level
through cigarette smoking. Cadmium has been found to accumulate in
tobacco plants. The average concentration of cadmium in a cigarette
is between 1 and 2 //g/cigarette. A cadmium exposure and absorption
level of 1 to 3 yug/d can result from smoking one pack of cigarettes
per day. Based on these data, the ATSDR states that, for the general
nonsmoking population, the highest levels of exposure to cadmium are
most likely to occur through consumption of food.- However, smoking
can result in double the exposure level. Individuals living near
emission sources may be exposed to above-average exposure levels
through multiple pathways, such as ingestion of contaminated drinking
water or garden vegetables, inhalation of airborne dust, and
incidental ingestion of contaminated soil.
A study that examines the transfer of metals to bovine milk
indicates that human exposure to cadmium through the consumption of
milk may not be of concern. Because the contribution to human
exposure through the food chain has not been thoroughly examined, a
study was undertaken to estimate the steady-state bovine milk
biotransfer factors (i.e., the rate at which the compounds are
-------
Table 8-2. Concentration of Cadmium in Various Food Products-
Food group
Potatoes
Leafy vegetables
Grain and cereal products
Root vegetables
Garden vegetables
Oils and fats
Sugars and adjuncts
Meat, fish, and poultry
Legume vegetables
Dairy products
Fruits
Beverages
Average concentration (ppm)
0.0421
0.0328
0.0237
0.0159
0.0171
0.0108
0.0109
0.0057
0.0044
0.0035
0.0021
0.0013
All groups
Range of concentrations (ppm)
0.016 to 0.142
0.016 to 0.061
0.002 to 0.033
trace-0.028
trace-0.093
trace-0.033
trace-0.053
trace-0.014
trace-0.016
trace-0.016
trace-0.012
trace
trace-0.142
transferred to milk) for six metals: arsenic, cadmium, chromium,
lead, mercury, and nickel.- Results from this study indicated that, of
the metals studied, lead and arsenic transferred to milk to the
greatest extent and cadmium to a lesser extent. The bioconcentration
factor estimated for cadmium is 1.3 x I0~s L/kg. To put this value
into perspective, the author points out the estimated biotransfer
factor of TCDD is 2.6 x 10"2 L/kg and that this particular food chain
pathway may be less important for these metals than it is for TCDD.
In general, human exposure to cadmium is most likely to occur
through the consumption of food products for nonsmokers and through
the consumption of food products and cigarette smoking for smokers.
Because cadmium emissions from utilities do not contribute
significantly to the total amount of cadmium released annually from
anthropogenic activities (i.e., approximately 1 percent) and because
there are numerous units dispersed throughout the country, it is
unclear whether the impact of the cadmium emissions on the background
concentration of cadmium in the various media is significant.
Exposure resulting from utility emissions of cadmium in excess of
background levels cannot readily be determined. Deposition rates were
estimated for cadmium using the RELMAP model. The methods and model
are described in detail in chapter 6 (section 6.6). Figures 8-7
through 8-9 show the results of the cadmium RELMAP modeling.
-------
00
I
0.005 to 0.1
0.1 to 0.2
0.2 to 0.5
0.5 to 1
> = 1
Figure 8-7. Predicted Cadimum Wet and Dry Deposition from Coal Utilities Based on
1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
0.005 to 0.1
0.1 to 0.2
0.2 to 0.5
0.5 to 1
> = 1
Figure 8-8. Predicted Cadimum Wet and Dry Deposition from Oil Utilities Based on
1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
0.005 to 0.1
0.1 to 0.2
0.2 to 0.5
0.5 to 1
> = 1
Figure 8-9. Predicted Cadimum Wet and Dry Deposition from Coal and Oil Utilities
Based on 1990 Emissions Estimates as Modeled with RELMAP, Units: iig/m2/yr
-------
8.4 OVERALL SUMMARY
Further quantitative analysis of noninhalation exposure to
HAPs that are persistent in the environment and that have the
potential to bioaccumulate may be appropriate in future studies.
Due, in part, to low emissions of these HAPs from utilities
relative to other anthropogenic sources, a quantitative
assessment of noninhalation exposure to lead and/or cadmium has
not been given as high a priority as arsenic, radionuclides,
dioxins, and mercury for multipathway assessment. However, lead
and cadmium are persistent in the environment, have a tendency to
bioaccumulate, and are toxic by the inhalation and ingestion
routes of exposure. Therefore, further assessment of the
multipathway exposures and risks may be appropriate in future
studies to more comprehensively evaluate the impacts of emissions
of lead and cadmium from utilities.
8-16
-------
REFERENCES
Agency for Toxic Substances and Disease Registry Toxicological
Profile for Lead. TP-92/12. U.S. Public Health Service.
Atlanta, GA. 1993.
U.S. Environmental Protection Agency. Air Quality Criteria for
Lead. EPA 600/8-83-028f. Washington, DC. 1983.
Stevens, J. B. Disposition of toxic metals in the agricultural
food chain. 1. Steady-state bovine milk biotransfer factors.
Environmental Science and Technology. Volume 25, No. 7. 1991.
pp. 1289-1294.
Bolger, P. M., Yess, N. J., Gunderson, E. L., Troxell, T. C., and
Carrington, C. D. Identification and reduction of sources of
dietary lead in the United States. Food Additives and
Contaminants, Volume 13, No. 1. 1996. pp. 53-60.
U.S. Environmental Protection Agency. 1990 Emissions Inventory
of Forty Section 112 (k) Pollutants. September 1997.
Agency for Toxic Substances and Disease Registry. Toxicology
Profile for Cadmium. TP-92/06. U.S. Public Health Service.
Atlanta, GA. 1993.
-------
9.0 MULTIPATHWAY EXPOSURE AND RISK ASSESSMENT FOR RADIONUCLIDES
9.1 SUMMARY OF RADIONUCLIDE ANALYSIS
The EPA assessed the exposure and risks due to radionuclide
emissions from coal-, oil-, and gas-fired utilities. The details of
this assessment are contained in an EPA report entitled: Estimates of
Health Risks Associated with Radionuclide Emissions from Fossil-Fueled
Steam Electric Generating Plants.1 This section summarizes that
report.
Shortly after the discovery of radioactivity at the turn of the
century, investigators became aware that nearly all natural materials
contained trace quantities of radioactivity. Natural radioactivity is
derived from two sources. A small percentage of natural radioactivity
is derived from the interaction of cosmic radiation with specific
elements (e.g., carbon-14, tritium). The majority of naturally
occurring radionuclides are classified as primordial radioisotopes or
their radioactive decay products. Primordial radionuclides are
believed to have been formed, along with all other terrestrial
elements except hydrogen by nuclear fusion reactions, neutron
absorption, and beta decay in a former star that exploded as a super
2
The behavior of primordial radionuclides and their decay products
in the environment is complex. Pathways leading to significant human
exposures include external radiation from the emission of gamma rays
from the ground and building materials. Internal exposure may result
from the transfer of radioactivity through root uptake by plants that
serve as food for domestic animals or are directly ingested by humans.
Internal exposure may also result from the inhalation of airborne
radioactivity.
The three major fossil fuels—coal, oil, and natural gas—contain
varying quantities of the naturally occurring radionuclides of the
uranium-238 and thorium-232 series and potassium-40. When these fuels
are burned to produce steam in the production of electricity,
radionuclides are entrained in the combustion gases and may be emitted
into the environment. As early as 1954, Anderson, Mayneord, and
Turner3 suggested that human activities, particularly the burning of
coal, might significantly perturb the natural radiation environment by
transferring additional radioactivity into the air, where it is more
readily available for human intake by inhalation.
Radionuclides are among the HAPs included in section 112(b).
Over the years, EPA has reviewed available information and provided
estimates regarding the radionuclide content of fossil fuels,
environmental emissions, human exposure, and health risks. This
information has been reported by the EPA in several earlier reports,
including the Background Information Document supporting the decision
not to regulate radionuclide emissions from coal-fired boilers issued
in 1989.4 The EPA report summarized in this chapter updates previously
published data and estimates with more recently available information
9-1
-------
regarding the radionuclide contents of fossil fuels, associated
emissions by utilities, and potential health effects to exposed
population groups.
9.1.1 Natural Radionuclide Content in Fossil Fuels: Coal
The decay series of uranium and thorium constitute the major
radionuclides contained in coal. Uranium-238 has 13 major radioactive
decay products and thorium-232 has 9. For coal, it is generally
assumed that primary members within each of the two decay series are
in secular equilibrium. Secular equilibrium means that the
radioactivity concentrations among primary decay chain members are
constant. A national database of nearly 7,000 coal samples was
analyzed with regard to uranium and thorium content of the major ranks
of coal used by utilities. Concentrations spanned a wide range of
values that were lognormally distributed. Table 9-1 summarizes the
data by providing the geometric mean concentration values expressed in
units of parts per million and identifies the relative percent utility
consumption of coal types.
Concentration values expressed in parts per million are readily
converted to radioactivity concentrations by means of the specific
activity values for uranium-238 and thorium-232. For U-238, 1 ppm is
equal to 0.33 pCi/g of coal; for Th-232, 1 ppm is equal to 0.11 pCi/g
of coal. For example, in bituminous coal with an average content of
1.24 ppm uranium and 2.18 ppm thorium there is a corresponding
activity of 0.41 pCi/g for each member of the U-238 series and 0.24
pCi/g for each member of the Th-232 series.
The radionuclide content of coal is not unique when compared to
other natural materials. In fact, it is generally assumed that the
average radioactivity of the earth's crust (i.e., soil and rocks) is
about twice that of coal.
9.1.2 Natural Radionuclide Content in Fossil Fuels: Natural Gas
Radioactivity in natural gas is almost exclusively radon-222,
which migrates from proximal geologic formations into gas reservoirs.
In 1989, the American Gas Association identified 262,482 production
wells that yielded more than 18 trillion cubic feet (Tcf) of natural
gas. An additional 1.53 Tcf of gas were imported primarily from
Canada. About 2.77 Tcf of gas were consumed by utilities to produce
electricity.
The radon content of natural gas at the wellhead has been
measured in thousands of wells over several decades. However, these
measurements are of limited use for estimating radon concentrations at
the point of consumption for several reasons. Radon concentrations
vary by geographic location and over time. Also, radon content is
markedly reduced when natural gas is processed to remove commercially
valuable heavier hydrocarbons (ethane, propane, butane). Further
reductions in radon concentrations reflect the natural decay that
occurs during the gathering, processing, and distribution/storage
9-2
-------
Table 9-1. Utilization and Radionuclide Content by Coal Rank
Coal rank
Bituminous
Subbituminous
Lignite
Percent
utilization
69.0
24.7
6.3
Average uranium
(ppm)
1.24
1.07
1.41
Average thorium (ppm)
2.18
2.28
2.38
of gas prior to consumption. The main radon isotope, Rn-222, has a
half-life of about 4 days; the other isotopes have half-lives of less
than 1 minute.
A more meaningful approach is to assess the radon content in gas
distribution lines. Analysis of gas in the distribution lines
eliminates well-to-well variations and accounts for radon reduction
from processing and natural decay. Radon measurements of natural gas
in distribution lines near the point of consumption suggest an average
value of 20 pCi/L. In this report, therefore, estimated radon
emissions from gas-fired boilers are based on a radon concentration of
20 pCi/L of processed gas.
9.1.3 Natural Radionuclide Content in Fossil Fuels: Oil
Residual fuel oil is a general classification of fuel obtained as
liquid still bottoms from the distillation of crude oil.
Nonradiometric analyses show that crude oil and various petroleum
products may contain as many as 60 different metals in measurable
quantities. Uranium and thorium are among the trace metals commonly
found in crude oil and petroleum products. The presence of these two
radioactive trace metals also implies the presence of their
radioactive decay products.
A comprehensive literature search, however, revealed that data
specific to the radionuclide content of residual fuel oil are not only
sparse but are considerably more difficult to interpret than those for
coal or gas. Contributing to the difficulty in data interpretation is
the absence of secular equilibrium among primary members of the U-238
and Th-232 decay chains. Due to the paucity of data, the EPA
concluded that there was a need for additional data and conducted its
own study.
The EPA enlisted the help of the Utility Air Regulatory Group
(UARG) and the EPRI to solicit the voluntary participation of
individual utilities in providing samples of residual oils for
radioanalysis. The selection of a utility was based on the utility's
geographic location, along with its generator nameplate capacity,
capacity factor, and/or annual fuel-oil consumption. Selection,
therefore, favored larger facilities with the highest capacity
factors/fuel consumption and accounted for radionuclide variability
based on origin of crude oil.
9-3
-------
In total, 12 utilities provided 42 samples of residual fuel oil
for analysis. Participating utilities represented major regions of
the United States where fuel oil serves as a primary fuel source.
Quantitatively, the 12 utilities had an annual consumption of about
2 billion gallons, which was estimated to be about 24 percent of the
fuel oil consumed by all U.S. oil-fired units.
Radionuclide analysis, data interpretation, and data verification
involved the efforts of a major commercial analytical laboratory, the
EPA's National Air and Radiation Environmental Laboratory (NAREL), and
the National Institute of Standards and Technology (NIST).
Table 9-2 provides estimates of the average radionuclide values
of the 42 residual fuel oil samples evaluated in the EPA study.
Values are well within the range of the limited study data reported by
others and support the conclusion that the radionuclide content of
residual fuel oil is low relative to coal.
9.1.4 Radionuclide Emissions from Fossil-Fueled Plants
Radionuclide emissions from utilities are affected by the
radionuclide content in fossil fuel, by plant design features, and by
operating parameters. Important design features involve the size of
the plant, type of furnace used, and the emission control systems
designed to remove pollutants from the flue gas. The most significant
operational factors, which dictate the rate of fuel consumption,
involve the percentage of time a plant is operating, the power level,
and the efficiency by which a plant converts thermal energy to
electric energy.
In this report, estimates of radionuclide emissions and
associated human health risks are based on fossil-fired boiler units
with generating capacities of 25 MWe or more. The 25-MWe selection
criterion reflects the low probability of significant emissions for
small plants, regardless of unit-specific operating parameters. Of
the Nation's 2,298 boiler units (Table 9-3), 1,748 units have a
generating capacity of 25 MWe or more.
From data reported to the EEI that include annual fuel
consumption and particulate removal efficiencies, emissions were
estimated for each of the 1,748 boiler units and aggregated by plant
affiliation. (The 1,748 fossil-fired boiler units represent a total
of 684 utility plants.) These unit- and plant-specific emission data
are contained in a separate addendum to the EPA report.5 Table 9-3
provides average annual emissions per operating boiler unit, as well
as per billion kilowatt-hour of electricity generated. For coal-fired
units, the average annual emissions for particulates range from a
fraction of a millicurie (mCi) to several millicuries among primary
radionuclides.
9-4
-------
Table 9-2. Estimates of Average Radionuclide Concentrations in
42 Residual Fuel Oil Samples
U-238 Series
U-238
Th-234
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232 Series
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Po-212
Concentration (pCi/g)
0.0018
0.0018
0.0018
0.0034
0.0068
0.0043
0.0043
0.0043
0.0043
0.0043
0.0043
0.44
0.44
0.44
Concentration (pCi/g)
0.0030
0.068
0.068
0.068
0.068
0.068
0.068
0.068
0.068
Although the average radionuclide content of residual fuel oil is
2 to 3 orders of magnitude lower than that of coal, Table 9-3 reveals
that average emission rates are nearly comparable. This is explained
by the fact that, unlike coal-fired units, the majority of oil-fired
units lack particulate emission control systems that remove
radionuclides from the flue gas with efficiencies of 95 percent or
more. Due to the fact that coal-fired units on average have a higher
capacity factor, the degree of comparability between coal-fired and
oil-fired units is further enhanced when emissions are defined per
unit of billion kilowatt-hours.
Particulate emissions for units designated as gas-fired are
generally small when compared to either coal- or oil-fired units.
Moreover, radionuclide emissions other than radon from units
designated as gas-fired principally result from the combustion of a
secondary fuel.
9.1.5 Summary of CAP-93 Model
The Clean Air Act Assessment Package-1993 (CAP-93) is the most
recent version of a computer model used for population dose and risk
assessment for radionuclide air emissions. For a given facility,
9-5
-------
Table 9-3. Average Annual Radionuclide Emissions per Operating
Boiler Unit and per Billion Kilowatt-Hour Electricity Generated
Radionuclide
U-238
Th-234
Pa-234m
Pa-234
U-234
Th-230
Ra-226
Rn-222
Po-218
Pb-214
Bi-214
Po-214
Pb-210
Bi-210
Po-210
Th-232
Ra-228
Ac-228
Th-228
Ra-224
Rn-220
Po-216
Pb-212
Bi-212
TI-208
K-40
Emission rates
Per operating unit (mCi/y)
Coal
2.3x10°
1.2x10°
1.2x10°
1.2x10°
2.3x10°
1.2x10°
1.7x10°
3.0 x102
5.6x10°
5.6x10°
1.2x10°
5.6x10°
5.6x10°
1.2x10°
5.6x10°
7.1 x10-1
1.0x10°
7.1 x10-1
7.1 x10-1
1.0x10°
1.6x102
3.5x10°
3.5x10°
7.1 x10-1
2.1 x10-1
7.8x10°
Gas
1.3x102
1.3x10-2
1.3x10-2
1.3x10-2
2.5x10-2
4.9 x10-2
2.9 x10-2
2.5 x103
3.1 x10-2
3.1 x10-2
3.1 x10-2
3.1 x10-2
3.1 x10°
3.1 x10°
3.1 x10°
2.1 x10-2
4.7 x10-1
4.7 x10-1
4.7 x10-1
4.7 x10-1
5.7 x10-1
4.7 x10-1
4.7 x10-1
4.7 x10-1
1.4 x10-1
6.2 x10-3
Oil
1.1 x10-1
1.1 x10-1
1.1 x10-1
1.1 x10-1
2.1 x10-1
4.1 x10-1
2.6 x10-1
3.8x102
2.7 x10-1
2.7 x10-1
2.7 x10-1
2.7 x10-1
2.7x101
2.7x101
2.7x101
1.8x101
4.1 x10°
4.1 x10°
4.1 x10°
4.1 x10°
8.4x10°
4.1 x10°
4.1 x10°
4.1 x10°
1.2x10°
5.2 x10-3
Per billion kWe-h generated (mCi/y)
Coal
1.5x10°
7.7 x10-1
7.7 x10-1
7.7 x10-1
1.5x10°
7.7 x10-1
1.2x10°
2.0 x102
3.8x10°
3.8x10°
7.7 x10-1
3.8x10°
3.8x10°
7.7 x10-1
3.8x10°
4.7 x10-1
7.1 x10-1
4.7 x10-1
4.7 x10-1
7.1 x10-1
1.1 x102
2.4x10°
2.4x10°
4.7 x10-1
1.4 x10-1
5.3x10°
Gas
2.6x10-2
2.6 x10-2
2.6 x10-2
2.6x10-2
4.9 x10-2
9.5 x10-2
5.7x10-2
4.9x103
6.0 x10-2
6.0 x10-2
6.0x10-2
6.0 x10-2
6.0x10°
6.0x10°
6.0x10°
4.1 x10-2
9.1 x10-1
9.1 x10-1
9.1 x10-1
9.1 x10-1
1.1 x10°
9.1 x10-1
9.1 x10-1
9.1 x10-1
2.7 x10-1
1.2x10-2
Oil
1.8x10-1
1.8x10-1
1.8x10-1
1.8x10-1
3.4 x10-1
6.7 x10-1
4.3 x10-1
6.2 x102
4.4 x10-1
4.4 x10-1
4.4 x10-1
4.4 x10-1
4.4 x101
4.4 x101
4.4 x101
3.0 x10-1
6.7x10°
6.7x10°
6.7x10°
6.7x10°
1.4x101
6.7x10°
6.7x10°
6.7x10°
1.9x10°
8.5 x10-3
9-6
-------
atmospheric releases may be modeled for as many as six independent
sources. Plume rise can be calculated assuming either a momentum- or
buoyancy-driven plume that reflects facility-specific plant
parameters. Plume dispersion is based on a modified Gaussian plume
equation and accounts for plume depletion that includes precipitation
scavenging and dry deposition. Primary model parameters for plume
dispersion and depletion are based on available site-specific
meteorological data. (A library of meteorological data that include
wind data files, annual precipitation, ambient temperatures, and lid-
height for all major cities is provided by the code of CAP-93).
From plume dispersion and plume depletion calculations, the CAP-
93 program computes radionuclide concentrations in air and rates of
deposition and buildup on ground surfaces and in soil. Estimates of
the radionuclide concentrations in produce, leafy vegetables, milk,
and meat are made by coupling the output of the atmospheric transport
models with the terrestrial food-chain models defined in the U.S.
Nuclear Regulatory Commission's Regulatory Guide 1.109. The
quantities of foodstuff produced locally are based on the average
agricultural productivity data of the State in which the assessment
area is located.
For dose and risk estimates, the population distribution at each
of the 684 assessed sites was developed by means of the GENPOP
computer code and 1990 Census Bureau data. Dose estimates reflect the
exposure from external (air immersion and ground surface) and internal
(inhalation and ingestion) sources. For low-LET external radiation,
CAP-93 employs the nominal risk coefficient of 3.9 x 10 "4 fatal cancers
per rem.
For internal exposures, dose and risk estimates are defined by
ICRP tissue/organ weighting factors that account for route of entry,
clearance class, and transfer factors within body compartments. In
summary, dose and cancer risks can be tabulated for individual
exposure pathways, radionuclides, and tissues/ organs. All risk
estimates pertain to the risk of fatal cancer and assume that exposure
occurs over the lifetime of individuals within the assessed
population.
EPA's methodology for estimating risks from Rn-222 emissions is
based on an extrapolation of epidemiologic findings of underground
miners exposed to radon.1'5 CAP-93 calculates working levels (WL), not
concentrations of specific radon daughter products. A WL is defined
as any combination of short-lived radon decay products in 1 liter of
air that will result in the emission of 1.3 x 105 MeV of alpha-particle
energy. Risk is not derived from dose but from time-integrated
exposure expressed in working level months (WLM). Under typical
residential exposure conditions, it is assumed that 1 WLM corresponds
9-7
-------
to 170 hours of exposure at 200 pCi/L of radon gas. CAP-93 employs a
risk coefficient of 3.6 x 10"4 fatal lung cancers per WLM.a
CAP-93 assesses risk for a circular grid that is defined by
sixteen sectors and up to 20 radial distances around a specified
facility. For this study radial distances of 400, 1,500, 3,500,
7,500, 10,000, 15,000, 25,000, 35,000, 45,000, and 50,000 meters were
used. Risk to the population is determined by summing individual
risks by distance and section for the 0- to 50-km grid around each
assessed facility. Risk to the maximally exposed individual(s)
corresponds to that location (i.e., distance and sector of highest
exposure) where individuals are believed to reside.
The population risk frequency distribution identifies the number
of people at various levels of risk. The risk categories are divided
into powers of 10, in which the individual lifetime cancer risk ranges
from one chance in ten to less than one chance in a million. Risk
data for each of the 684 assessed plants are provided in the
previously identified Addendum. Only a summary of these data is
provided below.
9.1.6 Estimates of Population Health Risks
Radionuclide emissions from utilities may result in public
exposure from multiple pathways that include (1) external radiation
from activity suspended in air or deposited on the ground and
(2) internal exposure from the inhalation of airborne contaminants or
ingestion of contaminated food products. Although the potential
health risks are essentially independent of whether a dose was
internal or external, the assessment of internal exposures is
considerably more complex. For ingested or inhaled radionuclides,
dose assessment requires biokinetic information that describes the
distribution and retention of individual nuclides, the type of
radiation emitted, and the amount of energy absorbed by individual
target tissues/organs.
Estimates of population doses from chronic atmospheric releases
require the use of a computer code that accounts for atmospheric
dispersion, radionuclide concentrations in environmental media, and
radionuclide intakes by inhalation and ingestion. In support of
Recently, the Agency revised its estimates of radiogenic cancer risks to
reflect the current epidemiological data and scientific consensus on
extrapolations from the available data to chronic low dose exposures.-
The revised estimates yield a nominal value of 5.1 x 10~4 fatal cancers per
rad for uniform whole body exposure to low-LET radiation and 2.2 x 10~4
fatal lung cancers per WLM for exposure to radon-222 and its decay
products. The radon risks reported in this study can be adjusted to the
new radon risk coefficient simply by applying a correction factor of about
0.6. No simple adjustment can be made to the non-radon risks to reflect
the Agency's current values. However, since the ground surface pathway
dominates the risk for maximally exposed individuals, an upward adjustment
of approximately 30 percent would bound their risks.
9-8
-------
National Emission Standards for Hazardous Air Pollutants, the EPA,
with support from Oak Ridge National Laboratory, developed the CAP-88
computer model. The CAP-88 (Clean Air Act Assessment Package-1988)
computer model is a composite of computer programs, databases, and
associated utility programs.
The CAP-88 programs are considered among the best available
verified models for population dose and risk assessment for
radionuclide air emissions. For a given facility, atmospheric
releases and dose assessment may be modeled for up to six independent
sources that take into account plant- and site-specific model
parameters.
Since it was first introduced, CAP-88 has been revised
periodically to reflect changes in database information and improved
risk methodologies. For this study, the most recent version of the
code, designated as CAP-93, was used. The CAP-93 contains a
correction to the procedure used to calculate wet deposition of
radionuclides from the plume.
For low doses of radiation, potential health effects may not
appear for years or even decades following exposure. Such delayed
effects are termed "stochastic" and are thought to result from highly
selective molecular changes in individual cell(s). Although these
highly selective changes occur rarely, when they do, the altered cell
may develop into cancer. Among the stochastic effects that have been
associated with radiation exposure, medical scientists consider cancer
induction the primary health effect of concern.
A key characteristic of a stochastic effect is that the severity
of the effect is not dose-dependent. However, the probability that a
stochastic event (i.e., cancer) may occur is dictated by the radiation
dose. The stochastic nature of low-dose radiation is not unique but
is universal to all carcinogenic agents that act by primary genetic
mechanisms.
The current method of estimating radiation risks relies on select
human studies in which cancer rates were observed at a higher
incidence among exposed individuals than would normally occur
spontaneously. The most intensely studied people are the Japanese
atomic bomb survivors of Hiroshima and Nagasaki. Data through 1985
show that, among the 76,000 individuals studied, 5,935 survivors have
died of cancer from all causes. It is estimated that about 340 of
these cancers (80 leukemias and 260 nonleukemias) were the result of
radiation exposure.
The data also define a dose response in which increasing doses
yielded an increased percentage of excess cancers, especially for
leukemia. However, some numerical estimates embody substantial
statistical uncertainties about the number of cancer deaths induced by
radiation. Thus, for doses less than 50,000 mrem (50 rem), the small
number of excess cancers above normal expected levels may reflect
9-9
-------
random fluctuations that are not linked to radiation exposure. When
doses exceed 50,000 mrem (50 rem), the number of excess cancers is
sufficient to support a causal link to human cancers.
For low-dose exposures, a causal link and a quantitative
relationship between radiation dose and cancer has not been
established. Yet, scientists conservatively assume that any dose of
radiation, no matter how small, may pose a risk to human health.
Estimates of health risks from low-level radiation are, therefore,
derived by extrapolating risks from high doses to lower doses using a
linear nonthreshold dose-response model contained in the CAP-88 and
CAP-93 computer codes.
9.2 RADIONUCLIDE UNCERTAINTY ANALYSIS
Uncertainties in the estimates of risk presented for radionuclides
emitted from fossil-fuel-fired steam electric- generating units were
assessed using both qualitative judgments and quantitative techniques.7
As in almost all assessments of environmental health risk, the risk
estimates were based on modeling rather than direct measurements of
exposure and risk; therefore, the results were subject to uncertainties
in modeling, completeness, and parameter values.
Modeling uncertainties pertain to the formulation of the
mathematical models used to predict risk and the degree to which they
accurately represent reality. Completeness uncertainties pertain to
whether or not all significant radionuclides and pathways of exposure
are addressed. Parameter uncertainties pertain to the specific values
assigned to the parameters that are input to the calculational models.
Census-tract, air dispersion, environmental transport, metabolic,
and dose-response models were used to predict the location of
individuals around the plants; the dispersion of the pollutants in the
environment; their concentrations in soil and air at receptor
locations; their accumulation and removal from soil; their uptake and
transfer from soil to foodstuffs; their intakes, translocations,
accumulations, and removal from the various organs and tissues of the
body; and the resulting risks to the individuals in the exposed
population.
The modeling uncertainty associated with the use of the GENPOP
census-tract model used to locate the individuals within 50 km of each
plant was limited to its ability to properly place individuals living
in proximity to the plant. The potential magnitude of this
uncertainty was partially assessed in a quantitative manner using
field-verification techniques to identify actual locations of nearby
individuals for the plants with the highest estimated maximum
individual risks. The results of these plant-specific assessments
indicated that, on a plant-by-plant basis, the reported MIR might be
high by an order of magnitude. However, they also demonstrated that
the reported MIR of about 10 ~5 is correct when viewed as an upper bound
for the entire population of electric utility steam-generating units.
9-10
-------
Modeling uncertainties associated with the air dispersion,
environmental transport, metabolic, and dose-response models were
considered via model input parameters. Significant model inputs were
included in a rigorous Monte Carlo analysis of parameter uncertainties
associated with two plants with the highest estimated MIRs. For the
dispersion and environmental transport models that were used, which
are widely recognized as appropriate for the physical processes that
govern dispersion and environmental concentration, assessment of
parameter uncertainty only was clearly appropriate. For the metabolic
and dose-response models, the parameter uncertainty relied on lumped
parameters. This approach reflected the limited data available on
the uptake and retention of radionuclides within the various organs
and tissues of the body and the necessity of extrapolating the
dose-response relationship from data reflecting much higher (orders of
magnitude) exposures.
Uncertainties in completeness are limited to source terms and
exposure pathways. Because the source terms for utilities are well
characterized, there is very little likelihood that significant
unaccounted for radionuclide releases are occurring at these
facilities. With respect to pathways of exposure, the analysis
assumed that four pathways of exposure (ingestion of milk, meat, and
vegetables; inhalation; immersion in contaminated air; and exposure to
contaminated ground) were present at all sites. The ground water
pathway was not included because the deposited material is on the
ground surface in a physical and chemical form that minimizes its
potential to leach to ground water.
Exposure to multiple sources is one potentially significant
exposure pathway that was not accounted for by the air dispersion and
environmental transport modeling. To evaluate the potential
significance of this pathway, explicit hand calculations were
performed for the plants with the highest estimated MIRs to estimate
the impacts from all plants within 50 km. The results of these
explicit calculations showed that omission of multiple plants from the
estimates resulted in less than a 5 percent error in the MIRs. A
related completeness uncertainty was the impact on individuals
residing beyond the 50-km assessment area around each plant. A
semiquantitative assessment of this uncertainty, which affects the
estimate of deaths per year in the exposed populations, indicated that
the risk was not understated by more than a factor of 3. Given these
results, completeness uncertainties are not judged to be a significant
contributor to the overall uncertainty in the analysis.
The largest uncertainties were associated with the parameter values
used in the assessment. As noted above, a rigorous assessment of
parameter uncertainties was conducted for the two plants with the highest
estimated MIRs. Nominal values used in the assessment were assigned a
distribution and range based on available data and expert judgment.
Based on this analysis, it was determined that the 90 percent confidence
interval for the reported MIR values of approximately 10"5 ranges from
about I0~s to ID'4.
9-11
-------
9.3 SUMMARY FINDINGS
Table 9-4 gives the distribution of fatal cancer risks to the
combined populations residing within the 50-km (35-mile) radii of the
684 utility plants. The aggregate of assessed populations living
within a 50-km radius of a plant is estimated to be 196.1 million,
which represents approximately 75 percent of the U.S. population. The
individual lifetime risk of fatal cancer to more than 99.9 percent of
the assessed population (i.e., 196.1 million) is less than one chance
in a million. The data further suggest that, under current operating
conditions, there are no instances in which the release of
radioactivity is likely to result in a lifetime fatal cancer risk to
any one person that is equal to or greater than 1 chance in 10,000.
It is estimated that about 1,027 individuals residing within a 50-km
distance of a plant may receive radiation exposures for which the
lifetime risk is between 1 in 10,000 and 1 in 100,000 (i.e., 1 x 10"4
to 1 x 1Q-5) .
It must also be pointed out that the distribution of individual
risks within each risk range is heavily skewed toward the lower value.
This is evidenced by the fact that the average individual lifetime
risk is a small fraction of the midpoint value within each of the risk
ranges. Correspondingly, the probability of a single fatal cancer
occurrence within the highest risk group of 1,027 individuals is less
than 2 chances in 10,000 per year. For the entire assessed population
of 196,100,000 within 50 km of these plants, the estimated cancer risk
attributable to radionuclide emissions from electric utility steam
generating units (SGUs) (includes coal-, oil-, and gas-fired utilities
)is less than 1 cancer death per year (i.e., 3.36 x 10^ deaths/year is
the risk equivalent of about 1 in 3 chances that a single cancer death
will occur in a year). Exposures and risks to individuals residing
beyond 50 km are not explicitly evaluated. However, using the
assumption that radionuclides dispersion and exposure beyond 50 km
would be similar to that of arsenic, which was modeled with the RELMAP
(see chapter 6), EPA estimates that the overall cancer incidence may
be seven times greater. That is, considering both local and long-
range transport, the cancer incidence could be roughly 2 cases per
year (i.e., 0.3 x 7). Most (approximately 99 percent) of the cancer
incidence is due to inhalation exposure. The EPA estimates that coal-
fired utilities are contributing about 25 percent of the cancer
incidence and oil-fired utilities the other 75 percent.
Based on radionuclide emissions and plant-specific/ site-specific
data, CAP-93 also calculates the MIR for each of the 684 plants.
Table 9-5 characterizes those plants with the highest estimated MIR
values expressed in lifetime fatal cancer risk. There were a total of
17 plants for which the lifetime risk of fatal cancer to the MIR is
estimated to exceed 1 x 10~5 due to multipathway exposures to
radionuclide emissions from utilities. The highest MIR value of
3 x 10"5 corresponds to a five-unit coal-fired facility that generated
3,340 MW of electricity in 1990. Of the 17 plants with the highest
MIR values, 11 are exclusively designated as coal boilers. Only
9-12
-------
Table 9-4. Frequency Distribution of Lifetime Fatal Cancer Risks
for All Plants
Lifetime cancer
risk range
1 x10°to 1 x1Q-1
1 x1Q-1to1 x10'2
1 x ID'2 to 1 x1Q-3
1 x1Q-3to1 x10'4
1 x1Q-4to1 x1Q-5
1 x10'5to 1 x 10'6
Less than 1 x10'6
Number of
people
0
0
0
0
1,027
95,745
196,000,000
Average
individual
lifetime risk
0
0
0
0
1.3x10-5
2.2 x10-6
1.2x10-7
Deaths per year in
this risk range
0
0
0
0
1.92x10-4
3.06x10-3
3.32x10-1
Death per year in this
risk range or higher
0
0
0
0
1.92x10-4
3.26 x10-3
3.36 x10-1
Table 9-5. Plants with the Highest Estimated Maximum Individual
Risk (MIR)
Plant name
Plant #222
Plant #247
Plant #60
Plant #301
Plant #251
Plant #406
Plant #256
Plant #17
Plant #133
Plant #318
Plant #672
Plant #668
Plant #82
Plant #207
Plant #253
Plant #489
Plant #651
MIR
3X10'5
3X10'5
2x10'5
2x10'5
2x10'5
2x10'5
2x10'5
2x10'5
2x10'5
1 x10'5
1 x10'5
1 x10'5
1 x10'5
1 x10'5
1 x10'5
1 x10'5
1 x10'5
Coal-fired
Units
5
4
4
2
4
4
3
2
6
8
7
3
4
MWe
3,340
900
3,160
750
1,540
2,777
1,728
1,135
1,100
1,965
2,304
2,052
1,872
Gas-fired
Units
3
MWe
262
Oil-fired
Units
2
2
2
2
6
MWe
1,112
66
804
558
372
MIR = maximum individual risk expressed as lifetime fatal cancer risk
9-13
-------
two facilities are identified as exclusively oil-fired plants.
The remaining four plants are represented by a combination of boilers,
where coal is at least one of the designated primary fuels.
The MEI risk due to inhalation exposure to radionuclides from the
highest risk oil-fired plant is estimated to be 1 x 10 "5. The MEI risk
due to inhalation exposure to radionuclides from the second highest
risk oil-fired utility is estimated to be 3 x 1Q~6. The other 135 oil-
fired utilities and all coal-fired utilities are estimated to pose
cancer risks less than 1 x I0~s due to inhalation exposure to
radionuclides.
Background Radiation Exposures
The risks due to exposure to radionuclide emissions from
utilities are substantially lower than the risks due to exposure to
natural background radiation. As shown in Tables 9-6 and 9-7 the
average exposure to natural background radiation (excluding radon
progeny) for the U.S. population has been estimated to be roughly
about 100 millirems (mRems) per year.8'9
Background radiation exposure can come from internal or external
sources. External sources include cosmic (extraterrestrial) and
terrestrial (radionuclides in soil and rock). Internal sources
include inhaled and ingested radionuclides retained in the body, with
inhaled radon progeny treated as a separate problem. Radiation from
consumer products and fallout from weapons tests make minor
contributions to background. Average doses are listed in Table 9-6.
Background radon exposure is assessed on the basis of exposure to
its progeny. The estimate is based on the distribution of short half-
life radioactive progeny of radon in the inspired air. Progeny
estimates must be calculated for each environment independently. The
values in Table 9-7 are average population values and do not reflect
the ranges possible.
Risks of background radiation exposure in the United States can
be calculated using the average annual dose from external and internal
sources and the average annual radon exposures and risk conversion
factors from appropriate references.
The age averaged lifetime risk of fatal cancer associated with
the average annual dose of 100 mrem (Table 9-6) is 5.7 x 10"5. 10
Continual lifetime exposure at 100 mrem/y yields a risk of 4.3 x 10 "3.
The lifetime risk of fatal lung cancer associated with the average
annual exposure of 0.257 WLM (Table 9-7) is 5.8 10"5.11 Continuous
lifetime exposure (about 75 years) at 0.257 WLM/y yields a risk of
4 .4 x 1Q-3.
9-14
-------
Table 9-6. Average Background Radiation Doses (effective dose
equivalent excluding inhaled radon progeny)
Source Annual Dose (mrem)
External
Cosmic12 28.4
Terrestrial13 23.6
Internal14
Ingested and inhaled 39
Fallout14 1
Building materials and consumer products14 8
Total 100
Notes:
An individual is expected to spend 87.9% of his/her time indoors (estimates of fraction of time in the indoor and
outdoor environments were adapted from data in the Draft Exposure Factors Handbook, U.S. EPA, Washington,
D.C., 1997.)
For terrestrial sources, the indoor dose rate in air is 0.8 times the outdoor value and the effective dose equivalent
(mrem) is 0.7 times the absorbed dose (mrad) in air.13
The range of background exposure in the United States from the sources listed in Table 9-6 is from about 75
mrem/year to about 200 mrem/year.
Table 9-7.
Progeny
Average Annual
Source
Residence11
Outdoors
Total
Background
Exposures Due to
Radon
Annual Exposure (WLM)
0.242
0.015
0.257
Notes:
WLM = working level month
This radon estimate is incomplete since there is no estimate of the average exposure level inside structures other
than residences. The estimated average exposure and risk can only be higher than listed here.
The estimates of average radon concentration used were: 1.25 pCi/l in residences and 0.3 pCi/l outdoors.
The estimated fraction of time in the outdoor environment (0.121) was adapted from data in the Draft Exposure
Factors Handbook, U.S. EPA, Washington, D.C., 1997. The equilibrium fraction was assumed to be: 0.5 indoors
and 0.8 outdoors.
9-15
-------
Because of limitations in the GENPOP computer code used for
identifying locations of individuals, the MIRs shown for each plant
should be viewed with caution; errors of a few hundred meters in the
location of individuals can result in an over- or underestimate of
risk by factors of 2 or more. The UARG reestimated the risks for the
17 plants with the highest MIRs using refined population grids. Their
results show lower MIRs for the majority of these plants, but their
highest MIR of 1 x 10"5 is consistent with the EPA's estimates. Thus,
the EPA believes the GENPOP methodology is sufficiently accurate to
establish the magnitude of MIRs for all utilities.
9-16
-------
9.4 REFERENCES
1. U.S. Environmental Protection Agency. Estimates of Health Risks
Associated with Radionuclide Emissions from Fossil-Fueled Steam-
Electric Generating Plants. EPA 402/R-95-16. Office of Radiation
and Indoor Air, Washington, B.C. August 1995.
2. Reeves, H. Stellar Evolution and Nucleosynthesis. Gordon and
Breach Science Publishers. New York. 1968.
3. W. Anderson, Mayneord W. V., and Turner R. C. The Radon Content
of the Atmosphere Nature. Volume 174. p. 424. 1954.
4. U.S. Environmental Protection Agency. Background Information
Document for Environmental Impact Statement on NESHAPS for
Radionuclides. EPA-550/1-89-006-1. 1989.
5. U.S. Environmental Protection Agency. Estimates of Health Risks
Associated with Radionuclide Emissions from Fossil-Fueled Steam-
Electric Generating Plants-Addendum. EPA 402/R-95-16a. Office of
Radiation and Indoor Air, Washington, B.C. August 1995.
6. U.S. Environmental Protection Agency. Estimating Radiogenic
Cancer Risks. EPA-402-R-93-076. Office of Radiation and Indoor
Air. Washington, B.C. 1994.
7. National Council on Radiation Protection and Measurements. A
Guide for Uncertainty Analysis in Dose and Risk Assessment
Related to Environmental Contamination. NCRP Commentary #14,
May 10, 1996. p. 54.
8. National Council on Radiation Protection and Measurements. NCRP
Report 93, 1993.
9. National Council on Radiation Protection and Measurements.
Exposures for the Population of the U.S. and Canada from Natural
Background Radiation. NCRP Report 94, 1994.
10. Health Risks from Low-Level Environmental Exposure to
Radionuclides, Federal Guidance Report No. 13-Part 1, Interim
Version, EPA 402-R-97-014, U.S. EPA, Washington, B.C. 1998.
11. Technical Support Document for the 1992 Citizen's Guide to Radon,
EPA 400-R-92-011, U.S. EPA, Washington, B.C. 1992.
12. Bogen, K. T. and Goldin, A. S. Population Exposures to External
Natural Radiation Background in the United States, Technical Note
ORP/SEPB-80-12, U.S. EPA, Washington, B.C. 1981.
9-17
-------
13. Sources and Effects of Ionizing Radiation, United Nations
Scientific Committee on the Effects of Atomic Radiation, United
Nations, New York, NY 1993, and K. M. Miller, Measurements of
External Radiation in United States Buildings. Radiat. Prot.
Dosim. Volume 45. 1992. pp. 535-539.
14. Ionizing Radiation Exposure of the Population of the United
States, NCRP Report No. 93, National Council on Radiation
Protection and Measurements, Bethesda, MD. 1987.
9-18
-------
10.0 SCREENING LEVEL ASSESSMENT OF MULTIPATHWAY EXPOSURES
AND RISKS TO ARSENIC EMISSIONS
10.1 PURPOSE AND SCOPE
Though arsenic is a naturally occurring element that is found in
environmental media (air, water, and soil), as well as in biota, it is
also released to the environment by anthropogenic sources, including
fossil-fuel-fired electric utility plants (i.e., utilities). Since
arsenic compounds are known to cause health effects in humans from
inhalation and ingestion exposures, and since arsenic has been found
at relatively high concentrations in animals and plants that are food
sources for humans, the potential impact of arsenic emissions from
utilities on human health was evaluated. Human exposure to arsenic
through multiple exposure pathways was assessed. Though inhalation
exposure assessments have been performed for arsenic emissions from
many types of sources, few assessments have examined non-inhalation
exposures to arsenic emissions from anthropogenic sources. Specific
objectives of this analysis were: (1) to assess the magnitude of the
contribution of arsenic emissions from utilities to concentrations in
environmental media and biota; (2) to assess potential human exposure
to arsenic emissions through multiple exposure pathways, including
ingestion; (3) to identify dominant pathways of potential exposure to
arsenic emitted from coal-fired and oil-fired utilities, and (4) to
characterize potential human health risks from exposure to arsenic
emissions from utilities.
10.1.1 Rationale and Usefulness of Model Plant Approach
Arsenic is generally present as a low-level contaminant in coal
and oil. During combustion arsenic is volatilized from coal and oil
and released to the atmosphere. For this assessment, it was not
possible to model the emission impact of every utility plant.
Consequently, the actual arsenic emission data and facility
characteristics for any specific source were not modeled. Instead, a
model plant approach was used to represent actual sources. The model
plants were designed to characterize the arsenic emission rates and
the atmospheric release processes exhibited by typical facilities in
each of the four source classes considered. The modeled facilities
were not designed to exhibit extreme sources (e.g., facilities with
the highest arsenic emission rates) but rather to serve as
representatives of the combustion source class.
In taking the model plant approach, it was realized that there
would be a great deal of uncertainty about the predicted fate and
transport of arsenic and about the ultimate estimates of exposure.
The uncertainty can be divided into modeling uncertainty and parameter
uncertainty. Parameter uncertainty can be further subdivided into
uncertainty and variability depending on the level to which a
particular model parameter is understood. A limited quantitative
analysis of uncertainty is presented. It is also hoped that the
direction of future research can be influenced toward reducing the
identified uncertainties that significantly impact key results.
10-1
-------
For the assessment, a series of fate, transport, and exposure
models were used to predict arsenic concentrations in environmental
media, pertinent biota, and arsenic contact with humans. An effort
was made to estimate the amount of receptor contact with arsenic as
well as the oxidative state and form of arsenic contacted. No attempt
was made to estimate an internal dose.
Three models were used to predict environmental arsenic
concentrations and exposure: the Regional Lagrangian Model of Air
Pollution (RELMAP), the Industrial Source Complex Short-Term Version 3
(ISCST3), and the Indirect Exposure Model 2 (IEM2).
10.2 BACKGROUND INFORMATION ON ARSENIC
10.2.1 Forms of Arsenic in the Environment
Arsenic has five electrons in its outer shell; hence, it has four
possible oxidation states: +5, +3, 0, and -3. Arsenic is rarely
found in the environment as a free element. The two primary valence
states of arsenic are the trivalent state, which is denoted by As3+/As3~
or As(III), and the pentavalent state, which is denoted by As5+ or
As(V). Arsenic in each of these valence states forms both organic and
inorganic compounds (an organic arsenic compound is one in which the
arsenic atom is covalently attached to at least one carbon atom).
Arsenic compounds are typically classified in two different manners:
(1) according to the oxidation state of arsenic (As3", As3+, and As5+) ,
or (2) according to whether or not arsenic is in the organic form.
Table 10-1 shows common arsenic compounds and their classification by
valence state and organic/inorganic form.
As (III) is more mobile and soluble than the other common form,
As(V).1 Organic arsenic is present in most soils2 and is found most
often in the environment in combination with oxygen, chlorine, and
sulfur. Inorganic arsenic occurs naturally in geologic formations,
where its most common form is arsenopyrite (FeAsS).
Figure 10-1 presents a generalized scheme for the geochemical
cycling of arsenic through various compartments of the environment.
The atmosphere is a major conduit for arsenic emitted from
anthropogenic sources to the balance of the cycle via the wet and dry
deposition process.3 Dry and wet deposition from the atmosphere onto
soils may be followed by movement through soils either into
groundwater or surface water. Passage of arsenic into surface waters
may then be followed by further transfer to sediments.
The existence of chemical and biochemical transformations that
occur within the cycle makes the environmental cycling of arsenic more
complicated. Trivalent arsenic in the atmosphere can undergo
oxidation to the pentavalent state. Such conversion can also occur in
aerated surface waters. On the other hand, pentavalent arsenic in an
aqueous medium which is somewhat acidic is an oxidant, and, in the
presence of oxidizable material, will react to form trivalent arsenic.4
10-2
-------
Table 10-1. Common Arsenic Compounds, and Classification by
Valence State and Organic/Inorganic5'6
Type
Inorganic
Organic
As(lll)
Arsenopyrite
Arsenite
Arsenous acid
Arsenic trioxide
Arsenobetaine
Arsenocholine
Tetramethylarsonium ion
As(V)
Arsenate
Arsenic acid
Arsenic pentoxide
Monomethylarsonic acid
Dimethylarsinic acid
Trimethylarsine oxide
Dimethylarsinylethanol
JNHALATIQNOF_DUST _
AND GASEOUS FORMS
OF ARSENIC
BIOSPHERE
PLANTS:
1 ANIMALS
DEGRADATION
DEGRADATION
AND
SOLUTION
ABSORPTION
AND
ADSORPTION
ATMOSPHERE
T
VAPORIZATION
PRECIPITATION
HYDROSPHERE
WATER
ISEDIMENTS
CHEMICAL
PRECIPITATION AND
SEDIMENTATION OF
SOLIDS
T
PRECIPITATION
SOLUTION
PEDOSPHERE
SOILS
GLACIAL MATERIALS
DUST
CHEMICAL
PRECIPITATION
LITHOSPHERE
ROCKS
ARSENIC-BEARING
DEPOSITS
SOLUTION AND
MECHANICAL
ENGINEERING
PRECIPITATION AND
CONSOLIDATION OF SOLIDS
Figure 10-1. The generalized geochemical cycle for arsenic.
,9
10-3
-------
10.2.2 Sources of Arsenic
The primary source of arsenic emissions in commercial processes
is as a by-product of the treatment of copper, lead, cobalt, and gold
ores. It is used in the production of commercial products such as
agricultural products, wood preservatives, animal feed additives,
medicine, ceramics, and glass, and in the dying and printing
processes.9 Estimates of the impact of anthropogenic activities that
release arsenic on the global arsenic cycle vary widely, but the
impacts on the local arsenic cycle can be significant.
As noted in Table 10-2, hazardous waste incineration is the most
significant source of anthropogenic emissions of arsenic to air for
the United States. Hazardous waste incineration is estimated to
result in 1,742 tons of arsenic emissions annually, 88 percent of the
total national atmospheric arsenic emissions from anthropogenic
sources. Industrial source combustion, followed by utility
combustion, are the second and third highest emitters of arsenic,
respectively. Industrial source combustors emit an estimated 67 tons
of arsenic annually, and utilities emit approximately 65 tons of
arsenic each year.
10.2.3 Arsenic in the Atmosphere
Arsenic air concentrations in unimpacted areas are generally
below several nanograms per cubic meter. However, values near some
combustion facilities can be significantly higher, as
evidenced by measurements taken outside of the United States.
Reported values in the literature are provided in Table 10-3.
Two studies document initial results from monitoring for wet and
dry deposition of selected trace elements at two Maryland
shoreline sites in close proximity to Chesapeake Bay waters.10'11 Both
sites are located in rural settings but receive trace metal inputs
from the Baltimore-Washington DC metro area and the Ohio Valley. The
Ohio Valley is heavily industrialized, and sources in the Baltimore
and Washington DC area include several coal and oil-fired utilities as
well as other non-utility sources. Measured deposition rates for
arsenic are summarized in Table 10-4.
10.2.4 Arsenic in Water
Arsenic concentrations in most U.S. rivers and lakes range from
less than 10 to over 1000 //g/L.12 A summary of measured values found
in precipitation, groundwater, lakes, rivers, and oceans is provided
in the EPA draft Screening Level Multipathway Exposure Analysis for
Arsenic .13
It is expected that As(V) will dominate fresh water bodies that
are approximately at equilibrium, and that a large fraction of the
arsenic carried to the water body will be deposited in the benthic
sediment.* As(V) is significantly more soluble in water than As(III)
(e.g., 630 g of arsenic pentoxide will dissolve in 100 g of water,
compared to only about 2 g for arsenic tridoxide at typical ambient
temperatures) .-
10-4
-------
Table 10-2. National Arsenic Atmospheric Emission Estimates by
Source Category14
Facility type
Hazardous waste incineration
Industrial combustion
Utility combustion
Primary copper smelting
Commercial/institutional combustion
Copper ores, mining
Pressed and blown glass and glassware manufacturing
Primary nonferrous metals
Landfill waste gas flares
Secondary nonferrous metals
Turbines - distillate oil
Agricultural chemicals manufacturing
Sewage sludge incineration
Wood preserving
Medical waste incineration
Municipal waste combustion
Industrial inorganic chemicals manufacturing
Other anthropogenic sources
Total
Tons/year
1,742
67
65
47
32
13
6.1
6.0
5.9
1.6
0.81
0.47
0.41
0.41
0.21
0.20
0.20
0.46
1,989
Percent
88
3.4
3.3
2.4
1.6
0.65
0.31
0.31
0.3
0.081
0.041
0.023
0.021
0.021
0.011
0.010
0.010
0.029
100
Table 10-3. Reported Arsenic Air Concentrations
Type
Remote areas
Kaarvatn, N.Europe
Nordmoen, N. Europe
Prestebaake, N. Europe
Chesapeake Bay (Wye, Elms)
Near coal-fired power plant (Czech)
Urban areas
Mean (ng/m3)
<21
0.5
1.2
0.5
0.69, 0.625
19000-69000
<160
Max (ng/m3)
1
2
1.8
1.96, 1.56
Reference
Eisler1994
Pacyna et al. 1989
Wu etal. 1994
Eisler1994
10-5
-------
Table 10-4. Measured Arsenic Deposition Rates3
Location
Lake Michigan
Chesapeake Bay
Chesapeake Bay
Arsenic dry deposition rate
(^g/m2/yr)
35
54-150
-
Arsenic wet deposition rate
(^g/m*/yr)
-
-
45-52
Reference
Wuetal. 1994
Wuetal. 1994
Scudlark et al.
1994
These deposition rates are the result of emissions from a combination of natural and anthropogenic sources.
Utilities are only one of the many source categories that may be contributing to the deposition rates.
10.2.5 Arsenic in Sediments
Arsenic concentrations in sediments of various rivers and harbors
in the United States have been reported by Brannon and Patrick (1987)
as 1.0 to 8.8 //g/g dry weight, and much higher concentrations have
been observed at contaminated sites. Sediments often act as a sink
for arsenic in natural systems and can hold high concentrations of
arsenic. At contaminated sites, such as Sugar Creek, arsenic
concentrations ranging from 4,470 to 66,700 //g/g dry weight were
measured.* A summary of measured sediment concentrations is provided
in the EPA draft Screening Level Multipathway Exposure Analysis for
Arsenic.—
10.2.6 Arsenic in Soil
The majority of soils contain levels of arsenic that vary between
1 and 5 mg/kg, with much higher concentrations being observed around
anthropogenic sources and areas with high background sources. Arsenic
also ends up in soil as a result of the historical direct applications
of arsenical pesticides (which are being phased out in some areas and
are prohibited in others), landfarming of sewage sludge, deposition
from air pollution, and as waste materials from industrial processes.
Soil concentrations have been found to decrease rapidly with distance
from an elevated point source and have been reported to decline to
background concentrations within 8 to 16 km of the source.-
Ambient soil concentrations given in Bowen range from 0.1 to 40
mg/kg, with a mean of 6;15 and Vinogradov16 estimated a range of 1 to 50
mg/kg with a mean of 5.— A range in the United States was given by
Stater, et al. of 1 to 20 mg/kg, with a mean of 7.5 mg/kg (n=52),17 and
by Conner and Shacklette as 1.6 to 72 mg/kg, with a mean of 7.5 mg/kg
(n=1215).18 Wood and Duda19 compiled background arsenic concentrations
in soils sampled from the southeastern United States (NC, SC, GA, and
AL) and found that mean arsenic concentrations ranged from 0.69 to
13.3 mg/kg." Cullen and Reimer report that the average arsenic
concentration in the continental crust is 3 //g/g.20
There are few data on either the speciation of arsenic in soil or
the determination of the fraction of arsenic in soil that is inorganic.
10-6
-------
Changes in season, temperature, pH, and relative reducing/oxidizing
conditions will initiate kinetically slow conversions between As(III)
and As(V) species. In aerobic soils, As(V) is the dominant species,
while As(III) may dominate in reduced soils (such as those that are
temporarily flooded or groundwater wells). In very reduced soils
(such as swamps), arsenic can be found in its elemental state, arsine,
or as methylated arsenic. The results in Takamatsu et al.21 and
Bombach et al.22 indicate that generally less than 20 percent of the
arsenic in soil is As (V)— •—
The bioavailability of arsenic in soils contaminated either
through smelter operation or mine tailings has been assessed.23 In a
recent study, immature swine were dosed with the contaminated soils
(immature swine were selected because of the similarity of their
gastrointestinal (GI) tracts to those of human children). The
absorption from the GI tract into the bloodstream was monitored. The
relative (mean) bioavailability of soil arsenic was 78 percent
(C.I. = 56-111 percent) and the absolute (mean) was 52 percent
(C.I. = 44-61 percent).
10.2.7 Arsenic in Terrestrial Plants
Arsenic does not readily translocate to the shoots of plants and
is found mostly in the roots. In general, arsenic uptake values are
low. Studies have found that, to achieve 1 mg/kg of arsenic in fresh
weight plants, a soil concentration of 200-300 mg/kg is necessary. 24-25-27
Measured concentrations in a variety of food and other plants are
shown in Table 10-5.
10.2.8 Arsenic in Aquatic Plants
Arsenic is not a major contaminant of aquatic plants, except in
severe cases of pollution. Highest reported concentrations in aquatic
plants are the result of mine and smelter wastes, reaching levels as
high as 1,450 mg/kg dry weight.28 Arsenic concentrations in some
untreated areas have been measured as ranging from 1.4 to 13 mg/kg dry
weight.—
10.2.9 Arsenic in Terrestrial Animals
Some studies have found arsenic can accumulate in meat/dairy
products with most arsenic compounds accumulating in the liver and
kidney.29'30 However, reported concentrations are rare. Table 10-6
shows values reported in the literature.
10.2.10 Arsenic in Fish
Arsenic does not appear to bioaccumulate or bioconcentrate in
freshwater finfish. However, it does appear to bioconcentrate in
freshwater bivalves. Lower trophic level organisms generally appear
to have higher concentrations of arsenic than predatory or omnivorous
aquatic and marine species.
10-7
-------
Table 10-5. Measured Arsenic Concentrations in Plants
Terrestrial plant
Concentration (dry
weight mg/kg) [mean,
(range)]
Reference
Edible plants
Lettuce
Spinach
Tomato
Carrot
Potato
Wheat
Barley
Oats
Apple
0.056 (0.01 2-0.68) a
0.72*
0.076 (0 - 0.20)a
0.056 (0.024 -0.1 2) a
1.1 6a
0.004 (0.001 2 -0.01 2) a
0.4a
0.12(0.024-0.44)a
0.064 (0.01 2- 0.20) a
0.04(0-1.16)a
0.24 (0.024 -1.44)a
< 0.24a
0.32 (0.024 - 1 .88)a
0.92 (0.44 -2.72)a
<0.24(0.16-0.4)a
0.08 (0.008 -1.08)a
0.12(<0.04-1.92)a
Wiersmaetal. 1986
Wiersma et al. 1986 (Barudi and Bielig 1980)31
Wiersma et al. 1986 (Jelinek and Corneliussen
1977)
Wiersmaetal. 1986
Wiersma et al. 1986 (Barudi and Bielig 1980)
Wiersmaetal. 1986
Wiersma et al. 1986 (Barudi and Bielig 1980)
Wiersma et al. 1986 (Jelinek and Corneliussen
1977)
Wiersmaetal. 1986
Wiersma et al. 1986 (Jelinek and Corneliussen
1977)
Wiersmaetal. 1986
Wiersma et al. 1986 (Varo et al. 1980)32
Wiersmaetal. 1986
Wiersmaetal. 1986
Varoetal. 1980
Wiersmaetal. 1986
Reinhard 1974
Impacted/arsenic-treated areas
Alfalfa from Montana smelter area
Various species from impacted soils (80
mg/kg)
(2.0-28.4)a
(<0.2-5.8)
Jenkins 1980«
Merry et al. 198633
Grass/fodder crops
Colonial bent grass on low arsenic soil
Grasses on non-treated areas
Scotch heather on low-arsenic soil
Grass (soil concentration 1-38 /^g/g dry
weight, mean 1 1 /4J/g)
Silage maize (soil concentration 1-110
/4j/g dry weight, mean 10 ,ug/g)
Sugar beet crowns & leaves (soil
concentration 1 -36 /^g/g dry weight,
mean 14 /^q/q)
(0.3-3)
(0.1-0.9)
0.3
1. 4 (0.36- 5.56) a
0.88 (0.28 - 2.36)
2.56 (0.44 - 8.0)
Jenkins 1980
NRC 1977
Jenkins 1980
Wiersmaetal. 1986
Wiersmaetal. 1986
Wiersmaetal. 1986
a Converted from freshweight values assuming water content of 0.8.
b These values are for the Netherlands. A range of mean soil concentrations of 5 - 12/^g/g dry weight is reported.
10-8
-------
Table 10-6. Measured Arsenic Concentrations in Meat and Other
Animal Products
Meat
product
Beef
Beef Liver
Milk
Pork
Poultry
Eggs
Lamb
Wet weight
(M9/9)
0.005 ±0.001
0.008 ±0.001
<0.001
<0.1
0-0.5
0-0.2
0.002 ±0.001
Assumed moisture
content
0.615a
0.70a
0.87a
0.615b
0.615b
?
0.67
Dry weight
(M9/9)
0.013 ±0.003
0.027 ±0.003
<0.003
<0.3
0-1.30
0.006 ±0.003
Reference
Vreman et al. 1986
Vreman et al. 1986
Vreman et al. 1986
Jelinek and Corneliussen
1977C
Jelinek and Corneliussen
1977C
Jelinek and Corneliussen
1977C
van der Veen and Vreman
1986
aBaesetal. 1984
b Assumed same as beef.
c Detection limit of 0.1 /^g/g. Measurement is for Asf>3
The species of arsenic present in fish appear to be somewhat
variable. Marine organisms can convert inorganic arsenic into organic
arsenic species, with arsenobetaine being both the major form
identified in fish and the suspected metabolic endpoint for arsenic in
the marine environment. Arsenobetaine, or "fish-arsenic," an organic
arsenic species, is a relatively nontoxic form of arsenic. There is
little evidence of human toxicity due to arsenobetaine exposure.
Arsenocholine and inorganic arsenic species have also been identified
as forms of arsenic in fish. Arsenocholine may be more toxic than
arsenobetaine, but these results are questionable. Inorganic
arsenicals are rapidly converted to less toxic organic forms of
arsenic in fish.34 The primary arsenic form in marine fish is
arsenobetaine, which comes from arsenosugars present in algae and not
from the formation of trimethylarsine oxide.
Table 10-7 shows a summary of measured freshwater fish
concentrations of total arsenic. Even though shellfish and other
marine foods contain the greatest arsenic concentrations, much of the
arsenic present in fish and shellfish exists in the less toxic organic
form.35 In shellfish, 0-2.9 percent of the total arsenic is generally
measured to be in inorganic forms; in finfish, 0-9.5 percent is
generally measured as inorganic.
10-9
-------
Table 10-7. Total Arsenic Concentrations in Freshwater Fish in
the United States
Freshwater fish
Bass - muscle
Bluegill- nationwide- whole
Catfish - native- muscle
Catfish - cultured- muscle
Common carp - nationwide - muscle
Northern pike - northern U.S. - muscle
Coho salmon- USA- muscle
Atlantic salmon - muscle oil
Rainbow trout- all tissues
Various species - USA, 1976 - 1984 -
White sucker - Muscle
Lake trout- nationwide- whole
Concentration range
(Wet weight mg/kg)
0.0-0.51
0.05-0.4
0.0-0.3
0.2-3.1
0.0-0.2
<0.01-0.1
0.07-0.17
0.8-3.1
<0.4
0.14- 2.9
0.03-0.13
0.06-0.68
Reference
Jenkins 1980*
Jenkins 1980; Wiener et al. 198436
Jenkins 1980
Jenkins 1980
Jenkins 1980
Jenkins 1980
Jenkins 1980
Jenkins 1980
Jenkins 1980
Lima et al. 198437; Schmitt & Brumbaugh
Jenkins 1980
Jenkins 1980
10.2.11 Speciation of Arsenic in Food Products
Though total arsenic levels have been measured in a variety of
food products, limited data are available on concentrations of
particular arsenic species in food products. The chemical forms of
arsenic in foods are varied and complex. Table 10-8 lists reported
percentages of inorganic arsenic in foods. The form found in fish and
marine foods was discussed above. Other food products, such as meats,
rice, and cereals, contain higher percentages, and often higher total
amounts, of inorganic arsenic.
10.2.12 Arsenic Near Anthropogenic Sources
Increased arsenic concentrations have been measured in
environmental media surrounding anthropogenic atmospheric
emission sources, including copper and lead smelters. The measured
data indicate an elevation of arsenic concentrations that is related
to the emission of arsenic and subsequent deposition to the
terrestrial environment. Table 10-9 shows a summary of values. Many
of these data were collected at sources that operated prior to the use
of modern particulate control technology, and the facilities are not
of the type considered in this report.38 However, they are included
because the data indicate that some arsenic sources can have an
observable impact on their local environment.
10.3 SUMMARY OF MODELS AND APPROACH
In order to address the objectives of this study, a series of
modeling efforts were performed. The Regional Lagrangian Model of Air
Pollution (RELMAP) model was used to model the long-range transport of
arsenic from coal- and oil-fired utilities. A separate effort
involved modeling the atmospheric dispersion and deposition of arsenic
on a local scale (within 50 km), using the Industrial Source Complex
10-10
-------
Table 10-8. Percentage of Inorganic Arsenic Compared to Total
Arsenic in Selected Foods
Food
Milk and dairy products
Meat- beef and pork
Poultry
Saltwater fish3
Freshwater fish (0-9.5 percent)3
Cereals
Rice
Vegetables
Potatoes
Fruits
Percent Inorganic As
75%
75%
65%
0%
10%
65%
35%
0.5%
10%
10%
3 The values presented for fish are not widely accepted.
Source: Weiler (1987) in Borum and Abernathy (1994) as cited in IF Kaiser, 1996.
Table 10-9. Environmental Concentrations near Facilities"
Facility Value(s) Reference
Air (ng/m3)
Former USSR
Texas
Tacoma, Washington
Romania
Germany
500-1900
max. 1400
max. 1 500
max. 1600
900-1500
Pershagen, G. and Vahter, M. 1979
Pershagen, G. and Vahter, M. 1979
Pershagen, G. and Vahter, M. 1979
Pershagen, G. and Vahter, M. 1979
Pershagen, G. and Vahter, M. 1979
Drinking water (,ug/L)
Mexico, from plant producing As2O3
Japan, near factory producing As sulfide
4000-6000
3000
Pershagen, G. and Vahter, M. 1979
Pershagen, G. and Vahter, M. 1979
Dusf (mg/kg or/sg/g)
Tacoma, WA, near smelter
1300 (remote from
smelter 70)
NRC 1977
Sediments (mg/kg dry weight, /^g/g)
Near sewer outfall
35
NRCC1978
So// (mg/kg or/sg/g) dry weight
Tacoma, WA, near smelter
Japan
Max. 380
max. 2470
Pershagen, G. and Vahter, M. 1979
Pershagen, G. and Vahter, M. 1979
Fish, near smelter (water arsenic 2.3-2.9 //g/L), ,ug/g freshweight
Total arsenic
Inorganic arsenic
0.05 - 0.24
0.01 - 0.02
Norinetal. 1985.
Norin, H., M. Vahter, A. Christakopoulos, and
M. Sandstroem 1985.
Source: Eisler(1994)
10-11
-------
Short-Term Model 3 (ISCST3). Finally, using the predicted air
concentrations and deposition rates, the Indirect Exposure Model 2
(IEM2) model was used to predict environmental concentrations and
subsequent exposure for several hypothetical exposure scenarios.
10.3.1 Source Classes Considered and Model Plant Approach
Four different arsenic source facilities were considered in the
local-scale aspect of this study: three coal-fired utilities (a
small, a medium, and a large facility) and an oil-fired utility. A
list of these facilities and their emission rates is provided in
Table 10-10. Details about these facilities and how the
parameterizations for the model plants were performed are provided in
Appendix A of the EPA Draft Screening Level Multipathway Exposure
Analysis for Arsenic."
10.3.2 Atmospheric Transport Modeling
10.3.2.1 Local Scale Modeling. The ISCST3 was used to estimate
the atmospheric dispersion and deposition of emitted arsenic within a
50 km radius of the facilities.39'40 The ISCST3 uses hourly
meteorological data to estimate the ambient air concentrations of an
emitted pollutant, as well as the wet and dry deposition rates. The
dry deposition of the air pollutant is calculated, based on particle
size, atmospheric conditions, and gravitational settling velocities.
Wet deposition of the pollutant is based on scavenging coefficients
that depend on particle size and the precipitation rate.
For each of the four facilities considered, ISCST3 was run in
both a humid and an arid site, with receptors placed at 16 directions
around the facility and 30 distances between 200 m and 50 km, for a
total of 480 receptors. Simple (i.e., flat) terrain was assumed.
Values at other locations were estimated using linear interpolation,
as discussed in Appendix A of EPA Draft Screening Level Multipathway
Exposure Analysis for Arsenic." Area-averaged air concentrations and
deposition rates were estimated, for use in the exposure modeling.
10.3.2.2 Regional Transport Modeling. Long-range transport was
modeled for arsenic. The methods, model, and approach are described
in detail in chapter 6 (section 6.6). Figures 10-2 through 10-4 show
the results of the arsenic RELMAP modeling.
10.3.3 Indirect Exposure Modeling
Because of its chemical and physical characteristics, arsenic
emitted to the atmosphere may be transported to other environmental
media (i.e., soil or water), thus allowing noninhalation exposures to
arsenic to occur. The IEM2 was used to predict the terrestrial and
aquatic fate and transport of arsenic deposited in the region of
interest, as well as the human exposure to arsenic. The IEM2
calculated arsenic concentrations in watershed soils; these
concentrations were then used in calculating concentrations in various
food plants. The waterbody component of the IEM2 model calculated
arsenic concentrations in surface water and in aquatic organisms. In
10-12
-------
Table 10-10. Summary of Model Plants and Emission Rates Used for
the Assessment
Model plant
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
Stack/source
Stack 1
Stack 2
Stack 3
Stack 1
Stack 2
Stack 1
Stack 2
Stack 1
Stack 2
Total arsenic emissions
kg/yr
85.0
74.0
98.0
32.0
30.0
15.4
8.8
14.6
18.8
g/s
2.70e-3
2.35e-3
3.11e-3
1.01e-3
9.516-4
4.88e-4
2.79e-4
4.63e-4
5.96e-4
addition, the IEM2 model calculated human exposures for selected
exposure scenarios through multiple exposure routes, including food
consumption, water ingestion, and inhalation.
10.3.3.1 Description of Exposure Scenarios. Three basic
exposure scenarios were considered: a subsistence farmer (adult and
child), a subsistence fisher (adult and child), and a pica child.
These scenarios were considered because they represent possible high-
end scenarios for exposure to arsenic. Table 10-11 summarizes the
exposure pathways considered for each of these scenarios.
Table 10-12 shows the default values for the scenario-independent
parameters for both the child and adult receptors, and Table 10-13 shows
the default values for the scenario-dependent exposure parameters (the
technical bases for these values are provided in Appendix B of the EPA
Draft Screening Level Multipathway Exposure Analysis for Arsenic).—
The subsistence farmer scenario consists of a subsistence farmer
and child who consume elevated levels of locally-grown food products.
It was assumed that each farm was located on approximately 10 acres.
The subsistence farmer was assumed to raise livestock and to consume
home-grown animal tissue and animal products, including chickens,
eggs, beef, and dairy products. All chicken feed was assumed to be
derived from non-local sources (and is, therefore, not contaminated
with arsenic). For bovine consumption of contaminated feed, 100
percent of the hay and corn used for feed was assumed to be from the
affected area. It was also assumed that the drinking water for the
subsistence farmer comes from rainwater collected in cisterns. Though
rainwater collected in cisterns may not be the primary source of
drinking water for most farmers, some are still expected to use
rainwater collected in cisterns as the primary source. Since
rainwater is likely to have the highest arsenic levels due to arsenic
emissions from anthropogenic sources, this assumption is consistent
with the high-end exposure scenario for this screening level assessment.
10-13
-------
0.05 to 2
2 to 5
5 to 10
10 to 20
> = 20
Figure 10-2. Predicted Arsenic Wet and Dry Deposition from Coal Utilities Based
on 1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
Wet + Dry Deposition - Arsenic from Oil Utilities
Micrograms per Square Meter per Year
o
I
0.05 to 2
2 to 5
5 to 10
10 to 20
> = 20
Figure 10-3. Predicted Arsenic Wet and Dry Deposition from Oil Utilities Based on
1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
Wet + Dry Deposition - Arsenic from Coal and Oil Utilities
Micrograms per Square Meter per Year
0.05 to 2
2 to 5
5 to 10
10 to 20
> = 20
Figure 10-4. Predicted Arsenic Wet and Dry Deposition from Coal and Oil Utilities
Based on 1990 Emissions Estimates as Modeled with RELMAP, Units: ug/m2/yr
-------
Table 10-11. Summary of Human Exposure Scenarios Considered
Exposure route
Air inhalation
Soil Ingestion
Animal ingestion
Vegetable ingestion
Fish ingestion
Water ingestion
Subsistence farmer
Adult
X
X
X
X
X
Child
X
X
X
X
X
Pica
child
X
X
Subsistence fisher
Adult
X
X
X
X
X
Child
X
X
X
X
X
Blank = Pathway not considered; X = Pathway considered.
Table 10-12.
Parameters
Default Values of Scenario-Independent Exposure
Parameter3
Body weight (kg)
Exposure duration (years)
Inhalation rate (m3/day)
Vegetable consumption rates (g/kg body weight/day)
Leafy vegetables
Grains and cereals
Legumes
Potatoes
Fruits
Fruiting vegetables
Animal product consumption rates (g/kg body weight/day)
Beef (excluding liver)
Beef liver
Dairy
Pork
Poultry
Eggs
Lamb
Soil ingestion rates (g/day)
Water ingestion rate (L/day)
Value"
Adult
70
30
20
0.028
1.87
0.381
0.17
0.57
0.064
0.341
0.066
0.599
0.169
0.111
0.093
0.057
0.1
2
Child
17
18
16
0.008
3.77
0.666
0.274
0.223
0.12
0.553
0.025
2.04
0.236
0.214
0.073
0.061
Scenario-dependent
1
a All human consumption rates except for soil and water are reported as dry weight.
b See Appendix B of the EPA draft Screening Level Multipathway Exposure Analysis for details regarding these parameter values.
10-17
-------
Table 10-13. Values for Scenario-Dependent Exposure Parameters3
Parameter
Finfish ingestion rates (g/day)
Soil Ingestion rate (g/day)
Contact time for inhalation (hr/day)
Contact fractions (unitless)0
Animal products
Leafy vegetables
Grains and cereals
Legumes
Potatoes
Fruits
Fruiting vegetables
Root vegetables
Drinking water"
Subsistence farmer
Adult
NA
0.1
24
1
1
1
1
1
1
1
1
1
Child
NA
0.2
24
1
1
1
1
1
1
1
1
1
Pica
Child
NA
7.5
24
NA
NA
NA
NA
NA
NA
NA
NA
NA
Subsistence fisherb
Adult
60
0.1
24
NA
0.058
0.667
0.8
0.225
0.233
0.623
0.268
1
Child
20
0.2
24
NA
0.058
0.667
0.8
0.225
0.233
0.623
0.268
1
a See Appendix B of the EPA draft Screening Level Multipathway Exposure Analysis for Arsenic for more details regarding these
values.
b The subsistence fisher scenario does not consider consumption of freshwater bivalves or marine organisms.
0 Contact fraction is the fraction of the total consumption of a food product from the study site.
d The source of the contaminated drinking water is different for the subsistence farmer and fisherperson scenarios.
NA - Not considered to be applicable to this assessment.
For the urban high-end scenario, a pica child was defined as
consuming 7.5 g of soil per day at the location of maximum deposition
of arsenic. Estimates of the rate of soil ingestion by pica children
range from 5 g/day to 50 g/day. The data of Calabrese et al. (1989)
show a range of 5-10 g/day. A value of 7.5 grams was considered
appropriate to represent this subpopulation of children.
The subsistence fisher scenario consisted of a subsistence fisher
and child whose fish and water consumption scenarios were associated
with the hypothetical lake setting. The high-end fish consumer
scenario represented an individual who was assumed to ingest large
amounts of locally-caught fish, as well as home-grown garden produce
(plant ingestion parameters identical to the rural home gardener
scenario) and drinking water from the affected lake. These
10-18
-------
consumption scenarios were thought to represent identified fish-
consuming subpopulations in the United States. No commercial
distribution of locally caught fish was assumed. Fish consumption
rates for the fish-consuming subpopulations were derived from the
Columbia River Inter-Tribal Fish Commission Report.41
All arsenic was assumed to be of the inorganic form when
estimating exposure, except for fish consumption. Based on the
estimates of 0.5 to 75 percent inorganic arsenic in a variety of food
products (see Table 10-8) , particularly percentages of 65-75 percent
for meat, dairy products, grains, and cereals, a conservative estimate
of 100 percent was selected, since the assessment is being performed
as a screening-level assessment. For freshwater fish, for which more
measurements have been taken, it is assumed that 10 percent of the
arsenic is inorganic arsenic. This is also a conservative estimate,
since lower levels of inorganic arsenic are generally measured in
fish, but it is appropriate to select a conservative number for this
screening-level assessment.
10.3.3.2 Description of Waterbody/Watershed Configuration. The
watershed was assumed to be a circular region with a radius of 3.5 km,
and the waterbody was assumed to be a small circular lake with a
radius of 0.9 km (i.e., the ratio of watershed to waterbody is 15).
Three different locations of the lake within the watershed were
considered; these locations are shown in Figure 10-5. The area-
averaged values for the watersheds and waterbodies were calculated at
five distances (0.2, 2, 5, 10 and 25 km; where "distance" is the
distance between the closest point on the watershed and the facility)
and for 16 directions around each facility; thus, there are a total of
240 different watershed/waterbody configurations per facility.
10.3.4 Determination of Background Values
To assist in determining the reasonableness of the IEM2 model
parameterization, model runs were performed using background soil
concentrations, air concentrations, and deposition rates. Of these
parameters, the soil concentration was found to be the most important
because it was critical to the estimation of arsenic concentrations in
plants and to the estimation of the flux to the water body. In this
section, the values assumed for arsenic background soil
concentrations, air concentrations, and deposition rates are
discussed.
In this assessment, background arsenic concentrations were
defined as the natural arsenic levels in the soil and air that would
exist without any anthropogenic input of arsenic. Determination of
such a background level presents a formidable challenge: it is
difficult to quantify; there can be considerable variability in what
constitutes background; and, at present, there is little guidance on
how background should be determined. However, due to the naturally
high concentrations of arsenic in many soils, it was considered
10-19
-------
Watershed
Local
Source
3 Lake /
Locations
Considered
Figure 10-5. Location of waterbody considered within watershed.
10-20
-------
critical that such background concentrations be addressed, even if
only in a simple, screening-level manner. It is hoped that this will
serve as a starting place for future arsenic assessments.
Natural and anthropogenic arsenic soil concentrations show a wide
range of variation. Even crustal levels of arsenic can range from 0.1
to several hundred mg/kg. Therefore, it is difficult to establish
typical ambient levels of arsenic in soil without respect to a
particular geological formation and geographical area, and general
comparisons must be made with caution. For the purpose of this
modeling exercise, a background soil concentration of 3 //g/g was
assumed. This is the average arsenic concentration for the
continental crust reported in Cullen and Reimer."
Chilvers and Peterson' reported that the current natural levels
of arsenic in the atmosphere account for approximately 60 percent of
the total atmospheric load, due primarily to volcanic sources and
biological methylation. Based on this, the mean air concentration of
0.7 ng/m3, reported in Wu et al.— for the Chesapeake Bay and assumed to
be representative for a relatively unimpacted area, was multiplied by
0.6 to obtain a background arsenic air concentration of 0.4 ng/m3. It
is important to note that the ultimate impact of the air concentration in
this case is minimal, as the assumed soil concentration drives the
concentrations in biota and the water body. The dry deposition rate was
assumed to be 60 //g/m2/yr (=0.6*100; see Table 10-4),"and the wet
deposition rate was assumed to be 30 //g/m2/yr ( = 0.6*50; see Table 10-4).42
Few data were located on the speciation of arsenic in the atmosphere.
Andreae reports that the As(V):As(III) ratio was 2:1 in rainfall,43 and
so it was assumed that this same relationship holds for the background
dry deposition rate and air concentration.
10.4 MODELING RESULTS
The following is a presentation of the results of this
assessment. Included is a discussion of the air modeling results, a
comparison of these results with measured data, and the potential
impact of these on hypothetical receptors.
10.4.1 Air Modeling Results/Comparison with Measured Data
10.4.1.1 Local Scale Modeling. The fraction of arsenic
emissions predicted to be deposited within 50 km of the source are
presented in Tables 10-14 and 10-15, representing the two different
meteorological regimes that were modeled. In general, less than 50
percent of the emissions are predicted to be deposited within 50 km,
and facilities with shorter stacks have a higher fraction of emissions
deposited within 50 km of the facility. Except for the large coal-
fired utility, dry deposition is greater than wet deposition. This is
the opposite of what has been concluded for regional models (e.g., the
regional results presented in this report and those in Alcamo et al.)44
10-21
-------
Table 10-14. Fraction of Arsenic Emissions Predicted to Be
Deposited Within 50 km in an Arid Site
Boiler type
Oil-fired utility boiler
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Stack"
Stack 1
Stack 2
Stack 1
Stack 2
Stack 3
Stack 1
Stack 2
Stack 1
Stack 2
Release
height
(m)
87
83
198
191
173
147
122
84
79
Fraction of emissions deposited within 50 km
Total Dry Wet
13% 12% 1%
14% 13% 1%
5% 4% 1 %
5% 4% 1 %
6% 5% 1 %
7% 6% 1 %
9% 8% 1 %
14% 13% 1%
14% 13% 1%
a Size distribution for stacks is mass based.
Table 10-15. Fraction of Arsenic Emissions Predicted to Be
Deposited Within 50 km in a Humid Site
Boiler type
Oil-fired utility boiler
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Stack"
Stack 1
Stack 2
Stack 1
Stack 2
Stack 3
Stack 1
Stack 2
Stack 1
Stack 2
Release
height (m)
87
83
198
191
173
147
122
84
79
Fraction of emissions deposited within 50 km
Total Dry Wet
21% 15% 6%
21% 16% 6%
10% 4% 7%
1 1 % 4% 7%
12% 5% 7%
14% 7% 7%
16% 10% 7%
23% 16% 7%
24% 17% 7%
a Size distribution for stacks is mass based.
For the large coal-fired utility, the high stacks result in a
significantly lower fraction deposited by dry deposition. The
predicted wet deposition is not as sensitive to stack height and,
therefore, does not change.
10.4.1.2 Regional Scale Modeling. RELMAP was used to model the
long-range transport of arsenic. Modifications to RELMAP for
atmospheric arsenic simulation were based on the assumption that all
arsenic emissions are in particulate form. Only the field mode of
RELMAP was used in this assessment (see Section 6.6 of Chapter 6 for
description of field mode and other details of the RELMAP). The RELMAP
regional-scale air modeling results are presented in Table 10-16.
10-22
-------
Table 10-16. RELMAP Air Modeling Results
Result
Arsenic air concentration (ng/m3)
Arsenic dry deposition rate (/4j/m2/yr)
Arsenic wet deposition rate (/^g/m 2/yr)
Western United States (>90° longitude)
50th percentile
9.00e-04
9.80e-03
7.70e-02
90th percentile
6.10e-03
8.50e-02
7.70e-01
Eastern United States
(>90° longitude)
50th percentile
1 .40e-02
7.40e-02
1 .90e+00
90th percentile
3.90e-02
7.60e-01
7.40e+00
10.4.2 Indirect Exposure Modeling
10.4.2.1 Comparison of Predicted Values with Measured Values.
Table 10-17 compares the results for the eastern and western sites.
The more humid climate of the eastern site results in slightly more
runoff, and, therefore, more influx of arsenic into the water body.
A critical factor in the water body calculations is the soil-water
partition coefficient. The value used here is 100 L/kg (see Appendix B
of the EPA draft Screening Level Multipathway Exposure Analysis for
Arsenic for Explanation)^- and is based on a combination of curve fits
to more sophisticated partitioning modeling and other literature data.
The surface water concentrations are essentially inversely
proportional to the soil-water partition coefficient. The surface
water and sediment concentrations are within the range of values
reported for lakes in Section 2. It is noted that the sediment
concentrations are not used for any subsequent modeling.
To assist in determining the reasonableness of the model
predictions, typical values for input parameters were used in the
model, and then the predicted concentrations were compared with
measured values. The predicted plant and animal concentrations are
shown in Table 10-18. The predicted concentrations are comparable
with those presented in (Tables 10-5, 10-6, and 10-7), although the
values are not consistently at one end of the range for all food
types. For example, the predicted beef values are slightly higher
(by about 6 ng/g), while most of the other food types are in the
middle or lower end of the range of concentrations reported in the
literature for what are assumed to be "unimpacted" values. Most of
the arsenic in the beef (60 percent) is predicted to come from the
ingest ion of grain/forage/silage, while the rest is predicted to come from the
ingest ion of soil.
The fish and bivalve concentrations are calculated from the
arsenic water concentration. The difference in predicted values
is due entirely to the difference in bioconcentration factors used
(1 L/kg for finfish, 350 L/kg for bivalves).
10-23
-------
Table 10-17. Predicted Surface Water and Benthic Sediment
Concentrations for the Hypothetical Water Bodies
Concentrations3
Total arsenic water concentration (//g/L)
Percent of arsenic dissolved
Predicted suspended sediment concentration (mg/L)
Total arsenic benthic sediment concentration (ug/g dry weight)
Eastern site
14
97
0.5
27
Western site
9
90
2.2
17
Assuming arsenic air concentration of 0.4 ng/m3, deposition rate of 90 ug/rrr/yr, and soil concentration of 3 ug/g .
Table 10-18. Modeled Arsenic Concentrations3
Biota
Grain
Fruits
Fruiting vegetables
Leafy vegetables
Beef
Dairy
Pork
Poultry
Freshwater fish (fresh weight)
Freshwater bivalves (fresh weight)
Modeled concentration (//g/g dry weight)
0.181
0.002
0.001
0.009
0.022
0.001
0.008
0.008
0.013
4.7
Background concentration
range (//g/g dry weight)"
0.024 - 2.72
0.008 - 1 .92
0.056-1.16
0.01 -0.016
< 0.003
<0.3
0 - 1 .30
0-3.1
a Assuming a background soil concentration (3 /zg/g), air concentration (0.4 ng/rrf ), and deposition rate (90
b Background concentration ranges were taken from data in Tables 5 and 6.
/yr).
The predicted exposures for the hypothetical receptors are shown
in Tables 10-19 (total arsenic) and 10-20 (inorganic arsenic) . The
intake of total arsenic is predicted to be dominated by the ingestion
of grains for all scenarios in which grains are consumed, except for
the adult subsistence fisher. In the case of the adult subsistence
fisher, most of the exposure to arsenic is predicted to be from the
ingestion of freshwater bivalves. Because only 10 percent of the
total arsenic in fish is assumed to be inorganic, ingestion of grains
is predicted to be the dominant food ingestion exposure to inorganic
arsenic, even for the adult subsistence fisher scenario. Exposure to
arsenic for the pica child scenario, in which only exposure to arsenic
through soil ingestion is considered, is predicted to be approximately
twice as large as that for the subsistence farmer child.
10-24
-------
Table 10-19. Predicted Total Arsenic Exposure for Hypothetical Receptors3
Scenario
Eastern site (humid)
Subsistence Farmer Adult
Subsistence Farmer Child
Pica Child
Subsistence Fisher Adult
Subsistence Fisher Child
Western site (arid)
Subsistence Farmer Adult
Subsistence Farmer Child
Pica Child
Subsistence Fisher Adult
Subsistence Fisher Child
Results for Total As
Total inhalation
intake (mg/kg/day)
1E-07
4E-07
4E-07
1E-07
4E-07
1E-07
4E-07
4E-07
1E-07
4E-07
Total ingestion
intake (mg/kg/day)
4E-04
7E-04
1E-03
5E-04
5E-04
4E-04
8E-04
1E-03
4E-04
5E-04
% for Receptor
Root_ Fruiting_ Leafy_ Beef_ Soil_
Water Grains Legumes Potatoes vegetables Fruits vegetables vegetables Beef liver Dairy Pork Poultry Eggs Lamb ingestion Fish Bivalves
0941 0 0 0 0 02100000100
0910 0 0 0 0 02000000500
0440 0 0 0 0 0000000012 52
0891 0 0 0 0 00000000730
1931 0 0 0 0 02100000100
1910 0 0 0 0 02000000500
1560 0 0 0 0 000000001 2 40
2891 0 0 0 0 00000000720
Assuming background arsenic soil concentration of 3 ^g/g, air concentration of 0.4 ng/m 3, and deposition rate of 90 ^g/m2/yr.
I-1
O
I
K>
Ul
Table 10-20. Predicted Total Inorganic Arsenic Exposure for Hypothetical Receptors3
Scenario
Eastern site (humid)
Subsistence Farmer Adult
Subsistence Farmer Child
Pica Child
Subsistence Fisher Adult
Subsistence Fisher Child
Western site (arid)
Subsistence Farmer Adult
Subsistence Farmer Child
Pica Child
Subsistence Fisher Adult
Subsistence Fisher Child
Results for Inorganic As
Total inhalation
intake (mg/kg/day)
1E-07
4E-07
4E-07
1E-07
4E-07
1E-07
4E-07
4E-07
1E-07
4E-07
Total ingestion
intake (mg/kg/day)
4E-04
7E-04
1E-03
3E-04
5E-04
4E-04
8E-04
1E-03
3E-04
5E-04
% for Receptor
Leafy_
Root_ Fruiting_ vegetable Beef_ Soil_
Water Grains Legumes Potatoes vegetables Fruits vegetables s Beef liver Dairy Pork Poultry Eggs Lamb ingestion Fish Bivalves
0941 0 0 0 0 02100000100
0910 0 0 0 0 02000000500
0861 0 0 0 0 0000000020 10
0911 0 0 0 0 00000000700
1931 0 0 0 0 02100000100
1910 0 0 0 0 02000000500
2891 0 0 0 0 00000000206
2901 0 0 0 0 00000000700
Assuming background arsenic soil concentration of 3 ^g/g, air concentration of 0.4 ng/m 3, and deposition rate of 90 ^g/m2/yr.
-------
10.4.2.2 Combined Results
10.4.2.2.1 Contribution of individual and regional sources to
total concentrations. One of the objectives of this analysis was to
assess the magnitude of the contribution of arsenic emissions from the
four types of utility boilers, as compared to concentrations in
environmental media and biota. This objective was addressed by
calculating the concentrations in media and biota with and without an
estimate of background, and by including the regional contribution of
nationwide utilities. These results are shown in Tables 10-21 through
10-25.
When background levels are included and watershed area-averaged
media concentration values are used, concentrations in media are
usually dominated by background estimates. This is because the
concentrations in most media are strongly dependent on the arsenic
soil concentration, and the build-up of arsenic in soil, due to
deposition of arsenic from the facilities over 30 years, is predicted
to be only a fraction of typical background soil concentrations. The
large coal-fired utility burner (LCUB) was predicted to account for
slightly more than 50 percent of the total deposition on the watershed
considered, but the build-up in the soil after 30 years is still
predicted to be less than 5 percent of the total soil concentration.
The facilities' contribution to the watershed air concentration is
small relative to the assumed background concentration of 0.7 ng/m3.
As mentioned above, none of the facilities is predicted to
contribute more than 10 percent to the total area-averaged arsenic
soil concentration in the watershed after a 30-year build-up period.
The LCUB has the largest contribution. A possibly important
uncertainty, though, is that it is not known if the deposited arsenic
from any of the facilities considered is more bioavailable than that
typically found naturally in the soil. If the deposited arsenic is
significantly more bioavailable, then the contribution to the total
exposure could be more than the predicted level.
The concentrations in grain were also examined because the
ingestion of grain is predicted to be a dominant pathway for exposure
to arsenic. The contribution of each facility to the arsenic
concentration in grain followed the same pattern as that for soil,
because the grain was predicted to accumulate arsenic mostly from the
soil.
It is important to note that the regional contribution of
utilities, as estimated by the RELMAP model, is generally less than 10
percent of the total background concentration or deposition. However,
this contribution is larger than that for some types of facilities
considered on the local scale when the upper 90th percentile is used.
10-26
-------
Table 10-21. Watershed Air Concentration
Watershed air concentration (jjg/m3)9
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
2.2E-06
1 .4E-06
2.3E-06
2.0E-06
4.0E-04 99
4.0E-04 100
4.0E-04 99
4.0E-04 100
4.4E-04 91 9
4.4E-04 91 9
4.4E-04 91 9
4.4E-04 91 9
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
1 .5E-06
1.1E-06
1 .9E-06
1 .9E-06
4.0E-04 100
4.0E-04 100
4.0E-04 100
4.0E-04 100
4.1E-04 98 1
4.1E-04 98 1
4.1E-04 98 1
4.1E-04 98 1
' Edge of watershed, 200 m from source
Table 10-22. Watershed Deposition Rate
Watershed deposition rate (ug/nrfyr)9
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
1.2E+02
3.1E+01
1.4E+01
1.5E+01
2.1E+02 42
1 .2E+02 74
1 .OE+02 86
1.1E+02 86
2.2E+02 41 4
1 .3E+02 70 6
1.1E+02 80 7
1.1E+02 79 7
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
1.6E+01
4.8E+00
3.9E+00
3.5E+00
1.1E+02 85
9.5E+01 95
9.4E+01 96
9.4E+01 96
1.1E+02 85 1
9.6E+01 94 1
9.5E+01 95 1
9.4E+01 95 1
' Edge of watershed 200 m from source
10-27
-------
Table 10-23. Surface Water Concentration
Surface water concentration (ug/L)a
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
3.1E-01
7.8E-02
3.4E-02
3.8E-02
1 .4E+01 97
1 .4E+01 98
1 .4E+01 99
1 .4E+01 99
1.4E+01 97 1
1 .4E+01 98 1
1 .4E+01 99 1
1.4E+01 98 1
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
1 .OE-01
2.9E-02
1 .9E-02
1 .8E-02
9.4E+00 96
9.3E+00 97
9.3E+00 97
9.3E+00 97
9.4E+00 96 3
9.3E+00 97 3
9.3E+00 97 3
9.3E+00 97 3
1 Edge of watershed 200 m from source.
Table 10-24. Untilled Soil Concentration
Untilled soil concentration (ug/g)a
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
4.7E-02
1 .2E-02
5.4E-03
5.8E-03
3.1E+00 97
3.0E+00 98
3.0E+00 99
3.0E+00 99
3.1E+00 97 1
3.0E+00 98 1
3.0E+00 99 1
3.0E+00 99 1
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
1 .5e-02
4.8E-03
3.8E-03
3.5E-03
3.1e+00 97
3.1E+00 97
3.1E+00 97
3.1E+00 97
3.1e+00 97 3
3.1E+00 97 3
3.1E+00 97 3
3.1E+00 97 3
1 Edge of watershed 200 m from source.
10-28
-------
Table 10-25. Grain Concentration
Grain (ug/g dry weight)9
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
6.4E-04
1 .6E-04
7.3E-05
7.9E-05
1.8E-01 99
1.8E-01 100
1.8E-01 100
1.8E-01 100
1 .8E-01 99 0
1.8E-01 100 0
1.8E-01 100 0
1.8E-01 100 0
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
7.2E-05
2.2E-05
1 .8E-05
1 .6E-05
1.8E-01 100
1.8E-01 100
1.8E-01 100
1.8E-01 100
1.8E-01 100 0
1.8E-01 100 0
1.8E-01 100 0
1.8E-01 100 0
a Edge of watershed 200 m from source.
10.4.2.2.2 Contribution of individual and regional sources to
inorganic arsenic exposure. The predicted exposures to inorganic
arsenic are summarized in Tables 10-26 through 10-33 for the
hypothetical exposure scenarios considered. Tables 10-26 through
10-29 represent the ingestion exposure pathway for the hypothetical
exposure scenarios, and Tables 10-30 through 10-33 represent the
inhalation exposure pathway for the exposure scenarios. In general,
less than 10 percent of the total inorganic arsenic exposure through
ingestion of food products is predicted to be attributable to any of
the local sources, for the hypothetical scenarios considered. This is
a reflection of the individual sources' contribution to the total soil
concentration, because the concentrations in media and biota, and,
hence, exposure to these media and biota, is strongly dependent on the
soil concentration. The exception to the trend is the pica child
scenario, in which a pica child is assumed to be exposed at the
location of maximum deposition. In this case, the LCUB is predicted
to contribute up to approximately 40 percent of the total inorganic
exposure. The other facilities are predicted to contribute less than
15 percent. While this scenario is considered to be rare, it does
indicate that is possible for some of the types of facilities
considered to result in exposures comparable to background exposure.
10-29
-------
Table 10-26. Inorganic Arsenic Intake via Ingestion for Pica
Child
Inorganic arsenic intake (mg/kg/day) for pica child9
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
7.3E-04
1.8E-04
7.7E-05
8.8E-05
2.1E-03 64
1 .5E-03 87
1 .4E-03 93
1 .4E-03 93
2.1E-03 64 1
1.5E-03 87 1
1.4E-03 93 1
1.4E-03 93 1
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
5.3E-05
3.5E-05
1 .6E-05
1 .7E-05
1 .4E-03 93
1 .4E-03 95
1 .4E-03 96
1 .4E-03 96
1.4E-03 93 3
1.4E-03 95 3
1.4E-03 96 3
1.4E-03 96 3
1 Receptor point based on location of maximum deposition.
Table 10-27. Inorganic Arsenic Intake via Ingestion for
Subsistence Farmer Adult
Inorganic arsenic intake (mg/kg/day) for subsistence farmer adult a
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
5.4E-06
1.3E-06
5.8E-07
6.3E-07
3.7E-04 98
3.6E-04 99
3.6E-04 100
3.6E-04 100
3.7E-04 98 0
3.6E-04 99 0
3.6E-04 99 0
3.6E-04 99 0
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
2.3E-05
5.2E-06
1 .8E-06
2.3E-06
3.9E-04 94
3.7E-04 98
3.7E-04 99
3.7E-04 99
3.9E-04 94 0
3.7E-04 98 0
3.7E-04 99 0
3.7E-04 99 0
1 Edge of watershed 200 m from source.
10-30
-------
Table 10-28. Inorganic Arsenic Intake via Ingestion for
Subsistence Farmer Child
Inorganic arsenic intake (mg/kg/day) for subsistence farmer child a
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
1 .OE-05
2.5E-06
1.1E-06
1 .2E-06
7.5E-04 98
7.5E-04 99
7.4E-04 100
7.5E-04 100
7.5E-04 98 0
7.5E-04 99 0
7.5E-04 99 0
7.5E-04 99 0
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
4.6E-05
1. OE-05
3.7E-06
4.7E-06
8.0E-04 94
7.6E-04 98
7.5E-04 99
7.5E-04 99
8.0E-04 94 0
7.6E-04 98 0
7.5E-04 99 0
7.6E-04 99 0
a Edge of watershed 200 m from source.
Table 10-29. Inorganic Arsenic Intake via Ingestion for
Subsistence Fisher Adult
Inorganic arsenic intake (mg/kg/day) for subsistence fisher adult9
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
5.0E-06
1.3E-06
5.3E-07
5.8E-07
2.7E-04 98
2.6E-04 99
2.6E-04 99
2.6E-04 99
2.7E-04 98 0
2.6E-04 99 0
2.6E-04 99 1
2.6E-04 99 1
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
2.3E-05
5.3E-06
1 .9E-06
2.4E-06
2.8E-04 91
2.6E-04 97
2.6E-04 99
2.6E-04 99
2.8E-04 91 1
2.6E-04 97 1
2.6E-04 99 1
2.6E-04 99 1
1 Edge of watershed 200 m from source.
10-31
-------
Table 10-30. Inorganic Arsenic Intake via Ingestion for
Subsistence Fisher Child
Inorganic arsenic intake (mg/kg/day) for subsistence fisher child9
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
9.3E-06
2.3E-06
9.7E-07
1.1E-06
5.1E-04 98
5.0E-04 99
5.0E-04 99
5.0E-04 99
5.1E-04 98 0
5.0E-04 99 0
5.0E-04 99 0
5.0E-04 99 0
Western site (arid)
Large Coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
4.7E-05
1.1E-05
3.7E-06
4.8E-06
5.5E-04 91
5.1E-04 97
5.1E-04 99
5.1E-04 99
5.5E-04 91 0
5.1E-04 97 1
5.1E-04 99 1
5.1E-04 99 1
1 Edge of watershed 200 m from source.
Table 10-31.
Child
Inorganic Arsenic Intake via Inhalation for Pica
Inorganic arsenic intake via inhalation (mg/kg body wt./day) for pica child9
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
4.1E-21
1.7E-18
1.7E-16
8.9E-16
3.8E-07 100
3.8E-07 100
3.8E-07 100
3.8E-07 100
4.1E-07 91 9
4.1E-07 91 9
4.1E-07 91 9
4.1E-07 91 9
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
4.9E-18
8.6E-24
6.8E-19
1.9E-18
3.8E-07 100
3.8E-07 100
3.8E-07 100
3.8E-07 100
3.8E-07 98 2
3.8E-07 98 2
3.8E-07 98 2
3.8E-07 98 2
1 Receptor point based on location of maximum deposition.
10-32
-------
Table 10-32. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Adult and Subsistence Fisher Adult
Inorganic arsenic intake via inhalation (mg/kg body wt./day) for subsistence farmer and fisher adults a
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
6.2E-10
4.1E-10
6.4E-10
5.7E-10
1.1E-07 99
1.1E-07 100
1.1E-07 99
1.1E-07 100
1.3E-07 91 9
1.3E-07 91 9
1.3E-07 91 9
1.3E-07 91 9
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
O.OE+00
4.3E-23
1.3E-18
4.7E-18
1.1E-07 100
1.1E-07 100
1.1E-07 100
1.1E-07 100
1 .2E-07 98 2
1.2E-07 98 2
1.2E-07 98 2
1.2E-07 98 2
a Edge of watershed 200 m from source.
Table 10-33. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Child and Subsistence Fisher Child
Inorganic arsenic intake via inhalation (mg/kg body wt./day) for subsistence farmer and fisher children a
Facility
Facility
only
Facility + %
background Background
Facility + background % %
+ RELMAPSOth Background RELMAP
Eastern site (humid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
2.0E-09
1 .4E-09
2.1E-09
1 .9E-09
3.8E-07 99
3.8E-07 100
3.8E-07 99
3.8E-07 100
4.2E-07 91 9
4.1E-07 91 9
4.2E-07 91 9
4.2E-07 91 9
Western site (arid)
Large coal-fired utility boiler
Medium coal-fired utility boiler
Small coal-fired utility boiler
Oil-fired utility boiler
O.OE+00
1 .4E-22
4.2E-18
1.5E-17
3.8E-07 100
3.8E-07 100
3.8E-07 100
3.8E-07 100
3.8E-07 98 2
3.8E-07 98 2
3.8E-07 98 2
3.8E-07 98 2
1 Edge of watershed 200 m from source.
10-33
-------
10.5 HAZARD IDENTIFICATION AND DOSE-RESPONSE FOR ARSENIC
10.5.1 Introduction
Inhalation exposure to inorganic arsenic in humans has been
strongly associated with lung cancer." Human oral exposure to
inorganic arsenic has been associated with an increased risk of
several types of cancer in humans, including skin, bladder, liver, and
lung cancer."'45 Oral exposure to inorganic arsenic has also been
associated with noncancer effects, including effects to the central
nervous system, cardiovascular system, gastrointestinal tract, liver,
kidney, and blood." Appendix E of this report contains more
information on health effects of arsenic. Chapter 4 contains general
information on terminology, definitions, and guidelines related to
risk assessments. A short summary of the health effects of arsenic is
presented here.
Arsenic in both the trivalent (+3) and pentavalent (+5) oxidation
states may form both inorganic and organic compounds. Both trivalent
and pentavalent species of inorganic arsenic may be taken up by humans
from the gut (as can organic forms); these forms may be found in urine
(after exposure), along with monomethylated arsenic and dimethylated
arsenic. When ingested by humans, pentavalent forms are reduced to
trivalent arsenic which is then subject to methylation-forming
monomethyl- and dimethyl arsenic. Trivalent forms appear to be more
toxic than pentavalent forms. Methylation was thought to result in
detoxification; however, this assumption has been called into question
recently. Despite its toxicity, there is some (weak) evidence to
indicate that arsenic is an essential element in the human diet.-
The toxicological data on organic arsenic are limited. In fish
and shellfish, organic arsenic is absorbed through the gastro-
intestinal tract and excreted rapidly (60-70 percent within a few
days) in the urine.- However, since the inorganic arsenic forms appear
to be of primary toxicological significance in humans, the following
discussion of health effects of arsenic and the risk characterization
focus only on inorganic arsenic.
10.5.2 Cancer Effects of Arsenic
There is clear evidence that chronic exposure to inorganic
arsenic in humans increases the risk of cancer. Various studies of
humans have reported that inhalation of arsenic results in an
increased risk of lung cancer. In addition, ingestion of arsenic has
been associated with an increased risk of non-melanoma skin cancer and
bladder, liver, and lung cancer. Animal studies have not clearly
associated arsenic exposure, via ingestion exposure, with cancer. No
studies have investigated the risk of cancer in animals as a result of
inhalation or dermal exposure.46
EPA has classified inorganic arsenic in Group A - Known Human
Carcinogen. The Group A classification was based on the increased
incidence in humans of lung cancer through inhalation exposure and the
10-34
-------
increased risk of skin, bladder, liver, and lung cancer through
drinking water exposure."
10.5.2.1 Inhalation Cancer Risk for Arsenic. EPA used the
absolute-risk linear extrapolation model to estimate the inhalation
unit risk for inorganic arsenic. Five studies on arsenic-exposed
copper smelter workers were modeled for excess cancer risk. Using the
geometric mean of these data, EPA calculated an inhalation unit risk
estimate (IURE) of 4.29 x 10"3 per (//g/m3) . The IURE is an upper-bound
estimate of the increased probability of a person developing cancer
from breathing air containing a concentration of 1 //g/m3 of air for
70 years.47 EPA has high confidence in the arsenic cancer IURE because
the studies examined a large number of people, the exposure
assessments included air measurements and urinary arsenic
measurements, and lung cancer incidence was significantly increased
over expected values." The inhalation cancer slope factor is
1.5 x 10+1 per (mg/kg/day). The inhalation slope factor is an upper
bound estimate of the increased risk of developing cancer due to an
average inhalation intake of 1 mg/kg/day of arsenic over a lifetime
(70 years).^
10.5.2.2 Oral Cancer Risk for Arsenic. To estimate the risks
posed by ingesting arsenic, EPA obtained data in Taiwan concerning
skin cancer incidence, age, and level of exposure via drinking water.
In 37 villages that had obtained drinking water for 45 years from
artesian wells with various elevated levels of arsenic, 40,421
individuals were examined for hyperpigmentation, keratosis, skin
cancer, and blackfoot disease (gangrene of the extremities caused by
injury to the peripheral vasculature). The local well waters were
analyzed for arsenic, and the age-specific cancer prevalence rates
were found to correlate with both local arsenic concentrations and age
(duration of exposure). Based on these data, although EPA has not
presented the calculations for the oral unit risk estimate for
arsenic, they did propose that a unit risk estimate of 5 x 10"5 (ug/L) ~1
from oral exposure to arsenic in drinking water be used." This
equates to an oral cancer slope factor of 1.5E + 00 per (mg/kg/day).
The oral cancer slope factor is an upper bound estimate of the
increased risk of developing cancer due to an average oral intake of
1 mg/kg/day of arsenic over a lifetime (70 years) . —'—
The Taiwan cancer data have several limitations: (1) the water
was contaminated with substances such as bacteria and ergot alkaloids,
in addition to arsenic; (2) total arsenic exposure was uncertain
because of intake from the diet and other sources; (3) early deaths
from blackfoot disease may have led to an underestimate of prevalence;
and (4) there was uncertainty concerning exposure durations. Due to
these limitations, and also because the diet, economic status, and
mobility of individuals in Taiwan are different from those of most
United States citizens, EPA has stated that "the uncertainties
associated with ingested inorganic arsenic are such that estimates
could be modified downwards as much as an order of magnitude, relative
to risk estimates associated with most other carcinogens.""
10-35
-------
10.5.3 Noncancer Effects of Arsenic
10.5.3.1 Chronic (Long-Term) Effects for Arsenic. The primary
noncancer effects noted in humans from chronic exposure to arsenic,
through both inhalation and oral exposure, are effects on the skin.
The inhalation route has resulted primarily in irritation of the skin
and mucous membranes (dermatitis, conjunctivitis, pharyngitis, and
rhinitis), while chronic oral exposure has resulted in a pattern of
skin changes that include the formation of warts or corns on the palms
and soles along with areas of darkened skin on the face, neck, and
back. Other effects noted from chronic oral exposure include
peripheral neuropathy, cardiovascular disorders, liver and kidney
disorders, and blackfoot disease."
EPA has established an RfD for inorganic arsenic of 0.0003
mg/kg/day, based on a NOAEL (adjusted to include arsenic exposure from
food) of 0.0008 mg/kg/day, an uncertainty factor of 3, and a modifying
factor of 1.— The EPA has not established a RfC for inorganic
arsenic.—
10.6 RISK CHARACTERIZATION
In this section, the information on hazard, dose-response, and
exposure are combined to characterize the potential risks due to
arsenic emissions from the model utility plants. As stated above,
inorganic arsenic is the form of arsenic considered to be of primary
concern for causing adverse health effects. Therefore, risks and
hazards have been estimated only for the inorganic arsenic.
Increased cancer risk for each hypothetical person, under each
hypothetical scenario, for four different model plants placed in two
locations has been estimated. The general method of estimating the
increased risk of cancer due to multipathway exposure to inorganic
arsenic, is to multiply the predicted intakes (mg/kg/day) from the
multipathway exposure modeling by the EPA's Oral Cancer Potency Factor
for inorganic arsenic (1.51E+00 per [mg/kg]/day) and then to multiply
by 30/70 to adjust for the assumed duration of exposure. The potency
factor is based on lifetime exposure (i.e., 70 years). For all of the
exposure scenarios, except for the pica child, it is assumed that the
hypothetical person is exposed for 30 years. A 30-year exposure
assumption is considered appropriate for most exposure scenarios
considered in this analysis. However, for the pica child, it is
assumed that exposure only occurs for 7 years (i.e., duration
adjustment factor was 7/70), since it highly unlikely that the pica
child would continue pica behavior for more than 7 years. (Note: the
7-year assumption may be high for pica behavior; however, other
exposures that may occur as the child gets older such as through food
consumption were not considered. Therefore, the 7-year assumption may
be a reasonable assumption for this screening assessment.) For the
subsistence fisher child and subsistence farmer child, a 30-year
exposure is assumed because it is quite possible that the child will
continue these consumption behaviors as they become adults. The doses
10-36
-------
may decrease some because of the changes in body size, however, the
consumption rate is also expected to increase as they become adults.
A 30-year exposure duration was assumed for the inhalation risk
calculations in this multipathway arsenic assessment. To estimate the
increased cancer risk due to inhalation exposure the predicted intakes
via inhalation (in mg/kg/day) were multiplied by the inhalation cancer
slope factor (1.5E+01 per [mg/kg]/day) and then multiplied by 30/70 to
adjust for the assumed duration of exposure.
To estimate the noncancer hazard, the predicted intakes are
compared to the EPA's oral RfD (3E-04 mg/kg/day). The estimated
intakes in mg/kg/day, were divided by the RfD to calculate a hazard
quotient (HQ). Therefore, a HQ greater than 1.0 indicates exceedance
of the RfD.
It is important to remember that this is a screening level
multipathway analysis. The analysis is based on the use of model
plants placed in hypothetical locations, as well as the use of
hypothetical exposure scenarios and various assumptions, to predict
exposure to arsenic emitted by the model plant. The results give some
indication of the potential hazards and risks that may occur due to
emissions from an electric utility plant. However, the results are
not applicable to any particular plant. There are uncertainties and
limitations to the analyses and results. The risk estimates could be
overestimates or underestimates of the true risk for any particular
facility. There are uncertainties and variabilities in the modeling
inputs and results, exposure estimates, cancer potency estimates, RfD,
and overall risk estimates. The risk estimates presented below are
predicted values based on hypothetical scenarios and are intended for
screening purposes only. The results are shown in Tables 10-34
through 10-41.
10.6.1 Discussion of Cancer Risk Assessment Results
This analysis of multipathway exposures to arsenic emissions is a
screening analysis. Thus, these quantitive exposure and risk results,
because of the many modeling and analytic uncertainties, are very
uncertain and do not, therefore, conclusively demonstrate the
existence of health risks of concern associated with exposures to
utility emissions either on a national scale or from any actual
individual utility. The lack of measured data around these sources
preclude a comparison with modeled results. These results do suggest
that exposures and risks of concern cannot at present be ruled out and
that there is a need for development of additional scientific
information to evaluate whether risk levels of concern may exist.
The cancer risks due to multipathway exposures to inorganic
arsenic from utility emissions alone (no background) are estimated to
be no greater than approximately 1 x 10"4 (for pica child) based on
this screening level analysis. The large coal-fired utility at the
eastern humid site was estimated to pose this highest risk for the
pica child. Considering background exposures alone, the risk for the
pica child is estimated to be as high as 3 x 10~4.
10-37
-------
Table 10-34. Inorganic Arsenic Intake, Predicted Cancer Risk,
and Noncancer Hazard Quotient (HQ) for Pica Childa
Electric utility model
Facility
Facility only
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility +backg round
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer
risk
Non-
cancer
HQ
Facility + background +
RELMAP90th
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
7.3E-04
1 .8E-04
7.7E-05
8.8E-05
1E-04
3E-05
1E-05
1E-05
2.4
0.6
0.3
0.3
2.1E-03
1 .5E-03
1 .4E-03
1 .4E-03
3E-04
2E-04
2E-04
2E-04
7
5
5
5
2.1E-03
1 .5E-03
1 .4E-03
1 .4E-03
3E-04
2E-04
2E-04
2E-04
7
5
5
5
Western site (arid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
5.3E-05
3.5E-05
1 .6E-05
1 .7E-05
8E-06
5E-06
2E-06
3E-06
0.2
0.1
0.05
0.06
1 .4E-03
1 .4E-03
1 .4E-03
1 .4E-03
2E-04
2E-04
2E-04
2E-04
5
5
5
5
1 .4E-03
1 .4E-03
1 .4E-03
1 .4E-03
2E-04
2E-04
2E-04
2E-04
5
5
5
5
a Receptor point based on location of maximum deposition.
Table 10-35. Inorganic Arsenic Intake, Predicted Cancer Risk,
and Noncancer Hazards for Subsistence Farmer Adult3
Electric utility model
facility
Facility only
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility + background
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility + background +
RELMAPSOth
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
5.4E-06
1 .3E-06
5.8E-07
6.3E-07
4E-06
1E-06
4E-07
5E-07
0.03
0.004
0.002
0.003
3.7E-04
3.6E-04
3.6E-04
3.6E-04
3E-04
2E-04
2E-04
2E-04
1.2
1.2
1.2
1.2
3.7E-04
3.6E-04
3.6E-04
3.6E-04
3E-04
2E-04
2E-04
2E-04
1.2
1.2
1.2
1.2
Western site (arid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
2.3E-05
5.2E-06
1 .8E-06
2.3E-06
1E-05
4E-06
1E-06
1E-06
0.08
0.02
0.006
0.008
3.9E-04
3.7E-04
3.7E-04
3.7E-04
3E-04
3E-04
3E-04
3E-04
1.3
1.2
1.2
1.2
3.9E-04
3.7E-04
3.7E-04
3.7E-04
3E-04
3E-04
3E-04
3E-04
1.3
1.2
1.2
1.2
1 Edge of watershed 200 m from source.
10-38
-------
Table 10-36. Inorganic Arsenic Intake, Predicted Cancer Risk,
and Noncancer Hazards for Subsistence Farmer Child3
Electric utility model
facility
Facility only
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility + background
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
Increased
cancer risk
Non-
cancer
HQ
Facility + background +
RELMAPSOth
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
1 .OE-05
2.5E-06
1.1E-06
1 .2E-06
1E-05
2E-06
1E-06
1E-06
0.03
0.008
0.004
0.004
7.5E-04
7.5E-04
7.4E-04
7.5E-04
5E-04
5E-04
5E-04
5E-04
2.5
2.5
2.5
2.5
7.5E-04
7.5E-04
7.5E-04
7.5E-04
5E-04
5E-04
5E-04
5E-04
2.5
2.5
2.5
2.5
Western site (arid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
4.6E-05
1 .OE-05
3.7E-06
4.7E-06
3E-05
1E-05
3E-06
3E-06
0.15
0.03
0.01
0.02
8E-04
7.6E-04
7.5E-04
7.5E-04
6E-04
5E-04
5E-04
5E-04
2.7
2.5
2.5
2.5
8.0E-04
7.6E-04
7.5E-04
7.5E-04
6E-04
5E-04
5E-04
5E-04
2.7
2.5
2.5
2.5
1 Edge of watershed 200 m from source.
Table 37. Inorganic Arsenic Intake, Predicted Cancer Risk, and
Noncancer Hazards for Subsistence Fisher Adult3
Electric utility model
facility
Facility only
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility + background
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility + background +
RELMAPSOth
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
5.0E-06
1 .3E-06
5.3E-07
5.8E-07
4E-06
1E-06
4E-07
4E-07
0.02
0.004
0.002
0.002
2.7E-04
2.6E-04
2.6E-04
2.6E-04
2E-04
2E-04
2E-04
2E-04
0.9
0.9
0.9
0.9
2.7E-04
2.6E-04
2.6E-04
2.6E-04
2E-04
2E-04
2E-04
2E-04
0.9
0.9
0.9
0.9
Western site (arid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
2.3E-05
5.3E-06
1 .9E-06
2.4E-06
1E-05
4E-06
1E-06
2E-06
0.08
0.02
0.006
0.008
2.8E-04
2.6E-04
2.6E-04
2.6E-04
2E-04
2E-04
2E-04
2E-04
0.9
0.9
0.9
0.9
2.8E-04
2.6E-04
2.6E-04
2.6E-04
2E-04
2E-04
2E-04
2E-04
0.9
0.9
0.9
0.9
1 Edge of watershed 200 m from source.
10-39
-------
Table 10-38. Inorganic Arsenic Intake, Predicted Cancer Risk,
and Noncancer Hazards for Subsistence Fisher Child3
Electric utility model
facility
Facility only
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility + background
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Facility + background +
RELMAPSOth
Inorganic
Arsenic
Intake
(mg/kg/d)
Estimated
increased
cancer risk
Non-
cancer
HQ
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
9.3E-06
2.3E-06
9.7E-07
1.1E-06
7E-06
2E-06
7E-07
1E-06
0.03
0.008
0.003
0.004
5.1E-04
5.0E-04
5.0E-04
5.0E-04
4E-04
4E-04
4E-04
4E-04
1.7
1.7
1.7
1.7
5.1E-04
5.0E-04
5.0E-04
5.0E-04
4E-04
4E-04
4E-04
4E-04
1.7
1.7
1.7
1.7
Western site (arid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
4.7E-05
1.1E-05
3.7E-06
4.8E-06
3E-05
1E-05
3E-06
3E-06
0.16
0.04
0.01
0.02
5.5E-04
5.1E-04
5.1E-04
5.1E-04
4E-04
4E-04
4E-04
4E-04
1.7
1.7
1.7
1.7
5.5E-04
5.1E-04
5.1E-04
5.1E-04
4E-04
4E-04
4E-04
4E-04
1.7
1.7
1.7
1.7
a Edge of watershed 200 m from source.
Table 10-39. Inorganic Arsenic Intake via Inhalation and
Predicted Cancer Risks for Pica Child3
Electric utility model
facility
Facility only
Inorganic
arsenic
intake
(mg/kg/d)
Estimated
increased
cancer risk
Facility + background
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased
cancer risk
Facility + background +
RELMAPSOth
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased cancer
risk
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
4.1E-21
1.7E-18
1.7E-16
8.9E-16
3E-20
1E-17
1E-15
7E-15
3.8E-07
3.8E-07
3.8E-07
3.8E-07
3E-06
3E-06
3E-06
3E-06
4.1E-07
4.1E-07
4.1E-07
4.1E-07
3E-06
3E-06
3E-06
3E-06
Western site (arid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
4.9E-18
8.6E-24
6.8E-19
1.9E-18
4E-17
7E-23
6E-18
1E-17
3.8E-07
3.8E-07
3.8E-07
3.8E-07
3E-06
3E-06
3E-06
3E-06
3.8E-07
3.8E-07
3.8E-07
3.8E-07
3E-06
3E-06
3E-06
3E-06
' Receptor point based on location of maximum deposition.
10-40
-------
Table 10-40. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Adult and Subsistence Fisher Adult and
Predicted Cancer Risks3
Electric utility model
facility
Facility only
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased
cancer risk
Facility + background
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased
cancer risk
Facility + background +
RELMAPSOth
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased
cancer risk
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
6.2E-10
4.1E-10
6.4E-10
5.7E-10
4E-09
3E-09
5E-09
4E-09
1.1E-07
1.1E-07
1.1E-07
1.1E-07
1E-06
1E-06
1E-06
1E-06
1 .3E-07
1 .3E-07
1 .3E-07
1 .3E-07
1E-06
1E-06
1E-06
1E-06
Western site (arid)
Large coal-fired boiler
Medium Coal-fired boiler
Small coal-fired boiler
Oil-fired Boiler
O.OE+00
4.3E-23
1.3E-18
4.7E-18
0.0
3E-22
1E-17
3E-17
1.1E-07
1.1E-07
1.1E-07
1.1E-07
1E-06
1E-06
1E-06
1E-06
1 .2E-07
1 .2E-07
1 .2E-07
1 .2E-07
1E-06
1E-06
1E-06
1E-06
a Edge of watershed 200 m from source.
Table 10-41. Inorganic Arsenic Intake via Inhalation for
Subsistence Farmer Child and Predicted Cancer Risks3
Electric utility model
facility
Facility only
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased
cancer risk
Facility + background
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased
cancer risk
Facility + background +
RELMAPSOth
Inorganic
arsenic intake
(mg/kg/d)
Estimated
increased
cancer risk
Eastern site (humid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
2.0E-09
1.4E-09
2.1E-09
1.9E-09
1E-08
1E-08
1E-08
1E-08
3.8E-07
3.8E-07
3.8E-07
3.8E-07
3E-06
3E-06
3E-06
3E-06
4.2E-07
4.1E-07
4.2E-07
4.2E-07
3E-06
3E-06
3E-06
3E-06
Western site (arid)
Large coal-fired boiler
Medium coal-fired boiler
Small coal-fired boiler
Oil-fired boiler
O.OE+00
1.4E-22
4.2E-18
1.5E-17
0
1E-21
3E-17
1E-16
3.8E-07
3.8E-07
3.8E-07
3.8E-07
3E-06
3E-06
3E-06
3E-06
3.8E-07
3.8E-07
3.8E-07
3.8E-07
3E-06
3E-06
3E-06
3E-06
1 Edge of watershed 200 m from source) and predicted cancer risks.
10-41
-------
The cancer risks due to multipathway exposures to inorganic
arsenic from utility emissions alone (no background) are estimated to
be no greater than approximately 1 x 10"4 (for pica child) based on
this screening level analysis. The large coal-fired utility at the
eastern humid site was estimated to pose this highest risk for the
pica child. Considering background exposures alone, the risk for the
pica child is estimated to be as high as 3 x 10~4.
The estimated cancer risks for the subsistence farmer adult and
subsistence fisher adult are estimated to be as high as 1 x 10~5 and
the estimated cancer risks for the subsistence farmer child and
subsistence fisher child are estimated to be as high as 3 x 10 ~5. In
all scenarios, it was the large coal-fired utility that was estimated
to pose the greatest multipathway risks, the medium coal-fired utility
was estimated to pose the next highest risks, and the small coal-fired
utility and the oil-fired utility were estimated to present very
similar magnitudes of exposures and risks. Background exposures were
estimated to dominate the exposures and risks in all scenarios.
Including background exposures increases the risks by approximately 2
to 200 times, depending on the receptor and plant scenario analyzed.
Inhalation risks are predicted to be significantly lower than the
risks from multipathway exposures. The highest inhalation risks due
to the model utility plant emissions only (no background) were
estimated to be 1 x 10~8 for the hypothetical receptors in this
analysis. The RELMAP results contributed slightly (approximately
9 percent) to the total inhalation exposures and risks. Similar to
multipathway exposures, the inhalation exposures to background were
estimated to be substantially higher than the inhalation exposures due
to utilities emissions based on this model plant analysis. However,
it is important to recall that the estimated risks for inhalation
exposures to arsenic from the HEM analyses, described in chapter 6 (in
which every actual plant [684 plants] in the United States was
modeled), were estimated to be up to 3 x 10~s, which is about 300 times
higher than the inhalation risks predicted in this model plant
analysis. This may be due to several factors. One likely factor is
that many more plants were modeled (684 actual plants, including some
outliers instead of 4 model plants) for the HEM analyses. Also, in
the HEM analyses, 349 meteorological sites were included, whereas only
2 meteorological stations were included in the multipathway analysis.
In addition, 70-year exposures were assumed for the HEM analyses;
30-year exposures were the primary assumption used for the
multipathway analysis. Moreover, the multipathway analysis was
intended to predict potential high-end exposures due to multipathway
exposures; therefore, the distances from plant were chosen that would
likely result in high-end multipathway exposures. The distances
chosen for this multipathway analysis were mainly based on the
locations of predicted high deposition rates, and not based on high
air concentrations. The 200-meter distance from stack used for most
of the scenarios is estimated to be a location of high deposition;
however, this distance is not likely to be the distance of highest air
concentrations. The distances of high air concentrations are likely
10-42
-------
to be much further from the stack. Therefore, it is not expected that
this model plant analysis would predict inhalation exposures and risks
as high as the HEM analysis.
10.6.2 Discussion of the Noncancer Risk Assessment Results
The highest predicted intake of arsenic due to utility emissions
only was 7E-04 mg/kg/day (predicted for the pica child/large coal-
fired utility/eastern site scenario), which is 2.4 times higher than
the RfD (3E-04 mg/kg/day) and equates to an HQ of 2.4. All other
scenarios had predicted HQs less than 1.0 when considering utility
emissions only. When background is included, the HQs range from 0.9
to 7. These results suggest that adverse noncancer effects due to
utility emissions alone could possibly be of concern for the pica
child. For all other scenarios analyzed, it appears that utility
emissions alone are not likely to be of concern for noncancer effects.
However, exposures to current background levels of arsenic, as well as
exposures to current background levels combined with utility emissions
of arsenic, could be a potential concern for adverse noncancer
effects.
10.7 CONCLUSIONS
10.7.1 Contribution of Arsenic Emissions from Utilities to
Concentrations in Environmental Media and Biota.
The facilities considered were found to contribute less than 10
percent of the total predicted arsenic concentrations in soil, water
and all biota when a background soil concentration of 3 //g/g and air
concentration of 0.7 ng/m3 were assumed, and area-averaged deposition
rates were used for a watershed 200 m downwind from the facility.
Soil and predicted water concentrations are dominated (e.g., >95
percent) by background when background is included.
10.7.2 Determination of Dominant Pathways of Potential Exposure to
Anthropogenic Arsenic Emissions
Using the hypothetical exposure scenarios discussed above, it was
found that exposure to inorganic arsenic through ingestion was mainly
through the ingestion of grain. This result holds true whether or not
anthropogenic sources are considered because there is not sufficient
evidence to support the assumption that the anthropogenic arsenic
varied chemically from that already in the environment. The highest
dietary exposures to total arsenic for adults not consuming marine
organisms or freshwater bivalves are the result of consumption of
freshwater fish and grains.
Exposure to inorganic arsenic through the ingestion of fish was
not predicted to be a major pathway of exposure because there is
considerable evidence that little of the total arsenic in fish tissue
is inorganic arsenic. For the subsistence fisher scenario considered,
exposure to inorganic arsenic from grain was larger than that from
fish, even though a high fish consumption rate was assumed.
10-43
-------
Exposure to arsenic through the ingestion of unusually high
amounts of soil (e.g., pica) was shown to be of potential concern.
Dominant exposure for the pica child is from background soil
concentrations of arsenic, when background levels are considered.
Since this is a screening level analysis, further detailed
analyses are needed to better characterize the risks posed.
10.8. UNCERTAINTIES AND LIMITATIONS
10.8.1 Limitations and Uncertainties for the Multipathway Exposure
Modeling
The following are uncertainties and limitations associated with
this assessment that may affect the results.
If a location is such that the background concentrations are
significantly lower than those assumed in the assessment, then the
contribution of any particular individual source to the total arsenic
concentrations and subsequent exposure will be larger than that
determined here. However, this will also reduce the total arsenic
exposure. If the background concentrations are significantly higher
than that assumed here, then although the contribution of a source to
the total is less, the additional contribution may result in the
exceedance of certain toxic thresholds. At this time, there is little
guidance on the determination of or incorporation of background
concentrations in an assessment. For chemicals which have a threshold
of action, inclusion of an appropriate estimate of background is
critical in the determination of the impacts to public health.
There are few data on the speciation or bioavailability of
arsenic in environmental media or anthropogenic emissions. In this
assessment, it was assumed that the arsenic emitted from utilities is
similar to that reported in the literature for various media. If the
form of the emitted arsenic is significantly different, then this can
have important impacts on the assessment results and conclusions.
The bioavailability of anthropogenic arsenic is assumed to be the
same as the bioavailability of arsenic that in the environment in this
assessment. If this is not true, then this could impact the
conclusions reached. More research is needed on the bioavailability
of anthropogenic arsenic.
An important uncertainty when identifying a dominant pathway is
the range of pathways which are considered. For example, not
considered here was exposure to arsenic through the ingestion of tap
or well water. Inclusion of this pathway would increase the primary
or background exposure (depending on the level of exposure), and hence
reduce the relative contribution of any particular source.
10.8.2 Limitations and Uncertainties for the Risk Characterization
As discussed in section 10.5 above and in Appendix E, there are
limitations in the health effects data for arsenic. For example,
10-44
-------
although human data are available on the carcinogenicity of arsenic,
there are no animal studies available on the carcinogenicity of
arsenic from inhalation exposure and the animal ingestion studies have
not clearly shown an association between arsenic ingestion exposure
and cancer.
The oral cancer potency factor (1.5 per mg/kg/day) is a
reasonable estimate based on available information. However, the
available toxicity data are limited and uncertain; hence the cancer
potency estimate is somewhat uncertain. Also, because of differences
in genetics, diet, lifestyle, stress management, and a number of other
host factors, the human population is likely to have considerable
variation in individual sensitivities to developing cancer per unit of
exposure. Therefore, although the oral potency factor may be a
reasonable estimate, it is not likely to accurately represent the
potency for all humans. Many humans may be either more or less
sensitive to developing cancer, than predicted by the oral potency
factor.
The human exposure estimates presented above are uncertain due to
modeling uncertainties (described above), and also because of human
variability and uncertainties associated with several factors such as
breathing rates, food consumption rates, activity patterns, body size,
metabolic differences, and other factors. Variation and uncertainty
in meteorology, environmental fate of arsenic, bioaccumulation rates,
and other factors also contribute to the overall uncertainty in the
risks estimates. Therefore, the risk estimates presented above should
be viewed as reasonable estimates for a screening level assessment,
but should also be viewed as containing significant uncertainties.
Further analyses are needed to more fully assess the potential risks
due to arsenic exposure from utilities.
10.9 RESEARCH NEEDS
To improve the multimedia, multipathway exposure assessment for
arsenic, additional data on the speciation of arsenic in air, soil,
and water are needed. Better information on arsenic bioavailability
is also needed. In addition, research to improve the understanding of
watershed dynamics would benefit this study, as well as other efforts
to model pollutant fate and transport. Additional health effects data
are also needed, especially for ingestion exposure route, for cancer
and noncancer effects, to better characterize the hazards of arsenic
exposures.
10-45
-------
10.10 REFERENCES
1. Deuel, L. E. and A. R. Swoboda. Arsenic solubility in a reduced
environment. Soil Sci. Soc. Am. Proc. Volume 36. 1972.
pp. 276-278.
2. Braman, R. S. Arsenic in the environment. In: Arsenical
Pesticides. American Chemical Society, Washington, DC. 1975.
pp. 108-123.
3. U.S. Environmental Protection Agency. Health Assessment Document
for Inorganic Arsenic: Final Report. EPA-600/8-83-021F. Office
of Health and Environmental Assessment, Washington, DC. 1984.
4. National Research Council (NRC). Medical and Biological Effects
of Environmental Pollutants: Arsenic. National Academy of
Sciences, Washington, DC. 1977. pp. 332.
5. IF Kaiser. Toxicity and exposure concerns related to arsenic in
seafood: as arsenic literature review for risk assessments.
Submitted in fulfillment of Region X EAST Work Unit Document 4038
under Technical Instruction Document 10-9601-815 as requested by
Patricia Crone. 1996.
6. Hindmarsh, J. T. and R. F. McCurdy. Clinical and Environmental
Aspects of Arsenic Toxicity. CRC Critical Reviews in Clinical
Laboratory Sciences. Volume 23, No. 4. 1984. pp. 315-347.
7. Chilvers, D. C., and P. J. Peterson. Lead, Mercury, Cadmium and
Arsenic in the Environment. Edited by T.C. Hutchinson and K. M.
Meema. Wiley, New York. 1987. pp. 279.
8. Union Carbide Corporation. Review of the Environmental Effects
of Arsenic. ORNL/EIS-79 Oak Ridge National Laboratory, Oak
Ridge, TN. 1977.
9. Radian Corporation. Draft Report- Locating and Estimating Air
Emissions from Sources of Arsenic and Arsenic Compounds. Office
of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC. 1995.
10. Wu, Z. Y, M. Han, Z. C. Lin, and J. M. Ondov. Chesapeake Bay
atmospheric deposition study, year 1: Sources and dry deposition
of selected elements in aerosol particles. Atmospheric
Environment. Volume 28, No. 8. 1994. pp. 1471-1486.
11. Scudlark, J. R., K. M. Conko, and T.C. Church. Atmospheric wet
deposition of trace elements to Chesapeake Bay: CBAD study year 1
results. Atmospheric Environment. Volume 28, No. 8. 1994.
pp. 1487-1498.
10-46
-------
12. Durum, W. H., J. D. Hem, and S. G. Heidel. Reconnaisance of
Selected Minor Elements in Surface Waters of the United States.
October 1970, Geologic Survey Circ. 643. U.S. Department of the
Interior, Washington, DC. 1971. p. 49.
13. U.S. Environmental Protection Agency. Screening Level
Multipathway Exposure Analysis for Arsenic from Several Model
Electric Utility Plants and Secondary Lead Smelters. EPA Draft
Technical Report. 1997.
14. U.S. Environmental Protection Agency. Locating and Estimating
Air Emissions from Sources of Arsenic and Arsenic Compounds.
Draft Report. July 6 1996.
15. Bowen. Elemental Chemistry of the Elements. Academic Press,
London and New York. 1979. pp. 60-61.
16. Vinogradov, A. P. The Geochemistry of Rare and Dispersed
Chemical Elements in Soils. New York, NY. 1959. pp. 65-70.
17. Stater, C. S., R. S. Holmes, and H. G. Byers. Trace elements in
the soil from the erosion experiment stations, with supplementary
data on other soils. U.S. Department Agric Tech Bulletin.
Volume 552. 1937. p. 23.
18. Conner and Shacklette. Background Geochemistry of Some Rocks,
Soils, Plants, and Vegetables in the Conterminous United States.
U.S. Government Printing Office, Washington, DC. 1975.
19. Wood, H. and S. Duda. Background risk for regional mean
inorganic concentrations from southeastern soils. Presented at
the Annual Society of Risk Analysis meeting, New Orleans, LA.
December 1996.
20. Cullen, W. R. and K. J. Reimer. Arsenic speciation in the
environment. Chemical Review. Volume 89. 1989. pp. 713-764.
21. Takamatsu, T., H. Aoki, and T. Yoshida. Determination of
arsenate, arsenite, monomethylarsenate, and dimethylarsinate in
soil polluted with arsenic. Soil Science. Volume 33. 1982.
pp. 239-246.
22. Bombach, G., A. Pierra, and W. Klemm. Arsenic in contaminated
soil and river sediment. Fresenius Journal of Analytical
Chemistry. Volume 350. 1994. pp. 49-53.
23. U.S. Environmental Protection Agency. Bioavailability of arsenic
and lead in environmental substrates. EPA 910/R-96-002.
Region 10. Seattle, WA. 1996.
10-47
-------
24. Aten, C. F., J. B. Bourke, J. H. Martini, and J. C. Walton.
Arsenic and lead in an orchard environment. Environmental
Toxicology. Volume 24. 1980. pp. 108-115.
25. Jones, J. S. and M. B. Hatch. Spray residues and crop assimilation
of arsenic and lead. Soil Science. Volume 44. 1945. Pp. 37-44.
26. Chisholm, D. Lead, arsenic, and copper content of crops grown on
lead arsenate-treated and untreated soils. Canadian Journal of
Plant Science. Volume 52. 1972. pp. 583-588.
27. Woolson, E. A. Effects of fertilizer material and combinations
on the phytotoxicity, availability, and content of arsenic in
corn. Journal of Science Food Agriculture. Volume 23. 1972.
pp. 1477-1481.
28. Jenkins, D. W. Biological monitoring of toxic trace metals.
Vol. 2 Toxic trace metals in plants and animals of the world.
Part 1. U.S. EPA Agency Report 600/3-80-090, 1980. pp. 30-138.
29. van der Veen, N. G. and K. Vreman. Transfer of cadmium, lead,
mercury and arsenic from feed into various organs and tissues of
fattening lambs. Netherlands Journal of Agricultural Science.
Volume 34. 1986. pp. 145-153.
30. Vreman, K., N. J. van der Veen, E. J. van der Molen and W. G. de
Ruig. Transfer of cadmium, lead, mercury and arsenic from feed
into milk and various tissues of dairy cows: chemical and
pathological data. Netherlands Journal of Agricultural Science.
Volume 3. 1986. pp. 129-144.
31. Barudi, W. and Bielig, H. J. Heavy metal content (As, Pb, Cd,
Hg) [arsenic, lead, cadmium, mercury] of vegetables which grow
above ground and fruits. Gehalt an Schwermetallen (Arsen, Blei,
Cadmium, Quecksilber) in oberirdisch wachsenden Gemuse- and
Obstarten. SOZ-Lebensm-Unters-Forsch.Munchen, J. F. Bergmann.
Volume 70, No. 4. 1980. pp. 254-257.
32. Varo et al. Acta Agric. Scand. Suppl. Volume 22. 1980. p. 27.
33. Merry et al. The effects of soil contamination with copper, lead
and arsenic on the growth and composition of plants. Effects of
source of contamination, varying soil pH, and prior waterlogging.
Plant and Soil. Volume 95, No. 2. 1985. pp. 255-69.
34. Moore, Edmonds, and Fracesconi. 1987.
35. Jelinek, C. F. and P. E. Corneliussen. Levels of arsenic in the
United States food supply. Environmental Health Perspectives.
Volume 19. 1977. pp. 83-87.
10-48
-------
36. Wiener et al. 1984. Longitudinal distribution of trace elements
(As, Cd, Cr, Hg, Pb, and Se) in fishes and sediment in the Upper
Mississippi River. In: J. G. Wiener, R. V. Anderson, D. R.
McConville (Eds.), Contaminants in Upper Mississippi River.
Butterworth, Stoneham, MA. 1984. pp. 139-170.
37. Lima et al. Acute and chronic toxicities of arsenic (III) to
fathead minnows, flagfish, daphnids, and an amphipod. Arch-
Environ- Contam-Toxicol . New York, NY: Springer-Verlag.
Volume 13, No. 5. 1984. pp. 595-601.
38. Eisler, R. A review of arsenic hazards to plants and animals
with emphasis on fishery and wildlife resources. Arsenic in the
Environment Part II: Human Health and Ecosystem Effects. Edited
by Nriagu, J. 0. Wiley Series in Environmental Science and
Technology. 1994.
39. U.S. Environmental Protection Agency. User's Guide for the
Industrial Source Complex (ISCST3) Dispersion Models, Volume I -
User's Instructions. EPA-454/B-95-003a. Office of Air Quality
Planning and Standards. Research Triangle Park, NC. September
1995.
40. U.S. Environmental Protection Agency. User's Guide for the
Industrial Source Complex (ISCST3) Dispersion Models, Volume II-
Description of Model Algorithms. EPA-454/B-95-003b. Office of
Air Quality Planning and Standards. Research Triangle Park,
North Carolina. September 1995.
41. Columbia River Inter-Tribal Fish Commission. A Fish Consumption
Survey of the Umatilla, Nez Perce, Yakima and Warm Springs Tribes
of the Columbia River Basin. Technical Report 94-3. October
1994.
42. Scudlark, J. R., and T. C. Church. The atmospheric deposition of
arsenic and association with acid precipitation. Atmospheric
Environment. Volume 22, No. 5. 1988. pp. 937-943.
43. Andraea, M. 0. Arsenic in rain and the atmospheric balance of
arsenic. Journal of Geophysical Research. Volume 85. 1980.
pp. 4512-4518.
44. Alcamo, J., J. Bartnicki, K. Olendrzynski, and J. Pacyna.
Computing heavy metals in Europe's atmosphere-I. Model
development and testing. Atmospheric Environment. Volume 26A,
No. 18. 1992. pp. 3355-3369.
45. Agency for Toxic Substances and Disease Registry (ATSDR).
Toxicological Profile for Arsenic. U.S. Public Health Service,
U.S. Department of Health and Human Services, Atlanta, GA. 1993.
10-49
-------
46. U.S. Environmental Protection Agency. Integrated Risk
Information System on Arsenic. 1995.
47. U.S. Environmental Protection Agency. Guidelines for the
preparation of Office of Water Health Advisories. U.S.
Environmental Protection Agency, Environmental Criteria and
Assessment Office-Cincinnati, OH. 1987.
10-50
-------
11.0 A MULTIPATHWAY SCREENING-LEVEL ASSESSMENT FOR DIOXINS/FURANS
11.1 INTRODUCTION
Polychlorinated dibenzo-p-dioxin (PCDD) and dibenzofuran (PCDF)
emissions were identified as a priority for multipathway exposure
analysis (see chapter 5). PCDDs and PCDFs, which will be referred to
collectively as dioxins, are ubiquitous in the environment.1 The
dioxin and furan compounds, chlorinated in the 2,3,7,8 positions on
the molecule, are highly toxic, environmentally persistent, and have a
tendency to bioaccumulate. Exposure to dioxins is a potential concern
for both cancer and noncancer effects, even at extremely low levels.
The EPA has concluded that 2,3,7,8-tetrachlorodibenzo-p-dioxin
(2,3,7,8-TCDD) and related compounds (congeners) are probable human
carcinogens.- In addition, EPA has concluded that there is adequate
evidence to support the inference that humans are likely to respond
with a broad spectrum of noncancer effects from exposure to dioxins,
if exposures are high enough.-
The occurrence of dioxin and dioxin-like compounds in the
environment appears to be primarily the result of human activities.i'2
The national estimated loading of these compounds from identified
sources into the environment is approximately 12,000 g toxicity
equivalents (TEQ)/yr. When this loading of 12,000 g TEQ/yr is
compared to annual loadings of other hazardous air pollutants (HAPs),
it appears to be relatively low. However, small quantities of dioxin
emissions can be of concern because of the high toxicity and
persistence, and tendency to bioaccumulate.- The draft dioxin
reassessment report estimated an average of 9,200 g TEQ/yr from known
combustion sources. Based on this study's estimate of 95 g TEQ/yr,
dioxin emissions from utilities represent about 1 percent of total
dioxin emissions. However, there are substantial uncertainties
associated with the dioxin emissions estimates due to several factors
including the following: emissions test data were available from only
13 utility units; data were not collected from all types of facilities
(e.g., no data were collected from units with hot-side ESPs); and many
measurements were below the minimum detection limit. Therefore, the
estimated emissions for the model plants described below are also
uncertain. The actual dioxin emissions could be greater than, or less
than, the predicted emissions presented in this chapter and other
chapters of this report.
There are several hypotheses on the chemistry and conditions for
the formation of PCDDs and PCDFs, including (1) contaminated feed
stock, (2) formation from precursors, and (3) formation de novo.- The
contaminated feedstock theory suggests that PCDDs and PCDFs in the
feed material survive the combustion process to become emissions.
However, this theory is not thought to be the principle explanation
for PCDD/PCDF formation. The second theory states that the formation
of PCDDs/PCDFs from precursors that are structurally related to
PCDDs/PCDFs occurs through thermal breakdown and rearrangement during
the combustion process. Precursors referred to in this theory include
11-1
-------
polychlorinated biphenyls, chlorinated phenols, and chlorinated
benzenes. This process is believed to occur when the precursor
condenses and adsorbs to the surface of fly ash particles in the
temperature range of 250-450° C. The third theory on the chemistry and
conditions of PCDD/PCDF formation states that creation is from
moieties that bear little structural resemblance to the PCDDs/PCDFs.
Such compounds may include petroleum products, chlorinated plastics,
non-chlorinated plastics, cellulose, lignin, coke, coal, particulate
carbon, and hydrogen chloride gas. This de novo formation is believed
to occur in the same temperature range used in the precursor theory,
but occurs as a reaction between a chlorine donor and another molecule
forms a chemical intermediate that then serves as a precursor in the
formation of PCDDs and PCDFs.
Emission data for various congeners of dioxins and furans were
obtained from 10 coal- and 2 oil-fired units using EPA Reference
Method 23. Eleven of these 12 tests resulted in one or more congeners
being identified at blank-corrected values above the minimum detection
level (MDL) for each individual test (there were usually three
replicate runs per test). Of the 188 sets of replicate analyses for
the various congeners, 109 (58 percent) had data values above the MDL
for all replicates. The dioxin/furan testing conducted at utility
boilers typically made use of field blanks, laboratory blanks,
calibration spiking, and strict measurement criteria during the
analyses.
As a larger portion of the dioxin/furan data sets were below the
MDL compared to other analyzed constituents, a summary of the protocol
for dealing with non-detect values is presented here. (See Appendix D
for further details.) Consistent with the methodology used for other
constituents, for those data sets containing at least one value above
the MDL, a run average was obtained by using one-half the MDL for
those values below the MDL in the averaging analysis. If all three
data values for a data set were below the MDL, that data set was not
used in the analyses. Using this approach resulted in a nationwide
dioxin emissions estimate of 95 g/yr TEQ for utilities.
There are other methodologies for handling non-detect values in
analyses. Reference Method 23, written for use on MWCs, indicates
that non-detects are to be handled as zero in the calculations. Use
of the Method 23 approach results in a nationwide dioxin/furan
emissions estimate for utility plants of 58 g/yr TEQ, which is roughly
2 times lower than the nationwide estimate of 95 g/yr TEQ reported
above. The EPA believes that the approach taken (i.e., using one-half
the MDL in the analyses) is appropriate for this analysis.
It has been hypothesized that the primary mechanism by which
dioxins enter the terrestrial food chain is through atmospheric
deposition.! The PCDDs and PCDFs have been found throughout the world
in all media, including air, soil, water, sediment, and in biota,
including fish and shellfish, and other plants and animals used as
food products. The ubiquitous nature of these compounds can be
11-2
-------
attributed to their stability under most environmental conditions and
also to the great number of sources located throughout the United
States. Once emitted into the atmosphere, the primary removal
mechanisms are degradation and deposition to environmental media, such
as soil, water, and vegetation. In general, the compounds have very
low water solubility and vapor pressures and high octanol/water
partition coefficients (Kow) and organic carbon partition coefficients
(Koc). These measures of chemical properties indicate that dioxins
tend to strongly adsorb to soils and, as a result, resist leaching or
volatilization. The fate of dioxins adsorbed to particles includes
burial in place, resuspension into the air, and erosion of soils into
waterbodies. In the aquatic environment, PCDDs and PCDFs remain
adsorbed to particulate matter, and those compounds that enter surface
waters and dissolve will tend to partition to suspended solids or
dissolved organic matter. The primary removal mechanism from the
water column is sedimentation and, ultimately, burial of sediments.
Once PCDDs and PCDFs are deposited and make their way into
various media and into biota, they are available for human uptake
through ingestion. The draft Dioxin Reassessment report states that,
with regard to average intake, humans are currently exposed to
background levels of dioxin-like compounds, including dioxin-like
polychlorinated biphenyls (PCBs) , on the order of 3 to 6 pg TEQ/kg
body weight/day.3 This value is more than 500-fold higher than the
EPA's 1985 risk-specific dose of 0.006 pg TEQ/kg body weight/day
associated with an upper-bound risk of 1 in a million (1 x 10 ~s) and is
several hundredfold higher than the revised risk-specific dose
estimates presented in the draft Dioxin Reassessment report.! An
inhalation exposure assessment for utility emissions was performed for
priority HAPs using the human exposure model (HEM), as described in
chapter 6 of this report. The cancer risk to the maximum exposed
individual in that assessment was estimated to be 1 x 10 ~7.
As part of the draft dioxin reassessment, methodologies for
conducting site-specific indirect exposure modeling are presented.1
These methodologies are intended for use in evaluating incremental
exposures (i.e., not background exposures) associated with specific
sources of dioxin-like compounds. In the dioxin reassessment
documents, example scenarios were developed and modeling was conducted
for six exposure scenarios to demonstrate these methodologies. Of the
six scenarios, scenarios 4 and 5 focused on indirect exposure
resulting from stack emissions from stationary combustion sources. Of
the emission sources modeled, scenarios 4 and 5 best represent
exposures that could possibly be roughly similar to exposures
resulting from the utility emissions. However, as stated in the draft
dioxin reassessment documents, in evaluating the results from this
modeling effort, it should be noted that the demonstration scenarios
were developed only to illustrate the site-specific methodologies and
that the exposure estimates generated for each scenario are not
generalizable to other sites.
11-3
-------
The results from the indirect exposure modeling effort presented
in the draft dioxin reassessment report indicate that exposure levels
due to the consumption of fish obtained from an impacted stream
dominated the results generated for scenario 4. Under scenario 5, the
highest levels of exposure occurred through the ingestion of fish and
homegrown beef and milk. In summary, results for the two scenarios
indicate that consumption of fish and consumption of beef and milk can
be significant routes of exposure of humans to dioxin and dioxin-like
compounds.
The transfer of dioxins from air to plants plays a major role in
the exposure of terrestrial animals to dioxins. A finding in the
draft dioxin reassessment report is that the principal cause for
terrestrial food chain contamination is the transfer of dioxins from
the air to vegetation that animals consume. It has been found that
dioxins in the vapor phase can transfer readily to plants and that
this is the primary pathway by which plants take up dioxins.- This is
of significance since the uptake of plants by foraging animals,
including cattle, is relevant to the concentrations of dioxins to
which humans are exposed through the ingestion of beef and other
animal products.
In addition to exposure pathways discussed above, exposure
through the consumption of breast milk appears to be of potential
concern for nursing infants. A study of 42 nursing mothers revealed
that an average concentration of 16 ppt of TEQ was found in the lipid
portion of their breast milk. A similar study conducted in Germany
revealed an average level of 29 ppt of TEQ in the lipid portion of the
breast milk. Based on the estimated adult intake of dioxin discussed
above, an exposure duration of 1 year (i.e., the infant nurses for 1
year), an average weight gain of 10 kg during the exposure period, and
a milk concentration of 20 ppt of TEQ, it is estimated that the
average daily dose to the infant over this period is approximately
60 pg of TEQ/kg/day, 20 to 60 times higher than the estimated range
for background exposure to adults (1 to 3 pg of TEQ/kg/day).
Because dioxins tend to accumulate in the environment and because
they are extremely toxic to humans and wildlife, even small amounts of
these compounds emitted from specific sources like utilities may be of
concern. Based on background media concentration data and other
assessments conducted by the Agency, including the one discussed
above, it can be hypothesized that the primary human exposure routes
of concern are those that are related to the ingestion of food
products (e.g., fish, meat, and dairy products). Animal exposure
through these routes is most likely to occur through the consumption
of animals or fish contaminated through the ingestion of contaminated
media or organisms or through the consumption of vegetation
contaminated by atmospheric deposition.
Dioxins were identified (in chapter 5) as high priority for
multipathway risk assessment. This chapter presents a screening-level
assessment of the multipathway exposures and risks associated with
11-4
-------
dioxins. The local-scale dioxin and furan emissions estimates used in
this screening-level assessment were from four model plant fossil-
fuel-fired utility boilers, which were developed for this analysis.
The cumulative effects of long-range transport of utility boiler
PCDD/PCDF emissions are also considered in the screening-level
analysis. The model plants were developed to represent actual utility
coal- and oil-fired boilers, and were designed to characterize
potential emission rates and atmospheric release processes for average
large and small boilers of each fuel type. The assessment presented
in this report used long-range and local atmospheric transport and
deposition modeling and direct and indirect multipathway exposure
modeling to predict cancer risks from inhalation and ingestion of
dioxin- and furan-contaminated air, water, soil, and food.
Polychlorinated dioxins and furans may also cause noncancer
health effects in humans. Developmental, reproductive, and immune
system endpoints have been reviewed as part of the dioxin reassessment
effort.- However, reference concentrations have not been established
by EPA for use in human risk assessment. As a result, only the cancer
endpoints have been examined in this study. Additionally, the breast
milk exposure pathway for infants has been recognized as a significant
source of exposure, as described above. However, quantitative
assessment of this pathway is not included in this analysis.
This analysis of noninhalation exposures to dioxin emissions is a
screening analysis. Thus, these quantitative exposure and risk
results, because of the many modeling and analytic uncertainties, are
very uncertain and do not conclusively demonstrate the existence of
health risks of concern associated with exposures to utility emissions
either on a national scale or from any actual individual utility. The
lack of measured data around these sources precludes a comparison with
modeled results. These results do suggest that exposures and risks of
concern cannot at present be ruled out and that there is a need for
development of additional scientific information to evaluate whether
risk levels of concern may exist.
EPA's Industrial Source Complex Short Term, Version 3 (ISCST3)
model was used to estimate individual congener ambient air
concentrations and wet/dry deposition rates associated with emissions
from each of the model plants.4 The ISCST3 uses Gaussian dispersion to
calculate air concentration and deposition in the local area (i.e.,
within 50 km of the emissions source). Particle size, atmospheric
conditions and gravitational settling velocities are used to predict
dry deposition. Precipitation rate and particle size are used to
estimate scavenging coefficients which are used to predict wet
deposition.
Long-range transport of dioxins emitted from multiple power
plants is also a potential cause of increased population risk.
Therefore, long-range modeling of utility dioxin emissions using the
RELMAP model are summarized below for consideration in the screening-
level multipathway exposure assessment.
11-5
-------
After the air concentrations and deposition rates were modeled,
multipathway exposures were estimated using a modified version of the
Indirect Exposure Methodology (IEM) - Spreadsheet provided by EPA
Office of Research and Development. The IEM methodology uses site-
specific exposure scenarios and dispersion model predictions of air
concentrations and deposition rates to calculate pollutant
concentrations in vegetation, soil, water, and the aquatic and
terrestrial food chain. The methodology then calculates human
exposure from data on ingestion and dermal contact.
The following section (section 11.2) summarizes the RELMAP long-
range dispersion modeling for dioxins. Section 11.3 describes the
methods and data used in this risk assessment, including emissions
sources, local dispersion modeling, and multipathway exposure and risk
modeling. In section 11.4 the results of the risk assessment are
presented for the different scenarios and pathways considered.
Section 11.5 describes how the results were tested for their
sensitivity to changes in various assumptions about the model
parameters. The final section (section 11.6) provides conclusions on
the analysis, limitations, and uncertainties.
11.2 LONG-RANGE TRANSPORT MODELING
Long-range atmospheric transport modeling was performed for the
continental United States using the RELMAP model. RELMAP modeling
allows consideration of emissions contributions from multiple sources
on a national scale. More detailed description of the RELMAP model,
methods, and uncertainties can be found in chapter 6 of this report.
The methods and emissions sources used in the RELMAP modeling of
dioxins were patterned after the RELMAP modeling described here in
chapter 6. The dioxin RELMAP results described here consider the
combined TEQ emissions from all U.S. coal- and oil-fired utilities.
The RELMAP model was used to generate air concentration, wet
deposition, and dry deposition results for dioxin emissions for the
contiguous 48 states. The results of all three analyses indicate that
long-range transport provides higher air concentrations and deposition
rates for dioxins in the Eastern United States, with the maximum for
each falling in the general region of the Ohio River Valley.
Figure 11-1 presents the RELMAP results for dioxin air
concentration. Long-range transport of emissions generated air
concentrations ranging from less than 10 to a maximum of 316
attograms/m3, with an average of 19 attograms/m? . Figure 11-2 provides
the RELMAP wet deposition results. The results show that long-range
transport accounted for wet depositions ranging from less than one to
a maximum of 93 picograms per square meter per year (pg/m2/yr) . The
overall average wet deposition for Figure 11-2 was 25 pg/m2/yr. The
dry deposition model results (Figure 11-3) are slightly less than the
wet deposition results. The dry deposition ranged from less than one
to a maximum deposition of 82 pg/m2/yr, with an average of 24 pg/m2/yr.
11-6
-------
5 to 20
20 to 50
50 to 100
100 to 200
>= 200
Figure 11-1. Results of the RELMAP Modeling Analysis from 1990 Emissions
Estimates for Total Dioxin (incl TEQ factors) from Coal and Oil
Utilities: Predicted Air Concentration of Total Dioxin,
Units: attograms (10~18 grams)/m3
-------
I
CO
1to5
5 to 10
10 to 20
20 to 50
>= 50
Figure 11-2. Predicted Total Dioxin (incl TEQ factors) Wet Deposition from Coal
and Oil Utilities Based on 1990 Emissions Estimates as Modeled with RELMAP,
Units: picograms (10~12 grams)/m3
-------
1to5
5 to 10
10 to 20
20 to 50
>= 50
Figure 11-3. Predicted Total Dioxin (incl TEQ factors) Dry Deposition from Coal
and Oil Utilities Based on 1990 Emissions Estimates as Modeled with RELMAP,
Units: picograms (10~12 grams) /m3
-------
11.3 RISK ASSESSMENT METHODOLOGY
11.3.1 Emissions Sources
A model plant approach was used to describe local-scale dioxin
emissions from utility boilers. Four model plants were selected, one
each for: large coal-fired, small coal-fired, large oil-fired, and
small oil-fired. The selection of large and small model plants
represents an upper- and lower-end depiction of dioxin emissions. The
dioxin emissions rates for the model plants were developed based on
dioxin emissions estimates from the EPA database of 426 coal-fired
plants and 145 oil-fired plants. The 426 coal-fired and 137 oil-fired
plants were ranked by 2,3,7,8-TCDD TEQ emissions (kg/y) and then
divided into thirds. The large model plants represent the average
dioxin emissions for the upper ranking third of the plants for each
fuel type (e.g., emissions for the large coal-fired utility model
plant are the average dioxin emissions for the 142 largest emitting
coal-fired utility plants), while the small model plants represent the
average dioxin emissions from the lower third of the plants for each
fuel type. The resulting dioxin and furan congener-specific emissions
rates and process parameters for each for each model plant are shown
in Tables 11-1 and 11-2.5
The emissions information for each of the four model plants was
applied for two locations. One location (case #1) was modeled using
the emissions from each of the model plants combined with high-end
meteorology and terrain conditions. Albany, New York, was selected to
be the case #1 meteorological location, as this location has
previously been shown to have high-end meteorology for the purpose of
dispersion modeling. Specifically, Albany's geographic location in
the Hudson River Valley limits wind flow patterns, which generally
limits the dilution effects of varied wind flow patterns. Complex
terrain was combined with the Albany meteorological data to complete
the case #1 application of the model plant data. Similarly, all model
plant data were applied to more average meteorology and simple
terrain. Springfield, Missouri, was selected as a location for
average meteorological conditions (case #2). The application of two
locations and terrain conditions to the model plant data gives a total
of eight model plants considered in the assessment (Table 11-3).
11.3.2 Local Air Dispersion Modeling
The EPA's ISCST3 dispersion model was used to predict the local-
scale (<50 km from the source) atmospheric dispersion of model plant
dioxin emissions. ISCST3 is a Gaussian plume model that has the
capacity to model wet and dry depletion with corresponding annual
deposition rates, as well as ambient air concentrations. A more
detailed description of the ISCST3 model may be found in the other
multipathway assessments of this report (Chapters 7 and 10), or in
other EPA publications.* The resulting annual depositions and ambient
concentrations are used as inputs to the IEM-spreadsheet model for the
purpose of multipathway exposure and risk analysis.
11-10
-------
Table 11-1
(kg/yr)
Congener-Specific Emissions Rates for Model Plants
Congener3
PCDDs:
2378
12378
123478
123789
123678
1234678
Octa
PCDFs:
2378
23478
12378
123478
123678
123789
234678
1234678
1234789
Octa
Large coal
1.24e-05
3.22e-04
7.326-04
5.546-04
4.396-04
4.346-04
4.416-03
2.996-04
7.886-04
1.856-04
9.776-04
3.026-04
6.406-04
1.236-03
1.496-03
1.316-02
1.256-03
Large oil
1.826-04
1.61e-04
3.466-04
2.336-04
1.526-04
5.546-04
6.476-04
1.286-04
1.346-04
1.21e-04
1.696-04
1.076-04
1 .62e-04
1.336-04
2.626-04
2.826-04
2.826-04
Small coal
1.676-04
4.346-05
9.876-05
7.466-05
5.926-05
5.856-05
5.946-04
4.026-05
1.066-04
2.506-05
1.326-04
4.076-05
8.636-05
1.656-04
2.016-04
1.776-03
1.686-04
Small oil
8.716-06
7.706-06
1.666-05
1.126-05
7.286-06
2.666-05
3.106-05
6.156-06
6.446-06
5.806-06
8.096-06
5.146-06
7.796-06
6.386-06
1.266-05
1.356-05
1.356-05
PCDDs = polychlorinated dibenzo-p-dioxins; PCDFs = polychlorinated dibenzofurans
a Congener: Numbers represent chlorine placement on the molecule.
Data in this table provided by EPA for this analysis.s
Table 11-2. Process Parameters for Model Plants
Model plant
Large coal-fired
(800 MW)
Small coal-fired
(232 MW)
Large oil-fired
(506 MW)
Small oil-fired
(174 MW)
Stack height (m)
211.8
119.5
157.0
70.10
Stack diameter (m)
8.290
4.938
6.584
3.962
Exit velocity (mis)
28
22
26
19
Exit temp. (K)
412.6
399.3
434.8
419.8
MW = megawatt
Data in this table provided by EPA for this analysis."
11-11
-------
Table 11-3. Naming Scheme for Eight Model Plants
LCH
SCH
LOH
SOH
LCC
sec
LOG
SOC
Large
Small
Large
Small
Large
Small
Large
Small
Coal-Fired Plant, High-End Meteorology and Complex Terrain
Coal-Fired Plant, High-End Meteorology and Complex Terrain
Oil-Fired Plant, High-End Meteorology and Complex Terrain
Oil-Fired Plant, High-End Meteorology and Complex Terrain
Coal-Fired Plant, Central-tendency Meteorology and Simple Terrain
Coal-Fired Plant, Central-tendency Meteorology and Simple Terrain
Oil-Fired Plant, Central-tendency Meteorology and Simple Terrain
Oil-Fired Plant, Central-tendancyMeteorology and Simple Terrain
The ISCST3 model requires inputs related to the processes and
location of modeled facilities. The ISCST3 dispersion modeling
requires various emissions source process parameters for dispersion
and deposition modeling, including stack height, stack diameter, exit
velocity, and exit temperature. The value of each of these inputs for
each of the model plants is presented in Table 11-2.- Other process-
specific dispersion parameters include particle size distributions,
mass fraction of particles, and wet and dry scavenging coefficients
(Table 11-4) . These data were obtained for utility boilers from AP-
42,6 Perry's Chemical Engineering Handbook,7 and the User's Guide for
the Industrial Source Complex Dispersion Models,* respectively.
Additional emissions source information, such as meteorology, complex
terrain, and land use factors, was varied between case #1 and case #2
(Table 11-5). The ISCST3 meteorological inputs were five years of
data from the Solar and Meteorological Surface Observation Network
(SAMSON) for each of the locations considered in this analysis.8 The
Albany, New York, terrain information was obtained through the United
States Geological Survey.9
ISCST3 modeling was completed using the "default" model options
specified in the Guidelines on Air Quality Models. 10 A unit emissions
rate (1 g/s) was used in modeling the ambient air concentrations and
wet and dry deposition rates. A polar array of receptors was used,
with receptors placed 22.5 degrees apart on concentric circles 500 m
(for case #1) and 5,000 m (for case #2) from the source. The case #1
and case #2 receptor distances were patterned after prevous dioxin
indirect exposure assessments.-'11 The receptor distances from the
source were patterned after these studies because experience has shown
that, in general, exposure is increased closer to the source, and
these locations were believed to be reasonable distances for potential
exposure scenarios. The locations were tested for sensitivity
regarding the modeled exposure/risk results (see section 11.5.3,
below), and determined to be adequate for this analysis.
11-12
-------
Table 11-4. Process-Specific Depletion Parameters
Parameter Value
Gas depletion
Gas Scavenging coefficient (s~1) 1.8 x 10~4
Particle depletion
Particle size distribution (microns) 15 10 6 2.5 1
Mass fraction per particle size: coal-fired3 .005 0.09 0.20 0.23 0.43
Mass fraction per particle size: oil-fireda 0.31 0.115 10.5 10.5 36.5
Particle density (g/cm3)3 1.4 1.4 1.4 1.4 1.4
Particle scavenging coefficient (s~1) 6.6 x10'4 6.6 x10'4 4.2 x10'4 1.8 x10'4 4.3 x10'5
Particle size distributions by mass taken from AP-42*data for dry bottom boilers burning bituminous coal and residual oil-
burning utility boilers. Particle density estimated for bituminous coal zand confirmed with EPA/ORD.
Table 11-5. Other Emissions Source Information
Setting type Case #1 Case #2
Location Data
Surface/upper air Albany, NY/Albany, NY Springfield, MO/Monett, MO
Anemometer height (m) 10.00 6.10
Preprocessor Inputs
Land use within 5 km Suburban/rural Suburban/rural
Minimum Monin-Obukov length (m) 2.00 2.00
Roughness height (m) 0.50 0.33
Noon-time albedo (fraction) 0.28 0.22
Bowen ratio (fraction) 0.50 0.51
Net radiation absorbed in ground (fraction) 0.15 0.15
Anthropogenic heat flux (watts/m2) 0.00 5.40
ISCST3 was used to generate a wet and dry deposition rate, as
well as an ambient air concentration for both particle and vapor phase
emissions. To determine the separate particle and vapor fractions of
the emissions, each model plant case was run twice: once for particle
phase emissions and once for vapor phase emissions. A particle
modeling run was performed for each model plant using the particle
information in Table 11-4, with the output options of wet and dry
deposition and air concentration. A vapor, or gas phase, modeling
run was also performed for each model plant by omitting particle
11-13
-------
information and allowing only air concentration and wet depletion
outputs. It was assumed that gaseous emissions behaved like small
particles for the purpose of gas scavenging and wet deposition
modeling. Dry deposition of vapors was approximated using a dry
deposition rate of 0.2 cm/s, which was estimated from previous EPA
research12 and cited in the secondary aluminum smelter indirect
exposure analysis for dioxins. " For each of the eight model plant
cases, the maximum vapor and particle air concentration and wet and
dry deposition rates were converted from the unit emissions results to
congener specific values using the model plants' individual dioxin
congener emissions rates (Table 11-1) . The vapor and particle
congener specific values were then multiplied by the appropriate
vapor/particle fraction for ambient conditions.- Additionally,
particle air concentrations were modified to reflect only particles
less than approximately 10 microns, based on the initial mass fraction
in Table 11-4, for the inhalation exposure.
11.3.3 Exposure Modeling and Risk Calculation
Exposure and risk modeling were performed using the IEM-
Spreadsheet. The spreadsheet model is based on the Methodology for
Assessing Health Risks Associated with Indirect Exposure to Combustor
Emissions13 and its Addendum.14 The I EM- spreadsheet model uses the
dispersion modeled annual air concentrations and wet and dry
deposition rates to estimate direct and indirect exposure and risks
associated with dioxin model plant emissions. Direct exposure and
risks are determined from the inhalation of airborne emissions
particles and vapors. Indirect (non-inhalation) exposure and risks
are those associated with human contact with the dioxins that have
accumulated in the environment. The model assesses indirect exposures
and risks by using a simplified mass balance approach to determine the
fate and concentrations of emissions in environmental media including
surface soils, surface water, fish, vegetation, and domestic animals.
Using estimates of human contact, these environmental media
concentrations are translated to the indirect (non-inhalation) human
exposures and the related cancer risk values. The different routes or
pathways of indirect human contact considered by the IEM-spreadsheet
model include dermal contact with soil and ingestion of water, soil,
fish, plants, and domestic animals. A flow chart diagram of the
processes involved in the multipathway analysis and the IEM-
spreadsheet model are presented in Figure 11-4.
The version of the IEM-spreadsheet used in this assessment was
developed for dioxins by the EPA Office of Research and Development
(EPA/ORD) .15 The spreadsheet was modified, as advised by ORD, to be
more consistent with the current assumptions and methods used by ORD
in their work on dioxins. The spreadsheet was modified with a single
soil loss rate constant, a biota sediment accumulation factor (BSAF),
and an increased no-till depth of soil incorporation, to suit the
recommendations of EPA/ORD. The single soil loss rate constant has a
value of 0.0277 y"1, which is based on a new estimate of a 25-year
11-14
-------
Model Power
Plant Dioxin
Emissions
Direct
Inhalation
Direct
Inhalation Risk
Total Direct
Risk
Vegetable
Ingestion Risk
Atmospheric
Processes
Food
Ingestion
Livestock
Ingestion Risk
Fish Ingestion \
Risk /
Total Indirect
Risk
Soil
Processes
Soil Ingestion
Risk
Dermal
Contact
Dermal
Contact Risk
S u rfa ce
Water
Processes
Water
Ingestion
Water
\ Ingestion Risk
Figure 11-4. Flow Chart of Multipathway Processes
half-life for dioxins in surface soils." The fish accumulation
pathway was adjusted to use BSAF calculations because dioxins are
"super-hydrophobic," and thus are not detected in the water column
even in highly contaminated environments where the aquatic life show
high tissue concentrations.1 In the IBM-spreadsheet model, the
hydrophobic compounds are assumed to be adsorbed to the hydrophobic
components of aquatic sediments. The BSAF is the preferred method of
calculation because it bases the concentration in fish on the
concentration in adsorbed to bottom sediments, thus providing a more
accurate relationship between aquatic environment concentrations and
fish tissue concentrations. The BSAF values used in this assessment
were based on those in Estimating Exposure to Dioxin-Like Compounds1
and additional data from the Great Lakes Water Quality Initiative
Technical Support Document for the Procedure to Determine
Bioaccumulation Factors,16 provided by EPA/ORD." The depth of no-till
soil incorporation was increased based on new data,17 which was
recommended by EPA/ORD." In addition to the above stated
11-15
-------
modifications, the animal bioaccumulation methodology was modified to
reflect Estimating Exposure to Dioxin-Like Compounds^ due to the
availability of data from this document. Also, the original version
of IEM-Spreadsheet did not calculate direct inhalation exposures and
risk, so the model was modified to calculate direct inhalation risk,
per the Addendum to the IBM Methodology.—
Aside from the parameters and recommended changes described
above, all other IEM-spreadsheet inputs were chosen specifically for
this study. Each input parameter, its modeled value, and the source
of the data are presented in Appendix G-l of this report. Every
attempt was made to keep the parameters used in this chapter
consistent with those used in the multipathway assessments for mercury
and arsenic (chapters 7 and 10 of this report). This includes the
description of the watershed, which was assumed to have the same
characteristics as the hypothetical watershed described in the mercury
and arsenic chapters. The IEM-spreadsheet inputs related to human
exposure were obtained from the Mercury Study Report to Congress18 due
to the parallel observed between the mercury multipathway analysis in
the Mercury Study Report to Congress and the mercury multipathway
analysis in Chapter 7. For parameters specific to dioxin and furan
congeners, the preferred source of input data was Estimating Exposure
to Dioxin-Like Compounds.- Additional site-specific information was
obtained for parameters related to the assumed model plant locations
(see section 11.3.2 above for location descriptions); and for
parameters not specified by the sources listed above, general defaults
were obtained from the methodology supporting the IEM-spreadsheet. ">"
The indirect exposure methodology and input parameters described
above were used to estimate exposures to dioxin and furans. The case
#2 exposure and risk modeling scenarios used the results of the
dispersion modeling (see section 11.3.2 above) and exposure
assumptions related to case #2 receptors. The case #2 receptors were
based on a resident adult and child scenario. The case #2 scenarios
were used in modeling exposure and cancer risks through the direct
inhalation, soil dermal contact, soil ingestion, home-grown vegetable
ingestion, and surface-supplied drinking water ingestion pathways.
The case #1 exposure and risk modeling scenarios employed the results
of the case #1 dispersion modeling (see section 11.3.2 above) and
exposure assumptions related to case #1 receptors. The case #1
receptors were based on subsistence fisher and subsistence farmer
scenarios. These case #1 receptors were considered for all the
pathways described for the case #2 receptors. Additionally, the
subsistence farmer scenario considered ingestion of home-grown animal
products and the subsistence fisher scenario considered ingestion of
locally caught fish. The case #1 scenarios also used different
vegetable ingestion contact fractions. The vegetable ingestion
contact fractions represent a subsistence farmer, a rural gardener,
and urban gardener" for the subsistence farmer, subsistence fisher,
and residents, respectively. Table 11-6 presents a summary of the
scenarios and pathways considered in this assessment, and Table 11-7
presents all 16 hypothetical scenarios included in this assessment.
11-16
-------
Table 11-6. Summary of Receptor Scenarios and Pathways
Receptor scenario
Pathways considered
Adult resident
(used for case #2 modeling scenarios only)
Child resident
(used for case #2 modeling scenarios only)
Subsistence fisher
(used for case #1 modeling scenarios only)
Subsistence farmer
(used for case #1 modeling scenarios only)
Direct inhalation, soil dermal, soil ingestion, vegetable ingestion,
and drinking water ingestion
Direct inhalation, soil dermal, soil ingestion, vegetable ingestion,
and drinking water ingestion
Direct inhalation, soil dermal, soil ingestion, vegetable ingestion,
drinking water ingestion, and fish ingestion
Direct inhalation, soil dermal, soil ingestion, vegetable ingestion,
drinking water ingestion, and animal product ingestion
Table 11-7. The 16 Hypothetical Scenarios Included in the
Screening Level, Model Plant, Dioxin Multipathway Exposure and
Risk Assessment.
Scenario
LCH-fisher
LOH-fisher
SCH-fisher
SOH-fisher
LCH-farmer
LOH-farmer
SCH-farmer
SOH-farmer
LCC-resident
LOC-resident
SCC-resident
SOC-resident
LCC-child
LOC-child
SCC-child
SOC-child
Description
Subsistence fisher living near large coal-fired plant, using case #1 modeling assumptions and
inputs
Subsistence fisher living near large oil-fired plant, using case #1 modeling assumptions and
inputs
Subsistence fisher living near small coal-fired plant, using case #1 modeling assumptions and
inputs
Subsistence fisher living near small oil-fired plant, using case #1 modeling assumptions and
inputs
Subsistence farmer living
inputs
Subsistence farmer living
inputs
Subsistence farmer living
inputs
Subsistence farmer living
inputs
Adult resident living near
Adult resident living near
Adult resident living near
Adult resident living near
Child resident living near
Child resident living near
Child resident living near
Child resident livinq near
near large coal-fired plant, using case #1 modeling assumptions and
near large oil-fired plant, using case #1 modeling assumptions and
near small coal-fired plant, using case#1 modeling assumptions and
near small oil-fired plant, using case #1 modeling assumptions and
large coal-fired plant, using case #2 modeling assumptions and inputs
large oil-fired plant, using case #2 modeling assumptions and inputs
small coal-fired plant, using case #2 modeling assumptions and inputs
small oil-fired plant, using case #2 modeling assumptions and inputs
large coal-fired plant, using case #2 modeling assumptions and inputs
large oil-fired plant, using case #2 modeling assumptions and inputs
small coal-fired plant, using case #2 modeling assumptions and inputs
larqe oil-fired plant, usinq case #2 modelinq assumptions and inputs
11-17
-------
11.4 DISPERSION, EXPOSURE, AND RISK RESULTS
This section presents the relationship between dioxin emissions
from the model plant stack and human cancer endpoints. Dispersion
modeling provided estimates of ambient air concentrations and wet and
dry deposition rates. The IEM-spreadsheet modeling estimated dioxin
concentrations in environmental media (i.e., soil, water, plant,
animal, and fish). The IEM-spreadsheet also estimated human exposure
to these environmental concentrations to approximate the corresponding
human exposure through various pathways (see Table 11-6) in terms of
lifetime averaged daily doses (LADDs) and cancer risk estimates. The
exposure and risk results of the local-scale and long-range modeling
were compared.
The dispersion modeling results for each model plant at both the
case #1 and case #2 receptor distances are presented in Appendix G-2.
The data in Appendix G-2 are the maximum vapor and particle phase
ambient air concentrations and wet/dry deposition rates for each model
plant, under both case #1 and case #2 conditions (see sections 11.3.1
and 11.3.2 above for description). The dispersion modeling results
indicate that the case #1 exposures stem entirely from wet deposition,
with the exception of very low air concentration and dry deposition
associated with the small oil-fired model plant. It is likely that
the air concentration and dry deposition rate are zero because the
modeled plume of emissions has not yet settled to ground level at this
distance. From this it can be concluded that the modeled case #1
receptors, located at 500 m, probably are not at the highest direct
inhalation exposure location because receptors are actually too close
to the very tall stacks. However, the 500 m distance has much higher
wet deposition rates, which may account for higher indirect exposures.
Following the observation of wet deposition versus plume impaction,
the sensitivity of the results to the receptor distance from the stack
was analyzed (see section 11.5.3 below).
The environmental media dioxin concentrations resulting from the
mass balance performed by the IEM-spreadsheet are presented in
Appendix G-3. The concentrations are reported for surface soil,
surface water, whole fish, and plants and animals produced on
contaminated lands. The results indicate that the bioaccumulated
concentrations in organisms (fish, plants, and animals) exceed those
that persist in soil and water. This is a reasonable result, given
the sequestration of the lipophilic dioxin and furan compounds in the
organisms' tissues.
The congener-specific LADDs for each of the exposure pathways and
scenarios modeled in this chapter are detailed in Appendix G-4. The
LADDs were calculated by the IBM-spreadsheet model, using the media
concentrations described above and the input parameters related to
human contact and exposure (Appendix G-l). The greatest LADD is seen
in the subsistence fisher fish consumption pathway, with animal
product and vegetable consumption following closely. This observation
is consistent with the modeled media concentrations because each of
11-18
-------
these pathways involves consumption of an organism that has
bioaccumulated dioxins from the environment.
Background exposures to dioxins in the United States have been
estimated.- The estimated exposures range from 0.8 pg/day for soil
ingestion to 37.0 pg/day for ingestion of beef and veal. The
background exposure from fish ingestion is listed as 7.8 pg/day, and
the background exposure from inhalation is 0.8 pg/day. After
converting the units on the modeled LADD exposures in Appendix G-3,
the corresponding maximum modeled exposures (which do not take into
account background) are 2.1 pg/day for soil ingestion, 140 pg/day for
animal product ingestion, 420 pg/day for fish ingestion, and 0.0014
pg/day for direct inhalation. In each of the indirect pathways, the
modeled exposures exceed background.
The congener-specific results of cancer risk calculation for each
pathway and each scenario are presented in Appendix G-5. The cancer
risk values were calculated using a slope factor for dioxin of 0.156
kg-d/ng, and the Toxicity Equivalency Factors (TEFs) for 2,3,7,8-
dioxin congeners (see Appendix G-l). This slope factor was adopted
from Estimating Exposure to Dioxin-Like Compounds;- the derivation of
the factor is described in the report Risk Analysis of TCDD
Contaminated Soil.19 The calculated risks show the same pattern as the
LADDs, with the highest risks generally associated with fish
consumption. A summary of the predicted dioxin TEQ cancer risks for
each hypothetical scenario is provided in Table 11-8. Note that the
zero direct risks in the case #1 exposure scenarios are due to the
selection of 500 m as the receptor location. As described above, at
this location the ambient air concentrations are zero, but the
deposition is near to maximum.
The TEQ cancer risks are greater for the combined indirect (non-
inhalation) exposures than the direct (inhalation) exposures in every
modeled scenario. In all cases, the indirect risks are at least an
order of magnitude larger. These results demonstrate the need to
consider indirect risks for environmental pollution that is persistent
in the environment, such as dioxins. This is especially evident in
the case #1 analyses where the direct risk is zero (because vapor and
particle air concentrations are zero), yet the indirect risks from the
multipathway exposures based on wet deposition are as high as 2 x 10 "4.
The highest risk, predicted to be 2 x 10 "4, is to the subsistence
fisher hypothetical scenario from the indirect (non-inhalation)
pathways for the large coal-fired model plant. Given the high level
of risk predicted for the subsistence fisher, this value is explored
further in a sensitivity analysis (see section 11.5.2 below). The
maximum direct (inhalation) risk for this model plant was 4 x 10 "10,
which is orders of magnitude less than the risk predicted for the
maximum exposed individual considered by the direct inhalation
assessment in chapter 6 of this report. This large difference in
inhalation risk estimates may be due to the selection of receptor
distance in the model plant analysis.
11-19
-------
Table 11-8. Summary of Predicted Cancer Risks from the Screening
Level Multipathway Assessment for Model Plants, for 16 Hypothetical
Scenarios
Scenario3
LCH-fisher
LOH-fisher
SCH-fisher
SOH-fisher
LCH-farmer
LOH-farmer
SCH-farmer
SOH-farmer
LCC-resident
LOC-resident
SCC-resident
SOC-resident
LCC-child
LOC-child
SCC-child
SOC-child
Predicted cancer risk due to direct
inhalation exposure1"
0
0
0
1E-13
0
0
0
1E-13
2E-10
2E-10
4E-10
7E-11
1E-10
1E-10
2E-10
5E-11
Predicted risk due to indirect (non-
inhalation) exposure1"
1 E-04 to 2E-04
5E-05 to 1 E-04
3E-05 to 6E-05
3E-06 to 6E-06
1 E-05 to 3E-05
1 E-05 to 2E-05
5E-06 to 1 E-05
5E-07 to 1 E-06
2E-08 to 5E-08
1 E-08 to 2E-08
1 E-08 to 3E-08
3E-9 to 6E-09
1 E-08 to 2E-08
1 E-08 to 2E-08
1 E-08 to 2E-08
1 E-09 to 2E-09
3 See Tables 11-3, 11-6, and 11-7 and text for description of scenarios and associated assumptions, methods and
inputs.
b The risks are presented as a range. For each scenario, the higher predicted risks are based on results of the
modeling method, assumptions, and data inputs as described in this chapter. The lower predicted risks for each
scenario are based on the assumption that the emissions estimates for the model plants (Table 11-1) could
possibly be overestimated by a factor of 2 (see discussion in section 11.1); therefore, it is predicted (based on the
assumption that the exposure and risk calculations are reasonably linear) that the risks for the hypothetical
scenarios could be about 2 times lower than those predicted using inputs and methods described in this chapter if
the emissions rates using in the calculations were 2 times lower than those shown in Table 11-1.
The predicted risks presented in Table 11-8 are screening level
estimates based on model plants and hypothetical scenarios. The
results do not apply to any actual utility plant. There are
substantial uncertainties in these predicted risks because of numerous
assumptions and data inputs with varying levels of uncertainty. For
example, the emissions estimates for the model plants are quite
uncertain and may be overestimated, or possibly underestimated (see
11-20
-------
section 11.1 for discussion of emissions data uncertainty).
Additional uncertainties and variability in the assessment and data
inputs are discussed in section 11.5.
The multipathway exposures and risks resulting from the local-
scale ISCST3 dispersion modeling results (described in the preceding
paragraphs) were compared to those resulting from the RELMAP modeling
of long-range transport (Appendix G-6). The comparison was made for
dioxin TEQ rather than the congener specific method used in the local-
scale modeling, due to the limited data available from RELMAP. To
obtain exposure and risk estimates, the average RELMAP data for air
concentration, and wet and dry deposition were used with the IEM -
spreadsheet model (thus paralleling what was done for the local-scale
ISCST3 dispersion result). It was assumed that the air concentration
results of the RELMAP model (Figure 11-1) were inhalable for the
purpose of inhalation exposure. A similar comparison using background
air concentration and deposition rates was not preformed for this
study due to limited data.
The comparison shows that in most cases the local-scale emissions
exposures are a greater percentage of the total exposures than the
long-range transport exposures. The importance of the local-scale
emissions generally decreases as distance from the facility increases.
Thus, long-range transport was more influential in the case #2
modeling scenarios, for which local-scale modeling was at a greater
distance from the facility. Long-range transport was most influential
in the direct inhalation pathway and the vegetable ingestion pathway
because the RELMAP modeled air concentrations were relatively large
portions of the total TEQ air concentration.
11.5 UNCERTAINTY AND SENSITIVITY ANALYSIS
The models and assumptions used in this study have been the
subject of significant review and revision as part of other EPA
studies.20 The reviews have identified limitations of the methods and
the data sets from which the parameters are derived. Reviewers have
also cautioned EPA against misuse of the indirect exposure
methodology, including both the dispersion and the intermedia transfer
modeling components. Every effort has been made in the present study
to use exposure models and parameter values consistent with the latest
EPA studies of dioxin and mercury risks.
Specifically, the review of IEM methodology performed by the
Science Advisory Board (SAB) 21 made several recommendations that were
considered in this analysis. The SAB recommended accounting for the
cumulative impacts of several sources on the regional or National
level. The national RELMAP modeling results included in sections 2.0
and 4.0 of this chapter respond to this recommendation. The SAB cited
lack of validation results as a scientific uncertainty in the IEM
methodology. The SAB recommends sensitivity analysis as a means of
determining the importance of input values. Sensitivity analyses were
performed for this assessment, and are described in the following
11-21
-------
subsections. However, the SAB notes problems in the model with
conservation of mass and chemical thermodynamics that were not
corrected in the model for this application. It should be noted that
these issues and others add uncertainty to the mass balance estimation
of media concentrations performed by the IEM spreadsheet, as well as
the corresponding modeled indirect exposures and risks.
11.5.1 Model Elasticity
A brief, initial sensitivity analysis was performed for selected
parameters employed in the risk modeling exercise. Using professional
judgement, a limited number of parameters were chosen for their
perceived importance in the model. Additionally, the chosen parameters
are known to have uncertainty in their value, generally due to limited
or variable empirical data. The selected parameters were individually
varied from 90 percent to 110 percent of their initial value, while
all other parameter values were held constant. The resulting changes
in the modeled risks for the pathway(s) impacted were analyzed. All
sensitivity analyses were performed using data from the large coal-
fired model plant because it demonstrated the largest indirect
exposures and risks.
The bioconcentration factor (BCF), or biota sediment accumulation
factor (BSAF), for fish was analyzed in the fisher scenario, as it is
recognized that fish consumption is an important factor for estimating
risks to the subsistence fisher, and is the overall driver of the risk
assessment. There is a limited set of data available for
consideration in the selection of BSAF;-'" and thus BSAF was analyzed
for impact on the fish consumption pathway by percentage variation of
the modeling input value.
Similarly, the BCFs for animal products (beef, chicken, dairy,
and eggs) were analyzed for the animal product consumption pathway of
the subsistence farmer scenario, as animal product consumption
provides significant risk in many of the modeled scenarios. Again,
there are limited data available for the bioconcentration of dioxin-
like compounds in animal products,-'21 and no true distribution of data
could be determined for analysis. Thus, in order to explore the
impact of changes in the BCFs, percentage changes in the modeling
input value were used.
The overall soil loss rate constant, which was recommended for
dioxin exposure modeling by EPA/ORD, was analyzed by expanding the
number of pathways considered in the adult resident scenario. The
soil loss rate constant was chosen because it influences the exposure
media concentrations in several pathways, including soil dermal
exposure, soil ingestion, plant ingestion, animal product ingestion,
water ingestion, and fish consumption. There was believed to be
uncertainty in the value of this parameter because the value
recommended by the EPA has changed from 0.0693 y"1 - to 0.0277 y'1.—
Again, due to the unavailability of a distribution for the soil loss
rate constant, the sensitivity analysis was performed by percent
change in the value.
11-22
-------
The results of this sensitivity analysis indicate that the IEM-
spreadsheet model is responsive to each of the parameters addressed in
this sensitivity analysis in differing amounts. None of the tested
cases create a change in risk greater than ten percent as a result of
the ten percent change in the tested parameter. The bioconcentration
and accumulation factors (Figure 11-5) show exactly a ten percent
change in modeled risk due to the ten percent change in the input
value, while the changes in risk, due to the ten percent change in the
soil loss rate constant, vary from 3.7 percent to 0.2 percent for the
different pathways considered (Figure 11-6).
These results indicate that none of the tested parameters will
create a greater percentage change in risk than the percentage change
in the input value. This conclusion is not unreasonable given the
general linearity of the model and the large number of input
parameters and equations playing into the calculated risks.
11.5.2 Fish Consumption Pathway Sensitivity Analysis
In consideration of the risk analysis results, it was observed
that the fish consumption pathway has the highest cancer risks. Thus,
a more extensive sensitivity analysis was undertaken for the input
parameters involved in the fish consumption pathway. The inputs into
the fish consumption pathway were considered, and five parameters were
chosen for this analysis: fish consumption rate of the subsistence
fisher, the fish lipid correction term, the BSAF, the fraction organic
carbon in the bottom sediments, and the organic carbon partition
coefficient (Koc) . The available data for each of these parameters was
examined and high and low values were selected, as described below.
The IEM-spreadsheet was then used to hold all other parameters
constant while individually considering each of the high and low
values for the fish consumption pathway input variables. This
analysis was performed using the large coal-fired model plant and the
subsistence fisher scenario because this combination generated the
greatest risk in the entire assessment.
The fish consumption rate for the subsistence fisher scenario in
the risk modeling exercise was 0.86 g fish/kg body weight/day. This
value was calculated from the 60 g fish/d recommendation of the
Mercury Study Report to Congress." This value is based on the
weighted mean of the data from a 1994 study by the Columbia River
Inter-Tribal Commission (as presented in the EPA Exposure Factors
Handbook) ,22-23 The units of g fish/d were converted by assuming a
70 kg body weight. It was also assumed that, as a subsistence
population, the fishers caught all of the consumed fish locally. High
and low consumption rates taken from the Columbia River data set have
the values of 0.04 g fish/kg body weight/day (10th percentile) and
2.0 g fish/kg body weight/day (90th percentile).
The fish lipid correction term is used as a correction for the
edible portion of fish lipid, as the BSAFs are developed for whole fish
lipid concentration. The fish lipid correction term in the risk modeling
exercise was 0.07, as recommended in Estimating Exposure to Dioxin-Like
11-23
-------
Indirect Risk Model Elasticity - Meat and Fish Factors
90.0
80.0
"Meat Bioconcentration Factor (BCF)
"Fish Biota Sediment Accumulation Factor
(BSAF)
Percent Change in Meat or Fish Factor
90% 100%
90.0 100.0
90.0 100.0
Percent Change in Risk Result
110%
110.0
110.0
Figure 11-5. Risk Model Sensitivity to Changes in Meat
and Fish Factors
Indirect Risk Model Elasticity - Soil Loss Rate Constant
106.0
104.0 ga
1 102.0 <^~^^!=Ss*^
* ^^^~-^*>**^
c 100.0 ^Tl'^j^^
I 98.0 ^5=Sass*=5=,^
0 ^**
2 96.0
a
a.
94.0
Percent Change in Soil Loss Rate Constant
90% 100%
-*-Soil Dermal 103.7 100.0
-H-Soil Ingestion 103.7 100.0
— *— Vegetation Ingestion 100.2 100.0
~~°~ Animal Product Ingestion 102.2 100.0
+ Water Ingestion 103.5 100.0
-S-pish Ingestion 103.5 100.0
Percent Change In Risk Result
A
^^>
^ffl
^8
110%
96.5
96.5
99.8
97.9
96.7
96.7
Figure 11-6. Risk Model Sensitivity to Changes in Soil
Loss Rate Constant
11-24
-------
Compounds.- This document notes that fish lipid content generally varies
from 5 percent to 20 percent, and thus these values were used as the high
and low sensitivity analysis values for this parameter.
The BSAFs used in the risk modeling exercise were the congener
specific values used in the demonstration of methodology in chapter 5
of Estimating Exposure to Dioxin-Like Compounds.- The limited set of
empirical values for dioxin congener BSAFs-'" was considered in the
determination of the high and low values for sensitivity analysis.
The highest and lowest empirical values for each congener were chosen
to use as the high and low values in this sensitivity analysis.
The fraction of organic carbon in the sediment used in the risk
modeling exercise was 0.03, as recommended by Estimating Exposure to
Dioxin-Like Compounds.- This value was assumed to be more than the
soil organic carbon fraction (0.01) and less than the organic carbon
fraction of suspended solids (0.05) . Thus, the assumed values for the
soil and suspended solids organic carbon fractions were used as the
low and high values for variation in sediment organic carbon content.
The Koc was considered in this sensitivity analysis because it
directly affects the concentration of dioxins and furans adsorbed to
the benthic sediments, which is used in the fish pathway. The
concentration of each congener that is adsorbed to the benthic
sediments is dependent on the sediment-water partition coefficient,
and this coefficient was determined by estimation using Koc. The
initial values for Koc were obtained from Estimating Exposure to
Dioxin-Like Compounds.- The Koc for each congener was varied, using the
high and low empirical values from Mackay, 24 and assumptions where data
was unavailable. The calculation of the adsorbed sediment
concentrations is complex and involves many input parameters;
therefore, the only attempt at analyzing sensitivity to the adsorbed
sediment concentration was through the analysis of Koc.
The results of the fish consumption pathway sensitivity analysis
are presented in Figure 11-7 and Table 11-9. The results indicate
that none of the single changes to low-end parameter values decreased
the fish ingestion risks below 9 x 1Q~6. However, the high assumption
for BSAF raised the modeled cancer risks to as high as 1 x 10 "3, or one
in one thousand for the subsistence fisher.
11.5.3 Sensitivity Analysis of Plume Impaction
As previously described in section 11.3, it was recognized that
the risks described for the case #1 farmer and fisher scenarios are
due to only wet deposition concentrations. The zero values for dry
deposition and air concentration at the case #1 location imply that
plume impaction has not occurred at 500 m. Thus, an analysis was
performed to identify areas of plume impaction, with the idea that
potential exposures may exceed even the case #1 modeling scenarios.
11-25
-------
BSAF
Organic carbon content of
sediment
Fraction lipid in fish
Fish consumption - Columbia [~
River L
Koc [
O.E+00 1.E-04 2.E-04 3.E-04 4.E-04 5.E-04 6.E-04 7.E-04 8.E-04 9.E-04 1.E-03
Modeled Total Risk
Nominal Risk = 2.E-04
Figure 11-7. Sensitivity of Predicted Risk to the Subsistence
Fisher to Changes in Parameter Values
Table 11-9. Fish Consumption Pathway Sensitivity Analysis Inputs
and Results
Parameter
BSAF (unitless)
Organic carbon content of
sediment (fraction)
Fraction lipid in fish (fraction)
Fish consumption - Columbia
River (kg/kg/d)
Koc(L/kg)
Low
congener
specific
1e-02
5e-02
4e-05
congener
specific
Parameter Value
Nominal
congener
specific
3e-02
7e-02
2e-03
congener
specific
High
congener
specific
5e-02
2e-01
2e-03
congener
specific
Low
2e-04
6e-04
1e-04
9e-06
2e-04
Risk Result
Nominal
2e-04
2e-04
2e-04
2e-04
2e-04
High
1e-03
1e-04
6e-04
3e-04
2e-04
BSAF = biota sediment accumulation factor
ISCST3 dispersion modeling for each of the case #1 model plants
was run again with receptors placed every 1000 m on a polar grid out
to a distance of 25,000 m. Flat terrain was used in the dispersion
remodeling because it was realized that variation in terrain would
influence the location of plume impaction and maximum concentration.
The resulting data has not been corrected for congener specific
emissions rates (unit emissions are used), and it has not been
proportioned for vapor and particle phase fractions. However, it
11-26
-------
still serves the analysis for determination of maximum values.
Table 11-10 presents the highest values for each model plant in terms
of wet and dry deposition and air concentration of both particles and
vapors. (Note that dry deposition of vapors is not included in this
table, as it is not an ISCST3 modeled value [see section 11.3.2.])
The location of each value in Table 11-10 is indicated by x and y
coordinates in meters. Table 11-11 presents modeled concentrations at
2.5 km, 10 km, and 25 km, as are reported in the mercury multipathway
analysis (chapter 7). The locations in Table 11-11 are described in
terms of degrees from the polar grid origin. For this analysis the
polar origin was North with degrees progressing clockwise.
Due to the disparate locations of the maximum modeled
concentrations for each modeled plant, it is difficult to determine
where the single site of maximum risk would be. It is likely that the
maximum risks would be observed at the 50 m location because that is
the location of maximum wet deposition. In all modeled cases, wet
deposition represents a significantly larger share, by an order of
magnitude or greater, of the environmental concentration than either
dry deposition or air concentration. This data does not indicate that
a re-analysis of the model plant risks is necessary because the effort
required to determine the exact location of a maximum and average
exposure scenario in a multipathway analysis is beyond the scope of
this screening-level analysis.
11.6 SUMMARY OF RESULTS
The results of this assessment indicate that the predicted cancer
risks associated with case #2 exposures to dioxin emissions from the
model plants are not greater than 1 x 10 ~7 for the model adult and
child resident scenarios. This is the case for both the direct and
the indirect exposure pathways. However, risks of 1 x 10 ~6 or above
are predicted for the indirect pathways of the case #1 scenarios. In
particular, the subsistence fisher scenario results in a predicted
risk of 2 x 10"4. When factors influencing the magnitude of the risk
to the subsistence fisher were examined in a sensitivity analysis, it
appeared that the assumptions about the BSAF had the greatest impact
on the predicted risk.
Taken one at a time, several other uncertainty factors have the
potential to raise the risk estimate for the subsistence fisher above
2 x 10"4; these include fish consumption rates, the organic carbon
content of sediment, and the fraction of fish tissue that is lipid.
This assessment has been conducted using model plant data and two
geographic locations selected to represent two potential scenarios.
The location of the case #1 receptors was arbitrarily placed at a
point of high wet deposition. This assumption was made for the
purpose of providing a conservative estimate of the case #1 risk.
The case #1 scenario represents a sub-population that has the
potential to receive high exposures due to relatively high ingestion
rates, and assumed to be in a location for which the meteorology would
11-27
-------
Table 11-10. Maximum Dispersion Modeling Locations and
Concentrations
Model plant
phase
LCH, vapor
LCH, particle
SCH, vapor
SCH, particle
LOH, vapor
LOH, particle
SOH, vapor
SOH, particle
Wet deposition
location in m Wet deposition
(x,y) (g/m2/y)
0, -500 0.07430
0, -500 0.09158
0, -500 0.08372
0, -500 0.10274
0, -500 0.07846
0, -500 0.14166
0, -500 0.09083
0, -500 0.16233
Dry deposition
location in m Dry deposition
(x,y) (g/m2/y)
Several 0.00026
8314, -3444 0.00210
Several 0.00092
4619, -1813 0.00997
Air
concentration Air
location in m concentration
(x,y) (/ig/m3)
0, 9000 0.00050
Several 0.00046
Several 0.00344
-3826, 9238 0.00302
Several 0.00105
Several 0.00084
-2296, 5543 0.00951
-1913,4619 0.00792
also tend to result in high exposures for given rates of release of
dioxins. The results of this analysis can be considered a potential
representative example for utility boilers, yet they do not represent
a specific plant or location. The analysis should be considered an
analysis of hypothetical scenarios used for demonstration.
This analysis of noninhalation exposures to dioxin emissions is a
screening analysis. Thus, these quantitative exposure and risk
results, because of the many modeling and analytic uncertainties, are
very uncertain and do not, therefore, conclusively demonstrate the
existence of health risks of concern associated with exposures to
utility emissions either on a national scale or from any actual
individual utility. The lack of measured data around these sources
precludes a comparison with modeled results. These results do suggest
that exposures and risks of concern cannot at present be ruled out and
that there is a need for development of additional scientific
information to evaluate whether risk levels of concern may exist.
11-28
-------
Table 11-11. Dispersion Modeling Concentrations at Specified
Distances
Model plant,
phase
Wet deposition
location (°)
Wet deposition
(g/m2/y)
Dry deposition
location (°)
Dry deposition
(g/m2/y)
Air
concentration
location (°)
Air
Concentration
(//g/m3)
2.5 km concentrations
LCH, vapor
LCH, particle
SCH, vapor
SCH, particle
LOH, vapor
LOH, particle
SOH, vapor
SOH, particle
180
180
180
180
360
360
180
180
0.01342
0.01414
0.01406
0.01535
0.01406
0.01970
0.01577
0.02141
-
360
-
360
-
360
-
112.5
-
0.00005
-
0.00088
-
0.00030
-
0.00768
360
360
360
360
360
360
360
360
0.00015
0.00014
0.00040
0.00196
0.00040
0.00036
0.00681
0.00607
10 km concentrations
LCH, vapor
LCH, particle
SCH, vapor
SCH, particle
LOH, vapor
LOH, particle
SOH, vapor
SOH, particle
180
337.5
180
180
180
180
180
180
0.00217
0.00121
0.00225
0.01890
0.00221
0.00211
0.00233
0.00210
-
337.5, 360
-
112.5
-
112.5
-
112.5
-
0.00019
-
0.00207
-
0.00084
-
0.00675
360
360
337.5
337.5
360
360
337.5
337.5
0.00049
0.00045
0.00344
0.00302
0.00100
0.00083
0.00852
0.00679
25 km concentrations
LCH, vapor
LCH, particle
SCH, vapor
SCH, particle
LOH, vapor
LOH, particle
SOH, vapor
SOH, particle
180
180
180
180
180
180
180
180
0.00046
0.00037
0.00047
0.00036
0.00046
0.00036
0.00047
0.00033
-
112.5
-
112.5
-
112.5
-
337.5
-
0.00026
-
0.00099
-
0.00080
-
0.00205
337.50
135
337.5
337.5
337.5
337.5
360
360
0.00046
0.00039
0.00256
0.00205
0.00103
0.00079
0.00444
0.00355
11-29
-------
11.7 REFERENCES
1. U.S. Environmental Protection Agency. Health Assessment for
2,3,7,8-TCDD and Related Compounds. EPA/GOO/BP-92/OOlc. 1994b.
2. U.S. Environmental Protection Agency. Estimating Exposure to
Dioxin Like Compounds, Vol. II: Properties, Sources, Occurrence,
and Background Exposures-External Review Draft. EPA/600/6-
88/005Cb. Office of Research and Development, Washington, DC.
1994 .
3. U.S. Environmental Protection Agency. Estimating Exposure to
Dioxin-Like Compounds, Vol. Ill: Site-Specific Assessment
Procedures. External Review Draft. EPA/600/6-88/005CC. Office
of Research and Development. Washington, DC. 1994a.
4. U.S. Environmental Protection Agency. User's Guide for the
Industrial Source Complex (ISC3) Dispersion Models, volumes I and
II. EPA-454/B-95-003a. Office of Air Quality Planning and
Standards, Research Triangle Park, NC. 1995b.
5. U.S. Environmental Protection Agency. Memo, from Jim Turner,
Research Triangle Institute, to Chuck French, EPA, regarding
Process Parameters for Utility Boiler Dioxin Modeling, Utility
Boiler Report to Congress. Dated July 14, 1997. 1997a.
6. U.S. Environmental Protection Agency. AP-42. Office of Air
Quality Planning and Standards. Research Triangle Park, NC.
January 1995a.
7. Perry, R. H., D. W. Green, and J. 0. Maloney. Perry's Chemical
Engineers' Handbook. McGraw-Hill Book Company, New York, NY.
1984.
8. SAMSON. Solar and Meteorological Surface Observation Network.
1961-1990 data, Version 1. United States Department of Commerce,
National Climatic Data Center, Asheville, NC. 1993.
9. USGS, United States Geological Survey. Earth Resources
Observation System Data Center digital elevation model data from
the online Global Land Information System (GLIS) at
http://edcwww.cr.usgs.gov/glis. 1997.
10. U.S. Environmental Protection Agency. Guideline on Air Quality
Models. Office of Air Quality Planning and Standards. Research
Triangle Park, NC. 40 CFR Parts 51 and 52. 1996b.
11-30
-------
11. U.S. Environmental Protection Agency. Secondary Aluminum
Smelters: Screening Level Indirect Risk Assessment. EPA Contract
No. 68-D3-0034, Work Assignment No. 111-26. 1997b.
12. U.S. Environmental Protection Agency. Risk Assessment Support to
the Development of Technical Standards for Emissions from
Combustion Units Burning Hazardous Wastes. 1996e.
13. U.S. Environmental Protection Agency. Methodology for Assessing
Health Risks Associated with Indirect Exposure to Combustor
Emissions. Interim Final Report. EPA/600/6-90/003. Office of
Health and Environmental Assessment. Washington, DC. 1990.
14. U.S. Environmental Protection Agency. Addendum to the
Methodology for Assessing Health Risks Associated with Indirect
Exposure to Combustor Emissions. External Review Draft.
EPA/600/AP-93/003. Office of Health and Environmental
Assessment. Washington, DC. 1993.
15. Lorber, Matthew. EPA Office of Research and Development.
Personal Communication. 1997.
16. U.S. Environmental Protection Agency. Great Lakes Water Quality
Initiative Technical Support Document for the Procedure to
Determine Bioaccumulation Factors. EPA-820-B-95-005. Office of
Water. March 1995c.
17. Brzuzy, L. P. and R. A. Hites. Estimating the atmospheric
deposition of polychlorinated dibenzo-p-dioxins and furans from
soils. Environmental Science and Technology. Volume 29. 1995.
Pp. 2090-2098.
18. U.S. Environmental Protection Agency. Mercury Study Report to
Congress, Volume III: An Assessment of Exposure from
Anthropogenic Mercury Emissions in the United States. SAB Review
Draft. EPA-452/R-96-001C. Office of Air Quality Planning and
Standards and Office of Research and Development. 1996c.
19. U.S. Environmental Protection Agency. Risk Analysis of TCDD
Contaminated Soil. EPA-600/8-84-031. Office of Health and
Environmental Assessment. 1984.
20. U.S. Environmental Protection Agency. Review of Draft Addendum
to the methodology for Assessing Health Risks Associated with
Indirect Exposure to Combustor Emissions. EPA-SAB-IAQC-94-009b.
Science Advisory Board, Washington, DC. 1994.
21. Stephens, R. D., M. S. Petreas, and D. G. Haward. Biotransfer
and Bioaccumulation of Dioxins and Furans from Soil: Chickens as
a Model for Grazing Animals. Science of the Total Environment.
Volume 175. 1995. Pp. 253-273.
11-31
-------
22. Columbia River Inter-Tribal Fish Commission. A Fish Consumption
Survey of the Umtilla, Nez Perce, Yakama and Warm Springs Tribes
of the Columbia River Basin. Technical Report 94-3. October
1994.
23. U.S. Environmental Protection Agency. Exposure Factors Handbook,
Volumes I, II, III. Draft Report. Office of Research and
Development. EPA/600-P-95/002Ba,b,c. Washington, DC. 1996a.
24. Mackay, Donald, Wan-Ying Shiu, and Kuo-Ching Ma. Illustrated
Handbook of Physical-Chemical Properties and Environmental Fate
for Organic Chemicals. CRC Press, Inc - Lewis Publishers. Boca
Raton, FL. 1995.
11-32
-------
12.0 LITERATURE REVIEW ON THE POTENTIAL IMPACTS OF
HYDROGEN CHLORIDE AND HYDROGEN FLUORIDE EMISSIONS
12 . 1 OVERVIEW
The information presented in this chapter and in Appendix H was
collected to expand the EPA's knowledge of the potential impacts of
HC1 and HF emissions from utilities. The details and references are
presented in Appendix H. This chapter presents a summary of the
findings. The EPA is updating its current state of knowledge of
potential health impacts; atmospheric chemistry (e.g., half-life,
impacts on the acid rain phenomenon); potential human exposure through
pathways other than direct inhalation; and possible ecological harm.
The EPA's goal is to understand the potential impacts from HC1 and HF
emissions to any and all health and environmental areas. This chapter
is not intended to provide a detailed, comprehensive treatise on the
above subject area; rather, it is designed to provide general
technical information that will identify possible problem areas that
may call for additional, more detailed research.
Published evidence for potential impacts of HC1 and HF was
evaluated from a wide variety of sources. Overall, there is extensive
information available on the toxicology of these two pollutants;
however, literature pertaining specifically to HF and HC1 atmospheric
chemistry is relatively scarce, especially that pertaining to fine
particulate matter and acid rain. Literature on HC1 and HF from
sources outside the United States and pertaining to emissions sources
other than utilities has also been evaluated.
This chapter is organized so that the findings for HC1 are
presented first, followed by the findings for HF. Within each
section, evidence from the literature for transport and transformation
through atmospheric, terrestrial, and aquatic processes is presented
first, followed by evidence for impacts on human health; vegetation;
and wild, domestic, and aquatic animals.
12 . 2 SUMMARY OF FINDINGS
This chapter provides a synopsis of the information of interest
found during the literature review on potential impacts (e.g., acid
rain, fine particulate matter, visibility, and toxicity to various
plant and animal species) of hydrogen chloride and hydrogen fluoride
emissions from utilities.
12. 2.1 Hydrogen Chloride
12.2.1.1 HC1 Emissions and Formation. Utilities emit a
substantial amount of the anthropogenic atmospheric emissions of HC1
in the United States.1 As shown in chapter 3 of this report, utilities
were estimated to emit 146,000 tpy of HC1 in the United States in
1990. Other important sources of HC1 are industrial coal combustion
and solid waste combustion. Natural sources of HC1 emissions include
12-1
-------
volcanic activity, marine plants and microorganisms, and land plant
combustion.
According to available information, ambient concentrations of HC1
in the United State ranged from none detected to 4 //g/m3. Rural sites
can be expected to be at the low end of this range and urban sites are
likely to be toward the high end. Human health effects are discussed
in Appendix E .
HC1 can be formed several ways in the atmosphere . Anthropogenic
chlorocarbons can react with OH radicals to produce small amounts of
HC1 . Nonanthropogenic HC1 can be formed from deliquescent sea-salt in
the marine environment. HC1 can also be created or destroyed through
the interaction between fogwater and aerosols. Although information
on HC1 formation by-products is scarce, available sources indicate
that reactions generating HC1 can produce the following by-products in
the atmosphere: NaN03, Na2S04, hydrocarbon radicals, and
12.2.1.2 HC1 Atmospheric Processes. The atmospheric lifetime of
HC1 is estimated to be between 1 and 5 days. HC1 is a highly reactive
gas that is removed from the atmosphere via wet and dry deposition.
12.2.1.3 HC1 Atmospheric Transport . In general, because of its
high solubility, HC1 will be removed from the atmosphere much faster
than S02 or N02 and will be deposited in close proximity to the
emissions source. However, conditions exist under which S02 is subject
to further transport. One study found evidence that HCl-enriched
plumes, believed to originate from coal-burning utilities, reached a
rural site two days after emissions release. Evidence was found that
HC1 may affect the atmospheric chemistry of mercury, and thus the
toxicity of mercury emissions from utilities. HC1 emissions are
believed to contribute to some limited degree to the formation of
atmospheric acidity and acid rain. In addition, HC1 appears to
indirectly contribute to some limited degree, to fine PM and
visibility problems. However, there are significant uncertainties as
to the extent of the impacts due to HC1 emissions in these areas.
Further research and evaluation are needed to determine if, and to
what extent HC1 contributes to acid rain, fine PM, and visibility
issues .
12.2.1.4 HC1 Terrestrial Processes. Information on the
terrestrial behavior of HC1 is scarce. The references found discuss
the evidence that HC1 can lower pH to the point that S02 oxidation is
delayed, possibly altering the spatial deposition of acid species, and
that gaseous HC1 damages limestone.
12.2.1.5 HC1 Aquatic Processes. The chemistry of ubiquitous
chlorinated compounds in natural waters is affected by a number of
factors that determine the persistence and toxicity of interim
species. Chloride cycling in watersheds was found to be more complex
than previously thought. The traditional view has been that
atmospherically deposited chloride is rapidly transported. Further
12-2
-------
research of chloride cycling would be needed to gain a better
understanding of the aquatic processes.
12.2.1.6 HC1 Human Health Impacts. Evidence of local irritation
to the upper respiratory tract by HC1 was found, and long-term
exposure may cause tooth erosion. The WHO concluded in a review on
HC1 that there are no mutagenic, carcinogenic, or teratogenic effects
related to HC1. Appendix E contains more information on health
effects.
12.2.1.7 HC1 Vegetation Impacts. Atmospheric and leaf chloride
levels were found to be closely correlated. Atmospheric chloride can
concentrate in and cause damage to foliar tissues. High chloride
concentrations resulted in tissue death in coastal vegetation.
12.2.1.8 HC1 Terrestrial Animal Impacts. HC1 that reaches
plants and soil via wet and dry deposition is then available for
uptake by animals. The adverse health symptoms of HC1 exposure in
animals include eye, nasal, and respiratory tract irritation, with the
respiratory tract being the primary target.
12.2.1.9 HC1 Aquatic Animal Impacts. Toxicity of chlorinated
compounds to aquatic biota varies widely. One study reported the 24-
hour LC50 for one marine species as 0.0018 mg/L, and the LG;0 for one
sensitive freshwater species as 0.003 mg/L.
12.2.2 Hydrogen Fluoride
12.2.2.1 HF Emissions and Formation. Anthropogenic sources are
responsible for most atmospheric fluoride. Anthropogenic emissions of
hydrogen fluoride (HF) originate from coal combustion and the
aluminum, phosphate, and steel-making industries. As shown in
chapter 3, utilities are estimated to emit approximately 19,600 tpy of
HF nationwide. Volcanoes are the primary natural sources of HF.
Ocean spray, fires, and dust from soil and rock weathering contribute
relatively minor amounts of fluoride to the atmosphere. Measured
concentrations of HF at one monitoring station in the United States
ranged between 1 //g/m3 and 8 //g/m3. Atmospheric concentrations of
fluoride in remote rural areas are reported to be approximately 0.1
//g/m3, which is at the limit of detection. Estimates of the relative
proportions of gaseous to particulate emissions of industrial fluoride
vary. Fluoride particulates range from distinct minerals to alumina
with HF adsorbed to its surface, and particle diameters range from
<0.1 urn to approximately 10 urn. Volcanic emissions are usually not
predominately HF, but rather other fluoride-containing compounds that
react in the atmosphere to form HF.
12.2.2.2 HF Atmospheric Processes. HF is described as
moderately persistent in the atmosphere, with an estimated lifetime of
approximately 1 to 5 days. Wet and dry deposition are the primary
routes of HF removal from the atmosphere. HF atmospheric reaction
products are primarily fluoride salts. HF does not biodegrade.
12-3
-------
12.2.2.3 HF Atmospheric Transport. Fluoride emissions from
utilities are transported on a regional scale. Fluoride has been used
as an atmospheric tracer because the fluoride to sulfur oxide ratio is
relatively constant in the coal-fired utilities examined, providing a
characteristic utility emissions fingerprint. Evidence was found that
coal-fired utility emissions contribute to measured concentrations of
atmospheric fluoride at distances of up to 500 km. Another study
found that elevated levels (18-21 ppm) of fluorides emitted from a
utility could be detected in grape leaves at distances of up to 37 km.
However, it has been reported that the measurement of fluoride
transport and deposition have problems of accuracy due to limitations
of analytical methods.
12.2.2.4 HF Terrestrial Processes. Fluoride is lost from the
various surfaces on which it is deposited and leaves ecosystems at a
rapid rate. The volatilization pathway as a route of fluoride export
from ecosystems needs further investigation. Soil can be both a sink
and source of fluoride, but fluoride is not usually available or
labile in soils. Soluble fluoride-containing process water and
leachate of phosphogypsum were shown to dissolve much of the fine soil
clay fraction, as well as the smectite of the coarse clay fraction of
soils. Several conditions are reported to facilitate rapid uptake of
water-soluble fluoride and transport. The natural buffering capacity
of soils or water, or dilution can reduce acidity added by the
presence of HF. Sulfate and fluoride were found to slightly retard
aluminum's mobility through soils.
12.2.2.5 HF Aquatic Processes. Fluoride is a major component of
seawater, and natural and anthropogenic fluoride may accumulate in
waterbody sediments. In freshwaters with pH greater than 5, fluoride
is mainly present as fluoride ion.
12.2.2.6 HF Human Health Impacts. Adverse effects of fluoride
on human health include dental fluorosis, gastric disturbances,
reductions in urinary concentrating ability, skeletal fluorosis, and
even death. Optimally fluoridated water has been shown not to be
associated with a detectable risk of cancer in humans. Fluoride
exposure is not associated with birth defects, and there is no
indication that organ systems are affected by chronic, low-level
fluoride exposure. Genotoxicity studies have yielded contradictory
results. There is disagreement about whether the increased prevalence
of dental fluorisis observed in the United States since the 1940s is a
toxic effect. Crippling skeletal fluorosis has not been and is not a
public health problem in the United States. Beneficial effects of
high fluoride regimens in reducing osteoporosis have not been
demonstrated. Further epidemiological studies are required to
determine whether or not an association exists between various levels
of fluoride in the drinking water and bone fractures.2 Appendix E
contains more information on health effects of HF.
12.2.2.7 HF Vegetation Impacts. Inherent differences in the
resistance of some tropical tree species to fluoride may be related to
12-4
-------
their capacity to accumulate aluminum. Although some plant species
are tolerant of elevated fluoride levels, the storage of large amounts
of fluoride in plant tissues may present a risk to ecosystems. Plant
uptake of fluoride was found to be limited to the smaller, water-
soluble, and labile fractions. The major pathway of fluoride to
plants is atmospheric deposition.
Atmospheric fluoride is capable of injuring certain plant species
at lower concentrations than any other air pollutant. However, most
plant species are relatively resistant to fluoride. No morphological
damage to lichen species exposed to high concentrations of fluoride
was found. Fluoride was found to be the most important pollutant
contributing to vegetation damage in one section of a tropical
rainforest.
12.2.2.8 HF Terrestrial Animal Impacts. Two studies on the
toxicity of fluoride to several rodent species found the animals to
exhibit visible incisor lesions after fluoride ingestion. Bone
fluoride loads in four predatory bird species were found to be greater
in males of all species examined and higher than average in more
industrial regions.
Conflicting information was found concerning whether fluoride
accumulates in food chains. Variations in fluoride concentrations
within plant organs can result in animal species with differing
feeding niches ingesting different amounts of fluoride. Contamination
of foliage with soil may constitute an important route of fluoride
transfer to large herbivores in situations where soil has been treated
with phosphate fertilizer or exposed to substantial airborne
deposition of fluoride.
12.2.2.9 HF Aquatic Animal Impacts. Biomagnification in aquatic
animals is reported to be negligible to very slight. Two trout
species were demonstrated to be more resistant to fluoride than
freshwater benthic macroinvertebrates. Diatoms appear to be tolerant
of, and stimulated to grow by, high fluoride concentrations; the
ecological significance of this is uncertain. Limited evidence exists
for fluoride-containing effluent effects on both abundance and
diversity of estuarine/marine organisms at relatively low fluoride
levels.
12-5
-------
12 . 3 REFERENCES
1. Placet, M., et al. Emissions Involved In Acidic Deposition
Processes, NAPAP Report 1, Nin: Acidic Deposition: State of
Science and Technology, National Acidic Precipitation Assessment
Program. 1990.
2. Whitford, G. M. The Metabolism and Toxicity of Fluoride (Revised
edition). Karger. 1996.
12-6
-------
13.0 ALTERNATIVE CONTROL STRATEGIES FOR HAZARDOUS AIR POLLUTANT
EMISSIONS REDUCTIONS
This chapter presents methods of reducing HAP emissions through
precombustion controls, combustion controls, postcombustion controls,
and alternative controls. Also, strategies for maximizing total HAP
control or minimizing total HAP emissions are reviewed.
The HAPs of concern include the trace elements identified in
chapter 5 as potential health risks. These consist of arsenic,
cadmium, chromium, lead, manganese, mercury, and nickel; dioxins and
furans (due to the toxicity of the organic chemical); and HC1 and HF
(due to the estimated emission quantities of the compounds).
13.1 PRECOMBUSTION CONTROLS
To reduce S02 emissions and thereby comply with the Phase II
requirements of the Acid Rain Provisions of the Act, some utilities
will switch to fuels that contain lower amounts of sulfur. The
effects of fuel switching on HAP emissions are briefly reviewed in
this section. Emissions of trace elements from coal-fired units may
be controlled through precombustion control techniques such as coal
cleaning and coal gasification. The effectiveness of these control
techniques is also reviewed.
13.1.1 Fuel Switching
Utilities that switch fuel may change from higher to lower
sulfur-containing coal (less than 1.5 weight percent sulfur) or elect
to burn a different type of fuel (e.g., switching from oil to natural
gas combustion). A potential concern with fuel switching is whether
or not it will increase HAP emissions, due to potentially increased
concentrations of trace elements in the fuel and different fly ash
characteristics that impact effective PM and HAP control with existing
APCDs (e.g., ESPs).
The qualitative effects of switching the type of fuel may be
noted through comparisons of the averages of trace element
concentrations in utility fuels. Table 13-1 lists the arithmetic
average, as well as the standard deviation of the average, for trace
element concentrations in coal, residual oil, and natural gas. As
indicated by the magnitude of the standard deviations listed in
Table 13-1, trace element concentrations vary considerably in coal and
residual oil. Some of the standard deviations are large enough that
comparable concentrations of trace elements may occur in some coal and
residual oil samples. For discussion purposes, any overlap in trace
element concentrations was ignored, and generalizations on the effects
of fuel switching were made from comparisons of average trace element
concentrations in the three fuels.
13-1
-------
Table 13-1. Comparison of Average Concentrations of Trace
Elements in Utility Fuelsa'b
Sulfur
average6
SD (mean)'
No. averages
Trace elements:
Arsenic
average
SD (mean)
No. averages
Cadmium
average
SD (mean)
No. averages
Chloride
average
SD (mean)
No. averages
Chromium
average
SD (mean)
No. averages
Fluoride
average
SD (mean)
No. averages
Lead
average
SD (mean)
No. averages
Mercury 1'2
average
SD (mean)
No. averages
Nickel
average
SD (mean)
No. averages
Coalc
(Ib/MMBtu)
1.24
0.19
26
Residual
(Ib/MMBtu)
0.31
0.07
13
Natural gas'1
(Ib/MMBtu)
0.00006
0.00006
2
Coalc
(Ib/trillion)
660
120
26
60
30
26
27,000
6,600
20
600
98
26
5,300
720
26
800
190
26
7.7
0.6
152
700
69
26
Residual
(Ib/trillion)
17
11
6
5.4
3.9
3
7,400
3,300
11
17
3
11
600
200
3
73
43
5
0.6
0.3
4
1,300
200
13
Natural gase
(Ib/trillion)
0.19
0.06
2
-
-
-
-
-
0.001
1
-
The coal data listed in Table 13-1 were not weighted for coal production by State of coal origin.
There were only two sets of data for concentrations of trace elements in natural gas in Table 13-1.
With the exception of the mercury data, coal values were determined from modified U.S. Geological Survey (USGS) data, by
State of coal origin, and coal shipment dat for coals that originated from three States. Modified USGS data are USGS data that
were modified to account for the effects of bituminous coal cleaning. Mercury data were reported by EPRI for samples of coal
shipments.2 None of the data were weighted for coal production.
Natural gas values were determined from the preliminary EPRI test reports for Sites 120 and 121. The listed values are detected
concentrations.
Averages of averaged data sets.
This is the standard deviation of the number of averages directly below.
13-2
-------
13.1.1.1 Switching to Natural Gas Combustion. As shown in
Table 13-1, natural gas has the lowest average concentrations, on a
Ib/trillion Btu basis, of sulfur, arsenic, and mercury when compared
with the corresponding values for residual oil and coal. The averages
listed for coal and residual oil exceed those listed for natural gas
by factors that range from approximately 100 (for the concentration of
arsenic in residual oil) to as much as approximately 21,000 (for the
concentration of sulfur in coal). Thus, of the three utility fuels,
natural gas contains the least amounts of the trace elements, and
switching from coal or residual oil to natural gas combustion would
ultimately reduce emissions of trace elements. Some total and
seasonal conversion of coal- and oil-fired units to natural gas firing
is expected to affect compliance with the various ozone and NOX control
provisions of Titles I and IV of the Act. However, a complete
conversion of all utility boilers to natural gas is not practical.
Even though the natural gas transmission network is expanding,
delivery of natural gas to each utility unit cannot yet be
accomplished. In addition, there is concern over the long-term
availability of natural gas (particularly with respect to other fuels)
given the projected usage (and increase in usage) in the residential,
commercial, and industrial sectors and the estimates of proved and
supplemental reserves of natural gas. Estimates of "proved reserves"
of natural gas have decreased each year (but one) for the past 10
years.
13.1.1.2 Switching from Coal to Residual Oil Combustion. As
shown in Table 13-1, with the exception of the average concentrations
of nickel, the average concentrations of trace elements listed for
coal exceed those listed for residual oil by factors that range from
approximately 4 (for the concentrations of sulfur and chloride) to as
much as approximately 40 (for the concentration of arsenic). However,
the average concentration of nickel in coal is approximately half the
corresponding value for residual oil. Thus, switching from coal to
residual oil combustion could result in increased emissions of nickel
and decreased emissions of the other trace elements.
13.1.1.3 Switching from Higher to Lower Sulfur Coals. The
effects of coal switching will be reviewed first for mercury and then
for the other trace elements.
Figure 13-1 shows the relationship between the concentrations of
mercury and sulfur in 153 samples of coal shipments.2 As shown in
Figure 13-1, there is no relationship between the sulfur and mercury
content in the sampled coal shipments; mercury concentrations below
approximately 15 Ib/trillion Btu are present in coal with both higher
sulfur concentrations (above 2.5 lb/MMBtu) and lower sulfur
concentrations (below 1.5 lb/MMBtu).
A conceivable control strategy would involve blending higher
mercury-containing coals with lower mercury-containing coals to reduce
mercury emissions. Such a practice would be comparable to blending
high and low sulfur-containing coals in order to meet S02 emission
13-3
-------
60 -r
50
40
S
c
o
30
o
^
-------
limits. However, coal blending for mercury control is not a proven
control strategy. Changes in the electrical resistivity and amount of
flyash resulting from coal blending could reduce PM capture
efficiencies by ESPs and subsequently lead to increased emissions of
PM and HAP metals. However, these effects on ESP performance can be
addressed by gas conditioning and/or modifications to the ESP.
Blending for mercury control could also increase levels of other HAPs
or sulfur. Another uncertainty with coal blending for mercury control
would be the possibility of changing the distribution of the elemental
and oxidized forms of mercury that could affect mercury control with
existing control devices. Another factor is that the blending of two
different coals might change the higher heating value of the resulting
mixture, with subsequent effects on the quantity of fuel required for
combustion.
The qualitative effects of switching to lower-sulfur-containing
coals on other metallic HAPs are examined in Figure 13-2(a-g) through
plots of the average concentrations of each HAP, excluding mercury,
with sulfur content in coal. As shown in Figure 13-2(a-g), the
average concentration of trace elements in coal shipments, as
approximated by the modified U.S. Geological Survey (USGS) data
(modified for the effect of coal cleaning on bituminous coals),
generally show no clear trends with sulfur content (i.e., decreasing
the sulfur content of coal does not generally lead to reduced
concentrations of trace elements in coal).
Based upon average concentrations of trace metals in coal from
the modified USGS data, fuel switching to lower-sulfur-containing
coals will not generally result in consistently reduced emissions of
the trace elements. Trace elements associated with the PM (e.g.,
arsenic, cadmium, chromium, lead, and nickel) could be removed from
coal-fired flue gas with a PM control device.
It should be stressed that the effects of coal switching were
drawn from comparisons of average concentrations of trace elements in
modified USGS coal data. The concentrations of trace elements in
actual coal shipments may vary from the USGS averages.
13.1. 2 Coal Cleaning
Approximately 77 percent of eastern and midwestern3 bituminous
coal shipments are cleaned to meet customer specifications on heat,
ash, and sulfur content. Subbituminous and lignite coals are not
routinely cleaned.4 Conventional coal cleaning removes mineral matter
and, in the process, may also remove some of the trace elements
contained in the mineral matter. The mineral matter is removed from
the coal by either crushing and screening or by coal washing (through
the difference in specific gravities of the constituents or by
surface-based floatation).5 In the process of removing the mineral
matter, coal cleaning generates solid refuse that contains trace
elements; the solid refuse must be disposed of properly. Any coal
cleaning liquid wastes will also contain trace elements, but the
liquid wastes may be properly clarified and then recycled.
13-5
-------
3,000
2,500
3 2,000
CO
1,500
.
o>
< 1,000
500
OJ
I
a\
140,000
120,000
100.000
"3"
£
I 80,000
•g 60,000
•c
o
6
40,000
20,000
(a) Arsenic
Data omitted from the graph: (3.5 Ib Sulfur/MMBtu,
707 Ib Cd/trillion Btu) and (3.5 Ib Sulfur/MMBtu,
405 Ib Cd/trillion Btu)
, i
1 2 3
Sulfur (Ib/MMBtu)
(c) Chloride
2 3
Sulfur (Ib/MMBtu)
140
120
_
&
i 80
I 60
I
O
40
20
3,000
2,500
S 2,000
c
g
1.1,500
O
1,000
500
(b) Cadmium
0.5 1 1.5 2 2.5
Sulfur (Ib/MMBtu)
(d) Chromium
3.5
1 2
Sulfur (Ib/MMBtu)
Figure 13-2 (a-g). Relation between concentration of selected
trace elements and sulfur in modified USGS data.
-------
(e) Fluoride
(f) Lead
14,000
12,000
10,000
m
| 8,000
i
^ 6,000
o
C
4,000
2,000
UJ
I
Sulfur (Ib/MMBtu)
1,600
1,400
1,200
5 1,000
c
£ 800
-Q
"5
o 600
•z.
400
200
0
4,000
3,000
i 2,000
1,000
Sulfur (Ib/MMBtu)
(g) Nickel
2 3
Sulfur (Ib/MMBtu)
Figure 13-2 (a-g). continued.
-------
Table 13-2 lists the limited amount of available data on trace
element reductions achieved through conventional coal cleaning. In
Table 13-2, some of the trace element reductions are negative.
Negative percentages occur when part of the coal is removed but the
element is not contained in the extracted portion of the coal, so that
the same weight of the element that was contained in the uncleaned
coal is contained within a relatively smaller weight of the cleaned
coal. Because the weight of the trace element does not change,
negative removal percentages are considered to indicate that no trace
element reduction occurred or that the trace element reduction was
effectively 0 percent.
As shown in Table 13-2, for the limited amount of available data,
trace element removal percentages may vary for coals obtained from the
same seam. The variability occurs because trace element
concentrations, in the mineral portion of coal, vary from coal to
coal. For the data listed in Table 13-2, the variation in trace
element reductions may also be due to the use of various cleaning
methods, the accuracy of the analytical techniques used to measure
concentrations of trace elements in cleaned and uncleaned coals, and
sample collection methods. With regard to analytical techniques,
CONSOL, Inc., noted the following difficulties in analyzing mercury
concentrations in coal: the volatility of mercury compounds, low
mercury concentrations in coal, large variability of approximately 50
percent in the interlaboratory reproducibility of mercury
concentrations, lack of certified mercury-in-coal standards, and lack
of standard sample preparation and analysis methods.6'7
The average trace element reductions, listed in Table 13-2 for a
limited amount of available data, were determined with the negative
percentages treated as 0 percent removal and the averages not weighted
by coal production since the EPA does not believe sufficient data
exist at this time to follow a production-weighted approach. The
average values for the limited amount of available data indicate that,
in general, lead concentrations were reduced the most (approximately
55 percent) while mercury concentrations were, on average, reduced the
least (approximately 21 percent). It should be stressed that better
and worse trace element reductions may be found for specific coals.
For instance, mercury removals of at least 50 percent should occur
during conventional cleaning of Upper Freeport coal, based upon the
modes of occurrence of mercury and available conventional coal
cleaning data.8 In Table 13-2, mercury removals were reported for
three samples of Upper Freeport coal; two indicated approximately
62 percent removal while the third indicated effectively no mercury
removal. More research is needed to establish trace element removal
efficiencies, achieved through conventional coal cleaning, on a
statistically representative sampling of commercially viable coal
seams. Additional information on advanced coal cleaning is provided
in section 13.6.1.
13-8
-------
Table 13-2. Trace Element Reductions Achieved Through Conventional Coal Cleaning3
w
Seam
C. App. A
C. App. B
IL
IL#6
IL#6
IL 2,3,5
IL 2,3,5
Ky#11
Ky#11
Ky#9&14
Lower Kittanning
Pittsburgh
Pittsburgh
Pittsburgh
Pittsburgh A
Pittsburgh B
Pittsburgh C
Pittsburgh D
Pittsburgh E
Pratt
Pratt
Pratt/Utley
Sewickley
Upper Freeport
Upper Freeport
Upper Freeport
Utley
All seams:
min
max
average
State
IL
IL
IL
IL
IL
Ky
Ky
Ky
Pa
Pa
Pa
Pa
Pa
Pa
Pa
Pa
Pa
Al
Al
Al
Pa
Pa
Pa
Pa
Al
Reference
a
a
e
a
c
d
d
d
d
c
c
b
c
c
a
a
a
a
a
d
d
d
c
b
d
d
d
% Removal
arsenic
39
22
54
47
3
30
48
52
13
40
66
74
53
27
65
67
67
77
50
29
7.1
29
36
38
73
74
23
3.4
77
45
% Removal
cadmium
14
33
59
76
32
53
43
57
20
40
58
63
55
50
0
0
0
52
40
42
0
0
76
38
% Removal
chromium
75
67
21
72
23
44
46
60
69
29
44
71
56
24
42
66
63
64
71
48
54
41
46
11
50
45
17
11
75
49
% Removal
fluoride
27
41
60
83
92
42
64
-459
67
30
19
54
65
57
47
80
75
6.5
-459
92
50
% Removal
lead
57
65
38
35
37
68
59
88
84
49
59
67
65
34
37
69
63
63
72
58
24
53
94
53
8.3
61
28
8.3
94
55
% Removal
mercury
-11
8.3
55
43
-8.3
17
42
20
13
23
20
15
7.7
27
36
7.1
-20
20
3.4
29
21
0
-200
64
60
21
-200
64
21
% Removal
nickel
50
39
24
40
21
58
37
68
73
33
35
33
51
31
51
62
53
69
69
46
21
54
52
26
15
30
24
15
73
43
Negative percentages are listed as entries. However, averages were determined with negative percentages treated as zero removal.
-------
Table 13-2. (continued)
References
Data for references a - d were taken from the report:
Akers, David, Robert Dospoy, and Clifford Raleigh, The Effect of Coal Cleaning on Trace Elements, Draft Report, Development of Algorithms, December 16,
1993, prepared for EPRI by CQ Inc.
Data for reference e were taken from the report:
Demir, llham, Richard D. Harvey, Rodney R. Ruch, Heinz H. Damberger, Chusak Chaven, John D. Steele, Wayne T. Frankie, Ken K. Ho, "Characterization of
Available Coals from Illinois Mines," draft report, December 28, 1993, Illinois State Geological Survey file number to be assigned.
Specific references that were mentioned in the report by Akers, Dospoy, and Raleigh:
a DeVito, M., L. Rosendale, and V. Conrad, "Comparison of Trace Element Contents of Raw and Clean Commercial Coals," presented at the DOE Workshop on
Trace Elements in Coal-Fired Power Systems, Scottsdale, AZ, April 1993.
c Ford, C. and A. Price, "Evaluation of the Effects of Coal Cleaning on Fugitive Elements: Final Report, Phase III," DOE/EV/04427-62, July 1982.
w
i
-------
Although there is variability in trace element reductions, the
data suggest that coal cleaning techniques may be useful in reducing
trace element concentrations in selected coals. More studies are
needed on diverse samples of coal to establish the effectiveness of
coal cleaning in reducing trace element concentrations and to
determine the causes of variability in cleaning effectiveness.
13.1.3 Coal Gasification
Coal gasification converts coal to a syngas-fuel form that emits
lower quantities of pollutants at the utility boiler than if the coal
were not converted. Although there are some disadvantages to this
process (the cost of gasification and the addition of another
combustion source), the total quantity of air pollutants emitted from
the combination of gasification and combustion is expected to be lower
than burning coal in a conventional system. 9'10
The gasification process typically described for near-term
generation projects uses integrated gasification combined cycle (IGCC)
technology and conventional cold-gas cleanup. In this process, gas
from coal is used to generate electricity from both a steam turbine
and a gas turbine. Steps in the process, shown in Figure 13-3,
include coal preparation, coal oxidation and gasification, gas
cooling, and gas cleanup. A large part of the pollutant mass is
transferred to the slag or ash produced during gasification, and more
of the impurities are transferred to water streams used in the gas
cleanup. Slag or ash from the gasification step may be treated for
recovery of salable products, and the stream from gas cleanup may be
treated for recovery of sulfur. Heat transferred from the cooling
step is used to produce steam for the steam turbine generator, while
fuel gas made from the coal is burned to produce more electricity from
the gas turbine generator. The IGCC technology can produce up to 25
percent more electricity from a given amount of coal than is currently
obtained from conventional boilers.9
Statements by the DOE9 suggest that IGCC technology is almost
certain to be one of the lowest-cost fossil fuel options for
generating electricity in the 21st century. When used to refurbish an
existing plant, the technology is less expensive than building a
conventional coal-fired plant. Other claims include higher thermal
efficiency (to about 40 or 45 percent from about 35 percent), higher
plant output (by 50 to 150 percent), and lower S02 and NOX emissions as
described above. The IGCC process has been demonstrated in a limited
number of commercial-scale projects. In addition, IGCC is being
utilized in several DOE CCT projects.
13-11
-------
OXIDANT
COAL
PREPARATION
GASIFICATION
COAL
GASIFICATION
ASH
DISPOSAL
SLAG/ASH
RECOVERY
SULFUR
RECOVERY
WATER
DISPOSAL
GENERATOR
HEAT
RECOVERY
STEAM
GENERATOR
GENERATOR
COMBINED CYCLE
Figure 13-3. Coal gasification combined cycle technology.
There are limited data available on the impact of IGCC on HAP
emissions. Experiments with a different gasifier, an air-blown,
fixed-bed gasifier coupled to a turbine simulator, produced trace
metal concentrations as shown in Table 13-3.10 The hot gas from the
gasifier was treated in a moving bed with zinc titanate sorbent,
13.2 COMBUSTION CONTROL
Combustion control deals with the effect of furnace type (firing
method and bottom type) and furnace modifications (such as the
addition of low-NOx burners) on HAP formation. Since the recent
emission testing on utility units provided a significant amount of
information on the generation and control of trace metals but
considerably less information on organic HAPs, trace metals are used
13-12
-------
Table 13-3. Emissions from an Air-Blown, Fixed-Bed Gasifier
Trace metal
Arsenic
Cadmium
Chromium
Mercury
Nickel
Selenium
Emissions to flare, ug/Nm3
639
16
155
20
1,530
68
Emissions from turbine
simulator, ug/Nm3
8
0.19
20
2
26
0.56
Total air, ug/Nm3
647
16.2
175
22
1,556
68.6
to analyze the effect of combustion control. The trace metals for
coal-fired units examined in this section are arsenic, beryllium,
cadmium, chromium, lead, manganese, and mercury; those for oil-fired
units are arsenic, lead, nickel, and mercury.
While the majority of recently collected HAP data has focused on
metals, some small-scale studies have been conducted to evaluate
changes in combustion conditions on organic HAPs. In one test, coal
was burned at normal and elevated excess air levels and with air
staging to simulate combustion modification NOX controls.11 A large
number of organic HAPs were sampled in each case. Some increases in
HAP emissions were noted for the air-staging conditions, but the
conclusion was that this increase would not result in emissions at
significant levels, even for a large utility boiler. During a second
small-scale test, combustion conditions were varied between very high
excess air and substoichiometric conditions.12 This study concluded
that low-NOx firing conditions did not necessarily exacerbate emissions
of organic HAPs. In both studies, the organic emissions were found to
be one or more orders of magnitude less than emissions of any of the
metallic HAPs, even under the worst combustion conditions tested.
Although these results are from small-scale units and are relatively
limited in their scope, they provide additional information supporting
the position that, in general, organic HAP emissions are not likely to
increase significantly due to the installation of low-NOx combustion
equipment.
The effect of NOX control on metallic HAP generation was examined
by developing an average emission output in Ib/trillion Btu from one
oil- and several coal-fired units. These units were tested before and
after the installation of NOX control or after the addition of greater
NOX control to an existing N0x-controlled unit. As shown in Table 13-4,
there appears to be a trend toward reductions in HAP emissions through
the addition of NOX control. However, this trend is neither uniform
13-13
-------
Table 13-4. Comparison of Electric Utility Emissions Before and After Application of NOxControl
or Application of Greater NOX Control on a Unit With Lesser NOX Control (Ib/trillion Btu) 13~16
Trace metal
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Nickel
Average
Overall
Percentage
Change
EPRI Site
110 (with
low NOX
burners)
141.73
43.87
448.83
745.12
878.75
1021.84
3.58
538.55
T
Lesser
NOX
control
EPRI Site 110
(with low NOX
burners and
offset air)
64.15
45.26
120.9
743.02
503.23
999.7
0.02
511.13
T
Greater NOX
control
EPRI Site
110
percentage
change
-121%
3%
-271%
-0.28%
-75%
-2%
-17981%*
-5%
-67%
EPRI Site
114
(without
NOX
control)
137.22
39.86
17.38
241 .77
1365.7
392.1
6.47
1073
T
Lesser
NOX
control
EPRI Site 114
(with overfire
air and reburn
burners)
151.4
36.21
20.41
164.57
1007.17
375.22
4.48
1437.58
T
Greater NOX
control
EPRI Site
114
percentage
change
9%
-10%
15%
-47%
-36%
-4%
-44%
25%
-11%
EPRI Site 13
(without NO
control)
7.1
ND
10.13
2.95
8.02
4.7
0.23
1827.15
T
Lesser NOX
control
EPRI Site 13
(with burners
out of
service)
3.7
ND
13.84
8.95
4.6
7.84
0.17
1355
T
Greater NOX
control
EPRI Si
EPRI Site 13
percentage
change
-92%
ND
27%
67%
-74%
40%
-36%
-35%
-15%
e16
EPRI Site 16
(wjth overfire
air)
1789.01
108.33
10.93
934.26
437.31
745.42
7.06
664.17
T
Lesser NOX
control
(with
overfire air
and low NO
burners)
1805.89
130.43
21.07
845.76
351.15
890.91
10.74
655.97
T
Greater NOX
control
EPRI Site 16
percentage
change
1%
17%
48%
-10%
-25%
16%
34%
-1%
10%
w
I
ND = This trace metal was below the detection limit in both the coal feed and the boiler exit emissions on both tests.
a This percentage difference seems too large and is not used in the average overall percentage change calculation.
-------
(see arsenic, beryllium, and cadmium percentage change between
Sites 110 and 114) nor universal (see Site 16 compared to all
other sites). The differences in the percentage change could be
due to the variability of trace metal concentrations in the oil
or in different sources of bituminous coal; differences in the
carbon, chloride, or ash content of the coal burned; differences
in the age or operating condition of the units; or a combination of
all of these or other factors.
The effect of the bottom type (furnace type) on HAP generation
was examined by analyzing an average emission output of coal-fired
units (see Table 13-5) with either a wet bottom furnace or a dry
bottom furnace, both burning pulverized coal. Emissions were further
segregated by coal type. Since there was only one test done on a
conventional, dry bottom, lignite-fired unit, no lignite-fired units
were analyzed. Site 111, which burned a mixture of bituminous and
subbituminous coal, was grouped with the subbituminous coal-fired
units for this analysis. To support the emission results, the EMFs
for these furnaces and the concentrations of the trace metals found in
feed coal are also included in Table 13-5. In this way, the effects
of furnace type and low-NOx operation and of trace metal concentration
in the coal can be observed. Furnace type, with and without low-NOx
operation, may affect the partitioning of ash between bottom ash and
fly ash. Trace metal concentration in the coal affects the trace
metal concentration in the flue gas in either vapor or solid form.
Oil-fired units could not be separated into wet or dry bottom
configurations, as all oil-fired units use dry bottom furnaces.
Therefore, the impact on HAP emissions of bottom type is not addressed
for oil-fired boilers.
When units firing bituminous and subbituminous coal were
analyzed, their averages showed that arsenic, lead, and mercury seemed
to be emitted in higher amounts by wet bottom units while beryllium,
cadmium, chromium, and manganese seemed to be emitted in higher
amounts by dry bottom units (see Figure 13-4a). When units firing
only bituminous coal were analyzed, the same effect was observed (see
Figure 13-5a). When units firing only subbituminous coal were
analyzed, their averages showed that emissions of almost all HAPs
stated above were emitted in higher amounts from dry bottom units than
from wet bottom units (see Figure 13-6a). Mercury was the exception,
being emitted in similar amounts by both bottom types. The percent
removal by bottom type and the concentrations of trace metals in the
feed coal show a logical relationship between the trace metal
concentration observed in the feed coal and the amount emitted from
the boiler. This relationship appears to be a consequence of the ash
and trace metal partitioning in the boiler.
13-15
-------
Compound
JottoirnVpJ
Furnace Type
EPRI Site 114
(without NOX
control)
WET
I
H
CTl
Cyclone
Fuel type | BIT
EPRI Site 114
flth overfire air
and reburn
burners)
WET
DOE Miles
test (with
SNOX SCR)
WET
Cyclone
BIT
Cyclone
—
DE Miles
"WET
Cyclone
BIT
Northern
States Power
Riverside 8
WET
Cyclone
SUB
EPRI Site 102
(same as NSP
A.S. King)
WET
Cyclone
SUB
Total bituminous &
subbitummous
Average Demissions
39
7
Average emissions
Average_emissions
620
28
Average percent removal by
boiler (1-average EMF)
59%
Average percent removal by
boiler (1-average EMF)
54°/
Average percent removal by
boiler (1 -average EMF)
56%
77%
Boiler emission modification fa tors
Average trace metal
concentrations in feed coal
(microgram/gram)
Average trace metal
concentrations in feed coal
(microgram/gram)
219
Average trace metal
concentrations in feed coal
(microgram/gram)
13.0
in feed coal (microgram/gram
34f
Trace metal concentrations
12.0
Beryllium
Cadmium
Chromium
Lead
Arsenic
Beryllium
Cadmium
[ead ~
Mercury
Boiler emissio
.
.
s(lb/trillionBtu)
137.2
sai
17.4
1365.7
39Z1
6.5
1007 2
3752
Arsenic
Beryllium
Cadmium
Chromium
Tea?
Manganese
Mercury
BIT = bituminous
-selective catalytic reduction process
-------
Table 13-5. Continued
Compound
Bottom Type
Furnace Type
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Fuel type
EPRISite 110
(w/LNBand
offset air)
DRY
Tangential
BIT
Boiler emissions (Ib/trillion Btu)
64.2
45.3
120.9
743.0
503.2
999.7
0.02
EPRI Site 110
(w/LNB)
DRY
Tangential
BIT
141.7
43.9
448.8
745.1
878.8
1021.8
Boiler emission modification factors
0.89
0.93
1.00
1.00
1.00
0.71
0.66
0.39
0.43
0.70
1.00
0.36
0.76
DOE Yates
DRY
Tangential
BIT
393.1
90.1
23.8
2798.8
688.6
2055.9
12.4
1.00
1.00
1.00
1.00
1.00
1.00
1.00
Trace metal concentrations in feed coal (microgram/gram)
2.0
1.3
2.1
9.0
17.0
16.0
0.08
1.9
0.6
1.7
5.5
5.0
17.0
2.3
1.1
0.3
24.8
8.0
23.4
0.08
EPRI Site 15
DRY
Tangential
BIT
598.9
45.8
7.4
1164.3
811.3
1752.6
0.60
0.54
0.01
0.58
1.00
0.81
1.1
8.0
26.0
4.0
28.0
DOE
Cardinal
DRY
Opposed
BIT
1043.2
113.0
32.4
976.0
623.0
1350.8
1.7
0.91
0.96
1.00
0.61
1.00
0.27
0.41
12.9
1.3
0.1
18.0
6.8
57.3
0.05
EPRISite 116
DRY
Front
BIT
61.5
4.1
0.6
71.8
23.9
71.8
2.2
0.70
0.35
0.12
0.27
0.26
0.21
0.97
5.0
0.7
0.3
15.0
5.3
19.0
0.13
EPRI Site 12
DRY
Opposed
BIT
498.4
58.6
33.8
1252.7
381.2
1723.6
7.6
1.00
1.00
0.14
1.00
1.00
1.00
0.74
6.2
0.7
3.4
17.0
2.4
18.0
0.14
EPRI Site 14
DRY
Opposed
BIT
230.4
30.1
5.3
735.6
115.2
1152.2
0.8
0.50
0.92
0.02
0.67
0.79
0.93
0.74
6.3
0.5
3.0
15.0
2.0
17.0
0.24
EPRI Site
115
DRY
Vertical
BIT
38.2
17.5
4.0
87.5
167.1
334.3
1.7
0.61
0.52
0.58
0.57
0.38
0.58
0.78
0.5
0.2
0.1
1.1
2.1
4.2
0.02
EPRISite 16
OFA
DRY
Opposed
BIT
1789.0
108.3
10.9
934.3
437.3
745.4
7.1
1.00
1.00
1.00
0.58
1.00
0.60
0.64
17.0
1.4
0.1
22.0
5.1
17.0
0.15
EPRISite 16
OFA/LNB
DRY
Opposed
BIT
1805.9
130.4
21.1
845.8
351.2
890.9
10.7
1.00
0.82
0.11
0.69
0.66
0.88
1.00
23.0
2.2
2.6
17.0
7.3
14.0
0.14
BIT = bituminous
SUB = subbituminous
SNOX = wet sulfuric acid-selective catalytic reduction process
-------
Table 13-5. Continued
Compound
Bottom Type
Arsenic
Beryllium
Cadmium
Lead
Manganese
Mercury
i 1 i i i i i i i i i i i i i i
1 1 1 ! 1 1 | ! I 1 | ! 1 I | i
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
M M M M M M M 1
! ! ! i ! ! : i : : i i : ! i ! :
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Fuel type
1 1 i i 1 1 ! i 1 1 i !
!:]!!:•:::!!
Boiler emissi
! 1 i M M M 1 ' !
Boiler emissic
! M M M M '' i i
! i i 1 ! , i ! i , i ! !
Trace metal c<
EPRISite 11
DRY
Tangential
SUB
3ns (Ib/trillion
124.1
9.9
58.7
248.2
143.1
1406.5
I M '' M '• h ' i i
n modification
0.92
0.79
0.35
0.72
1.00
0.98
! i j ' : i M : i ! !
i i ' ! i M ! i M i
jncentrations
1.6
0.2
2.0
4.1
1.4
17.0
EPRISite 111
DRY
Opposed
SUB (BIT Mixed)
!!l|!il|!!ll!ill
' i ! i ! i !!!•!!:•!! I
3tu)
5.8
16.6
i M M i i M M M ; i i
1 ! i • ' : ! • '• i i '• '• : ; !
factors
0.11
0.05
i M : i ! ' : i ! ! M : ! I
i ' ' i i ' ! i i I ! i i ! i i
n feed coal (microc
0.6
4.0
DOE
Springerville
DRY
Tangential
SUB
' i i i ; i i i '• i i
: 1 1 i : 1 1 i ; 1 1
81.0
93.4
172.4
607.0
91.0
5627.6
5.4
! I I i ! I I I i h I
0.29
0 87
1.00
068
0.19
0.72
1.00
i i M i i M i i i i
jj!!!!!;!!!;
M I | I I i | I I i |
ram/gram)
1.5
1.1
0.5
92
50
80.8
0.04
NSP
Sherburne 3
DRY
Opposed
SUB
n i M
: 1 1 i i 1 1 i : 1 1 i !
644
21.8
7.0
253.6
525.4
49782
6.0
i i i j i ! i !::!!•
: 1 1 : : 1 1 : : 1 1 : !
0.79
0.58
0.99
0 49
0.49
0.00
1.00
i i i i i i i i i i i i i
i | i i i | i I | I i I i
0 7
0.3
0.1
4 4
9.1
0.02
NSP
Riverside 6,7
DRY
Front
SUB
i i ; i i i ' i i i ! i i
47/I
9.4
10.3
326.7
236.1
480.0
4.3
H M H M i i M i
• ! ! ! • '• ! M ! ! ! !
0.99
0.40
0.25
1 00
0.19
0.77
1.00
M i i i i i i M i i i
! i j ; : j ! ] : : ! ! !
0.4
0.2
0.4
? K
10.5
5.4
0.001
i ! I I i ! I I i ! I I i
! I i I i I i I i i i I
: | | | I | | | | | j |
i i i i i i i i i i M
i i M i i M h M
Total bituminous &
subbituminous
| i h i | j ] | i| | i i i! ( i ! I i i ! I ! |
Average emissions
437
55
61
786
398
1639
5
i i i ! i i i M M M M ! i M M M M i
Average percent removal by
boiler (1 -average EMF)
27%
24%
39%
30%
24%
19%
jiliiilHI MMMMMM!!!
Average trace metal
concentrations in feed coal
(microgram/gram)
5.5
0.9
1.8
12.7
6.1
23.9
0.09
Bituminous
i ! !: i !!••:::•:::•!: I !:: | !:
Average emissions
606
62
64
941
453
5
Average percent removal by
boiler (1 -average EMF)
22%
23%
48%
27%
23%
29%
23%
M M i i M M M i i M i i M i i M i i
Average trace metal
concentrations in feed coal
(microgram/gram)
7.7
1.0
2.0
15.5
59
21.0
0.11
Subbituminous
i i i i i i ! ! i I ; i i ! ' i i ! ! i i ! i ! i
: i ! i : i ! i ! ! I i I ! I i ! ! : j i ! : i i !
Average emissions
64
34
53
359
249
3123
5
i ! ; : i ! ! i i i : : ! ! : : ! ! ! : : : ! I : :
Average percent removal by
boiler (1 -average EMF)
38%
34%
47%
28%
53%
38%
0%
Average trace metal
concentrations in feed coal
(microgram/gram)
1.0
0.4
1.4
5.1
6.5
34.4
0.02
BIT = bituminous
SUB = subbituminous
SNOX = wet sulfuric acid-selective catalytic reduction process
-------
Figure 13-4a. Average boiler emissions
1639
Arsenic Beryllium Cadmium Chromium
Lead Manganese Mercury
Figure 13-4b. Average trace metal removal by boiler
77%
79%
75%
Arsenic
Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 13-4c. Average trace metal concentration in feed coal
Arsenic Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 13-4 (a-c). Average coal-fired boiler emissions, trace metal removal, and average trace element
concentration in feed coal vs. bottom type (bituminous and subbituminous coal)
Note: Data taken from Table 13-5.
13-19
-------
Figure 13-5a. Average boiler emissions
1100
Arsenic
Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 13-5b. Average trace metal removal by boiler
92%
82%
Arsenic Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 13-5c. Average trace metal concentration in feed coal
Arsenic Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 13-5 (a-c). Average coal-fired boiler emissions, trace metal removal, and average trace element
concentration in feed coal vs. bottom type (bituminous coal-fired only)
Note: Data taken from Table 13-5.
13-20
-------
Figure 13-6a. Average boiler emissions
s.
3500
3000
2500
i
12000
|
11500
3
1000
500
0
3123
64 39
34 7
53 5
5 5
Arsenic
Beryllium Cadmium Chromium
Lead Manganese Mercury
Figure 13-6b. Average trace metal removal by boiler
76%
66%
°-20%
10%
0%
0%
4%
Arsenic
Beryllium Cadmium Chromium
Lead Manganese Mercury
Figure 13-6c. Average trace metal concentration in feed coal
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese Mercury
Figure 13-6 (a-c). Average coal-fired boiler emissions, trace metal removal, and average trace element
concentration in feed coal vs. bottom type (subbituminous coal-fired only)
Note: Data taken from Table 13-5.
13-21
-------
Based on this analysis, wet bottom furnaces seem to have better
trace metal removal than dry bottom furnaces. Trace metal removal in
a furnace is likely due to the partioning mentioned above. A possible
explanation for the effect would be that bottom ash in a wet bottom
furnace is kept in a molten state, and, thus, the trace metals in the
bottom ash are less likely to reentrain into the fly ash. It needs to
be emphasized that these analyses are based on limited data and may
not hold true for all units and coals. More data and analyses are
needed.
It needs to be noted that the averages in Figures 13-4(a and b),
13-5(a and b), and 13-6(a and b) were computed from data with EMFs
limited to a maximum of 1.0, meaning that no more HAP could exit a
device than entered it. All of the data used in these figures had at
least one instance in which an EMF of 1.0 was used. This situation
did not occur in Figures 13-4 (c), 13-5 (c), or 13-6 (c) because these
data were taken directly from the coal feed without modification. The
result of this methodology is several sets of data averages where more
HAP is emitted than was present in the feed coal. These data averages
were composed of a large number of EMFs of 1.0.
13.3 POSTCOMBUSTION CONTROL
To comply with various local, State, and Federal requirements,
utilities routinely use postcombustion technologies for the control of
PM and S02. The following sections assess how different APCDs affect
removal of selected HAPs from fossil-fuel-fired electric utility flue gas.
13.3.1 Particulate Phase Controls
Figures 13-7 through 13-14 and Tables 13-6 through 13-9 show the
relationship between the HAP metal removal and PM collection efficiency of
different particulate controls (namely ESPs and FFs). The HAP removal
effectiveness is shown in the tables in this section as percent removal.
Percent removal is equivalent to 1 minus the EMF (see chapter 3, section
3.4.6). A 90 percent removal indicates that 90 percent of that HAP has
been collected by a PM control device. The HAP metals that exist
primarily in particulate form are readily controlled by PM control
devices. These HAPs include arsenic, beryllium, cadmium, chromium, lead,
and manganese. Table 13-10 shows the percentage of data, for all listed
HAPs other than mercury, with a control device HAP removal efficiency
greater than 90 percent. For example, 90 percent of the particulate from
metallic HAPs data points for cold-side ESPs fall into the 90 percent or
better removal category. For the two oil-fired sites for which ESP
removal data were available, the control of particulate metallic HAPs was
not clear. It should be noted that the concentrations of metallic HAPs in
oil, with the exception of nickel, which is not discussed here, are
significantly lower than those in coal, and the fuel-ash characteristics
are also quite different. These factors could explain the spread of PM
HAP removal of 51 to 93 percent.
13-22
-------
co
>
o
E
CD
100 —i
75 -
50
25 -
Hi I
I «
to
00
92 96 100
PM collection efficiency (%)
Figure 13-7. Removal of Metallic HAPs by Electrostatic
Precipitators (Cold-side, Coal) (Includes, Arsenic, Beryllium,
Cadmium, Chromium, Lead, and Manganese)
o
100 -,
80 —
60 -
40 —
20 -
92 96 100
PM collection efficiency (%)
Figure 13-8. Removal of,mercury by electrostatic
precipitators (cold-side, coal)
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Mean
98
94
80
97
93
98
25
Median
99
98
92
97
97
97
16
Sdev
2.1
17
31
2.3
15
1.5
26
Min
94
27
0
93
34
95
0
Max
>99
>99
98
>99
>99
>99
82
Count
18
18
18
18
18
18
17
14, 16, 17, 19-21, 23, 25, 26, 32, 34-38
Table 13-6. Descriptive Statistics for HAP Removal Efficiencies Shown in Figures 13-7 and 13-8
-------
100 —i
100
co
>
o
E
75 -
50
CO
>
o
E
99
98
98
99
0
Count
2
2
2
2
2
2
2
13
Table 13-7. Descriptive Statistics for HAP Removal Efficiencies Shown in Figures 13-9 and 13-10
-------
100 — |
75 -
CO
0
E 50 —
CD
^p
o^
25 -
— •
D
*
D
*
+
100 —
Compounds
+ Arsenic
0 Lead
Q Nickel
v J
75 -\
"cc
o
E 50 —
CD
^p
0^
25 -
n
u
•
•
I I
OJ
i
to
Ul
60 80 100
PM collection efficiency (%)
Figure 13-11. Removal of Metallic HAPs by an Electrostatic
Precipitator (Oil)(Includes, Arsenic, Lead, and Nickel)
60 80 100
PM collection efficiency (%)
Figure 13-12. Removal of Mercury by an
Electrostatic Precipitator (Oil)
Arsenic
Lecd
Mercurv
Nickel
Mecn
48
55
62
83
Medcn
48
55
62
83
SDev
4.6
2.9
29
14
Min
45
5?
42
73
MCK
51
§7
83
93
Count
2
2
2
2
39, 40
Table 13-8. Descriptive Statistics for HAP Removal Efficiencies Shown in Figures 13-11 and 13-12
-------
100
:i
100
o
E
o>
75 -
50
75 -
03
>
O
e
05
50
25 -
25 -
u>
to
en
96 98 100
PM collection efficiency (%)
Figure 13-13. Removal of Metalic HAPs by a Fabric Filter
(Coal) (Includes, Arsenic, Beryllium, Cadmium, Chromium,
Lead, and Manganese)
96 98 100
PM collection efficiency (%)
Figure 13-14. Removal of Mercury by a
Fabric Filter (Coal)
Arsenic
Beryllium
Cadmium
Chromium
Lead
Manganese
Mercury
Mean
99
99
72
94
99
98
36
Median
99
>99
95
99
99
99
34
Sdev
1.4
2.4
48
10
1.3
1.8
37
Min
97
94
0
75
97
95
0
Max
>99
>99
99
>99
>99
>99
73
Count
5
5
4
5
5
5
4
Table 13-9.
18, 27-29, 41
Descriptive Statistics for HAP Removal Efficiencies Shown in Figures 13-13 and 13-14
-------
Table 13-10. Particulate Metallic HAP Removal Percentage from ESPs
and FFs (Excluding Mercury)
Particulate control device
(coal)
ESP (cold-side)
ESP (hot-side)
FF
Number of data
points
108
12
33
Percentage of data with a HAP
removal efficiency
greater than 90 percent
89
92
89
ESP = electrostatic precipitator
FF = fabric filter
HAP = hazardous air pollutant
Mercury, however, is not well controlled by PM APCDs. This
situation would be expected because mercury is emitted as a mixture of
solid and gaseous forms. Mercury removals and current investigations
on the control of mercury are further discussed in section 13.6.
Dioxin removal in utility boiler PM control equipment has been
measured at one coal-fired boiler and one oil-fired boiler. In both
cases, measurements contained many values of uncertain accuracy.
However, removal efficiencies could be estimated for one dioxin and
three furans at the coal-fired boiler. This unit was a 615-MWe boiler
firing Pennsylvania bituminous coal that had an ESP with an overall PM
efficiency of 99 percent. The ESP's apparent efficiency for the
following compounds was:
• 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin 7 percent
• 2,3,7,8-tetrachlorodibenzofuran 38 percent
• 2,3,4,6,7,8-hexachlorodibenzofuran 35 percent
• 1, 2 , 3,4,6,7,8-heptachlorodibenzofuran 29 percent.
The oil-fired boiler was an 850-MWe unit firing residual oil and
had an ESP with an overall PM collection efficiency of 92 percent.
All of the measurements for this site indicated negative efficiencies
for dioxins and furans in the ESP.
Under certain conditions in MWC systems, dioxins and furans can
be formed in the particulate-laden flue gas stream upstream of or
within the associated PM control equipment at temperatures in the
range of 480° to 1,020° F (250° to 550° C) . 42'44 Units that are equipped
with hot-side ESPs (ESPs upstream of the air preheater operating at
temperatures in the range of 600 to 750°F [316° to 400° C] ) are of
13-27
-------
particular concern with regard to this formation mechanism since their
operating temperatures typically fall within this range. Little
information is available regarding dioxin formation in utility
particulate removal equipment, however, and additional information is
needed to adequately quantify the potential for dioxin formation in
utility pollution control systems. Thus, at utility plants, any
strategy for dioxin control must consider adequately treating large
volumes of gas in order to remove relatively small concentrations of
dioxin.
13.3.2 Vapor Phase Controls
Figures 13-15 through 13-18 and Tables 13-11 and 13-12 show the
relationship between HAP metal removal and the inlet temperature for
S02 control devices. The correlation between FGD scrubber inlet
temperature and HAP metal removal is difficult to determine. This
difficulty is compounded by having a maximum of eight data sites at
which four of the eight test sites employ flue gas bypass in their
design. A bypass means that part of the flue gas is diverted around
the FGD or SDA/FF while the majority of the flue gas is treated by
these control devices. Bypass is used to minimize the size and the
lime/limestone cost of the FGD unit while still meeting S02 emission
limits. Another factor is that FGDs usually follow an ESP or an FF;
thus, the concentrations of metallic HAPs that reach the FGDs are
generally less than 10 percent of the amounts intercepted by primary
PM control devices.
The HAP metal removal by SDA/FF-equipped units seems to follow
the results found in FF PM-controlled units, i.e., metallic HAPs are
controlled to approximately 90 percent or better. This situation
follows because an SDA/FF contains an FF.
13.3.3 Acid Gas Control
There was a limited amount of data (using EPA Method 26a)
available on the removal efficiencies for HC1 and HF from air
pollution control devices. Since utilities operate with varying
amounts of bypass, removal percentages for FGDs and SDA/FFs were
estimated for operations with 15 percent and 14 percent bypass,
respectively. The test report data indicated that: ESPs removed less
than 6 percent of the acid gases; FFs removed approximately 44 percent
of the HC1 and essentially none of the HF; an FGD with 15 percent
bypass was estimated to remove approximately 80 percent of the HC1 and
approximately 29 percent of the HF; and an SDA/FF with 14 percent
bypass was estimated to remove approximately 82 percent of each acid
gas.47 Despite the inconsistencies in removal efficiencies achieved
for HC1 and HF with FFs and FGDs, the data indicate that the S02
control devices remove more of the acid gases than do PM controls.
13-28
-------
100 1
s
I
100 -i
CO
>
o
E
CD
75 -
50 —
25 -
80 —
« 60 H
o
E
o
40
20 -
See Note 1
U)
I
to
200
300
400
FGD inlet temperature (F)
Figure 13-15. Removal of Metallic HAPs by an FGD
(Coal) (Includes, Arsenic, Beryllium, Cadmium,
Chromium, Lead, and Manganese)
200 250 300
FGD inlet temperature (F)
350
Figure 13-16. Removal of Mercury
by an FGD (Coal)
Note 1 - This unit (EPRI Site 12) was
retested for mercury, but was tested as
a combined ESP/FGD system. Since
there was no way of determining which
component was responsible for the
mercury removal, the ESP was given
the full credit for removal. This explains
the "zero" data point in Figure 13-16.
(The ESP for this site was given an
"82" percent removal.)
Arsenic
Bervllium
Cadmium
Chromium
Lead
Manaanese
Mercurv
Mean
69
73
38
55
57
50
31
Median
79
87
25
57
61
68
23
Sdev
26.0
36
35
35.8
39
47.4
23
Min
24
3
0
0
0
0
0
Max
96
98
90
97
98
99
62
Count
8
6
7
8
8
7
9
21, 23, 25, 38, 45, 46, 48-50
Table 13-11. Descriptive Statistics for HAP Removal Efficiencies Shown in Figures 13-15 and 13-16
-------
100
• •
100 -|
75 H
80 —|
as
>
o
E
CD
50
o
E
i
ui
o
280
320
SDA inlet temperature (F)
Figure 13-17. Removal of Metallic HAPs by a
Spray Dryer Adsorber/fabric Filter (Coal)
(Includes, Arsenic, Beryllium, Cadmium, Chromium,
Lead, and Manganese)
260 280 300
SDA inlet temperature (F)
320
Figure 13-18. Removal of Mercury by a Spray Dryer
Adsorber/Fabric Filter (Coal)
Arsenic
Bervllium
Cadmium
Chromium
Lead
Mcnacnese
Mercury
Mecn
99
> 99
90
98
99
> 99
26
Median
> 99
> 99
90
97
> 99
> 99
24
SDev
1.8
0.3
8 2
1.5
0.7
0.1
29
Min
96
99
81
97
99
> 99
0
Mcx
> 99
?> 99
> 99
> 99
> 99
> 99
55
Count
4
3
4
3
3
3
4
22, 30, 31, 45
Table 13-12. Descriptive Statistics for HAP Removal Efficiencies Shown in Figures 13-17 and 13-18
-------
13.3.4 Carbon Adsorption
A possible way of further reducing the amount of vapor phase HAPs
emitted from utilities is through the use of carbon adsorption.
Activated carbon (AC) is a specialized form of carbon produced by
pyrolyzing coal or various hard, vegetative materials (e.g., wood) to
remove volatile material. The resulting char then undergoes a steam
or chemical activation process to produce an AC that contains multiple
internal pores and has a very high specific surface area. With this
internal pore structure, the AC can adsorb a broad range of
contaminants.
Activated carbon can be introduced through the use of a fixed-
carbon filter bed, a moving bed, a fluidized carbon filter bed, or
through direct AC injection into the flue gas stream.
Other than for mercury (see section 13.6.1.2), no utility data
were found for AC removal of HAPs. However, for other industries
(e.g., MWCs and medical waste incinerators [MWIs]), dioxin removal
from the flue gas stream has also been achieved by AC injection.
13.4 ALTERNATIVE CONTROLS
An alternative to pollution control systems applied directly to
boiler systems is to use alternative, nontraditional control methods
such as demand side management (DSM) and energy conservation. Demand
side management addresses the issue of reducing potential HAP
emissions by reducing the amount of electricity needed. This
objective can be accomplished by several different methods.
Through the use of progressively higher efficiency electrical
devices (e.g., more efficient home appliances, lighting systems, and
industrial machinery), the overall tonnage of fossil fuel burned would
be reduced.51 Also, campaigns to reduce the use of fossil-fuel-
generated electricity have a positive effect on reducing HAPs by
slowing down the necessity of building additional plants. Research
into more efficient electrical generators and transmission equipment
could produce units that deliver the same amount of energy for less
fuel.52
Another potential fuel option is the use of liquid or gaseous
fuels derived from biomass. Currently, fuels such as ethanol and
methanol, derived from corn, grains, and other crops, are being used
to produce "gasohol" (a blend of up to 10 percent ethanol in
gasoline). Future use of these fuels by utilities, as well as
hydrogen fuels derived from biomass, could reduce (but perhaps not
eliminate) HAP emissions. However, DOE has estimated that development
of technologies to produce sufficient quantities of biomass-derived
fuels may not be commercially viable until 2030. 53 Other assessments
indicate that with realistic investments in research and development
13-31
-------
in both crop production and generation techniques, biomass could be
competitive (without subsidies) in niche markets within 5 years (with
high energy prices, which are not currently forecast) and within 10
years in larger markets.54
Another method would be to switch to a source of renewable energy
(e.g., wind, solar), but to date these methods have been limited to
certain geographic locations only. (However, such campaigns may also
lead to slower introduction of new, more efficient fossil fuel-burning
technologies.)
Future potential electrical transmission equipment could include
the development and use of superconductive power lines, which could
substantially reduce the amount of electricity that needs to be
generated to move the high voltage electricity through power lines
because of the negligible power loss (due to lower resistance).55
13.5 POLLUTANT TRADEOFFS
13.5.1 HAP Increase/Decrease
The various strategies for limiting HAP emissions, discussed in
sections 13.1 through 13.4, have different effects in controlling air
emissions of all HAPs. Table 13-13 presents the qualitative effects
of the different control strategies on air emissions.
Table 13-13 provides a comparison of HAP removal effectiveness of
different existing and alternative control strategies. As shown in
Table 13-13, the effect on HAP emissions of: (1) switching from a
higher- to a lower-sulfur coal, (2) NOX controls, and (3) boiler types
cannot be predicted from the available data. Techniques that would
reduce emissions of the HAPs of concern include: (1) switching from
coal or oil to natural gas; (2) coal gasification; and (3) alternative
controls, such as energy conservation or DSM. The remaining
strategies control certain HAPs. Existing PM control devices, such as
ESPs and FFs, generally do not remove the vapor-phase HAPs (i.e.,
organics, elemental mercury, HC1, and HF). (However, these controls
do provide some reduction of ionic mercury emissions.) Emissions of
the vapor-phase HAP, dioxin, are not controlled by ESPs but are
controlled to some extent with FFs because dioxins adsorb onto the
filter cake. As noted in section 13.3.1, there is the potential for
dioxins to be produced in hot-side ESPs due to the temperature and
catalytic effects of the fly ash. With the exception of elemental
mercury emissions, the existing S02 control devices, namely FGDs and
SDA/FFs, tend to reduce emissions of the HAPs and provide some
reduction of emissions of ionic mercury.
13-32
-------
Table 13-13. Qualitative Effects of Different Control Strategies on Air Emissions of HAPs
M
w
1
w
w
Type of alternative control strategy
PRECOMBUSTION CONTROL
Conventional coal cleaning
Coal gasification
Fuel switching
From a higher to lower sulfur coal
To western and certain eastern coals
Coal or oil to gas
COMBUSTION CONTROLS
NOX controls
Boiler type
POSTCOMBUSTION CONTROLS
F 'articulate phase controls
ESP
FF
Vapor phase controls
FGD
SDA/FF
Carbon adsorption
NON-TECHNOLOGY-BASED CONTROL OPTIONS
Demand management
Effect on particulate HAPs
Primarily
organically
bound
No removal
Primarily
inorganically bound
Effect on mercury
Primarily
oxidized
mercury
Primarily
elemental
mercury
Effect on acid
gases Effect on dioxin
(HCI, HF)
Decrease in emission
Decrease in emission
Decrease in emission
(highly coal-specific)
Cannot predict the effect
Decrease in emission if chloride or fluoride is
reduced
Decrease in emission
Cannot predict the effect
Cannot predict the effect
Decrease in emission
Cannot predict the effect
Cannot predict the effect
Decrease in emission
Cannot pred ct the effect
Cannot pred ct the effect
Decrease in emission
Cannot be determined without further testing
Cannot be determined without further testing
Decrease in emission
Decrease in emission
Some decrease
in emission
Some decrease
in emission
No effect
No effect
No effect"
Decrease in
emission
No effect
because of filter
cake adsorption
Decrease in emission
Decrease in emission
Decrease in
emission
Some decrease
in emission
No effect
No effect
Decrease in emission
Decrease in emission
Decrease in emission. (Elemental and ionic mercury were removed, respectively, by impregnated and unimpregnated carbon adsorption, in
pilot-scale studies.)
Decrease in emission
There is the potential for dioxins to be produced in hot-side ESPs due to the temperature and
catalytic effects of the fly ash.
-------
Emissions of HAPs could be reduced through energy conservation or
DSM. Finally, pilot-scale studies suggest that ionic and elemental
mercury emissions could be controlled, respectively, with
unimpregnated- and impregnated-carbon adsorption, but full-scale
testing is needed to establish the effectiveness of these techniques.
13.5.2 Water/Solid Waste Considerations
Coal cleaning can produce a variety of waste problems.56 The
process creates a liquid waste containing fine coal particles and
inorganic elements and compounds dissolved from the ash in the coal.
The large volumes of water used in the cleaning process and the large
amount of suspended solids generated dictate that process water be
clarified and recycled. The usual means of clarification is retention
in large sedimentation ponds. Contamination of surface water or
groundwater can occur from coal wastepiles or storage piles if water
is allowed to infiltrate them. Contaminants such as iron, manganese,
and heavy metals (such as cadmium and silver) may leach from the
wastes.5S
Precombustion controls such as fuel switching could have an
effect on reducing this waste. If coal with lower amounts of ash and
sulfur was mined, there would be less need to clean the coal and
therefore fewer tailings would be created at the coal washing site.
Switching from coal to natural gas would reduce the need to mine and
wash the coal.
In either case, the amount of bottom ash and captured fly ash is
quite large. Because the metals are not destroyed in the combustion
process, the ash will have a higher concentration of metals than the
coal, and water contamination may occur if water runoff from the ash
is not controlled. The amount of metals in ash disposal pits is not
likely to increase significantly if particulate controls are already
in place. If new controls are installed, larger amounts of ash will
require disposal, leading to an increase in the potential for water
contamination from ash leachate.
Coal gasification would not necessarily reduce the waste
potential of coal, but the processing of coal into a gasified form
would tend to localize the waste and transfer it from a gaseous stream
to a solid stream that could be more easily disposed of after being
properly treated.10 For example, gasifier residue may contain
contaminants in a nearly vitreous matrix that is not easily leached.
In conventional coal combustion, these contaminants would largely
appear in the flue gas stream.
Combustion controls, such as different boiler/furnace types or
the adding of NOX controls, would have the effect of changing
combustion conditions in the coal-fired furnace and, thus, changing
13-34
-------
the ratio of bottom ash to fly ash. Increasing the amount of bottom
ash in the furnace would consequently reduce the amount of fly ash
exiting the furnace. However, if the amount of fly ash exiting the
furnace was increased, there would be greater fly ash loading on the
PM controls. If the PM controls were not designed to accept this
additional load, excess fly ash could cause stack opacity problems and
higher HAP emissions.
The wet FGD processes use a liquid absorbent to absorb S02 gases.
This absorbent is most likely an alkaline slurry composed of lime or
limestone slurried with water that can be used in a regenerable
process or in a nonregenerable process. Both processes produce a
calcium sludge effluent that must be disposed of properly. The sludge
can be stored in a settling pond or dewatered to take up less space in
a landfill.57 In addition to removing sulfur oxides, regenerable
processes generate a usable product from the sludge, such as gypsum,
which can be used in the manufacture of wallboard.
In a dry FGD system, the flue gas is contacted with an alkaline
material to produce a dry waste product for disposal. The alkaline
material can be added either in the fuel prior to combustion, through
dry injection into the flue gas, or as an alkaline slurry.58 For
example, SDAs inject a lime/limestone alkaline slurry into the flue
gas steam. The reagent droplets absorb S02 while simultaneously being
dried. In all these methods, the resulting dried reagent and fly ash
are then captured by an FF or ESP and, thus, do not represent an
additional liquid or solid waste problem except that the sorbent may
contaminate salable fly ash.
The use of all forms of carbon adsorption have the potential to
add additional PM loading to existing PM controls. Also, the spent AC
either has to be disposed of as an additional solid waste or
regenerated and reused. Studies on other industries indicate that the
adsorbed HAP (e.g., mercury, dioxin) does not have a tendency to leach
out of the AC.
Section 3001 (b)(3)(c) of the Resource Conservation and Recovery
Act (RCRA) required that the EPA determine, based on the results of a
study required by section 8002(n) of RCRA, whether RCRA subtitle C
regulation of fossil fuel combustion wastes is warranted. On August
9, 1993, the Agency determined that regulation of four large-volume
waste streams resulting from fossil fuel combustion (specifically,
coal combustion in utility steam-generating units) was inappropriate.
These waste steams are fly ash, bottom ash, boiler slag, and flue gas
emission control waste. However, any change in the characteristics of
these wastes resulting from actions taken to specifically control HAPs
could necessitate a review of that decision by the Agency.59
13-35
-------
Changes in all elements of the alternative controls, from
conservation to technology improvements, can reduce the amount of
waste produced by the utility industry.
13.6 AVAILABLE CONTROL TECHNOLOGY AND STRATEGIES FOR MERCURY
CONTROL
Typical mercury removal efficiencies for conventional emission
controls are discussed in section 13.3. Conventional controls are
generally inconsistent in their effectiveness, and range from 0 to
more than 83 percent removal (based on emissions testing at full-scale
utility boilers).
Strategies for further reducing mercury emissions from electric
power generation include demand reduction to decrease overall fossil
fuel use, use of other forms of generation (e.g., nuclear power,
biomass), switching to fuels having less mercury (e.g., natural gas),
improving the mercury removal efficiency of conventional controls, and
adding controls that remove mercury more effectively than conventional
controls. The mercury content in utility flue gas ranges from 0.2
//g/dry standard cubic meter (//g/dscm) to 25 /ug/dscm at 7 percent
oxygen and standard conditions of 20° C and 1 atmosphere, while utility
flue gas flow rates may range from 10,000 to 4,000,000 dscm/min.60
Thus, at utility plants, any strategy for mercury control must
consider adequately treating large volumes of gas in order to remove
relatively small concentrations of mercury as well as addressing any
resulting impacts on power plant equipment operations (such as
particulate control devices) and on waste disposal issues.
There has been some evidence that a higher carbon content in fly
ash may lead to lower levels of stack mercury emissions, with the
hypothesis being that the mercury is adsorbed by the carbon in the
flue gases. There is other evidence that the chloride content of the
coal impacts on the form and suitability of mercury removal by
conventional control systems. This evidence has led to research on
improving mercury removal from utility flue gas streams.
This section briefly reviews one pre-combustion technology (coal
cleaning) and three classes of post-combustion technologies
(enhancement of existing APCDs, carbon injection, and novel
techniques). Additional information may be found in Appendix I.
13.6.1 Pre-Combustion Strategies
As mentioned in section 13.1.2, conventional coal cleaning may be
effective for reducing mercury concentrations only in specific coals
and, at this time, cannot be considered a mercury control technique
for all coals.63 Advanced coal cleaning techniques are being
13-36
-------
investigated for improved mercury removal potential. Like
conventional cleaning techniques, the advanced cleaning techniques
cannot be considered a mercury control technique for all coals at this
time.
Advanced coal cleaning methods such as selective agglomeration
and column froth flotation have the potential to increase the amount
of mercury removed by conventional cleaning alone. Bench-scale
studies indicate that the combination of conventional and advanced
coal cleaning techniques removed from 40 to 82 percent of the mercury
contained in eight samples of raw coal.61'62
Advanced cleaning methods, such as column froth flotation, are
starting to emerge. Microcel™ is a type of column froth flotation
available through ICF Kaiser and Control International. The company
is the exclusive licensee for the technology in the coal fields east
of the Mississippi River and has sold units for commercial operation
in Virginia, West Virginia, and Kentucky. Ken-Flote™ is another type
of column froth flotation cell.
13.6.2 Post-Combustion Strategies
13.6.2.1 Impact of Fuels and Temperature on Mercury Emissions
Fuels and the temperature of the flue gas can have a significant
impact on the quantity of mercury emissions in the flue gas from a
boiler and on the ability of control systems to remove the mercury.
13.6.2.1.1 Fuels and Mercury Speciation. Mercury is contained
in the coal and oil fuels burned in utility boilers. During
combustion, mercury readily volatilizes from the fuel and is found
predominantly in the vapor phase in the flue gas63 in one of three
forms: (1) elemental, (2) ionic, or (3) organic.
In the early 1990s, test methods were developed to quantify the
species of mercury present in utility flue gas. Since that time, more
accurate speciation test methods (e.g., Ontario Hydro, EPA draft 101B)
have been developed. Mercury speciation testing indicates that the
distribution of ionic mercury, most likely mercuric chloride (HgCl2),
and elemental mercury each varied in the sampled coal-fired utility
flue gas.
Preliminary test results suggest that the chloride concentration
in the coal and the type of coal (e.g., bituminous, subbituminous, or
lignite) may be associated with a particular speciation of mercury in
the flue gas, but more data are needed to verify these associations.
Specifically, higher concentrations of ionic mercury were associated
with tested coals containing high chloride concentrations (0.1 to 0.3
weight percent) , 64~66 while 149° C (300° F) flue gas from tested
13-37
-------
subbituminous coals appeared to contain approximately ten times the
percentage of elemental mercury as compared with flue gas from tested
bituminous coals.61'69 The variability in the speciation of vapor-phase
mercury in coal-fired flue gas may explain the variation in mercury
removal that is seen with existing control devices.70
The association between fuel chloride content and the
concentration of ionic mercury in the flue gas may also apply to fuel
oil, but this association has not been examined. Analysis of two
samples of flue gas suggests that mercury is predominantly in the
elemental form when the fuel is oil.
It is important to understand mercury speciation because it will
indicate potential ways to reduce mercury emissions. For example, the
ionic mercury form (i.e., Hg++) is water soluble. Wet scrubbing of the
flue gas may result in increased ionic mercury removal.
The scrubbing affinity for ionic mercury has been observed in
pilot-scale studies. Preliminary results from tests of pilot-scale
wet and dry scrubbers indicate that at least 90 percent of the ionic
mercury was captured, while the removal of elemental mercury ranged
from 0 to approximately 70 percent. S5'ss'71"73 These preliminary test
results suggest that a scrubbing system will have a low mercury
removal efficiency if it treats flue gas from a boiler that fires
subbituminous coal (containing relatively more elemental mercury) and
a high mercury removal efficiency if it treats flue gas from a boiler
that fires bituminous coal (containing relatively more ionic
mercury).74 Full-scale studies are needed to verify these
observations.
Speciation of mercury is important in planning control strategies
but is still under investigation. The speciation information is also
needed to understand what is emitted from utility stacks, how it is
affected by atmospheric chemistry, and the subsequent deposition of
mercury.
13.6.2.1.2 Temperature. Utility flue gas typically has a
temperature range of 121° to 177° C (250° to 350° F) after leaving an
air preheater, which is a heat exchanger commonly used to heat
incoming combustion air.60 Mercury is found predominantly in the vapor
phase in utility flue gas.63 If the vapor-phase mercury was condensed
onto PM, the PM could be removed with existing PM control devices.
Theoretically, cooler temperatures will give relatively more mercury
condensation onto PM63 and, subsequently, increased mercury removal
with existing PM control devices.
There is limited, preliminary evidence for the temperature
dependence of mercury removal in a pilot-scale FF study. The pilot
13-38
-------
study suggests that mercury removal efficiencies apparently increased
from 27 percent to 51 percent as the temperature of the flue gas
decreased from 107° C (225° F) to 96° C (205° F) . 75
13.6.2.2 Effectiveness of Mercury Capture Effectiveness of
Existing Pollution Control Technology. Typical mercury removal
efficiencies for conventional emission controls are discussed in
section 13.3. Since conventional controls are generally inconsistent
in their effectiveness, with a range from 0 to more than 83 percent
removal (based on emissions testing at full-scale utility boilers),
conventional controls cannot be considered a mercury control strategy
for all electric utility plants at this time. Due to the limited
effectiveness of conventional controls, research continues on ways to
improve mercury capture by conventional emission control devices,
sorbent injection, and the development of novel techniques. In order
to develop low cost post-combustion mercury strategies for full-scale
utility operation, fundamental research must continue on the chemistry
and interactions of flue gas constituents, fly ash, and mercury-
species encountered at various flue gas conditions across the utility
industry.
The following sections briefly describe research on enhancing
mercury removal with existing control technologies, activated carbon
injection, and development of novel mercury control technologies.
13.6.2.3 Enhancement of Existing Control Devices.
13.6.2.3.1 Enhanced Particulate Control. Several approaches to
improving the capture of fine particles and mercury with existing
control devices are being investigated. Studies are underway on
enhancing particulate control with a new Asea Brown Boveri (ABB)
precharger, a wet ESP, flue gas cooling and humidification, and an
advanced power supply (the ABB Switched Integrated Rectifier). Flue
gas cooling and humidification tests are currently in progress. This
approach shows promise in improving the collection of particulate-
bound mercury, and may also cause vapor-phase mercury to condense on
particulate matter and be captured in the ESP.76 Research is also
underway on optimizing the mercury removal capability of conventional
pollution control technologies.77
13.6.2.3.2 Enhanced Wet Scrubbers. Several approaches to
improving the capture of total and elemental mercury are being
investigated. These include optimizing the liquid-to-gas ratio, wet
FGD tower design, and improved elemental mercury removal with
scrubbing liquid additives and catalysts. Brief descriptions of these
approaches follow.
13-39
-------
Liquid-to-gas ratio. The liquid-to-gas (L/G) ratio of a wet FGD
system is dictated by the desired removal efficiency of S02 and impacts
the removal efficiency of oxidized mercury. In general, high
efficiency (95 percent S02 removal) systems are designed with L/G
ratios of 120 gal/1000 acf to 150 gal/1000 acf. In two separate
pilot-scale studies, increasing the L/G ratio from approximately 40
gal/1000 acf to approximately 125 gal/1000 acf increased the removal
efficiency of oxidized mercury from 90 percent to 99 percent.73'78 Test
data suggest that increasing the L/G ratio did not affect the removal
efficiency of elemental mercury, which was close to zero percent.79
Wet FGD tower design. Most of the existing U.S. wet FGD systems
have open spray tower or tray tower designs.80 Recent research has
shown that tray tower designs are more effective in removing oxidized
mercury from boiler flue gas than tower designs. In one study of wet
FGD systems, where the composition of the flue gas was mostly oxidized
mercury, the tray tower design removed from 85 to 95 percent of the
total mercury, whereas the open spray tower design removed from 70 to
85 percent of the total mercury.81
Improved elemental mercury removal. Since two studies noted
higher concentrations of elemental mercury in the outlet of a wet FGD
system compared to the inlet concentrations of elemental mercury, 82'83
research is in progress on ways to convert and capture elemental
mercury. Research currently is investigating scrubbing liquid
additives and catalysts to convert elemental to ionic mercury.
Several scrubbing liquid additives that combine strong oxidizing
properties with relatively high vapor pressures are being investigated
as techniques to enhance the capture of mercury in a wet scrubber. Of
three halogen (i.e., chlorine, bromine, and iodine) solutions tested
to date, the chlorine solution appears to remove the most elemental
mercury in the presence of S02 and NO. Further testing of these and
possibly other oxidizing reagents is planned. 84
Due to a much higher solubility compared to elemental mercury,
oxidized mercury is readily removed in a wet scrubber.
Investigations are underway on the conversion of vapor-phase elemental
mercury to more soluble Hg++ in bench and pilot-scale studies. The
effect of flue gas temperature and residence time on the oxidation
potential of a number of catalysts and coal-based fly ashes is being
studied. To date, pilot-scale tests of three iron-based catalysts, a
carbon, a bituminous, and lignite fly ash have shown the carbon-based
catalyst to be the most effective in converting elemental mercury to
Hg++. Further testing of the carbon catalysts is planned.85
13.6.2.4 Activated Carbon Injection. Activated carbon (AC)
injection is considered a potential control technology for mercury
13-40
-------
emitted from electric utilities, since a form of this technology has
been successfully demonstrated on medical waste incinerators and
municipal waste combustors .8S"89
The level of mercury control that might be achieved with AC
injection into utility flue gas may depend upon flue gas
characteristics such as temperature, chloride content (in the fuel),
mercury content (in the flue gas), and the volume of flue gas. As
shown in Table 13-14, these properties distinctly differ from those in
MWC flue gas. In particular, mercury concentrations in MWC flue gas
streams may be up to several orders of magnitude greater than those
seen in utility flue gas streams.
Due to the differences between the flue gas characteristics at
MWCs and utility units, the application of AC injection to utility
flue gas has not been directly scaled from the application at MWCs.
At utility plants, the small concentrations of mercury are contained
in a large volume of flue gas, and large amounts of AC may be needed
to provide adequate contact between the carbon particles and mercury.
Pilot-scale studies of AC injection on utility flue gas have been
conducted, but full-scale testing is needed to determine the
feasibility of using AC at utility plants.
13.6.2.4.1 Factors Affecting Mercury Removal Efficiency.
Preliminary data from various pilot- and bench-scale studies suggest
that factors besides the optimum amount of AC that is injected may
affect mercury removal. These factors are temperature, the speciation
of the vapor-phase mercury and type of activated carbon67 injected into
the flue gas, and flue gas composition.
Temperature. A pilot-scale study of AC injection upstream of an
FF suggests that mercury removal efficiencies and the required amount
of AC injection were apparently temperature dependent.75 In reducing
the temperature from approximately 116°C (240°F) to approximately 93°C
(200°F), the mercury removal increased from approximately 80 percent
(with an injection rate of approximately 3,500 //g carbon///g of inlet
mercury) to approximately 98 percent (with an injection rate of
approximately 155 //g carbon///g of inlet mercury) . (The high mercury
removal percentages suggest that flue gas contained mostly ionic
mercury.)
These test results suggest that more mercury is removed and less
carbon is needed at lower flue gas temperatures. However, it may not
be possible to lower the flue gas temperature sufficiently at a given
utility plant because utility plants typically operate with a stack
gas temperature between 121° and 177° C (250° and 351° F) upstream of
any PM control device to avoid acid condensation and, consequently,
equipment corrosion. The stack gas temperature may be lowered below
13-41
-------
Table 13-14. Comparison of Typical Uncontrolled Flue Gas Parameters
at Utilities and MWCsa
Uncontrolled flue gas
parameters
Temperature (°C)
Mercury content (^g/dscm)
Chloride content (^g/dscm)
Flow rate (dscm/min)
Coal-fired utility boiler 60'90
121 -177
1 -25
1,000-140,000
11,000-4,000,000
Oil-fired utility boiler 91'92
121 -177
0.2-2
1,000-3,000
10,000-2,000,000
MWCb 93'94
177-299
400-1,400
200,000-400,000
80,000-200,000
MWC = municipal waste combustion
aStandard conditions are 0° C and 1 atmosphere.
"Moisture content in the MWC flue gas was assumed to be 13.2 percent.
96° C (205° F) and still avoid acid condensation, provided low-sulfur
coals (less than about 1 weight percent sulfur) are burned.95 If the
utility burns low-sulfur coal and uses an ESP for PM control, the flue
gas will probably require conditioning to reduce the high resistivity
of the fly ash since high resistivity makes the fly ash hard to
collect in an ESP.
Speciation of mercury. The effectiveness of AC injection in
recovering different forms of mercury is still under investigation.
The available data indicate iodide- and sulfur-impregnated AC are
needed for significant elemental mercury removal.
Studies without AC injection at a pilot-scale SDA/ESP system in
Denmark and a full- and pilot-scale SDA/FF system indicated that
essentially all of the ionic mercury was removed (with greater than 97
percent removal efficiencies) while essentially none of the elemental
mercury was removed (with 0 to approximately 3 percent removal
efficiencies).65 Studies indicated that the removal of elemental
mercury was increased to approximately 50 percent with AC injection
ahead of the SDA/ESP and SDA/FF and to approximately 93 percent with
injection of iodide- and sulfur-impregnated AC ahead of two pilot-
scale test systems (a SDA/FF system and the University of North Dakota
Energy and Environmental Research Center (UNDEERC) system consisting
of a boiler and baghouse) . S5'9S Pilot-scale testing at the UNDEERC
system indicated that the percent-removal of elemental mercury with
lignite-based AC was temperature dependent.97
Since mercury speciation affects total mercury removal from
utility flue gas with AC injection and because the speciation of
13-42
-------
mercury is not understood at this time, more data are needed to
establish the factors that affect, and to characterize, mercury
speciation in utility flue gas.
Flue gas composition. Flue gas components such as sulfur oxides,
water, and chlorine compounds can affect the mercury removal
efficiency of carbon. A recent bench-scale study investigated the
effects of S02 and HC1 on the adsorption of elemental mercury and
mercuric chloride (HgCl2) by a lignite-based activated carbon.98
Removing S02 from the flue gas increased the equilibrium adsorption
capacity for elemental mercury by a factor of about 5.5 compared to
3.5 for mercuric chloride. Removing HC1 from the flue gas did not
affect the equilibrium adsorption capacity of the carbon for mercuric
chloride; however, it did prevent the carbon from adsorbing elemental
mercury. With no HC1 in the gas, the carbon adsorption capacity for
mercuric chloride was larger than that for elemental mercury. Other
carbons may not be affected by the presence of HC1 and S02 if the
mercury adsorption mechanism is different.
Research continues on the chemistry and interactions of flue gas,
fly ash, and mercury species. This fundamental research is needed at
various flue gas conditions encountered across the utility industry in
order to develop low cost mercury strategies for full-scale utility
operation. Thus, while AC injection shows promise as a mercury
control technology, more data and research are needed to understand
the factors that affect mercury removal.
13.6.2.5 Emerging Technologies for Controlling Mercury Emissions
from Utilities. Research continues on developing potential
technologies for mercury emission reduction from utility plants. This
research is aimed at either the addition of some type of sorbent
technology or novel technology for mercury control. Emerging
technologies are described below.
13.6.2.5.1 Sorbent technology. Although AC injection has been
shown to be a promising technology, research with impregnated ACs,
sodium sulfide (Na2S) injection, and an AC circulating fluidized bed
suggest that greater mercury removal is possible.
Sulfur-impregnated carbon. In sulfur-impregnated AC injection,
the carbon-bound sulfur reacts with mercury to form mercuric sulfide
(HgS) on the carbon and the carbon is removed by a PM control device.
In a pilot-scale study, sulfur-impregnated carbon increased mercury
removal to 80 percent, an increase of 25 percent over results achieved
with an equal amount of nonimpregnated AC.65
Iodide-impregnated carbon. With iodide-impregnated AC injection,
the carbon-bound iodide reacts with mercury to form mercuric iodide
13-43
-------
(HgI2) on the carbon and the carbon is removed by a PM control device.
In pilot-scale studies, iodide-impregnated carbon achieved
approximately 99 percent mercury removal.65'99
While all testing to date has shown that iodide-impregnated AC
injection has a substantial effect on the mercury removal capability
of AC, further testing has shown that, under certain conditions (with
certain coal types and at temperatures of 177° C [350° F] and higher),
a portion of the captured mercury (postulated to be mercuric iodide)
may be revolatilized as oxidized mercury.100
Chloride-impregnated carbon. Chloride-impregnated AC injection
has been tested only on MWCs in Europe. The chloride reacts with
mercury to form HgCl2 on the carbon and the carbon is removed by a PM
control device. European MWC experiments have shown that impregnating
AC with chloride salts increases the adsorptive capacity of the AC
300-fold.108 Although the amount is small, chloride-impregnated AC
injection would introduce additional chlorine (a HAP) into the flue
gas stream.
Sorbalit. Another potential method of improving mercury
collection efficiency is to combine calcium hydroxide (lime, Ca(OH)2)
with AC. This reagent, consisting of approximately 95 to 97 percent
lime and 3 to 5 percent AC, is known under the product name Sorbalit.93
Sorbalit has only been tested on European MWCs and MWIs.
Sulfur-, iodide-, chloride salt-, and Ca (OH) 2-impregnated ACs
show promise for increasing the mercury removal efficiency, but
further testing is needed. The cost of these modified carbons can be
as much as 20 times higher than that of unmodified AC.101
Other sorbents. Numerous studies are underway to develop other
sorbents as economical alternatives to activated carbon.102'103 Some of
the sorbents under investigation include volcanic pumice, sulfur- and
iodide-impregnated carbons, several proprietary sorbents, high-carbon
fly ash, Darco FGD (an activated carbon derived from lignite), an
activated carbon prepared from a bituminous coal, steam-activated
lignite, thermal-activated bituminous coal, chemical-activated
hardwood, iodine impregnated steam-activated coconut shell, and
sulfur-impregnated steam-activated bituminous coal.104
Sorbent Technologies is marketing a sorbent called Mercsorbent.105
The company claims that the sorbent is effective in removing elemental
mercury at high temperatures typical of utility flue gas, and is
unaffected by common co-existing flue gases, such as S02, HC1, and H20.
Mercsorbent can be used for sorbent injection or it can be used as a
coating on an FF. A bench-scale duct-injection system at Sorbent
Technologies facilities is now being used to test Mersorbent with this
13-44
-------
approach. The company is also scheduled to demonstrate the sorbent at
the refuse incinerator in Fort Dix, New Jersey; prior compliance
sampling at this facility suggests that a significant amount of its
mercury is in the elemental form. A coal-fired boiler or slipstream
is also being sought for a test of the new sorbent material.
Sodium sulfide injection. Mercury reduction has been achieved at
MWCs through injection of Na2S solution into the flue gas prior to the
acid gas control device. The resulting solid, HgS, can be collected
by an FF.10S There are several potential limitations to Na^S injection.
These include reaction of Na2S with calcium in the sorbent (as found in
Sorbalit) to form calcium sulfide (CaS); reduction of the amount of
sulfur available to react with mercury (CaS can also cause scaling of
the sorbent feed line); corrosion of ductwork (Na2S is a corrosive
material); clogging and plugging of the screw conveyor due to
solidification of Na2S; and sludge formation due to the presence of
inorganic salts in the mixing water.107 At present, full-scale
operational injection of Na2S has been done only in MWCs. No plans
have been announced to test this technology on utility units.
Carbon with circulating fluidized beds (CFBs). Another potential
process for the reduction of mercury emissions is the use of AC in a
CFB.63 In a CFB, the AC is continuously fed to the reactor, where it
is mixed with the flue gas at a relatively high velocity, separated in
the subsequent FF, and recycled to the reactor. A small part of the
used AC is withdrawn from the process and replaced by fresh material.108
The main advantages of CFBs over fixed carbon beds are the
increased flue gas-to-carbon contact area and the smaller overall
pressure drop. No pilot or full-scale utility boiler testing has yet
been performed with this system although it has been used in Germany
for MWC operation.
In the United States, Environmental Elements Corporation is
developing a CFB that promotes agglomeration of fine particulate
matter, allowing for its capture in an ESP. In addition, activated
carbon is added to the fluid bed to adsorb mercury vapor. High
residence time, due to the recirculation of the particles, allows for
effective utilization of the carbon. Water sprayed within the
circulating bed further promotes the removal of mercury. Results from
bench-scale testing indicate that mercury was significantly reduced
when passed through the fluidized bed of fly ash and activated carbon.
Based on these tests, a carbon-to-mercury usage was determined for the
system that compares favorably with other sorbent-based mercury
control techniques. There are plans to install a pilot unit and test
at Public Service Electric and Gas's Mercer Station.109
13-45
-------
13.6.2.5.2 Novel technologies. Additional potential processes
for controlling mercury emissions include advanced coal cleaning, a
condensing heat exchanger, gold sorbent technology, other sorbent
injection processes, a corona reactor, and mercury amalgamation.
These technologies are briefly described below.
Condensing heat exchanger. Based on condensing heat exchanger
technology, McDermott Technology (formerly Babcock & Wilcox) is
developing an integrated flue gas treatment system for recovering
waste heat and removing S02, S03, particles, and trace elements from
coal combustion flue gas. The condensing heat exchanger is a two-
pass, counter-flow shell and tube heat exchanger. The hot flue gas
enters the top and flows downward through the first cooling stage,
across a horizontal transition region, then upward through the second
cooling stage. An alkali reagent is sprayed from the top of the second
stage to aid in the removal of S02. Testing of the technology is being
conducted at McDermott Technology's research facility in Alliance,
Ohio. Preliminary results indicate that total mercury removal across
both stages of the condensing heat exchanger is about 62 percent when
firing a blend of Ohio coals. Additional testing is planned on two
other bituminous coals.110
Gold sorbent technology. ADA Technologies has begun development
and testing of a process, called Mercu-RE, for the removal and
recovery of vapor-phase mercury from coal-fired utility boilers. The
process takes mercury from flue gases and produces liquid, elemental
mercury with no secondary wastes. Noble metals are used to adsorb
mercury at typical flue gas temperatures. The mercury is then
thermally desorbed.
Preliminary results from laboratory tests indicate that a gold-
coated monolith captured virtually all of the elemental mercury
injected into a simulated flue gas. Pressure drop through the
monolith was low, which is critical to full-scale use. Further
testing of the gold monoliths will include repeated sorption and
desorption cycles. This phase will be followed by testing on a pilot-
scale coal combustion flue gas at Consol's research facility in
Library, Pennsylvania.11:L
Sorbent injection processes. The Enhanced Limestone Injection
Dry Scrubbing (E-LIDS™) process combines furnace limestone injection
with dry scrubbing to achieve high efficiency S02, particulate, and
trace element emissions control. Dry, pulverized limestone is
injected into the upper furnace region of the boiler. The limestone
is calcined to lime and a portion of the sorbent reacts with S02 in the
flue gas. The flue gas passes through a particulate matter collector
ahead of the dry scrubber to remove some of the solids from the gas
13-46
-------
stream. The solids are mixed with material collected in a baghouse to
produce the S02 scrubbing reagent for the spray dryer.
Application of the E-LIDS™ system, when firing an Ohio bituminous
coal in the Clean Environment Development Facility (CEDF) at the
Alliance Research Center of McDermott Technology, Incorporated, has
shown efficient emissions control performance. Sulfur dioxide
emissions generated from firing the nominal 3-percent sulfur coal were
reduced by more than 99 percent to less than 0.10 Ibs S02/10S Btu.
Total mercury emissions were reduced from an uncontrolled level of
17.6 //g/dscm to less than 0.2 //g/dscm for an average total removal
efficiency of greater than 98 percent. The measured performance
confirmed earlier results obtained in the 5 x 10s Btu/hr small boiler
simulator (SBS) facility. Mercury measurements upstream of the dry
scrubber indicated that both the limestone injection and operation of
the spray dryer/baghouse system at close to the saturation temperature
contributed to the observed total mercury emissions reduction. The
furnace limestone injection alone reduced mercury emissions to an
average of 3.1 //g/dscm.112
Corona reactor. Environmental Elements Corporation is developing
a process for mercury control through DOE's Small Business Innovative
Research program. The first concept utilizes an intense corona
discharge to convert Hg° to mercuric oxide. The process also produces
S03 to serve as a conditioner for high-resistivity fly ash. A corona
discharge in coal combustion flue gas will produce oxidizing radicals,
such as OH and atomic oxygen. Bench-scale results indicate that the
corona reactor, operating at relatively low power levels and short
residence time, yielded high elemental mercury vapor oxidation. The
mercuric oxide, in the form of solid particles, was removed using
conventional particulate control technology. The corona reactor may
also convert mercuric chloride to mercuric oxide, allowing for its
capture as well. The system is currently being tested on a slipstream
at Alabama Power's Plant Miller.113
Mercury amalgamation. There are plans to investigate the
interaction of mercury with metals such as zinc, silver, tin, and
cadmium. Mercury has been shown to amalgamate with certain metals.
Both experimental and modeling efforts are planned to determine the
suitability of metals for the capture of mercury.114
13-47
-------
13.7 REFERENCES
1. Memorandum from Baker, Samuel S., RMB Consulting, to Participants
in the EPRI/UARG Mercury in Coal Study, August 17, 1994. Mercury
in Lignite, Table 3.
2. Baker, Samuel S. EPRI Mercury in Coal Study. A Summary
Report for Utilities That Submitted Samples Update.
Prepared for Utility Air Regulatory Group by Systems
Applications International, June 1994. pp. D-l to D-4.
3. Akers, David, Clifford Raleigh, Glenn Shirey, and Robert
Dospoy. The Effect of Coal Cleaning on Trace Elements,
Draft Report, Application of Algorithms. Prepared for EPRI
by CQ, Inc, Homer City, Pennsylvania. February 11, 1994.
4. Letter from Burke, F. P., CONSOL, Inc., to W. H. Maxwell, EPA.
May 28, 1993. Use of USGS data in estimating the emissions of
air toxics.
5. U.S. Environmental Protection Agency. Assessment of
Physical Coal Cleaning Practices for Sulfur Removal. Final
Report. EPA-600/7-90-013, Research Triangle Park, NC. June
1990.
6. Letter from Burke, F. P., CONSOL Inc., to W. H. Maxwell, EPA.
March 9, 1994. CONSOL Inc.'s round-robin study of the variation
in determinations of mercury concentrations in coal.
7. Lengyel, John, Matthew S. DeVito, and Richard A. Bilonick.
Interlaboratory and Intralaboratory Variability in the Analysis
of Mercury in Coal. CONSOL, Inc., Library, Pennsylvania.
March 9, 1994.
8. Letter from Finkelman, R. B., of U.S. Geological Survey, to W. H.
Maxwell, EPA. January 21, 1994. Comments on the concentration
of mercury in USGS and as-fired coal samples.
9. U.S. Department of Energy, Advanced Power Generation Future
Bright With Coal Gasification-Combined Cycle, Clean Coal Today,
DOE/FE-0215P-5, Issue No. 6. Spring 1992.
10. Baker, D. C. Hazardous Air Pollutants and Other Trace
Constituents in the Syngas From the Shell Coal Gasification
Process. IGTI-Vol. 7, ASME COGEN-TURBO, 6th International
Conference on Gas Turbines in Cogeneration and Utility Industrial
and Independent Power Generation, Houston, TX, as modified for
submission to the U.S. Environmental Protection Agency.
September 1992.
13-48
-------
11. Miller, C. A., R. K. Srivastava, and J. V. Ryan. Emissions of
Organic Hazardous Air Pollutants from the Combustion of
Pulverized Coal in a Small-Scale Combustor. Environmental
Science and Technology. 28:1150-1158. 1994.
12. Nsakala, N., D. Raymond, R. Patel, and M. Cohen. Measurement of
Organic Air Toxics Emissions from Coal Firing in a Laminar-Flow
Reactor. Presented at the Pacific Rim International Conference
on Environmental Control of Combustion Processes, Maui, Hawaii.
October 16-20, 1994.
13. Southern Research Institute. Preliminary draft emissions report
for EPRI Site 110 (baseline and with NOX control) for the EPRI
PISCES Study. SRI report No. SRI-ENV-92-796-7496. October 1993.
14. Radian Corporation. Preliminary draft emissions report for EPRI
Site 114, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. EPRI report No. DCN 92-
213-152-51. May 1994.
15. Radian Corporation Preliminary draft emissions report for EPRI
Site 13, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. EPRI report No. DCN 93-
213-152-36. February 1993.
16. Electric Power Research Institute. Preliminary draft emissions
report for EPRI Site 16 (OFA and OFA/Low NOX) for the Clean Coal
Technology Project (CCT). Prepared for the Department of
Energy/Pittsburgh Energy Technology Center (DOE/PETC), EPRI
report No. DCN 93-209-061-01. November 1993.
17. Battelle. Preliminary draft emissions report for Niles Station
Boiler No. 2 (Ohio Edison) for the Comprehensive Assessment of
Toxic Emissions from Coal-Fired Power Plants. Prepared for the
Department of Energy/Pittsburgh Energy Technology Center
(DOE/PETC). DOE contract # DE-AC22-93PC93251. December 1993.
18. Battelle. Preliminary draft emissions report for Niles Station
Boiler No. 2 with NOx (Ohio Edison) for the Comprehensive
Assessment of Toxic Emissions from Coal-Fired Power Plants.
Prepared for the Department of Energy/Pittsburgh Energy
Technology Center (DOE/PETC). DOE contract # DE-AC22-93PC93251.
December 1993.
19. Radian Corporation. Preliminary draft emissions report for EPRI
Site 102, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. EPRI report No. DCN 92-
213-152-35. February 1993.
13-49
-------
20. Interpoll Laboratories, Inc. Results of the Air Toxic Emission
Study on the No. 8 Boiler at the NSPC Riverside Plant. Prepared
for Northern States Power Company. Report No. 2-3590. September
1992 .
21. Electric Power Research Institute. Preliminary draft emissions
report for Plant Yates Unit No. 1 (Georgia Power Company) for the
Comprehensive Assessment of Toxic Emissions from Coal-Fired Power
Plants. Prepared for the Department of Energy/Pittsburgh Energy
Technology Center (DOE/PETC). EPRI Report No. DCN 93-643-004-03.
December 1993.
22. Southern Research Institute. Preliminary draft emissions report
for Springerville Generating Station Unit No. 2 (Tucson Electric
Power Company) for the Comprehensive Assessment of Toxic
Emissions from Coal-Fired Power Plants. Prepared for the
Department of Energy/Pittsburgh Energy Technology Center
(DOE/PETC). DOE contract # DE-AC22-93PC93254, SRI Report No. SRI-
ENV-93-1049-7960. December 1993.
23. Radian Corporation. Preliminary draft emissions report (and
mercury retest) for EPRI Site 11, Field Chemical Emissions
Monitoring Project. Prepared for Electric Power Research
Institute. EPRI report Nos. DCN 92-213-152-24 and DCN 92-213-
152-48. November 1992/October 1993.
24. Radian Corporation. Preliminary draft emissions report for EPRI
Site 15, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. EPRI report No. DCN 93-
213-152-26. October 1992.
25. Radian Corporation. Preliminary draft emissions report (and
mercury retest) for EPRI Site 12, Field Chemical Emissions
Monitoring Project. Prepared for Electric Power Research
Institute. EPRI report Nos. DCN 92-213-152-27 and DCN 93-213-
152-49. November 1992/October 1993.
26. Energy and Environmental Research Corporation. Preliminary draft
emissions report for Cardinal Station - Unit 1 (American Electric
Power) for the Comprehensive Assessment of Toxic Emissions from
Coal-Fired Power Plants. Prepared for the Department of
Energy/Pittsburgh Energy Technology Center (DOE/PETC), DOE
contract # DE-AC22-93PC93252. December 1993.
27. Roy F. Weston, Inc. Preliminary draft emissions report for
Boswell Energy Center - Unit 2 (Minnesota Power Company) for the
Comprehensive Assessment of Toxic Emissions from Coal-Fired Power
Plants. Prepared for the Department of Energy/Pittsburgh Energy
Technology Center (DOE/PETC), DOE contract # DE-AC22-93PC93255,
Weston project # 10016-011, Weston Report # DOE017G.RP1.
December 1993.
13-50
-------
28. Interpoll Laboratories, Inc. Results of the Air Toxic Emission
Study on the No. 6 & 7 Boilers at the NSPC Riverside Plant.
Prepared for NSPC, Report No. 1-3468A. February 1992.
29. Carnot. Preliminary draft emissions report for EPRI Site 115,
Field Chemical Emissions Monitoring Project. Prepared for
Electric Power Research Institute. Carnot report No. EPRI E-
10106/R022C855.T.
30. Radian Corporation. Preliminary draft emissions report for EPRI
Site 14, Field Chemical Emissions Monitoring Project, Prepared
for Electric Power Research Institute. EPRI report No. DCN 93-
213-152-28. November 1992.
31. Interpoll Laboratories, Inc. Results of the Air Toxic Emission
Study on the No. 3 Boiler at the NSPC Sherburne Plant. Prepared
for Northern States Power Company. Report No. 0-3005. June 1990/
October 1991.
32. Preliminary draft emissions report for EPRI Site 116, Field
Chemical Emissions Monitoring Project, prepared by Radian
Corporation for EPRI. EPRI report No. DCN 94-213-152-55.
October 1994.
33. Radian Corporation. Final emissions report for EPRI Site 111,
Field Chemical Emissions Monitoring Project. Prepared for
Electric Power Research Institute. EPRI report No. TR-105631.
December 1995.
34. Interpoll Laboratories, Inc. Results of the Air Toxic Emission
Study on the No. 1, 3, & 4 Boilers at the NSPC Black Dog Plant.
Prepared for Northern States Power Company. Report No. 1-3451.
January 1992.
35. Interpoll Laboratories, Inc. Results of the Air Toxic Emission
Study on the No. 2 Boiler at the NSPC Black Dog Plant. Prepared
for Northern States Power Company. Report No. 2-3496. May 1992.
36. Interpoll Laboratories, Inc. Results of the Air Toxic Emission
Study on the No. 3, 4, 5 & 6 Boilers at the NSPC High Bridge
Plant. Prepared for Northern States Power Company. Report No.
1-3453. January 1992.
37. Roy F. Weston, Inc. Preliminary draft emissions report for
Baldwin Power Station - Unit 2 (Illinois Power Company) for the
Comprehensive Assessment of Toxic Emissions from Coal-Fired Power
Plants. Prepared for the Department of Energy/Pittsburgh Energy
Technology Center (DOE/PETC). DOE contract # DE-AC22-93PC93255,
Weston project # 10016-011, Weston report # DOE018G.RP1.
December 1993.
13-51
-------
38. Battelle. Preliminary draft emissions report for Coal Creek
Station - Unit 2 (Cooperative Power Association) for the
Comprehensive Assessment of Toxic Emissions from Coal-Fired Power
Plants. Prepared for the Department of Energy/Pittsburgh Energy
Technology Center (DOE/PETC). DOE contract # DE-AC22-93PC93251.
December 1993.
39. Carnot. Preliminary draft emissions report for EPRI Site 112,
Field Chemical Emissions Monitoring Project. Prepared for
Electric Power Research Institute. Carnot report No. EPRI E-
10106/R016C374.T. March 1994.
40. Carnot. Preliminary draft emissions report for EPRI Site 118,
Field Chemical Emissions Monitoring Project. Prepared for
Electric Power Research Institute. Report No. EPRI
E-10106/R140C928.T. January 1994.
41. Radian Corporation. Preliminary draft emissions report for EPRI
Site 10, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. EPRI report No. DCN 92-
213-152-35. October 1992.
42. Environment Canada. Environmental Characterization of Mass
Burning Incinerator Technology at Quebec City, Summary Report.
EPS 3/UP/5. National Incinerator Testing and Evaluation Program,
Ottawa, Canada. June 1988.
43. Goldfarb, T. D. Evidence for Post-Furnace formation of PCDDs and
PCDFs -- Implications for Control. Chemosphere. 18:1051-1055.
1989.
44. Steiglitz, L., G. Zwick, J. Beck, W. Roth, and H. Vogg. On the
De-Novo Synthesis of PCDD/PCDF on Fly Ash of Municipal Waste
Combustors. Chemosphere. Vol. 18, pp. 1219-1226, 1989.
45. Interpoll Laboratories, Inc. Results of the Mercury Removal
Tests on Units No. 1 & 2, and Unit 3 Scrubber System at the NSPC
Sherburne Plant. Prepared for Northern States Power Company,
Report No. 1-3409. October 1991.
46. Interpoll Laboratories, Inc. Results of the May 1, 1990 Trace
Metal Characterization Study on Units No. 1 & 2 at the NSPC
Sherburne Plant. Prepared for Northern States Power Company,
Report No. 0-3033E. July 1990.
47. Memorandum from Cole, Jeffrey, RTI, to William Maxwell, EPA.
February 20, 1998. Update of May 11, 1994 memorandum to Systems
Application International (SAI) about HC1 & HF emission factor
output.
13-52
-------
48. Preliminary draft emissions report for EPRI Site 20, Field
Chemical Emissions Monitoring Project, prepared by Radian
Corporation for EPRI. EPRI report No. DCN 93-213-152-54. March
1994 .
49. Preliminary draft emissions report for EPRI Site 101, Field
Chemical Emissions Monitoring Project, prepared by Radian
Corporation for EPRI. EPRI report No. DCN 94-643-015-02.
October 1994.
50. Draft final report for Paradise Fossil Plant for the
Comprehensive Assessment of Air Toxic Emissions, prepared by
Southern Research Institute for the Department of
Energy/Pittsburgh Energy Technology Center (DOE/PETC), SRI report
No. SRI-ENV-95-338-7960. May 1995.
51. Department of Energy. National Energy Strategy, Powerful Ideas
for America. First Edition. Washington, DC. February 1991.
p. 44.
52. Ref. 51, p. 8.
53. Ref. 51, p. 127.
54. Memorandum from Gibbons, Jack and Rosina Bierbaum, OSTP, to
William H. Maxwell, EPA, May 6, 1996, Interim Report.
55. Ref. 51, p. 39.
56. Kilgroe, J. D. Cleaned Coal (Chapter 19). In: Handbook on Air
Pollution Control, John Wiley & Sons, Inc. April 1983. pp. 38-39.
57. U.S. Environmental Protection Agency. Air Pollution Training
Institute Course 415, Control of Gaseous Emissions (Student
Manual). EPA 450/2-81-005. December 1981. pp. 8-7 through 8-9.
58. Ref. 57, p. 8-23.
59. 58 FR 42466. August 9, 1993. Final Regulatory Determination on
Four Large-Volume Wastes from the Combustion of Coal by Electric
Utility Power Plants.
60. Radian Corporation. Preliminary Draft Report on Field Chemical
Emissions Monitoring Project. Prepared for Electric Power
Research Institute from the following reports: Site 10,
October 6, 1992; Site 11, October 6, 1992; Site 12, November, 23,
1992; Site 15, October 6, 1992; Site 21, May 14, 1993.
13-53
-------
61. Smit, F. J., Gene L. Shields, Mahesh, C. Jha. "Reduction of
Toxic Trace Elements in Coal By Advanced Cleaning." Presented at
the Thirteenth Annual International Pittsburgh Coal Conference,
September 3-7, 1996.
62. "Topical Report No. 5 Trace Element Removal Study." Prepared for
U.S. Department of Energy's Pittsburgh Technology Center by ICF
Kaiser Engineers, March 1995.
63. Clarke, L. B., and L. L. Sloss. Trace Elements-Emissions from
coal combustion and gasification. IEACR/49. IEA Coal Research,
London. July 1992. pp. 29 - 32, 48 - 58.
64. Bloom, Nicolas S., Eric M. Prestbo, and Vesna L. Miklavcic.
Flue Gas Mercury Emissions and Speciation from Fossil Fuel
Combustion. Second International Conference on Managing
Hazardous Air Pollutants, Washington, DC. July 1993.
65. Felsvang, Karsten, Rick Gleiser, Gary Juip, and Kirsten Kragh
Nielsen. Air Toxics Control by Spray Dryer Absorption Systems.
Second International Conference on Managing Hazardous Air
Pollutants, Washington, DC. July 1993.
66. Noblett, Jr., J. G., F. B. Meserole, D. M. Seeger, and D. R.
Owens. Control of Air Toxics from Coal-fired Power Plants using
FGD Technology. Second International Conference on Managing
Hazardous Air Pollutants, Washington, DC. July 1993.
67. Letter from Chang, Ramsay, EPRI, to Martha H. Keating, EPA.
February 7, 1994. Comments on the draft report Mercury Control
Technologies and Costing of Activated Carbon Injection for the
Electric Utility Industry, prepared by RTI, September 1993.
68. Letter from Boyce, P. L., Northern States Power Company, to
Martha Keating, EPA. January 19, 1994. Comments on the draft
report Mercury Control Technologies and Costing of Activated
Carbon Injection for the Electric Utility Industry, prepared by
RTI, September 1993.
69. Chang, R, and D. Owens, 1994. "Developing Mercury Removal
Methods for Power Plants." EPRI Journal, July/August, 1994.
70. DeVito, Matthew S., Prasad R. Tumati, Rachel J. Carlson, and
Nicolas Bloom. Sampling and Analysis of Mercury in Combustion
Flue Gas. Second International Conference on Managing Hazardous
Air Pollutants, Washington, DC. July 1993.
71. Chow, W., et. al., 1994. "Pathways of Trace Elements in Power
Plants: Interim Research Results and Implications." Trace
Element Transformations in Coal-Fired Power Systems, Fuel
Processing Technology, August, 1994, pp. 5-20.
13-54
-------
72. Letter from Kevin E. Redinger of Babcock & Wilcox to William
Maxwell, Emission Standards Division, U.S. EPA. February 7,
1996.
73. Redinger, K. E., A. P. Evans, R. T. Bailey and P. S. Nolan, 1997.
"Mercury Emissions Control in FGD Systems," Presented at the
EPRI-DOE-EPA Combined Air Pollutant Control Symposium,
Washington, DC. August 25-29, 1997.
74. Chang, R., and D. Owens, 1994. "Developing Mercury Removal
Methods for Power Plants." EPRI Journal, July/August, 1994.
75. Chang, Ramsay, C. Jean Bustard, Gordon Schott, Terry Hunt, Howard
Noble, and John Cooper. Pilot Scale Evaluation of AC for the
Removal of Mercury at Coal-fired Utility Power Plants. Second
EPRI International Conference on Managing Hazardous Air
Pollutants, Washington, DC. July 1993.
76. Feeley, T. J., Ill, "An Overview of the U.S. Department of
Energy's Electric-Utility Mercury Emissions R&D Activities," Acid
Rain & Electric Utilities II Conference, Scottsdale, AZ,
January 21-22, 1997. p. 4.
77. Ref. 76
78. Electric Utility Trace Substances Synthesis Report - Volume 3:
Appendix 0, Mercury in the Environment." EPRI TR-104614-V3,
Project 3081,3297, November, 1994.
79. Ref. 73
80. Ref. 73
81. Ref. 73
82. Hargrove, 0. W., 1994. "A Study of Toxic Emissions from a Coal-
Fired Power Plant Demonstrating The ICCT CT-121 FGD Project."
Tenth Annual Coal Preparation, Utilization, and Environmental
Control Contractors Conference, Pittsburgh, PA, July, 1994,
pp. 267-274.
83. Ref. 73
84. Ref. 76, p. 3.
85. Ref. 76, p. 3.
86. Report, Municipal Waste Combustors-Background Information for
Proposed Standards: Post-Combustion Technology Performance,
EPA:ISB, September 22, 1989.
13-55
-------
87. Report, Emission Test Report OMSS (Ogden Martin Systems of
Stanislaus, Incorporated) Fields Test on Carbon Injection for
Mercury Control, EPA:OAQPS, Publication No. EPA-600/R-92-192,
September 1992.
88. Report, Emission Test Report Field Test of Carbon Injection for
Mercury Control Camden County Municipal Waste Combustor,
EPA:OAQPS, Publication No. EPA-600/R-93-181, September 1993.
89. Report, Medical waste incinerators - background information for
proposed standards and guidelines: Control technology
performance report for new and existing facilities, EPA:OAQPS,
Publication No. EPA-453/R-94-044a, pp. 98 to 99 and B-7 to B-8,
July 1994.
90. Report, A Comprehensive Assessment of Toxic Emissions from Coal-
Fired Power Plants: Phase I Results from the U.S. Department of
Energy Study, DOE:FETC Contract No. DE-FC21-93MC30097 (Subtask
2.3.3). pg. A-14. September 1996.
91. Temperature and flow rate data taken or determined from
Preliminary Draft Report on Field Chemical Emissions Monitoring
Project: Emissions Report for Sites 103 - 109, prepared for
Electric Power Research Institute by Radian Corporation, March 3,
1993. Mercury content taken from Preliminary Draft Report on
Field Chemical Emissions Monitoring Project, prepared for
Electric Power Research Institute by Radian Corporation, from the
following reports: Site 13, February 12, 1993; Site 112,
December 30, 1993; Site 117, January 20, 1994; Site 118,
January 20, 1994.
92. Chloride content data determined from Electric Power Research
Institute (EPRI) test site 13, in the Preliminary Draft Report on
Field Chemical Emissions Monitoring Project: Site 13 Emissions
Report, prepared for Electric Power Research Institute by Radian
Corporation, February 12, 1993.
93. Mercury content data taken from Nebel, K. L. and D. M. White. A
Summary of Mercury Emissions and Applicable Control Technologies
for Municipal Waste Combustors. Radian Corporation. Research
Triangle Park, NC. June 1991. p. 2-1.
94. Temperature, chloride content, and flow rate data taken or
determined from Brown, B. and K. S. Felsvang, "Control of Mercury
and Dioxin Emissions from United States and European Municipal
Solid Waste Incinerators by Spray Dryer Absorption Systems,"
proceedings of the ASME/EPRI/AWMA 5th Integrated Environmental
Control for Power Plants Conference, Figure 2.
95. McKenna, John D., and James H. Turner. Fabric - Filter -
Baghouses I, Theory, Design, and Selection (A Reference Text).
ETS, Inc., Roanoke, VA. 1989. p. 6-9.
13-56
-------
96. Miller S. J., Dennis L. Laudal, Ramsay Chang, and Perry D.
Bergman. "Laboratory-Scale Investigation of Sorbents for Mercury
Control." Presented at the 87th Annual Meeting of the Air &
Waste Management Association, Cincinnati, OH, June 19-24, 1994.
97. Ref. 96
98. Carey, T. R., Oliver W. Hargrove Jr., Carl F. Richardson, Ramsay
Chang, Frank B. Meserole. "Factors Affecting Mercury Control in
Utility Flue Gas Using Sorbent Injection." Presented at the Air
& Waste Management Association's 90th Annual Meeting & Exhibition,
June 8-13, 1997, Toronto, Ontario, Canada.
99. University of North Dakota Energy and Environment Research Center
(UNDEERC), "Effective Sorbents for Trace Metals," A study
performed from March 1994 to March 1995. Abstract taken from the
UNDEERC internet web page.
100. Ref. 96, p. 3.
101. Memo from Maxwell, William H., EPA, to Kenneth R. Durkee, EPA.
August 3, 1993. Second International Conference on Managing
Hazardous Air Pollutants.
102. Ref. 76
103. Ref. 99
104. Ref. 99
105. Nelson Jr., S., Jon Miller, and Deborah Summanen. "Innovative
Mercury Emission Control.," Presented at the Air & Waster
Management Associations's 90th Annual Meeting & Exhibition,
June 8-13, 1997, Toronto, Ontario, Canada.
106. Ref. 93, p. 3-1. (equations given in this reference were balanced)
107. Ref. 93, pp. 3-6 and 3-7.
108. New Jersey Department of Environmental Protection and Energy.
Task Force on Mercury Emissions Standard Setting Preliminary
Report. Volume III, Technical and Regulatory Issues. July 1993.
p. 6.9.
109. Ref. 76, p. 5.
110. Ref. 76, p. 5.
111. Ref. 76, p. 5.
112. Ref. 73
13-57
-------
113. Ref. 76, p. 5.
114. Ref. 76, p. 7.
13-58
-------
14.0 SUMMARY OF RESULTS, TECHNICAL FINDINGS, AND RESEARCH NEEDS
The following summary of results, technical findings, and
research needs is based on this study and the currently available
scientific data.
14.1 INDUSTRY GROWTH AND HAP EMISSIONS
1. Utility units emit a significant number of the 189 HAPs
included on the section 112(b) list, although in most cases
they are responsible for very small percentages of total
anthropogenic emissions. Coal-fired units emit the largest
number of utility-originated HAPs. Coal-, oil-, and gas-
fired utilities emit a mix of HAPs, including organics
(e.g., polycyclic aromatic hydrocarbons, dioxins) and heavy
metals (e.g., arsenic, lead, cadmium, chromium, nickel,
mercury). Utilities are responsible for approximately 34
percent of the United States anthropogenic airborne
emissions of mercury but no more than 4 percent of other
measured HAPs.
2. Under the assumptions made in this Report, HAP emissions are
predicted to increase during the period 1990-2010. Over
this period, utility coal consumption is estimated to
increase by approximately 29 percent, oil consumption is
estimated to decrease by approximately 48 percent, and
natural gas consumption is expected to increase by about 61
percent. Coal combustion accounts for the increase in HAP
emissions.
3. Actions anticipated to be taken by the utility industry to
comply with other provisions of the Act (e.g., acid rain,
NAAQS revisions) or with other initiatives (e.g.,
electricity restructuring, global warming abatement) may
impact HAP emissions.
4 . The emission test program provided valid and representative
results for the purpose of this study. The utility industry
is composed of a wide variety of units employing a mix of
fuel types, burner types, boiler types, and control device
configurations. The HAP emission test data were obtained by
a variety of organizations using common test and analytical
protocols. The EPA helped to establish these protocols.
14.2 INHALATION RISK ASSESSMENT
5. For the vast majority of the more than 196 million persons
living within 50 km of any utility unit, the lifetime cancer
risk due to inhalation exposure to HAP emissions is likely
to be less than 1 x 10~s.
14-1
-------
6. Out of 426 coal-fired plants, EPA's modeling indicates that
only 2 coal-fired plants pose high-end excess lifetime
cancer risks greater than 1 x I0~s due to inhalation
exposure to HAP emissions. For no plant does the inhalation
MIR exceed 4 x 10~6. More than 50 percent of the inhalation
cancer risk is attributable to arsenic. The average
inhalation MIR across all plants is estimated to be roughly
l/10th to 1/100 as large as the highest inhalation MIR.
Central tendency inhalation risks for all exposed
individuals are predicted to be approximately one to three
orders of magnitude lower than the highest inhalation MIR.
The population cancer incidence due to inhalation exposure
to coal-fired utility HAP emissions, considering the results
of both the local and long-range transport modeling, is
estimated to be no greater than 1.3 cases per year
nationwide.
7. For the year 2010, the cancer MIRs from coal-fired utilities
are not expected to be significantly different. However,
due to uncertainties about future demand, industry
operations, government regulation, etc., the EPA has low
confidence in this projection.
8. Out of 137 oil-fired plants, EPA's modeling indicates that
only 11 oil-fired plants pose high-end excess lifetime
cancer risks greater than 1 x I0~s due to inhalation
exposure to HAP emissions. For no plant does the inhalation
MIR exceed 6 x 10"5. More than 50 percent of the inhalation
cancer risk is attributable to nickel and the assumption
that nickel emissions are 50 percent as carcinogenic as
nickel subsulfide. The average inhalation MIR across all
plants is estimated to be roughly l/10th to l/100th as large
as the highest inhalation MIR. Central tendency inhalation
risks for all exposed individuals are predicted to be
roughly two to three orders of magnitude lower than the
highest inhalation MIR. Population cancer incidence due to
inhalation exposure to oil-fired utility HAP emissions
(considering local and long-range transport) is estimated to
be no greater than 0.5 cases per year nationwide.
9. For the year 2010, the cancer MIRs from oil-fired utilities
are predicted to be approximately 30 to 50 percent lower
than the 1990 estimates. The EPA has good confidence in
this projection because of the well-established trend of
declining oil use by utilities.
10 . Based on the quantitative parameter uncertainty analysis
conducted for the inhalation risk assessment, the EPA
estimates that the high-end inhalation risk estimates
presented in this report are conservative (i.e., more likely
to be overestimating rather than underestimating the true
MIR). The quantitative variability and uncertainty of many
14-2
-------
of the input parameters such as emission estimates, stack
height, breathing rates, and exposure duration were
considered in the uncertainty analysis. This resulted in an
estimated range and distribution of potential risks due to
inhalation exposure only. The median ("central tendency
estimate") inhalation MIR estimates are predicted to be
roughly 2 to 10 times lower than the high-end MIR estimates.
However, there are limitations to the uncertainty analysis
and limitations in available data and the range of
uncertainty is, most likely, larger than estimated by this
study.
11. The risk estimates contain additional uncertainties that are
not represented in the quantitative uncertainty analysis.
The impact of using different high-to-low dose extrapolation
models was not quantitatively assessed in this study. In
addition, there are other factors, such as variation in
population sensitivity (e.g., sensitive subpopulations),
residence time, activity patterns, and other uncertainties
and variabilities, that were not quantitatively assessed.
12. With regard to noncancer effects, the highest hazard
quotient for any HAP, considering both short- and long-term
exposures, is approximately 0.1 for HC1 from coal-fired
utilities. That is, the highest exposure concentration for
HC1 was estimated to be 10 times lower than the Inhalation
Reference Concentration (RfC) for HC1. The highest hazard
index for all HAPs is about 0.2.
14.3 MERCURY
13. Mercury cycles in the environment as a result of natural and
human (anthropogenic) activities. Most of the mercury in
the atmosphere is elemental mercury vapor, which circulates
in the atmosphere for up to a year, and hence can be widely
dispersed and transported thousands of miles from sources of
emissions. Even after it deposits, mercury commonly is
emitted back to the atmosphere to be re-deposited elsewhere.
The modeling of long-range transport of mercury suggests
that about one-third of United States utility emissions is
deposited within the lower 48 States. The remaining two-
thirds are transported outside of United States borders
where it diffuses into the global reservoir. Finally,
predictions of the ISC3 and RELMAP models indicate that most
of the mercury emitted by utilities is transported further
than 50 km from the emission source.
14. The analysis of mercury fate and transport completed for
this study, as well as the analyses completed in the EPA's
1997 Mercury Study Report to Congress, in conjunction with
available scientific knowledge, supports a plausible link
between mercury emissions from anthropogenic combustion and
14-3
-------
industrial sources and mercury concentrations in air, soil,
water and sediments. The critical variables contributing to
this linkage are these:
• the species of mercury that are emitted from the
sources, with Hg° mostly contributing to
concentrations in ambient air and Hg2+ mostly
contributing to concentrations in soil, water, and
sediments;
• the overall amount of mercury emitted from a
combustion source;
• the watershed soil loss rates, including reduction and
erosion;
• the water body loss rates, including outflow,
reduction, and settling; and
• the climate conditions.
In addition, the analysis of mercury fate and transport
supports a plausible link between mercury emissions from
anthropogenic combustion and industrial sources and
methylmercury concentrations in freshwater fish. However,
these fish methylmercury concentrations also result from
existing background concentrations of mercury (which may
consist of mercury from natural sources, as well as mercury
which has been re-emitted from the oceans or soils) and
deposition from the global reservoir (which includes mercury
emitted by other countries). Given the current scientific
understanding of the environmental fate and transport of
this element, it is not possible to quantify how much of the
methylmercury in fish consumed by the United States
population is contributed by current United States emissions
relative to other sources of mercury (such as natural
sources and re-emissions from the global pool). The
critical variables contributing to the plausible link
include:
• the species of mercury that are emitted, with emitted
divalent mercury mostly depositing into local
watershed areas and, to a lesser extent the
atmospheric conversion of elemental mercury to
divalent species which are deposited over greater
distances;
• the overall amount of mercury emitted from a source;
• the watershed soil loss rates, including reduction and
erosion;
14-4
-------
• the water body loss rates, including outflow,
reduction, and settling;
• the extent of mercury methylation in the water body;
• the extent of food web bioaccumulation in the water
body; and
• the climate conditions.
From the analysis of deposition and on a comparative basis,
the deposition of Hg2+ close to an emission source is
greater for receptors in elevated terrain (i.e., terrain
above the elevation of the stack base) than from receptors
located in flat terrain (i.e., terrain below the elevation
of the stack base). The critical variables are parameters
that influence the plume height, primarily the stack height
and stack exit gas velocity.
14.4 DIOXINS AND ARSENIC
15. Based on a screening level multipathway analysis, the
highest estimated individual risks due to utility arsenic
emissions are predicted to be no higher than 1 x 10"4 and
are probably much lower for the vast majority of the exposed
population. The increased cancer risks due to multipathway
exposures to arsenic emissions, based on screening level
model-plant analysis, using hypothetical scenarios, were
predicted to range from less than 4 x 10 "7 up to 1 x 10 "4.
The highest predicted risk (i.e., 1 x 10 ~4) was for a
hypothetical scenario pica child assumed to be living at the
point of maximum deposition, which is considered an upper
bound, conservative scenario. When the risk from background
exposure (2 x 10"4) is added to the maximum risk from
utility exposure, the risk for the pica child is estimated
to be up to 3 x 10"4. Background exposures were estimated
to dominate the exposures and risk. There are substantial
uncertainties associated with this screening level arsenic
multipathway assessment, and the results do not apply to any
specific existing utility plant. Further assessment is
needed to more fully evaluate the risks due to arsenic
emissions from utilities.
16. Based on a screening level multipathway assessment for
dioxins, total modeled lifetime cancer risks related to
indirect exposure to dioxins, based on model-plant analyses,
are predicted to range from 1 x 10 "10 to 2 x 10 "4. The
results of the analyses indicate that the exposures and
risks due to fish consumption are the highest of all
pathways considered. In all scenarios, the noninhalation
(e.g., ingestion) exposures were predicted to be at least
one order of magnitude larger than the inhalation exposures
14-5
-------
and modeled exposures exceed the background exposures for
dioxins. There are substantial uncertainties associated
with this dioxin screening level multipathway assessment and
the results do not apply to any specific existing utility
plant. Further assessment is needed to more fully evaluate
the risks due to dioxin emissions from utilities.
14.5 RADIONUCLIDE ANALYSIS
17. For the vast majority of the more than 196 million persons
living within 50 km of any utility unit, the lifetime fatal
cancer risk due to radionuclide emissions is less than
1 x IP"6.
18. The highest MIR to any individual, within a 50-km radius,
resulting from multipathway exposure to radionuclide
emissions from utility units is estimated to be up to
3 x IP"5, and 17 of the 684 plants were estimated to
potentially pose an MIR greater than 1 x 10 "5.
19. Based on the multipathway exposure modeling conducted with
the CAP-93 model, which estimates exposure within 50 km of
each utility unit, the estimated deaths/year resulting from
multipathway exposure to radionuclide emissions from utility
units is approximately 0.3/year.
20. The quantitative uncertainty analysis performed for the
radionuclide analysis indicates that the population risk
estimates are central values of the true probability
distribution.
14.6 ALTERNATIVE CONTROL STRATEGIES
21. There are a number of alternative control strategies that
are effective, based on the data obtained for this report,
in controlling some of the HAPs emitted from utility units.
These strategies are summarized below.
• Conversion of coal- and oil-fired units to natural gas
firing effectively eliminates emission of HAPs.
• Conversion of coal-fired units to oil combustion will
effect decreases in emissions of some HAPs but could
lead to increased emissions of others (e.g., nickel).
• Because of the wide variability in the trace metal
contents of coals, switching from one coal to another
will not generally result in consistently reduced
overall HAP emissions.
• Current methods of coal cleaning are able to remove
portions of the trace metals contained within the
14-6
-------
coal. These emission reductions range from
approximately 20 percent for mercury to approximately
50 percent for lead. Advanced coal cleaning
technologies show promise in reducing mercury from
coal from approximately 30 to greater than 60 percent.
Further research is needed in methods of effecting
greater trace metal removals during coal cleaning and
in assessing the various impacts of these methods.
Newer forms of fuel combustion (e.g., coal
gasification) show promise as being cleaner sources of
electricity but available data are limited and further
research is needed.
The impact of combustion controls on HAP emissions is
inconclusive given the current level of knowledge.
While available data appear to indicate that
installation of low-NOx burners results in a trend
toward lower HAP emissions, the trend is neither
universal nor uniform.
Particulate matter control devices (i.e., FFs, ESPs)
generally effect good control (i.e., greater than 90
percent removal) of the trace metallic HAPs, with the
exception of mercury. Research is underway to further
enhance fine particle removal, including trace
metallic HAPs, from these control devices. Organic
HAPs do not appear to be well controlled by PM control
devices but these compounds are generally found near
the detection limit. Fabric filters appear to
moderately control HC1 (i.e., 50 percent removal) but
not HF, and ESPs do not generally effect good acid gas
HAP control.
Wet acid gas control devices (e.g., FGDs) by
themselves do not uniformly effect good control of the
trace metallic HAPs, including mercury. Based on
limited data, SDA/FF combinations, however, appear to
be as effective as do FFs alone. Research is underway
to further enhance the mercury removal capability of
FGD systems. Flue gas desulfurization units (as
operated on utility units) generally effect good
control (i.e., greater than 80 percent) of HC1, but
control of HF is not uniform.
Add-on technologies for the control of mercury have
not been demonstrated on utility units in the United
States. Pilot-scale work on activated carbon
injection indicates that mercury removal is possible
but that such removal is inconsistent and variable and
that further research is needed.
14-7
-------
• Pollution prevention methods (i.e., DSM, energy
conservation, repowering for energy efficiency) have
the potential to result in reduced HAP emissions. The
extent that these methods will be utilized by the
industry is not known and, thus, the extent of any
emission reductions cannot be forecast at this time.
Each of these alternative control strategies may have significant
cost, economic, technical, and research implications before they can
be widely utilized in the utility industry.
14.7 AREAS FOR FURTHER RESEARCH AND ANALYSIS
There are numerous uncertainties and data gaps described
throughout this report. This section identifies several of the
important areas in which further research or scientific and technical
work is needed.
14.7.1 Emissions Data for Dioxins
Emissions data for dioxins were available from only eight of the
tested utility plants. Therefore, there are greater uncertainties
with the dioxin emissions than for many of the other HAPs. All types
of utility units were not tested (e.g., there are no data available
from coal-fired units with hot-side ESPs).
14.7.2 Speciation of Nickel and Chromium
There are significant uncertainties regarding the forms of nickel
being emitted from oil-fired utilities and the health effects
associated with those different forms. Therefore, further research
and evaluation of the emissions to determine what forms are being
emitted and the health effects associated with those different forms
would be of value. Further evaluation of chromium speciation is also
needed.
14.7.3 Multipathway Risk Assessment
As mentioned previously, further work is needed to study the
risks due to multipathway exposure to HAPs that are persistent and
bioaccumulate. Arsenic and dioxins are two HAPs identified as
priority for further multipathway assessment.
14.7.4 Long-range Transport Exposures
Uncertainties remain regarding long-range transport of HAPs.
Further modeling and evaluation could be helpful to assess the impacts
of long-range transport of HAPs from utilities.
14.7.5 Mercury Issues
There are numerous areas regarding mercury that may need further
research, study, or evaluation. A number of potential areas for
further study include the following:
• review the estimates of the levels of exposure to mercury
associated with subtle neurological endpoints;
14-8
-------
• quantify and/or evaluate the relationship between a change
in United States mercury emissions and the resulting change
in methylmercury levels in fish;
• evaluate actual consumption patterns and estimates of the
methylmercury exposure of the subpopulations of concern;
• gather additional data on the mercury content of various
types of coal;
• establish improved methods for measuring mercury
concentrations in water;
• study the occupational, dietary, and behavioral factors that
affect mercury exposures for people who are determined to be
exposed above a threshold of concern;
• study the public health and environmental benefits that
would be expected by reducing mercury emissions from
utilities;
• evaluate and/or research control technologies or pollution
prevention options that are available, or will be available,
that could potentially reduce mercury emissions and what are
the costs, economic impacts, and feasibility of those
options;
• evaluate how other regulations, programs, and activities
(e.g., acid rain program, electricity restructuring, NAAQS,
climate change) affect mercury emissions;
• gather additional data on mercury emissions (e.g., how much
is emitted from various types of utility units, how much is
divalent vs elemental mercury, and how do factors such as
control device, fuel type, and plant configuration affect
emissions and speciation); and
• study how much mercury is emitted from natural sources and
past anthropogenic sources.
14.7.6 Projections to the Year 2010
There are significant uncertainties and unknowns in the emissions
and risk projections made to the year 2010 (e.g., impact of industry
restructuring; impact of State efforts to regulate restructuring;
impact of any climate change initiatives). Further research and
evaluation in this area is needed.
14.7.7 Ecological Risks
The effects of HAPs on wildlife, endangered species, and
terrestrial and aquatic ecosystems were not evaluated in this study.
Although not mandated by section 112(n)(1)(A), further evaluation of
14-9
-------
ecological risks due to HAP emissions would be needed to fully
evaluate the impacts of utility HAP emissions.
14.7.8 Criteria Pollutant and Acid Rain Programs
Further evaluation is needed to assess the impacts of the Acid
Rain and Criteria Pollutant programs (e.g., impact of revisions to the
PM-fine and ozone NAAQS; impact of Ozone Transport Assessment Group
[OTAG] activities) on HAP emissions.
14.7.9 Short-term Emissions
A limited assessment of short-term exposures was completed.
However, further evaluation of short-term releases, especially high-
end, peak releases, could be useful to fully assess the potential
impacts to public health due to emissions of HAPs (particularly HC1
and HF) from utilities.
14-10
-------
-------
TECHNICAL REPORT DATA
Please read Instructions on reverse before completing)
1. RE PORT NO.
EPA-453/R-98-004a,-b
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Study of Hazardous Air Pollutant Emissions from Electric
Utility Steam Generating Units - Final Report to Congress
5. REPORT DATE
February 1998
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
Emission Standards Division/Air Quality Strategies and
Standards Division
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report has been prepared pursuant to section 112(n)(1)(A) of the Clean Air Act, and provides the
Congress and the public with information regarding the emissions, fate, and transport of utility HAPs. The
primary components of this report are: (1) a description of the industry; (2) an analysis of emissions data;
(3) an assessment of hazards and risks due to inhalation exposures to 67 HAPs; (4) assessments of risks
due to multipathway (inhalation plus non-inhalation) exposures to four HAPs (radionuclides, mercury,
arsenic, and dioxins); and (5) a discussion of alternative control strategies. The assessment for mercury in
this report includes a description of emissions, deposition estimates, control technologies, and a dispersion
and fate modeling assessment which includes predicted levels of mercury in various media (including soil,
water, and freshwater fish) based on modeling from four representative utility plants using hypothetical
scenarios. The EPA has not evaluated human or wildlife exposures to mercury emissions from utilities in
this report. With regard to non-inhalation exposures (e.g., ingestion) to other HAPs, this report presents a
limited qualitative discussion of arsenic, cadmium, dioxins, and lead.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Atmospheric Dispersion Modeling
Electric Utility Steam Generating Units
Hazardous Air Pollutants/Air Toxics
18. DISTRIBUTION STATEMENT
Release Unlimited
b. IDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
19. SECURITY CLASS (Report)
Unclassified
20. SECURITY CLASS (Page)
Unclassified
c. COSATI Field/Group
21. 787
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
------- |