United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-453/R-96-013a
October 1996
Air
EPA Study °f Hazardous Air Pollutant
Emissions from Electric Utility Steam
Generating Units -- Interim Final Report
Volume 1.
-------
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Flow
Chicago, IL 60604-3590
-------
Table of Contents
VOLUME I
Section
GLOSSARY
1.1
1.2
1.3
1.4
2.0 CHARAI
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0 EKCSS:
3.1
3.2
3.3
3.4
LEGISLATIVE MANDATE
ACT PROVISIONS ANT) STUDIES RELATED TO THIS fTTTPV ,.,,--.
1.2.1 Section 112 of the Act - Hazardous Air Pollutants
1.2.2 Title I - Nonattainment Provisions
1.2.3 Title IV - Acid Deposition Control
1.2.4 New Source Performance Standards (NSPS)
OVERVIEW AND APPROAfH OF KLF.TPTC UTILITY HAP STRIPY ,......-.
RKFERENOF-P ,....,-..,---,,
*^HR!ZATZQB? OF THE INDUSTRY . ....... ™ . .
INDUSTRY BACKGROUND
FOSRTIi-FUETi-FTREP KLETPTC1 tprTT'TTY .TFAM-fiFKnspATTNn TJNTTS .
2.2.1 Types of Electric Utility Facilities
2.2.2 Types of Ownership
DESIGN np Ej.RrfRTr UTTLTTY TTNTTS ,,..... -----
2.3.1 Furnace Types
2.3.2 Bottom Types
2 3.3 Cogeneration ....
2.3.4 Combined-Cycle System
PARTICULATE 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 SO, Control . . .
NOX CONTROL
2.6.1 Combustion Control
2.6.2 Postcombustion Control
TITT'M'T'Y rNW-9TRY AFTER TMPT..KMFNTATTON r>p 1 99° AMT!WPMENT.'S ........
2.7.1 Industry Growth
2.7.2 Title I and Title IV, Phase I and Phase II, Compliance Strategy
Impact ,
DTSPnsSTON OF FEPKRAT, TfH'ERAPENtT FKVTKW rnMMFNTP ..-..- .
p (•>' KK^Mr'F.s . ...... ..-,,- ,.-,-.,.
ION DATA QATHERXNS AND ANALYSIS .
LITERATURE REVIEW AND BACKGROUND ,
POLLUTANTS STUDIED
DESCRIPTION OF EMISSION TEST PROGRAMS ,
DEVELOPMENT OF HAP EMISSION TOTALS ,
3.4.1 Program Operation
3.4.2 Data Sources
3.4.3 Operational Status of Boilers .
3.4.4 Trace Element Concentration in Fuel
3.4.5 HC1 and HF Concentration in Fuel
3.4.6 Emission Modification Factors for Inorganic HAPs
3.4.7 Organic and Mineral Acid HAPs ,
Page
xiii
XV
ES-1
1-1
1-1
1-2
1-2
1-4
1-4
1-5
1-6
1-7
2-1
2-1
2-1
2-3
2-3
2-5
2-5
2-7
2-9
2-9
2-9
2-11
2-12
2-12
2-13
2-14
2-14
2-17
, 2-17
2-22
2-24
2-25
2-27
2-28
2-31
2-33
2-36
3-1
3-1
3-3
3-4
3-4
3-4
3-8
3-8
3-9
3-11
. 3-11
. 3-13
-------
Table of Contents (continued)
VOLUME I
Section Page
3.4.8 Model Estimates for the Year 2010 3-14
3.5 SELECTED ESTIMATED NATIONWIDE HAP EMISSIONS 3-14
3.6 COMPARISON OF EFP ESTIMATES WITH TEST T5ATA 3-14
3.7 CHARACTERISTIC PLANT EMISSIONS 3-17
3.8 UNCERTAINTY ANALYSIS OF THE EMISSION FACTOR PROGRAM 3-20
3.9 REFERENCES 3-21
4.0 INTRODUCTION FOR THE HEALTH HAZARD RISK ANALYSIS 4-1
4.1 INTRODUCTION AND BACKGROUND 4-1
4.1.1 Principals of Risk Assessment 4-1
4.1.2 U.S. EPA Risk Assessment Guidelines 4-3
4.1.3 Risk Assessment Council (RAO 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 Doseresponse 4-6
4.3.3 Dose-Response Evaluation for Carcinogens 4-7
4.3.4 Long-term Noncancer Health Effects Data 4-9
4.3.5 Short-term Noncancer Health Effects Data 4-10
4.3.6 Summary of Health Effects Data Sources 4-10
4.4 METHODOLOGY FOR ESTIMATING INHALATION EXPOSURE FOR THE LOCAL ANALYSIS . 4-10
4.4.1 Emissions Characterization 4-11
4.4.2 Atmospheric Fate and Transport 4-11
4.4.3 Characterization of Study Population 4-12
4.4.4 Exposure Calculations 4-12
4.5 METHODOLOGY FOR ESTIMATING QUANTITATIVE INHALATION RISKS 4-13
4.5.1 Estimating Cancer Inhalation Risks 4-13
4.5.2 Individual Risk 4-13
4.5.3 Population Cancer Risk 4-14
4.5.4 Distribution of Individual Risk within a Population 4-14
4.5.5 Aggregate Inhalation Cancer Risk 4-15
4.5.6 Estimating Noncancer Inhalation Risks 4-15
4.5.7 Inhalation Hazard Quotient (HQ) 4-15
4.5.8 Total Risk for Noncancer Effects 4-16
4.5.9 Direct Inhalation Exposure and Risk Default Options 4-16
4.6 REFERENCES 4-18
5.0 SCREENING RISK 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 5-11
5.5.1 Overview 5-11
5.5.2 Prioritization of HAPs for Multipathway Exposure Assessment . . 5-12
5.6 SELECTION OF HAPS FOR FURTHER ANALYSIS 5-16
5.7 LIMITATIONS OF SCREENING ASSESSMENT 5-17
5.8 REFERENCES 5-19
6.0 INHALATION RISK ASSESSMENT 6-1
6.1 BASELINE ASSESSMENT OF INHALATION EXPOSURES AND RISKS FOR 14 PRIORITY
ii
-------
VOLUME I
Table of Contents (continued)
Section
Page
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 VERSUS RURAL LOCATIONS 6-14
6.3 INHALATION RISK ESTIMATES FROM THE YEAR 2010 6-15
6.4 ASSESSMENT OF RISKS DUE TO SHORT-TERM EXPOSURE 6-17
6.4.1 Methodology 6-17
6.4.2 Results 6-19
6.5 OVERLAPPING PLUMES/DOUBLE COUNTING 6-21
6.6 ASSESSMENT OF EXPOSURE DUE TO LONG-RANGE TRANSPORT 6-22
6.6.1 History and Background Information 6-22
6.6.2 pKTjfl&p Modeling Strategy for Atmospheric Arsenic 6-22
6.6.3 Model Parameterizations 6-25
6.6.4 Exposure and Risk Estimates 6-26
6.7 DISCUSSION OF BACKGROUND EXPOSURES 6-31
6.7.1 Arsenic 6-31
6.7.2 Chromium, Nickel, Manganese, and HC1 . . '. 6-32
6.8 CHROMIUM SPECIATION UNCERTAINTY AND IMPACT ON RISK ESTIMATES 6-33
6.9 ISSUES WITH THE ARSENIC CANCER UNIT RISK ESTIMATE AND IMPACT ON THE
INHALATION RISK ESTIMATES 6-33
6.10 NICKEL SPECIATION UNCERTAINTY AND IMPACT ON RISK ESTIMATES 6-35
6.11 POTENTIAL INCREASED DIOXIN EMISSIONS FROM UTILITIES WITH ELECTROSTATIC
PRECIPITATORS 6-37
6.12 DISCUSSION OF UNCERTAINTY AND ASSUMPTIONS FOR DOSE-RESPONSE ASSESSMENT
FOR CARCINOGENS 6-37
6.12.1 Default Options 6-39
6.12.2 Models, Methods, and Data 6-40
6.12.3 Discussion of Uncertainty in lUREs 6-43
6.12.4 Variability in Cancer Dose-Response Assessment 6-44
6.13 PRELIMINARY QUANTITATIVE UNCERTAINTY AND VARIABILITY ANALYSIS FOR
INHALATION EXPOSURE AND RISK ASSESSMENT 6-46
6.13.1 Introduction 6-46
6.13.2 Approach to Quantitative Uncertainty Analysis 6-47
6.13.3 Discussion of Results of the Quantitative Uncertainty Analysis . 6-58
6.14 QUALITATIVE DISCUSSION OF ADDITIONAL UNCERTAINTIES 6-60
6.14.1 Uncertainty using lUREs 6-60
6.14.2 Residence Time and Activity Patterns 6-60
6.15 DISCUSSION OF FEDERAL INTERAGENCY REVIEW COMMENTS 6-60
6.16 REFERENCES 6-62
7.0 MERCURY KDUTIPATBHAY ASSESSMENT 7-1
7.1 OVERVIEW 7-1
7.1.1 The Mercury Cycle 7-1
7.1.2 Sources of Mercury 7-1
7.1.3 Deposition of Mercury 7-5
7.1.4 Mercury in Soil 7-7
7.1.5 Plant and Animal Uptake of Mercury 7-8
7.1.6 Mercury in the Freshwater Ecosystem 7-8
7.2 MERCURY CONCENTRATION IN BIOTA 7-10
7.3 MEASUREMENT DATA NEAR UTILITIES 7-15
111
-------
Table of Contents (continued)
VOLUME I
Section
8.0
9.0
10.0
11.0
7.4
7.5
7.6
7.7
7.8
QuAK
8.1
8.2
8.3
8.4
8.5
8.6
8.7
MOLT:
9.1
9.2
ALTE
10.1
10.2
10.3
10.4
10.5
10.6
10.7
10.8
PREL
11.1
11.2
11.3
MODELING THE FATE OF MERCURY EMISSIONS FROM UTILITIES
7.4.1 Long-range Transport Analysis
7.4.2 Local Analysis
PRELIMINARY OBSERVATIONS FROM LOCAL ANALYSIS
GENERAL FINDINGS FOR MERCURY FROM UTILITIES
DISCUSSION OF FEDERAL INTERAGENCY REVIEW COMMENTS .... . . .
REFERENCES .... .
CTATIVE MDLTIPATHWAY ASSESSMENT FOR ARSENIC, DIOXINS, LEAD, AND CADMIUM . .
BACKGROUND . .
ARSENIC COMPOUNDS
DIOXIN AND DIOXIN-LIKE COMPOUNDS
LEAP COMPOUNDS . .
CADMIUM COMPOUNDS
OVERALTi SUMMARY
REFERENCES ,,--..,,,...,, , , . t , , , ,
SUMMARY OF RADIONUCLIDE ANALYSIS
9.1.1 Natural Radionuclide Content in Fossil Fuels-Coal
9.1.2 Natural Radionuclide Content in Fossil Fuels-Natural Gas . . .
9.1.3 Natural Radionuclide Content in Fossil Fuels-Oil
9.1.4 Radionuclide Emissions from Fossil-Fueled Plants
9.1.5 Summary of CAP-93 Model
9.1.6 Estimates of Population Health Risks
RADIONUCLIDE UNCERTAINTY ANALYSIS
9.2.1 Summary Findings
REFERENCES ----..---...--.,,.,...,..---,
RNATTVE CONTROL STRATEGIES FOR HAZARDOUS AIR POLLUTANT EMISSIONS REDUCTIONS
PRECOMBUSTION CONTROLS
10.1.1 Fuel Switching
10.1.2 Coal Cleaning ... . . . ... ....
10.1.3 Coal Gasification
COMBUSTION CONTROL
POSTCOMBUSTION CONTROL ...
10,"?,! psp-t-inilat-C" ph^pe> font-mi p --,,,, - -
10.3.2 Vapor Phase Controls
10.3.3 Acid Gas Control
10 3.4 Carbon Adsorption ...
ALTERNATIVE CONTROLS
POLTiTJTANT TRADEOFFS , , , . , ,
10.5.1 HAP Increase/Decrease
10.5.2 Water /Solid Waste Considerations
AVAILABLE CONTROL TECHNOLOGY AND STRATEGIES FOR MERCURY CONTROL . . .
10.6.1 Impact of Fuels and Temperature on Mercury Emissions ....
10.6.2 Developing Technologies: Activated Carbon
DISCUSSION OF FEDERAL TNTERAGENCY REVIEW COMMENTS ,--,---
XMINARY OBSERVATIONS --..,,- ....
INDUSTRY GROWTH AND HAP EMISSIONS . .
INHALATION RISK ASSESSMENT
MERCURY
Page
7-17
7-18
7 30
7-37
7-38
7-39
7-46
8-1
8-1
8-2
8-5
8-10
8-12
8-16
8-17
9-1
9-1
9-2
9-2
9-3
9-4
9-7
9-9
9-11
9-13
9-17
10-1
10-1
10-1
10-8
10-11
10-14
10-23
10-23
10-29
10-32
10-32
10-33
10-34
10-34
10-34
10-38
10-39
10-41
10-48
10-51
11-1
11-1
11-3
11-5
IV
-------
VOLUME X
Table of Contents (continued)
Section
Page
11.4 QUALITATIVE ASSESSMENT OF DIOXINS, ARSENIC, CADMIUM, AND LEAD 11-6
11.5 RADIONUCLIDE ANALYSIS 11-6
11.6 ALTERNATIVE CONTROL STRATEGIES 11-7
11.7 AREAS FOR FURTHER RESEARCH AND ANALYSES 11-9
11.7.1 Emissions Data for Dioxins 11-9
11.7.2 Speciation of Nickel and Chromium 11-9
11.7.3 Multipathway Risk Assessment 11-9
11.7.4 Long-Range Transport Exposures 11-9
11.7.5 Mercury Issues 11-9
11.7.6 Projections to the Year 2010 11-10
11.7.7 Ecological Risks 11-10
11.7.8 Criteria Pollutant and Acid Rain Programs 11-10
11.7.9 Short-Term Emissions 11-10
11.8 FEDERAL INTERAGENCY REVIEW COMMENTS 11-10
11.9 REFERENCES 11-12
VOLUME II
Appendix A Median emission factors, determined from test report data, and
total 1990 and total 2010 emissions, projected with the computer
emissin progam A-l
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 E-l
Appendix F Documentation of the Inhalation Human Exposure Modeling for the Utility
Study F-l
Appendix G Priliminary Uncertainty Analysis for the Characterization of
Human Health Risks from Direct Inhalation Exposures to Electric
Utility HAP EmissionsSuinmary of Speciation, Environmental Chemistry,
and Fate of Eight HAPs Emitted from Utility Boiler Stacks G-l
VOLUME III
Appendix H Summary of Speciation, Environmental Chemistry, and Fate of Eight HAPs
Emitted from Utility Boiler Stacks H-l
Appendix I Summary of EPRI's Utility Report 1-1
Appendix J Parameter Justifications: Scenario Independent Parameters J-l
Appendix K Parameter Justifications Scenario-Dependent Parameters K-l
-------
Table of Contents (continued)
VOLUME I
Section Page
Appendix L Mercury Partition Coefficient Calibrations L-l
Appendix M Description of Exposure Models M-l
VI
-------
LIST OF TABLES
Table Page
2-1. Comparison of Particulate Matter Collection Systems 2-15
2-2. Distribution of SO2 Control Technologies in 1990 2-18
2-3. Distribution of NOX Control by Fuel Burned, by Unit, in 1990 . 2-26
3-1. Average Higher Heating Values of Coal 3-10
3-2. Assigned Chloride ppmw and HCl ppmw Concentrations in coal,
by State of Coal Origin 3-12
3-3. Selected Nationwide HAP Emissions (estimated) in tons/year
for 1990 and 2010 3-15
3-4. Comparison of Utility Boiler Emissions from EFP Estimates
and from Tests 3-16
3-5. Emissions from an Characteristic Coal-fired Electric Utility
Plant (1990) 3-18
3-6. Emissions from an Characteristic Oil-fired Electric Utility
Plant (1990) 3-19
3-7. Emissions from an Characteristic Natural Gas-fired Electric
Utility Plant (1990) 3-19
4-1. Weight-of-Evidence (WOE) Classification 4-7
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
5-4. 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 5-10
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 5-10
5-7. Inhalation Screening Assessment For HAPS Emitted From
Gas-Fired Utilities 5-11
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 5-14
5-9. Comparison of Cancer and Noncancer Effects Benchmarks and
Emissions estimates for the 13 Selected HAPs 5-15
5-10. Pollutants Considered Priority For Further Analysis Based on
The Results of The Screening Assessment 5-18
6-1. Summary of Baseline Risk Estimates from Chronic Inhalation
Exposure by HAP for 424 U.S. Coal-fired Utilities 6-3
vii
-------
LIST OF TABLES (continued)
Table Page
6-2. Summary of Population Exposed at Various Levels of Inhalation
Risk or Greater by HAP: Coal-Fired Utilities 6-7
6-3. Summary of Baseline Risk Estimates from Inhalation Exposure for
Priority HAPs for 137 U.S. Oil-Fired Utilities 6-8
6-4. Summary of Population Exposed at Various Levels of Risk or
Greater from Oil-Fired Utilities 6-12
6-5. Summary of Baseline Inhalation Risk for Gas-Fired Utilities . 6-14
6-6. Comparison of Inhalation Cancer Risk Estimates Based on
(1) HEM Modeling Using Urban Default Assumption and
(2) HEM Modeling Using Urban vs. Rural Distinction 6-16
6-7. Comparison of Inhalation Noncancer Risk Estimates Based on
(1) HEM Modeling Using Urban Default Assumption and
(2) HEM Modeling Using Urban vs. Rural Distinction 6-16
6-8. Estimated Inhalation Cancer Risks for the Year 2010 Compared
to 1990 for Coal- and Oil-Fired Utilities 6-18
6-9. Estimated Inhalation Noncancer Risks for Coal-Fired
Utilities for the Year 2010 Compared to the Year 1990 .... 6-18
6-10. Noncancer Reference Exposure Levels (Acute) from CAPCOA . . . 6-19
6-11. Sample Stack Parameters for Typical Utility Plant 6-20
6-12. Stack and Emission Values Input to TSCREEN 6-20
6-13. Results of the TSCREEN Model 6-20
6-14. Comparison of Risk Estimates for Single Count Versus Double
Count Runs to Assess the Impact of Overlapping Plumes .... 6-23
6-15. Windspeeds Used for Each Pasquill Stability Category in GARB
Subroutine Calculations 6-27
6-16. Roughness Length Used for Each Land-Use Category in the CARB
Subroutine Calculations 6-27
6-17. Exposure and Risk Estimates Based on RELMAP Modeling of Arsenic
Emissions from All Oil- and Coal-fired Utilities in the U.S. . 6-30
6-18. Chromium Speciation for Coal-fired Utilities: Inhalation
Risk Estimates due to Chromium Emissions Based on Various
Assumptions of Percent Cr VI. 6-34
6-19. Chromium Speciation Analysis for Oil-fired Utilities:
Inhalation Risk Estimates due to Chromium Based on Various
Assumptions of Percent Chromium VI 6-34
6-20. Arsenic Inhalation Risk Estimates: Comparison of Results
Using the EPRI, EPA-verified, and Canadian IURE 6-36
6-21. Nickel from Oil-Fired Utilities: Inhalation Cancer Risk
Estimates Based On Various Assumptions of Speciation and Cancer
Potency 6-38
6-22. Comparison of Nickel Exposure to Various Noncancer Health
Benchmarks 6-38
6-23. Summary of Basic Parameters Used in Risk Assessment for
Electric Utilities 6-48
6-24. Summary of Results for Monte Carlo Simulation of HAP Emissions
(kg/year) from Oil-Fired Plant No. 29 6-54
6-25. Distribution of MIR: Plant No. 29:
Comparison of FCEM and SGS Concentration Data 6-59
viii
-------
LIST OF TABLES (continued)
Table
7-1. Measured Mercury Concentrations in Freshwater Sport Fish
(Total Mercury, ng/g wet wt.) 7-11
7-2. Mercury Concentrations Measured in Some Marine Fish and
Shellfish (//g/g wet wt.) 7-13
7-3. Mercury Concentrations in the Atmosphere and Mercury Measured
In Rainwater Collected in Broward County, Florida 7-18
7-4. Mercury Concentrations Measured at Two Sites in the Atmosphere
over Detroit, Michigan 7-18
7-5. Models Used to Predict mercury Air Concentrations, Deposition
Fluxes, and Environmental Concentrations 7-19
7-6. Factors Potentially Important in Multipathway Modeling of
Mercury How They are Addressed in This Assessment 7-19
7-7. Combination of Local and Regional Impacts: Contribution
of Regional Sources to Key Output at Eastern Site 7-26
7-8. Combination of Local and Regional Sources: Contribution
of Regional Sources to Key Output at Western Site 7-27
7-9. Comparison of Assumed Deposition Parameters for Emitted
Forms of Mercury 7-32
7-10. Process Parameters for Model Plants 7-34
7-11. Summary Mercury Concentrations in Media and Biota 7-35
8-1. Estimated TEQ Background Dioxin Exposures in the United States 8-8
8-2. Concentration of Lead in Various Food Products 8-13
8-3. Concentration of Cadmium in Various Food Products 8-15
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 9-14
9-5. Plants with the Highest Estimated Maximum Individual Risk . . 9-15
10-1. Comparison of Average Concentrations of Trace Elements in
Utility Fuels 10-2
10-2. Trace Element Reductions Achieved Through Conventional Coal
Cleaning 10-9
10-3. Emissions from an Air-Blown, Fixed-Bed Gasif ier 10-13
10-4. 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 (lb/trillion Btu) .... 10-15
10-5. Comparison of wet bottom vs. dry bottom electric utility unit
metallic HAP emissions, trace metal removal, and trace metal
concentrations in feed coal 10-17
10-6. Descriptive statistics for removal efficiences shown in
Figures 10-7 and 10-8 10-24
10-7. Descriptive statistics for removal efficiences shown in
Figures 10-9 and 10-10 10-25
10-8. Descriptive statistics for removal efficiences shown in
Figures 10-11 and 10-12 10-26
ix
-------
LIST OF TABLES (continued)
Table Page
10-9. Descriptive statistics for removal efficiences shown in
• Figures 10-13 and 10-14 10-27
10-10. Particulate Metallic HAP Removal Percentage From ESPs and
FFs (Excluding Mercury) 10-29
10-11. Descriptive statistics for removal efficiences shown in
Figures 10-15 and 10-16 10-30
10-12. Descriptive statistics for removal efficiences shown in
Figures 10-17 and 10-18 10-31
10-13. Qualitative Effects of Different Control Strategies on Air
Emissions of HAPs 10-35
10-14. Comparison of Typical Uncontrolled Flue Gas Parameters at
Utilities and MWCs 10-44
x
-------
LIST OF FIGURES
Figure Page
2-1. Fossil fuel use in the utility industry in 1990 2-2
2-2. Unit types in the utility industry by fuel type in 1990 . . . 2-8
2-3. Particulate control in the utility industry by fuel type in 19902-10
2-4. SO2 control in the utility industry type in 1990 (coal-fired
boilers only) 2-16
2-5. Nitrogen oxide control in the utility industry by fuel type
in 1990 2-23
2-6. Fuel use in the utility industry by fuel type in 1990 and
projections for the year 2010 2-29
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. Maximum Individual Risk Posed by HAPs Emitted from All
U.S. Coal-Fired Electric Utilities 6-4
6-lb. Maximum Individual Risk posed by HAPs emitted from All
U.S. Coal-Fired Electric Utilities 6-5
6-2a. Maximum Individual Risk Posed by HAPs Emitted From All
U.S. Oil-Fired Electric Utilities 6-10
6-2b. Number of Oil-Fired Utilities Posing Various Levels of
Maximum Individual Risk (by Levels of MIR) 6-11
6-3. Estimates of Annual Cancer Incidence Due to Inhalation Exposure
to HAP Emissions from Oil-Fired Electric Utilities Based on
the Local Analysis 6-13
6-4. Results of the RELMAP Analysis for Arsenic for All Utilities:
Air Concentration of Arsenic, Units: ng/m3 6-29
6-5. Depiction of Combining Component Uncertainty Distributions
Into an Overall Distribution on Uncertainty 6-52
6-6. Summary of Results of Monte Carlo Simulation of HAP Emissions
from Oil-Fired Plant #29 6-55
7-1. The Mercury Cycle 7-2
7-2. Average Ionic Mercury (Hg42) Concentration from Electric
Utilities Only (Base) 7-21
7-3. Average Elemental Mercury (Hg°) Concentration from Electric
Utilities Only (Base) 7-22
7-4. Average Particulate Mercury Concentration from Electric
Utilities Only (Base) 7-23
7-5. Total Mercury Wet and Dry Deposition from Electric Utilities
Only (Base) 7-24
7-6. Overview of the IEM2 watershed modules 7-32
10-1. Relation between the concentrations of mercury and sulfur in
153 samples of coal shipments 10-4
10-2(a-g). Relation between concentration of selected trace
elements and sulfur in modified USGS data 10-6
10-3. Coal gasification combined cycle technology 10-13
10-4(a-c). Average coal-fired emissions, average trace metal removal,
and average trace element concentration in feed coal vs.
bottom type (bituminous and subbituminous coal-fired) . 10-20
xi
-------
LIST OF FIGURES (continued)
Figure Page
10-5(a-c). Average coal-fired emissions, average trace metal removal,
and average trace element concentration in feed coal vs.
bottom type (bituminous coal-fired only) 10-21
10-6(a-c). Average coal-fired emissions, average trace metal removal,
and average trace element concentration in feed coal vs.
bottom type (subbituminous coal-fired only) 10-22
10-7. Removal of metallic HAPs by electrostatic precipitators
(cold-side, coal) 10-24
10-8. Removal of mercury by electrostatic precipitators (cold-side,
coal) 10-24
10-9. Removal of metallic HAPs by electrostatic precipitators
(hot-side, coal) 10-25
10-10. Removal of mercury by electrostatic precipitators (hot-side,
coal) 10-25
10-11. Removal of metallic HAPs by electrostatic precipitators (oil) 10-26
10-12. Removal of mercury by electrostatic precipitators (oil) . . 10-26
10-13. Removal of metallic HAPs by a fabric filter (coal) .... 10-27
10-14. Removal of mercury by a fabric filter (coal) 10-27
10-15. Removal of metallic HAPs by an FGD (coal) 10-30
10-16. Removal of mercury by an FGD (coal) 10-30
10-17. Removal of metallic HAPs by an spray dryer adsorber/fabric
filter (coal) 10-31
10-18. Removal of mercury by an spray dryer adsorber/fabric filter
(coal) 10-31
10-19. Comparison of mercury removal efficiencies with activated
carbon injection 10-45
XII
-------
Glossary
HAPs
MW
Ib/MMBtu
Ib/ trillion Btu
Nm3
dscm
m/s
MEI
HEM
WOE
IURE
RfC
TEQ
HQ
CNS
AALG
PEL
MIR
ng/L
ng/m3
pg/m3
the Act
a utility
Hazardous Air Pollutants
Megawatt
Pounds per Million British Thermal Units
Micron (Micrometer)
Microgram
Pounds per Trillion British Thermal Units
Dry Normal Cubic Meter (1 atm, 0°C)
Dry Standard Cubic Meter (1 atm, 20°C)
Meters per Second
Maximum Exposed Individual
Human Exposure Model
Weight of Evidence
Inhalation Unit Risk Estimate
Inhalation Reference Concentration
Toxic Equivalent Emissions
Hazard Quotient
Central Nervous System
Ambient Air Level Guidelines
Permissible Exposure Limit
Maximum Individual Risk
Nanograms per Liter
Nanograms per Cubic Meter
Picograms per Cubic Meter
Clean Air Act as amended in 1990
a fossil-fuel-fired electric utility steam
generating plant
Xlll
-------
This page intentionally left blank.
xiv
-------
PREFACE
Overview of Report
This interim final report on hazardous air pollutant (HAP)
emissions from fossil fuel-fired electric utility steam
generating units (i.e., utilities) has been prepared by the
United States Environmental Protection Agency (EPA) pursuant to
section 112(n)(1)(A) of the Clean Air Act, as amended in 1990
(the Act). This report provides the Congress and the public with
information regarding the emissions, fate, and transport of
utility HAPs.
The primary components of this interim report are the
following: (1) a description of the industry; (2) an analysis of
emissions data; (3) an assessment of hazards and risks due to
inhalation exposures to numerous HAPs (e.g., arsenic, nickel,
cadmium, chromium, beryllium, and others [but excluding
mercury]); (4) an assessment of risks due to multipathway
(inhalation plus non-inhalation) exposure to one class of HAPs
(radionuclides); (5) a general assessment of the fate and
transport of mercury through various environmental media; and (6)
a discussion of alternative control strategies.
The assessment for mercury in this interim report includes a
description of mercury 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}. These predicted levels are
based on modeling of mercury emissions from four representative
utility plants using hypothetical scenarios. The EPA has not
evaluated exposures to mercury emissions from utilities for
humans or wildlife in this interim report. If appropriate and
feasible, the EPA may include these analyses in the final report.
To provide general information regarding potential
background levels of several HAPs (i.e., mercury, arsenic,
cadmium, lead, and dioxins) in the environment due to all sources
(natural and anthropogenic), this interim report presents
measured levels in various media (e.g., soil, air, water, and
food products) as reported by various studies.
Assessments of human exposures to mercury and the associated
risks of health effects were included in previous drafts of this
report and in a related draft EPA report (Mercury Study Report to
Congress [i.e., mercury study]). However, during external review
of these draft reports, several critical issues related to the
mercury risk assessment, including the impending release of new
xv
-------
mercury health data, were raised. As a result of that review,
the Agency plans to complete the mercury study when two important
on-going human health studies are published and reviewed. At
this time, the EPA believes that it is appropriate to exclude
such assessments for mercury until after the mercury study is
issued. However, this issue is still under consideration and
negotiation, and may be dependent on results of additional peer
review and other factors.
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. However, non-inhalation
exposures were not estimated for these four HAPs because of the
complexity and the intensive data requirements of such analyses.
The EPA recognizes that non-inhalation exposures could be
important for these HAPs. Therefore, the EPA has initiated a
multipathway assessment for arsenic, and may consider conducting
assessments for additional HAPs in the future.
This report is not a final report because the assessment of
impacts to public health is not yet complete. For example, as
indicated above, the evaluation of risks due to non-inhalation
exposures was limited. In addition, conclusions regarding the
significance of the risks, as well as the regulatory
determination required in section 112(n)(1)(A), are not provided.
The EPA plans to publish a final utility HAP report at a
later date which will include a more complete assessment of the
exposures, hazards, and risks due to utility HAP emissions, and
will include conclusions, as appropriate and feasible, regarding
the significance of the risks and impacts to public health. In
addition, the EPA plans to include in the final report a
determination as to whether regulation of HAPs from utilities
under section 112 is appropriate and necessary, as required by
section 112(n)(1)(A) of the Act and a court order. This court
order was issued pursuant to litigation filed against the EPA for
failing to meet the statutory deadline for the utility report.
The EPA intends that this regulatory determination would be a
decision, based on the estimated impacts to public health,
whether or not to pursue a regulatory development program under
section 112. During any regulatory development process, the EPA
would evaluate a range of potential control technologies and
emission reduction options and their associated costs.
There are uncertainties, data gaps, and limitations to the
current analyses, which are discussed throughout this interim
report. If new data become available or improvements are made to
the analyses, these changes will be included in the final report.
xvi
-------
Peer Review
Draft versions of this report were reviewed during the
slimmer of 1995 by numerous non-EPA scientists representing
industry, environmental groups, academia, and other
organizations. In the Spring of 1996, the draft report underwent
additional review by EPA, State and local air pollution agencies,
and other Federal agencies. In addition, a revised draft interim
report underwent an expedited review (1 week) by State and local
air pollution agencies and other Federal agencies during
September 1996.
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. At the end of each Chapter, the EPA has
included comments received from other Federal Agencies (e.g.,
Department of Energy, Food and Drug Administration, National
Marine Fisheries Service) that were not fully addressed, along
with relevant explanations, as appropriate.
Draft versions of this report, along with all the comments
received, have been submitted to the docket (A-92-55) and are
available for public inspection.
xvn
-------
This page intentionally left blank.
-------
EXECUTIVE SUMMARY
ES.1 BACKGROUND
This interim final report on emissions of hazardous air
pollutants (HAP) from fossil fuel-fired electric utility steam
generating units (i.e., utilities) was prepared by the United
States (U.S.) Environmental Protection Agency (EPA) pursuant to
section 112(n)(1)(A) of the Clean Air Act (the Act), as amended
in 1990. The primary components of this interim report are: (!)
a description of the industry; (2) an analysis of emissions data;
(3) an assessment of hazards and risks due to inhalation
exposures to numerous HAPs (e.g., arsenic, nickel, chromium); (4)
an assessment of risks due to multipathway (inhalation plus non-
inhalation) exposure to one class of HAPs (i.e., radionuclides) ,-
(5) a general assessment of the fate and transport of mercury
through environmental media; and (6) a discussion of alternative
control strategies.
The study was based on two scenarios: (1) 1990 base year
emissions; and (2) 2010 emissions. 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. 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 on-going and future regulatory activities under
other provisions of the Act (e.g., ambient air quality and acid
rain programs) are in place. The 2010 scenario also included
estimated changes in HAP emissions resulting from projected
trends in fuel choices and electric power demands.
ES.2 DESCRIPTION OF INDUSTRY
A total of 684 utility plants were identified 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 (i.e., boilers) and several plants
burn more than one type of fuel (e.g., contain both coal- and
oil-fired boilers). There are 426 plants that burn coal as one
of their fuels, 137 plants that burn oil, and 267 plants that
burn natural gas.
There are many different types of facilities, varying in
boiler type, emission control devices (controls), and other
characteristics. Based on data for 1990, all coal-fired units
and about one-third of oil-fired units use some form of
particulate matter (PM) control. Approximately 15 percent of
coal-fired units utilize add-on controls for sulfur dioxide
(S02) . Approximately 70 percent of oil- and gas-fired units
-------
employ controls for nitrogen oxides (NOX) ; and 80 percent of
coal-fired units have NOX controls.
ES.3 EMISSIONS DATA ANALYSIS
Emission estimates for the years 1990 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.
These test data provided the basis for estimating average
annual emissions for each of the 684 plants. A total of 67 HAPs
were identified in the emissions testing program as potentially
being emitted by utilities. Tables ES-1 and ES-2 present
estimated emissions for a subset of HAPs.
The average annual emissions estimates are considered
appropriate for assessing long-term exposures on a national
basis. However, since the EPA did not have emissions test data
for each utility in the U.S., there may be individual plants for
which the EPA either underestimated or overestimated emissions.
Based on an uncertainty analysis, the average annual emissions
estimates are predicted to be roughly within a factor of plus or
minus three of actual annual emissions. However, this analysis
had limitations. For example, the analysis did not include data
on potential upsets or unusual operating conditions; therefore,
the range of uncertainty could be greater. The range of
uncertainty for short-term emissions has not been determined.
ES.4 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. Non-inhalation exposures are presented for one
class of HAP (radionuclides).
For many of the 67 HAPs, inhalation exposure is believed to
be the dominant exposure pathway. However, for HAPs that are
persistent, bioaccumulate, and are toxic by ingestion, the non-
inhalation exposure pathways are likely to be more important. In
addition to radionuclides, the EPA also identified five other
ES-2
-------
Table ES-1. Nationwide Utility Emissions for a Subset of HAPs
HAP
Arsenic
Cadmium
Chromium
Lead
Mercury
Nickel
Hydrogen chloride
Hydrogen fluoride
Dioxins0
Nationwide HAP emission estimates (tons per year)'
Coal (426 plants)
1990
54
1.9
70
72
51
48
140,000
20,000
0.00015
2010
54
2.3
83
83
65
57
1 50,000
26,000
0.00020
Oil (137 plants)
1990
5
1.7
4.7
11
0.25
390
2,900
140
1 x 10'5
2010
3
0.9
2.4
5.6
0.13
200
1,500
73
5x 10'6
Natural gas (267 plants)
1990
0.16
0.054
1.2
0.44
0.0016
2.3
NMb
NM
NM
2010
0.25
0.086
1.9
0.68
0.0024
3.5
NM
NM
NM
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.
Based on an uncertainty analysis conducted for this study, the EPA predicts that the emissions
estimates for individual plants are generally within a factor of roughly three of actual emissions.
NM = Not measured.
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 for toxicity relative to
2,3,7,8-tetrachlorodibenzo-p-dioxin (i.e., 2,3,7,8-TCDD).
HAPs (mercury, arsenic, dioxins, cadmium, and lead), that could
present additional impacts due to non-inhalation exposures. The
dispersion, fate, and environmental concentrations of mercury
were evaluated; however, exposures and risks were not estimated.
The other four HAPs (arsenic, dioxins, lead, and cadmium) were
examined qualitatively for their potential for multipathway
hazards. However, multipathway exposure assessments were not
conducted for these four HAPs. The EPA recognizes that for
mercury, as well as these other four HAPs, non-inhalation
exposures could be important. Quantitative analyses were not
performed for arsenic, cadmium, dioxins, and lead because of the
complexity of such analyses, the intensive data requirements of
such analyses, and because of the limited chemical-specific data
available (e.g., chemical-specific air-to-plant biotransfer
factors, bioconcentration factors, chemical-specific plant uptake
rates) for conducting such analyses. The EPA plans to continue
assessing the multipathway exposures and hazards for mercury.
The EPA has initiated a multipathway assessment for arsenic.
ES-3
-------
Table ES-2. Estimated Emissions From Characteristic Utility
Units (1990; tons per year)
Fuel:
Unit size (MWe):
Arsenic
Cadmium
Chromium
Lead
Mercury
Hydrogen chloride
Hydrogen fluoride
Dioxins0
Nickel
Coal
325
0.081
0.00051
0.086
0.075
0.05
190
14
0.00000014
NC
Oil
160
0.016
0.0077
0.018
0.053
0.0012
9.4
NC
0.000000035
2.1
Natural gas
240
0.0003
NCb
NC
NC
NC
NC
NC
NC
0.004
There are uncertainties in these numbers. Based on an uncertainty analysis conducted for this study,
the EPA predicts that the emissions estimates are generally within a factor of roughly three of actual
emissions.
NC = Not calculated.
See footnote b of Table ES-1.
Multipathway analyses may be undertaken for some of the other
HAPs (e.g., dioxins) in the future should the EPA determine that
such analyses are feasible and warranted, and as resources allow.
ES.5 SCREENING ASSESSMENT
Initially, the EPA 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) . If the MEI risk was above a minimum
measure (e.g., exposure greater than one-tenth the inhalation
reference concentration [RfC] 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 four criteria: (1) persistence; (2) tendency to
ES-4
-------
bioaccumulate; (3) toxicity by ingestion; and (4) quantity of
emissions.
Based on this assessment, 15 HAPs (arsenic, beryllium,
cadmium, chromium, lead, manganese, mercury, nickel, hydrogen
chloride [HCl], hydrogen fluoride [HF], acrolein, dioxins,
formaldehyde, n-nitrosodimethylamine, and radionuclides) were
identified as priority based on their potential to pose impacts
to public health due to inhalation or non-inhalation exposures.
The other 52 HAPs were not evaluated beyond the screening
assessment.
ES.6 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 individually (i.e., local analysis). For 14 of
the 15 HAPs, the HEM was used; for radionuclides, the Clean Air
Act Assessment Package-1993 (CAP-93) model was used. The cancer
risks for gas-fired plants were less than one chance in one
million (i.e., 1 x 10"6) 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 (i.e., conservative,
but not overly conservative) estimate of the risks due to
inhalation exposure within 50 km of plants. That is, the risk
estimates from the local analysis are estimated to represent
approximately the 90th to 95th percentile. Conservative
estimates are considered appropriate so that errors are on the
side of public health protection.
ES.6.1 Inhalation Cancer Risks for Coal-fired Utilities Based on
Local Analysis
The large 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 chance in 1 million (i.e., 1 x 10"6)
due to inhalation exposure. Only two of the 426 plants are
estimated to pose inhalation risks greater than 1 x 10"6 (see
Figure ES-1).
ES-5
-------
^^
^^
Q)
W
O
H
w
£
ro
-3
•D
'>
T3
C
C
E
X
CO
0
C
0
^
.Q
£
(0
^
•
i
CO
UJ
CD
D
O
•_
I *
U—
.2
«M
"M
D
•D
(D
il
"m
0
O
E
0
•D
+J
4-*
E
LU
(A
Q.
-j-
x
^
d
^f
I
0
^j
O
3
C
0
l_
1
c
o
c
o
'E
o
0)
o
' c
'o
CO
o
<
a
i^m
sHm
^WSSiWBHSSBSSBlSSBi
ggggggggg^^ggggg
,1,1,1,1,
C
0
-H
4J
(0
1 1
0
c
O
•H
•H
*
^*
3
2
KH
^ i
HIM
^p i
H§l 1
1
«HH
?HI
I
P
, i ,
in
UJ
T-
0
(D
111
f
CD
111
r-
0
4-1
s
1
w
T-
D
s
111
r-
0
4->
00
1
UJ
T—
00
1
LLJ
i—
V
i-,
CD
•H
U
W
CU
JJ
•H
Q)
C
-H
0)
c
0)
U)
0)
a
a
0)
o
K
•H
c?
H
S3
•H
M
rH
(C
S
-H
C
•H
1
OOOOOOOO -H
in o in o m o m x
CO CO CM CM T- r-
•^
0 fl
k (
• *•
^ *" i
C ^ t
h<
z u a
g
0)
<~\
£-i
• •
OJ
JJ
O
J3
(d
u
£^
O
•H
rH
rH
•H
g
a>
§
•H
Q)
U
d
(C
u
CU
o
o
w
•H
M
ncer
o
Q)
rO
CU
0
d
-H
a
CU
'a
X
CU
o
.
jj
c
0)
o
p<
CU
£
(tS
w
(0
J_l
o
1
w
rH
W
(C
0
0
X
rH
CQ
(C
73
0)
w
en
0)
SH
a
0)
CU
45
ES-6
-------
The increased lifetime cancer risk due to inhalation
exposure to HAP emissions for the highest MEI, based on the local
analysis, is estimated to be 5 x 10'6. Arsenic and chromium are
the HAPs contributing most to the inhalation cancer risks (Table
ES-3). All other HAPs, including radionuclides, were estimated
to present inhalation risks less than 1 x 10"6.
The cancer incidence in the U.S. due to inhalation exposure
to HAP emissions (including radionuclides) from all 426 coal-
fired utility plants based on the local analysis is estimated to
be approximately 0.2 cancer case per year (cases/yr), or 1 case
every 5 years.
ES.6.2 Inhalation Cancer Risks for Oil-fired Utilities Based on
Local Analysis
The majority of the oil-fired plants (more than 114 of the
137 plants) are estimated to pose inhalation cancer risks less
than 1 x 10"6. However, up to 22 of the 137 oil-fired plants are
estimated to present inhalation risks above 1 x 10"6 (see Figure
ES-2). Nickel, arsenic, radionuclides, and chromium are the
primary contributors to these cancer risks.
The highest contribution to the MEI risk is nickel. The
range in MEI risk (see Table ES-4) reflects a range in
assumptions regarding the form of nickel being emitted and the
associated cancer potency. Nickel subsulfide is a known human
carcinogen and appears to be the most carcinogenic form based on
available data. Several other nickel species are also
potentially carcinogenic but the potencies are not known.
To evaluate the range of potential risks due to nickel
emissions, the EPA estimated risks due to nickel emissions using
various assumptions for nickel cancer potency. For example,
assuming the nickel mix is 100 percent as carcinogenic as nickel
subsulfide, the highest MEI inhalation cancer risk due to the
aggregate of HAP emissions from the highest risk oil-fired
utility plant is estimated to be 1 x 10"4. Assuming the nickel
mix is 10 percent as carcinogenic as nickel subsulfide, the
highest MEI inhalation risk is approximately 3 x 10"5. The
values in Figure ES-2 are based on the assumption that the nickel
mix is 100 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 10"6 for all 137 oil-fired plants.
ES-7
-------
Table ES-3. Summary of 1990 Inhalation Cancer Risk Estimates
from Local Analysis for Coal-fired Utilities
HAP
Arsenic
Chromium
Total" (Aggregate of HAPs)
MEI lifetime
risk"
3x TO'6
2x TO"6
5x TO"6
Population with lifetime risk
> 1 x 10"*
2,400
110
2,400
Number plants with MEI
lifetime
risk > 1 x 10"6
2
1
2
Estimated MEI risk due to inhalation exposure for the 'highest risk" coal-fired plant. Based on an
uncertainty analysis, these estimates are considered reasonable high-end estimates (roughly the 90th
to 95th percentile) of the risks for the MEI due to inhalation exposure (see section ES.6.3).
Estimated risk due to inhalation of the aggregate of HAPs assuming additivity of risk for 26 individual
carcinogenic HAPs.
Table ES-4. Summary of 1990 Inhalation Cancer Risk Estimates
Based on Local Analysis for Oil-fired Utilities
HAP
Nickel"
Arsenic
Radionuclides
Chromium
Cadmium
Total0 (aggregate)
Highest MEI lifetime risk"
1 x 10'5to9x 10'5
1 x 10'6
1 x 10'5
5x 10'6
2x 10"6
3x 10'5to 1 x 10"4
Population with lifetime risk
> 1 x 1 0"6
2,400 to 1 ,600,000
2,400
2,400
2,300
45
2,400 to 1,600,000
Number plants with MEI
lifetime risk > 1 x 1 0'6
2 to 20
2
2
1
1
2 to 20
Estimated MEI risk due to inhalation exposure to HAPs for the "highest risk" oil-fired plant. Based on
an uncertainty analysis, these estimates are considered reasonable high-end estimates (roughly the
90th to 95th percentile) of the risks for the MEI due to inhalation exposure. See section ES.6.3 for
discussion.
These estimates are presented as a range because of the uncertainties associated with the nickel risk
assessment. If the nickel mix is assumed to be 10% as carcinogenic as nickel subsulfide, then the
MEI risk for nickel is estimated to be 1 x 10's. If the nickel mix is assumed to be 100% as
carcinogenic as nickel subsulfide, the estimated MEI risk for nickel is 9 x 10's.
Estimated risk due to inhalation of the aggregate of HAPs assuming additivity of risk for 14 individual
carcinogenic HAPs. The low end of the range is based on assumption that the mix of nickel
compounds is 10% as carcinogenic as nickel subsulfide. The high-end of the range is based on
assumption that the mix of nickel compounds is 100% as carcinogenic as nickel subsulfide.
ES-8
-------
T3
0)
(0
0
a.
. ,»
K G)
_ '^
CD E
3^
p -^
'> T3
H
it
E =
x<
1E
^ 0
§«
.2s
IE
Suj
•S to
CSI
CO
LU
£
3
HI
T-
0
4-1
in
ill
U)
I
UJ
0
4-1
(0
UJ
(D
UJ
UJ
S
i
UJ
0
*J
GO
UJ
00
UJ
(0
CJ U
o
•H fl
•u) O
-l
-— U
oi d
H -H
03 -
-H 0)
(ti (0
3 X ^
T3 Q) O
•H O O
T3 Cn -i
a
-MX
Q) (0 (D
4-1
O
Q)
ES-9
-------
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
between 0.3 and 0.7 cancer cases/yr. The high end of this range
is based on the assumption that the nickel mix is as carcinogenic
as nickel subsulfide. The low end of the range assumes that the
mix of nickel is 10 percent as carcinogenic as nickel subsulfide.
ES.6.3 Inhalation Cancer Risks Based on Long-Range Transport
Analysis
In addition to the above analyses, the EPA conducted long-
range transport analyses to assess emissions dispersion and
exposures on a national scale. 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.
The RELMAP modeling was conducted for all coal- and oil-
fired utilities, but was limited to mercury and arsenic. Only
arsenic is discussed in this section; the modeling for mercury
is presented in section 7. The long-range transport modeling of
arsenic indicates that the local HEM analysis alone does not
account for a substantial percentage of the population exposures
due to utility emissions. A comparison of the HEM results to the
RELMAP results for arsenic indicates that a significant portion
of emissions disperse further than 50 km, apparently due to the
tall stack heights and other dispersion factors. Based on the
RELMAP analysis, the nationwide dispersion of arsenic emissions
leads to an estimate of population exposure and cancer incidence
that is approximately seven-fold greater than the population
exposures and cancer incidence predicted by the HEM when only
local dispersion is considered (see Table ES-5).
The RELMAP results for arsenic (which is emitted mainly as
PM) were used to estimate the potential long-range transport
inhalation exposures for cadmium, chromium, nickel, and
radionuclides since it is believed that these other HAPs are also
emitted as PM and exhibit proportional emission rates and
atmospheric dispersion behavior similar to that of arsenic.
Because the estimated population exposures resulting from the
long-range transport analysis for arsenic were about seven times
greater than the population exposures predicted by the local
analysis alone, it was also assumed that this ratio also could
hold true for nickel, chromium, cadmium, and radionuclides.
Using this methodology, the cancer incidence for coal-fired
utilities considering both local and long-range transport is
estimated to be roughly 1.4 cases/yr (i.e., 0.2 x 7 = 1.4). The
ES-10
-------
Table ES-5. Summary of Inhalation Risk Estimates Due to Local and
Long-range Transport
LOCAL IMPACTS (dispersion within 50 km of each utility plant)d
Pollutant
Radionuclides
Nickel'
Chromium
Arsenic
Cadmium
All Others"
Total0
. OIL-FIRED PLANTS
Maximum exposed
individual (MEI)
1 x 10'5
9 x 10s
5x 10*
1 x 10'5
2x 10'6
8x 1C'7
1 x 10-4
Annual increased
cancer Incidence
0.2
0.4
0.02
0.04
0.005
0.005
0.7
COAL-FIRED PLANTS
Maximum exposed
individual (MEI)
2x 10'8
7x TO'7
2 x 10'6
3x 10-6
2x 10'7
8x 10'7
5x 10-6
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.)8
Pollutant
Radionuclides
Nickel'
Chromium
Arsenic
Cadmium
All Others'1
Total5
OIL-FIRED PLANTS
Maximum exposed
individual (MEI)
Not estimated
9 x 10'5
5x TO'6
1 x 10'5
2x 10-6
8 x 10'7
1 x 10"4
Annual increased
cancer incidence
1.4
2.8
0.14
0.28
0.035
0.035
4.8
COAL-FIRED PLANTS
Maximum exposed
individual (MEI)
Not estimated
9x 10'7
3x 10-6
4x 10'6
3x 10'7
1 x 10'6
7 x 10'6
Annual increased
cancer incidence
0.7
0.035
0.14
0.35
0.04
0.03
1.3
Assumes that the nickel mixture is as carcinogenic as nickel subsulfide.
Estimated risks due to exposure to all remaining HAPs analyzed (i.e., excluding nickel, arsenic,
chromium, cadmium, and radionuclides).
This is the aggregate risk (i.e., the risk due to inhalation exposure to all carcinogenic HAPs, assuming
additivity of risks).
There are uncertainties associated with these risk estimates. See sections 6.4 for discussion.
These risk estimates are based on an extrapolation of RELMAP modeling results for arsenic to other
HAPs. Therefore, there are considerable uncertainties associated with these results. See sections 6.3
and 6.4 for discussion.
ES-11
-------
cancer incidence for oil-fired utilities (assuming the nickel mix
is 100 percent as carcinogenic as nickel subsulfide) is estimated
to be as high as 5 cases/yr (i.e., 0.7 x 7 = 4.9). These
estimates should be viewed as highly uncertain high-end estimates
(particularly the estimate of five cases/yr for oil-fired
utilities) because of modeling uncertainties and extrapolations
(e.g., using the modeling results for arsenic to predict
dispersion and exposure for the other HAPs) and because of the
assumption for nickel carcinogenicity.
For risks to the MEI, a comparison between the HEM local
dispersion results and the long-range transport modeling results
indicates that long-range transport is not as important for the
MEI risks as it is for cancer incidence. For example, the MEI
risk from the local analyses for coal-fired utilities (i.e.,
inhalation risk of 5 x 10"6) is increased by approximately 40
percent to roughly 7 x 10"6 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 MEI inhalation
risks because of the remote location of the two highest risk oil-
fired plants. Table ES-5 presents a comparison of results from
the local versus long-range transport analyses.
ES.6.3 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 quantitative uncertainty analysis indicates that the MEI
inhalation cancer risk estimates presented above from the local
analysis 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., roughly
the 90th to 95th percentile). Conservative assumptions are
considered appropriate so that errors are on the side of public
health protection. The uncertainty analysis suggests that the
most likely estimated inhalation risks for MEIs (i.e., central
tendency MEI risk estimates) may be roughly 5 to 10 times lower
than the MEI estimates presented above.
ES-12
-------
ES.6.4 Summary of the Inhalation Cancer Risks
For the majority of utility plants (approximately 662 of the
684 plants), the estimated inhalation cancer risks due to HAP
emissions are less than 1 x 10~6. However, several plants (2
coal-fired plants and between 2 and 22 oil-fired plants) are
estimated to pose inhalation cancer risks above 1 x 10~6, and one
oil-fired plant is estimated to pose an MEI inhalation cancer
risk between 3 x 10'5 and 1 x 10~4. The cancer incidence in the
U.S. due to inhalation exposure to HAP emissions from all
utilities (coal-, oil- and gas-fired combined) is estimated to be
between 0.5 and 6 cases/yr. Further research and evaluation is
needed to more comprehensively assess the inhalation cancer
risks, especially the impacts of long-range transport of HAPs and
speciation of nickel.
ES.6.5 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.
Based on modeling HAPs for all 684 plants with the HEM, the
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. 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.
Using a short-term air dispersion model that considers all
reasonable meteorological conditions, the 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.7 MULTIPATHWAY ASSESSMENT
The utility HAPs were prioritized for potential non-
inhalation exposures. The following characteristics were
considered: (1) persistence, (2) toxicity, and (3) potential to
bioaccumulate. Mercury, radionuclides, arsenic, dioxins,
cadmium, and lead were selected as priority for multipathway
assessment.
ES-13
-------
ES.7.1 Mercury Modeling Assessment
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 fate and transport modeling, (2) local scale
dispersion modeling, and (3) modeling of environmental
concentrations. The RELMAP was used to predict long-range
dispersion and deposition across the continental U.S. For the
local analysis, a model designed to predict deposition of HAPs
within 50 km was used. The Indirect Exposure Model (IBM) was
used to estimate environmental concentrations.
Three types of hypothetical locations were considered in the
modeling analyses: (1) agricultural, (2) lacustrine (near lakes),
and (3) urban. Using four model utility plants, and various
assumptions and scenarios, mercury concentrations in various
environmental media were estimated.
There are significant uncertainties in the models, data
inputs, assumptions, and the quantitative results. However, the
analyses were useful for gaining a better understanding of the
fate and transport of mercury in the environment, and for
estimating plausible levels in environmental media.
The modeling also provided information on whether local
and/or long-range transport of mercury is important in a variety
of scenarios. The models indicate that most of the mercury from
utilities is transported further than 50 km from the source.
ES.7.1.1 General Findings for Mercury. Mercury emissions
disperse in the atmosphere and deposit to land and water bodies.
Deposition is of potential concern because mercury persists in
the environment, and bioaccumulates in the food web (especially
in the aquatic food web). The form of mercury found in fish
tissue is predominantly methylmercury. Of all the media and
biota studied, fish have the highest concentrations of mercury in
the environment.
ES.7.1.2 Summary of Mercury Assessment Results for
Utilities. For the year 1990, coal-fired utilities were
estimated to emit approximately 51 tons per year (tpy) of mercury
nationwide, which is approximately 21 percent of the 248 tpy of
anthropogenic emissions of mercury estimated to be emitted in the
U.S. for the years 1990 to 1992. The EPA also estimates that
utility mercury emissions will increase to 65 tpy by the year
2010. If one assumes that current anthropogenic activity
represents between 40 and 75 percent of the total emissions
(anthropogenic plus other emissions [e.g., natural emissions]),
ES-14
-------
one can calculate that U.S. utilities emit roughly 8 to 15
percent of the total emissions of mercury in the U.S.
Recent estimates of global anthropogenic mercury emissions
are about 4,400 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
this global estimate, U.S. anthropogenic mercury emissions could
account for about 6 percent of the global total, and U.S.
electric utilities would account for about 1 percent of global
anthropogenic emissions (using 1990 emission estimates).
Based on the RELMAP modeling analysis, approximately 30
percent (i.e., 15 tpy) of 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. The highest
deposition appears to occur 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 data, the RELMAP model seems to
over- and underestimate mercury values within a factor of two and
appears to be relatively unbiased in its predictions.
Although the amount of mercury being emitted from any single
utility may seem relatively small, these emissions are of
potential concern for a number of reasons. First, mercury is
persistent. It is not degraded, but continually accumulates in
the environment. Consequently, over time there is potential for
concentrations in the environment to build up. Second, mercury
bioaccumulates in the food web. Third, current scientific
evidence indicates that most of the mercury emitted to the
atmosphere from sources such as utilities, which have tall
stacks, does not deposit near the source but is deposited farther
away. As a result, even though the ambient concentration of
mercury around a single source may not be elevated, there are
sufficient data from which to conclude that the cumulative impact
of many small sources may lead to the accumulation of mercury in
the soils and sediments, and bioaccumulation in freshwater fish.
Therefore, the incremental emissions of mercury from utilities,
added to the mercury emissions from all of the other sources,
contribute to overall environmental loadings, and thus, may
contribute, to some degree, to the mercury levels in freshwater
fish.
ES-15
-------
The modeling assessment in conjunction with available
scientific knowledge, suggests that there is a plausible link
between emissions of mercury from utilities and the mercury found
in soil, water, air, and freshwater fish. As noted above, there
are many sources of mercury emissions worldwide, both natural and
anthropogenic. The fish methylmercury levels are probably due,
in part, to mercury emissions from all of these various sources
over time. The coal-fired utilities are one category of the
mercury sources. The EPA has not yet determined whether the
mercury emissions from utilities are a concern for public health.
The EPA recognizes that there are significant uncertainties
regarding the extent of the exposures and risks due to utility
mercury emissions, and that further research and evaluation is
needed to reduce uncertainties and to characterize the exposures
and risks. Areas of uncertainty include the following: (1) what
exposure levels are likely to result in adverse health effects;
(2) what percent of mercury emissions are elemental versus
divalent mercury; (3) how much mercury is emitted from natural
sources; (4) how much mercury is removed during coal cleaning;
and (5) what affects the bioaccumulation of methylmercury in
fish. The EPA plans to continue evaluating the exposures and
public health impacts due to mercury emissions., In addition, the
EPA plans to review new data (e.g., health and exposure data) as
they become available and will consider the new data, as
appropriate, in future assessments.
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,
mercury removal by existing PM control devices on utilities
varies considerably, ranging from 0 to 82 percent removal, with a
median efficiency of 16 percent removal. Existing flue gas
desulfurization (FGD) units exhibit poor mercury control, ranging
from 0 to 59 percent removal, with a median removal of 17
percent. 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-16
-------
ES.7.2 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 MEI cancer risk due to multipathway exposure to
radionuclide emissions from utilities is 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. Including consideration of long-range
transport (based on extrapolation from the arsenic RELMAP
results), the cancer incidence is estimated to be roughly as high
as 2 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.
ES.7.3 Qualitative Multipathway Exposure Assessment
Other than radionuclides, the EPA has not assessed the non-
inhalation exposures of HAPs emitted from utilities. The EPA
recognizes that non-inhalation exposure pathways could be
important for other HAPs (e.g., mercury, arsenic, dioxins,
cadmium, lead) that are persistent and tend to bioaccumulate. As
indicated above, further evaluation of mercury is planned. The
other four HAPs are discussed below.
ES.7.3.1 Arsenic. Multipathway exposures potentially could
increase the total arsenic risks. Inhalation cancer risks are
estimated to be above 1 x 10'6 for arsenic for 4 plants (2 coal
and 2 oil) . Arsenic is persistent and has a tendency to
bioaccumulate. Ingestion of arsenic can pose a cancer risk, and
utilities emit approximately 59 tpy of arsenic nationwide. For
these reasons, the EPA has initiated a multipathway assessment
for arsenic.
ES-17
-------
ES.7.3.2 Dioxins. The EPA estimates that coal-fired
utilities emit 0.4 pounds per year (Ib/yr) of dioxin (toxic
equivalents, TEQ) and that oil-fired utilities emit 0.02 Ib/yr.
These estimates combined are roughly 1 to 2 percent of the
nationwide anthropogenic dioxin emissions. However, dioxin
emissions data were only available for eight test utility plants;
therefore, the emissions data for dioxins from utilities are
considered more uncertain than the emissions data for many of the
other HAPs.
The highest MEI inhalation cancer risk due to dioxin
emissions from any utility was estimated to be 1 x 10"7. The
qualitative multipathway exposure assessment indicates that
dioxins are highly persistent, tend to bioaccumulate in the food
chain, are highly toxic by low-dose ingestion exposure, and
present the greatest exposure through ingestion of contaminated
foods. Thus, although the inhalation risks are low, the EPA
believes that further evaluation of multipathway exposure for
dioxins may be needed to more comprehensively evaluate the risks.
ES.7.3.3 Cadmium and Lead. Cadmium emissions from the vast
majority of plants (i.e., 683 of the 684 plants) are estimated to
pose inhalation risks less than 1 x 10"6, and the highest modeled
air concentration of lead was 200 times below the national
ambient air quality standard (NAAQS) for lead. Yet, cadmium and
lead are persistent, may bioaccumulate, and are toxic by
ingestion. Therefore, the EPA may consider conducting further
evaluations of multipathway exposures of cadmium and lead
emissions from utilities in future analyses.
ES.7.3.4 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, at relatively low ingestion doses
(below the toxic threshold), nickel and chromium are considered
to be relatively nontoxic. Also, it is highly uncertain whether
they pose a carcinogenic risk by ingestion. Therefore, these
metals appear to be mainly a concern from inhalation exposure.
Hence, the EPA does not plan to assess multipathway exposures for
nickel and chromium for utilities.
ES.8 ALTERNATIVE CONTROL STRATEGIES
There are numerous potential alternative control
technologies and strategies for 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
ES-18
-------
nontraditional controls (e.g., demand side management [DSM],
pollution prevention, energy conservation). The degree of
feasibility, costs, 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 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.9 OTHER ISSUES AND FINDINGS
ES.9.1 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 interim report, HAP
emissions from coal-fired utilities are predicted to increase by
10 to 30 percent by the year 2010. However, based on EPA's
analysis, 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 other provisions of the Act 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 and acid rain precursors (e.g.,
add-on controls to reduce SOX and NOX emissions), these actions
could result in reductions in HAP emissions. Other potential
(but unknown) actions (e.g., fuel switching, repowering) may have
a significant impact on HAP emissions; however, these unknowns
were not included in the 2010 projection.
The approach EPA utilized to estimate emissions for the year
2010 is one of several possible approaches for making such
projections. Other organizations have made projections that
ES-19
-------
differ from EPA's. For this interim report, the EPA did not
conduct alternative approaches and did not compare its results
with projections made by other organizations. However, if
feasible, the EPA will consider evaluating alternative approaches
and comparing the EPA's results with those from other
organizations in future analyses.
ES.9.2 Peer Review
Draft versions of Chapters 1 through 10 of this report (not
including the Executive Summary) and draft technical support
documents were reviewed by numerous non-EPA scientists
representing industry, environmental groups, academia, and other
parties. The EPA held a scientific 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. 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.
Comments received by other Federal agencies that could not be
substantially addressed are presented at the end of each Chapter.
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 reasonable fee) on request by calling
the above number.
ES.9.3 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 10"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.
ES-20
-------
However, it should be noted that in the EPRI analysis, exposures
to mercury through fish consumption were only considered for two
of the four plants studied.
The EPRI risk estimates are generally similar to, but in
several cases lower than, those of EPA. Differences between the
two studies include: (1) EPA's use of a higher unit risk factor
for arsenic; (2) EPA's assumption that nickel emissions were
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) the EPRI
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,
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.9.4 Potential Environmental Impacts Not Included in Study
There are other potential environmental issues associated
with utilities not assessed in this report. First, this study
did not assess 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. Second, this study did not include an assessment of
ecological impacts. Third, this study did not assess the impacts
of carbon dioxide emissions. Fourth, this study did not assess
the impacts resulting from 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 the Act.
ES.9.5 Link to Particulate Matter (PM)
Arsenic, cadmium, chromium, lead, nickel, and radionuclides
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 (roughly two-thirds of oil-fired units are
uncontrolled for PM) . The impacts for PM were not addressed in
this study, but are being studied under Title I of the Act.
However, if additional controls of PM emissions are utilized,
this could result in reductions in HAP emissions.
ES-21
-------
ES.10 OVERALL SUMMARY
Based on this study, cancer risks due to inhalation exposure
to HAP emissions from the large majority of utility plants are
less than 1 x 10"6. However, 2 coal-fired plants and up to 22
oil-fired plants are estimated to present inhalation cancer risks
above 1 x 10"6 (primarily due to nickel, arsenic, radionuclides,
chromium, and cadmium). The inhalation cancer risks due to
exposure to the remaining HAPs emitted from utilities are
estimated to be less than 1 x 10"6. The EPA estimates that
between 0.5 and 6 cancer cases/yr occur in the U.S. each year due
to inhalation exposure to HAP emissions from utilities.
With regards to noncancer effects from inhalation exposure,
the modeling assessment indicates that HAP emissions from
utilities are not expected to result in any exceedances of the
RfCs or similar inhalation benchmarks.
Further evaluation of the impacts of the long-range
transport of HAPs and the speciation of nickel, and also the
potential impacts of short-term peak emissions of certain HAPs
(e.g., HC1, HF) , may be needed to more comprehensively evaluate
the inhalation exposures and risks.
Available information indicates that mercury emissions from
utilities may contribute to the mercury levels in the
environment, including the levels in freshwater fish. However,
at this time, the EPA has not yet determined whether the mercury
emissions from utilities are a concern for public health. The
EPA plans to continue evaluating the potential exposures and
potential public health concerns due to mercury emissions from
utilities. In addition, the EPA plans to evaluate information on
the various potential control technologies for mercury, including
pollution prevention options, and the costs, technical
feasibility of such measures, and resulting economic impacts.
The EPA plans to issue a final Report to Congress at a later date
which will include a more complete assessment of the exposures,
hazards, and risks due to utility HAP emissions, and will include
conclusions, as appropriate, regarding the significance of the
risks and impacts to public health. In addition, the EPA plans
to include in the final report a determination as to whether
regulation of HAPs from utilities under section 112 is
appropriate and necessary.
ES-22
-------
1.0 INTRODUCTION
This chapter presents an introduction to the study of
hazardous air pollutant (HAP) emissions from electric utility
steam-generating units (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 (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 mandated that
the Environmental Protection Agency (EPA) "... 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."
The list of HAPs is presented in section 112(b) of the Act.
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
• The EPA is to proceed with rule-making activities under
section 112 to control HAP emissions from utilities if
it determines from the study that such regulation is
appropriate and necessary.
The Act, in section 112(a)(8), 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 MW output to
any utility power distribution system for sale is also considered
an electric utility steam-generating unit under section
112(a) (8) .
The Code of Federal Regulations (CFR), chapter 40, part
60.41a, defines fossil fuels 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.
-------
.The wording of section 112(n)(1) (A) did not include a
mandate to include an analysis of the cost(s) of alternative
controls in the study. Therefore, no cost analyses (e.g.,
control costs, economic, cost-benefit) were performed as a part
of this interim study. These analyses would be conducted during
any regulatory effort should regulation be determined to be
appropriate and necessary in determining the level of any
standard(s).
When this study began, only a small amount of reliable data
on HAP emissions from utilities were available; most of the new
data did not become available until the beginning of 1994.
Because of the lack of data, the submission of this interim final
report for this study was delayed until October 1996. The EPA
plans to issue a final report at a later date.
This study addresses the impact of pollution controls
mandated by other sections of the Act, determines which HAPs are
present in utility unit emissions, and partially estimates
exposures and risk to humans from the emission of these HAPs.
1.2 ACT PROVISIONS AND STUDIES RELATED TO THIS STUDY
There are several other provisions in the Act that relate to
the utility industry and may have impact in the future. This
section describes these provisions and their relevance to the
study.
1.2.1 Section 112 of the Act - Hazardous Air Pollutants
The 1990 amendments to the Act also mandated 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
assessment methodologies study, (4) Great Waters study, and
(5) the Presidential Risk Commission.
1.2.1.1 Mercury Study. Section 112(n)(l)(B) required that
the EPA complete a study of mercury emissions from utilities,
municipal waste combustion units, and other sources, including
area sources. The study was 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. Section
112(n)(1)(B) mandated that the EPA submit a report to Congress by
November 15, 1994, reporting the results of the study. However,
the EPA has decided to delay significantly the release of the
mercury study report to allow time to incorporate the results of
two major studies on the impact of methylmercury on children in
1-2
-------
fish-eating populations in the Faroe Islands and Seychelles
Islands. The delay is necessary so that the EPA can completely
fulfill the mandate to evaluate the health effects of mercury.
1.2.1.2 NIEHS Health Effects of Mercury Study. In section
112(n)(1)(c), Congress gave the NIEHS the task of identifying the
threshold level of mercury exposure that would not adversely
affect human health. A report on the NIEHS study was published
in 1993.x
1.2.1.3 NAS Risk Assessment Methodologies Study. In
January 1995, the National Academy of Sciences finalized a
report2 on the risk assessment methodologies used by the EPA.
The results of the NAS study were used to help develop the
methodologies for the risk assessment portions of this study.
1.2.1.4 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.3 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 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. The May 1994 report was based on
a limited amount of existing data and will be updated in the
future.
1.2.1.5 Presidential Risk Commission. In section 303 of
Title III of the 1990 amendments to the Act, Congress directed
i •
1-3
-------
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 meets monthly
and is currently working on the draft of its report. Because the
report is not due until November 1996, the results were not
available for this report. However, the results may be
considered in the final utility report to be published at a later
date.
1.2.2 Title I - Nonattainment Provisions
Title I includes requirements for attaining and maintaining
the national ambient air quality standards (NAAQS). The NAAQS
are designed to protect public health and welfare and have been
established for six criteria pollutants (ozone, carbon monoxide
[CO], particulate matter [PM], lead, sulfur dioxide [S02] , and
nitrogen oxides [NOX] ) . Sources in ozone nonattainment areas,
including utilities, may need to install NOX controls to reduce
NOX emissions. These new NOX controls may affect HAP emissions.
Future ozone NAAQS may set even lower ozone concentration limits,
and the lower limits could result in the need for additional
utility NOX reductions. Changes in the ambient PM standard
(e.g., from PM10 to PM2.5 or PMj.) may also affect HAP emissions.
1.2.3 Title IV - Acid Deposition Control
Title IV of the Act addresses control of the pollutants
associated with acid rain in two phases. The pollutants covered
by Title IV are S02 and NOX.
Phase I and Phase II requirements of Title IV grant utility
units "allowances" to emit S02. Emission allowances are
allocated to existing utility units based upon historical
operating conditions. One allowance equals the right to emit
1 ton of S02. Affected units are required to turn in to the EPA
one allowance for each ton S02 emitted in a calendar year.
Unused or "excess" allowances may be sold on the open market. To
comply with the requirements, utilities may do many things, such
as: (1) install flue gas scrubbers, (2) switch to a fuel that
contains less sulfur ash, 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.
The Phase I requirements affect 202 boiler units at
110 utility plants. These high-S02-emitting, coal-fired utility
units must comply with the Phase I requirements by January 1995.
Under Phase II, all utility units will be covered by 2000. Both
1-4
-------
Phase I and Phase II require facilities to install continuous
emission monitoring systems for S02, N0x; and volumetric flow to
ensure compliance and provide an accurate basis for allowance
trading.
Under Phase I, the EPA required that tangential-fired and
dry-bottom wall-fired boilers meet annual average NOX emission
limits of 0.45 pound per million British thermal units (Ib/MMBtu)
and 0.50 Ib/MMBtu, respectively, by January 1, 1996. Utilities
could meet the limits by installing low-NOx burner technology or
other combustion control technology or by averaging emissions
among several units. This rule was issued as a direct final rule
on April 13, 1995 (60 FR 18751).
Under Phase II, the EPA will establish NOX emission limits
for all other boilers, including wet-bottom wall-fired boilers
and cyclones, by January 1, 1997, and the affected units must be
in compliance by 2000. The EPA will also reevaluate and revise,
if necessary, the standards established under Phase I to
implement any new technologies that could meet more stringent NOX
emissions limits (57 FR 55633). Units that do not meet the NOX
emission limits may install controls or average emissions among
several units. The effects of Title IV on utility HAP emissions
were estimated in this study.
1.2.4 New Source Performance Standards (NSPS)
Emissions of S02, NOX, and PM from utilities are subject to
NSPS for new or modified sources, pursuant to section 111 of the
Act. Units greater than 73 MW heat input that commenced
construction or modification after August 17, 1971, are subject
to requirements of the NSPS (40 CFR part 60, subparts D or Da).
Under section 407 of the Act, the EPA must revise the NSPS
requirements for NOX emissions from utility and nonutility units
to reflect improvements in emission reduction methods.
Furthermore, future NOX emission limits could be set to minimize
the multiple environmental effects of NOX on ground-level ozone
formation, ozone and nitrous oxide formation in the atmosphere,
nitrogen enrichment in water, and acid rain.
The NSPS are technology-based standards and are designed to
reflect the degree of emission limitation achievable through
application of the best demonstrated technology that is also
cost-effective. The NSPS program also 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, installation of
1-5
-------
S02 scrubbers will also control some vapor-phase HAPs such as
hydrogen chloride (HC1) and hydrogen fluoride (HF).
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 plant configurations and obtain
as much information as possible for the assessment. This report,
and the data and methodologies utilized, was 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.a
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, 8, and 9
address mercury, arsenic, dioxins, lead, cadmium, and
radionuclides noninhalation exposure. Alternative control
strategies for HAP emissions reductions are given in Chapter 10.
Chapter 11 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 and will
continue to be considered during preparation of the final report.
1-6
-------
1.4 REFERENCES
1. U.S. Department of Health and Human Services, National
Institutes of Health. Report to Congress on Methylmercury.
NIEHS, Research Triangle Park, NC. 1993.
2. National Academy of Sciences, National Research Council.
Science and Judgement in Risk Assessment. Washington DC.
1994.
3. 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.
1-7
-------
2.0 CHARACTERIZATION OF THE INDUSTRY
This chapter presents a characterization of the 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 10.
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 1990. 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 (i.e., utility) is
defined (section 112(a)(8) of the Act) as any fossil-fuel-fired
combustion unit of more than 25 MW 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 MW electrical (MWe) output to any utility power distribution
system for sale.
Utilities are fueled primarily by coal, oil, or natural gas.
Figure 2-1 shows the 1990 distribution of fossil fuels burned by
the electric utility industry by unit (boiler) and by total
megawatts.1 Coal-fired boilers account for the largest portion
of the industry by number of units (1,097 units, 56 percent),
representing 66 percent of the industry's total megawatts.
Gas-fired boilers make up 33 percent of the industry's units
(663 units) and account for 24 percent of the total megawatts.
Oil-fired boilers account for 11 percent of the units (227 units)
and represent 10 percent of the megawatts. This characterization
excludes 151 units that are effectively shut down but that still
retain operating permits.
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.
-------
V?
s^
o>
C\J
i
I
O)
o
55
CM
CO
<
CD
OS
OS
t-
c
•p*
>>
€
e
O
O
2-2
-------
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 while 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 (U.S.). In 1990,
there were 1,893 conventional utility steam-generating units in
the U.S., with 1,043 burning coal of some type. The total output
was 493.1 gigawatts (GW) electrical.
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 U.S. 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 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
2-3
-------
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 260 separate companies, investor-
owned utilities provide 75.4 percent of kilowatt hour (kWh)
generation of electric power to the Nation. Publicly owned
utility companies, which consist of approximately 2,017 separate
companies, represent 10.6 percent of the Nation's electric power
supply. The 10 Federal power agencies generate 8.6 percent of
the Nation's electric power supply. Rural electric cooperatives,
numbering approximately 939 separate companies, serve 5.4 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
The fastest growth in the production of electricity (by
unit) for the 1990s has been projected for nonutility
generators.6 Ownership of nonutility generators can be further
divided into ownership by:
• Units that cogenerate steam and electricity (qualifying
facilities7)a
• Small power producers (<80 MWe) that generate
electricity primarily from a renewable source
• Other nonutility generators (e.g., independent power
producers [IPPs], units that cogenerate steam and
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."
2-4
-------
electricity [nonqualifying facilities], and other
commercial and industrial units).
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.
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 (FBC).
2.3.1.1 Stoker-fired. 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.8 The 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.9 Because of their design,
stokers are used only for smaller furnaces firing coal.
2.3.1.2 Cyclone-fired. 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 furnace. Because of this slagging system,
cyclone-firing furnaces are almost exclusively coal-fired;
however, some units can fire oil.10
2.3.1.3 Tangential-fired. 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
2-5
-------
to fill the furnace.11 Tangential-fired boilers can fire coal,
gas, or'oil.
2.3.1.4 Wall-fired. 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 more than one wall that
face each other). Circular register burners and cell burners are
types of burner configurations found in single-wall or
opposed-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.12 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.13 "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.14
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, while the BFB combustors have relatively low velocities and
coarse bed-particle size.15'16
Most FBCs are of the atmospheric fluidized-bed (AFBC) type,
which, as the name suggests, operate at atmospheric pressure. A
newer and potentially promising 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.17
2-6
-------
.2.3.1.6 Distribution of Furnace Types. Figure 2-2 shows
the 1990 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 (50.2 percent), which represents 47.8 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 39.8 percent of the units (43.6
percent of the total megawatts), and cyclone firing is used in
8.4 percent of the units (8.3 percent of the total megawatts).
Stoker-fired boilers and FBCs account for about 1.6 percent of
designs among the coal-fired units (0.3 percent of the total
coal-fired megawatts). Wall-fired designs represent the largest
portion of gas- and oil-fired units by number of units
(68 percent), which represents 64 percent of the total megawatts.
The second most common design is the tangential-fired unit.
Tangential-fired units represent 30 percent (34 percent of the
total megawatts) of the gas- and oil-fired units, and combined-
cycle gas turbine units account for about 2 percent (2 percent of
the total megawatts) of designs for 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. In comparison,
essentially all elemental HAPs leaving the furnace enter into 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 10 provides a discussion, from
limited data, suggesting that, for example, organic HAP emissions
are increased as furnace conditions are changed. Similarly for
elemental HAPs, chapter 10 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 modeling described later in this report. Appendix D
describes the construction of the models 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
2-7
-------
D
LU
GENTIAL FIR
2
<
a
Q
LU
CC
u_
_J
_J
1
H
a
LU
CC
u.
LU
CD
o
O
0
Q
UJ
CC
li-
CC
UJ
X.
o
fe
0
Q
UJ
CO
Q
LU
N
Q
LL
•
LU
IBINED-CYCl
TURBINE
«£ UJ
0<
oo
m
1
0)
CD
E
15
o
O
1
CO
O)
a>
E
as
w
CO
O
c
3
•a
CO
CO
CO
CD
ON
fH
c
• V*
V
a
^
»HM
0)
I?
,g
>%
•PN
c/j
C
P
=5
ex
2-8
-------
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.18
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.19"21
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, such as a combustion gas turbine with
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). Figure 2-3 illustrates the 1990 distribution of PM
control by fuel in the utility industry by unit and by total MW.1
In 1990, 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
2-9
-------
CO
UJ
1 1
—
lo"
UJ
CO
Fs (BAGHOU
u.
•B
^^
DC
UJ
CD
CD
rr
ARTICLE SCI
Q.
\7\
\£~l
lECHANICAL
2:
5^1
l£\J
O CONTROL
z
s
Li^J
w
o
o
,Q-
strostatic prec
-------
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 (1990) the largest portion of
the oil-fired units (60 percent) and accounted for 52 percent of
the oil-fired industry's total MWs. In 1990, ESPs were used on
22 percent of the oil-fired units or at 26 percent of the
capacity of the oil-fired industry. Mechanical controls
(cyclones) were used on 18 percent of the oil-fired units
(22 percent of the total MWs). 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-um to 50-um range22 (on a mass basis), HAPs tend to concentrate
preferentially on particles smaller than about 7 urn, and
especially on those around 0.3 urn.23 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. Many of the efficiency data by
particle size originate 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 um.
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-um particles,
but may drop to less than 20 percent for 1-um particles.24 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.25
2-11
-------
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 which 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 um are typically
collected with efficiencies from 95 to 99.9 percent.26 Particles
near the 0.3 um size are in a charging transition region that
reduces collection efficiency." These particles have been shown
to have lower collection efficiency (about 50 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.28 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.29 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 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
2-12
-------
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 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 yum, efficiency
may be reduced to less than 50 percent.30 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.31
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 /urn (the range of the
measuring equipment).32 Because of its high collection
efficiency for small particles, the baghouse should be
particularly effective for removing particles that have been
enriched with HAPs.33'34 However, further study is required to
determine if baghouses can remove significantly greater
quantities of HAPs than are removed by other control systems.
2-13
-------
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 10-6 and 10-9, 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.
2.5 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 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 1990, all oil-
and gas-fired units burn compliance fuel, while 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, arid 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 SO2 from
five coal-fired units, representing a total capacity of 610 MWe.
These units are FBCs and control S02 in the combustion zone by
using limestone as a sorbent.
2-14
-------
Table 2-1. Comparison of Particulate Matter Collection Systems
35
Collector
Multicyclone
ESPs
Particle scrubber
FFs
Typical mass
efficiency, %
70-90
99 - 99.7
95-99
99-99.9
Efficiency at
0.3 Mm, %
0- 15
80-95
30-85
99 - 99.8
Energy consumption, in. H,0
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.
Because of the 1990 amendments, oil- and natural-gas-fired
units now burn compliance fuels that combust with lower S02
emissions. Using compliance fuels allows the units to avoid
postcombustion scrubbing.1 However, approximately 15 percent of
the coal-fired units use postcombustion control of S02
emissions.1 The rest of this section describes precombustion
techniques and postcombustion S02 control devices, namely
scrubbers, that are used in the coal-fired utility industry.
Figure 2-4 shows S02 control devices used in coal-fired
utilities in 1990 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 scrubber was used at approximately
14 percent of the units (approximately 21 percent of the
coal-fired total electric capacity), while 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 SO2 emission limits of 1.2 Ib
S02/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
2-15
-------
LU
O
^>
LU
D
O
CC
1-
O
O
o
Z
D
CC
LU
CO
CO
CC
0
CO
Q
o
LL
m
LU
CO
O
I
3
m
LL
^
Q
CO
H
o
05
CO
D)
CD
E
o
o
OS
t-
T3
c
cy
=
•p*
"o
o
c
3
>»
CO
2-16
-------
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. 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.36
Approximately 77 percent of the eastern and midwestern bituminous
coal shipments are subjected to some physical cleaning process.37
Subbituminous and lignite coals are not routinely cleaned.38-39
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.40
Bituminous coals from the eastern U.S., 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.41
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.37-42 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, 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 10.1.2.
2.5.2 Postcombustion Control: Flue Gas Scrubbing for S07 Control
According to the 1992 compilation of the Edison Electric
Institute's (EEI) Power Statistics database (examining 1990
data), scrubbers were installed on 152 boiler units (out of about
1,043 coal-fired units in the U.S.) with a total rated capacity
of 68,695 MWe.1 Table 2-2 lists the different types of scrubbing
installations used in U.S. utility power plants. As shown in
Table 2-2, wet limestone/lime slurry scrubbing represents the
2-17
-------
Table 2-2. Distribution of S02 Control Technologies in 19901
Wet limestone
Wet lime
Dry lime/SDA
Sodium carbonate
Dual-alkali
Wellman-Lord
Mag-Ox
Dry aqueous carbonate
No. of boiler units
69
45
15
9
6
4
3
1
Installed FGD capacity,
MWe
34,521
19,977
5,626
3,181
2,267
1,779
895
450
Total
Total percent of
installed FGD capacity,
%
50.3
29.0
8.2
4.6
3.3
2.6
1.3
0.7
100.0
FGD = Fluidized gas desulfurization.
SDA = Spray dryer adsorber.
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.43
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
2-18
-------
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.44
The basic wet limestone scrubbing process is simple and well
established. Limestone sorbent is cheap and generally locally
available in the U.S. The S02 removal efficiencies of existing
wet limestone scrubbers range from 52 to 95 percent, with an
average of 85 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.
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 1990, wet limestone scrubbers were used at 69 units,
or at 34,521 MWe of the total scrubbing capacity.
2.5.2.2 Wet Lime. In a wet lime scrubber, flue gas
containing S02 is contacted with hydrated lime-water slurry; the
S02 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.45
Wet lime scrubbing is a proven technology; considerable
operating experience has been gained in 45 utility units.1 These
units represented 19,977 MWe of the total scrubbing capacity in
1990. The S02 removal efficiencies of existing wet lime
scrubbers range from 72 to 99 percent. Recent advances include
the use of additives to improve performance, reduce scaling
problems, and produce a salable gypsum byproduct. Lime scrubbing
processes require consideration of 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 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.46 The water is evaporated by the heat of the flue
gas. The dried solids are entrained in the flue gas, along with
2-19
-------
fly ash, and are collected in a PM collection device. Most of
the SO2 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 90 percent.1 Spray dryers were used at 15 units and
constituted 5,626 MWe of scrubbing capacity in 1990.
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
1990) using wet sodium carbonate scrubbing in the U.S.,
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 53 to 91 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 U.S., 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.47 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 (1990) on six
boiler units in the U.S. with a combined capacity of 2,267 MWe.
2-20
-------
The S02 removal efficiencies at these plants range from 89 to 95
percent.1 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 Wellman-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-t8 The treated flue gas is
demisted and reheated before it is exhausted through a stack.
The Wellman-Lord process has been installed on four U.S. boiler
units with a combined capacity of 1,779 MWe (1990), with S02
removal efficiencies ranging from 85 to 90 percent.1
2.5.2.7 Magnesium Oxide. Similar to Wellman-Lord, the
magnesium oxide (MAG-OX) fluidized gas desulfurization (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.49
Because of the high-temperature regeneration step at 800° to
1,000° C (1,472° to 1,832° P), 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 U.S. with
a combined capacity of 895 MWe (1990) . The S02 removal
efficiencies at these plants range from 92 to 95 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, consisting of 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
2-21
-------
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.50 Only one unit, of
450 MWe capacity, uses the dry aqueous carbonate system for FGD
(1990) ; it has a 70 percent S02 removal efficiency.1
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. 51~54
Figure 2-5 shows NOX control approaches used in 1990 based
on the number of units and total MW capacity.1 Around 81 percent
of coal-fired plants, representing about 67 percent of the coal-
fired MW capacity, had no NOX control, while around 19 percent of
the units, representing about 33 percent of the coal-fired MW
capacity, used some kind of NOX control. Approximately
73 percent of the gas- and oil-fired units, with about 62 percent
of the MW capacity, did not use NOX control, while approximately
27 percent of the units, representing about 38 percent of the
gas- and oil-fired MW capacity, used some kind of NOX control.
The chemical species nitrogen dioxide (N02) and nitrogen
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 N02.55 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.56
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) ,57 Fuel switching,
then, may yield reduced NOX emissions.
The formation of NOX for coal-fired units depends on factors
such as the type of boiler, type of burner, and facility
operation.58 Any of these factors that increase temperature or
residence time at high temperature will promote NOX formation.59
In general, cyclone and other wet-bottom boilers have relatively
higher NOX emissions, with an approximate range of 1 to 2
2-22
-------
D
HI
8
Q
LU
d
cr
8
n
OS
ON
vH
C
•**
a
§
co
D)
V
E
5
o
O
O)
"CO
O
O
co
1
D)
0
T3
C
CO
CO
CO
O
>»
.S
"o
o
c
3
'6
T3
CO
CO
CO
CD
fi
0)
Dfi
O
2-23
-------
Ib/MMBtu, than do dry-bottom boilers, which have an approximate
range of 0.4 to 1.5 Ib/MMBtu.60 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 lb/106 Btu, 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.61 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 emissions.62
The reduction of NOX emissions is important for controlling
acid rain and ozone formation.63 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.64 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.65 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).66 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.67
Low NOX burners may be used in coal-, oil-, and gas-fired
boilers to lower NOX emissions by about 25 to 55 percent.68
Overfire air may be used as a single NOX control technique, with
NOX reductions of 15 to 50 percent.69'70 When OFA is combined with
low NOX burners, reductions of up to 60 percent may result.71 The
actual NOX reduction achieved with a given control technique may
vary from site to site.72
2-24
-------
. 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.73 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.74'75 Reburn with other fuels, primarily coal, is currently
under development, as are improvements in the process.76
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.77 By
returning part of the flue gas to the primary combustion zone
(FGR), the flame temperature and the concentration of oxygen in
the primary combustion 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.
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),78
However, this technique is not used for utility boilers because
it has a high efficiency penalty (about 10 percent).7S As shown
in Table 2-3, approximately 36 percent of the combined-cycle
turbine units used steam or water injection for NOX control in
1990, while only approximately 2 percent of the boilers reported
using this technique. Temperatures may also be reduced in the
primary combustion zone by increasing the spacing between burners
for greater heat transfer to heat-absorbing surfaces.79 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 and soot emissions.80
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
2-25
-------
Table 2-3. Distribution of NOX Control by Fuel Burned, by Unit,
in 1990 l-a
Fuel
Coal, Boiler bottom
Dry
Wet
Oil
Gas
Combined-cycle turbine
Percent NOX control
None
79 (64)
97 (97)
73 (66)
73 (60)
64 (63)
Staged
combustion
18(32)
3(3)
26 (33)
25 (38)
-
Boiler design
3(4)
-
-
< 1
-
Staged combustion with
injection of water or steam
_
-
--
2(2)
36 (37)b
8 Values listed in parentheses are the percent distribution by MWe for each type of fuel.
Only steam or water injection.
by reducing NOX emissions only (selective noncatalytic reduction
[SNCR],) or by reducing combined emissions of CO, hydrocarbons,
and NOX (selective catalytic reduction [SCR]).81 Postcombustion
control has, to date, 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.82 Selective catalytic reduction has also
been implemented at an IPP Cogen plant in New Jersey.3
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.83 Selective catalytic reduction units provide up to
70 to 90 percent NOX reduction84 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) .8S
Selective noncatalytic reduction is currently achieved
commercially in one of two ways: THERMAL DeNOx®, an Exxon
process, or NOXOUT®, an EPRI process. THERMAL DeNOx® 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.86
2-26
-------
. 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.
Table 2-3 presents a general breakdown of utility industry
NOX control usage according to the 1990 EEI power statistics
database.1 As shown in Table 2-3, most of the utility industry
has no NOX control; 79 percent of the dry-bottom coal-fired
boiler units, 97 percent of the wet-bottom coal-fired boiler
units, 73 percent of the oil- and gas-fired boiler units, and
64 percent of the combined-cycle turbine units had no NOX control
in 1990. 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 18
percent of the dry-bottom coal-fired units, 3 percent of the
wet-bottom coal-fired units, 26 percent of the oil-fired units,
and 25 percent of the gas-fired boiler units. As previously
noted, steam or water injection was used for NOX control in
approximately 36 percent of the combined cycle units. Table 2-3
also shows that approximately 3 percent of the dry-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 1990 amendments 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 the Acid Rain Provision
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, re-powering, and
penalties. Although these provisions affect the manner in which
2-27
-------
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.
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.87 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 Acid Rain Division of the
EPA's Office of Air and Radiation by ICF Resources Incorporated.
This single projection was used by the Office of Air Quality
Planning and Standards (OAQPS) to maintain consistency with the
Acid Rain Division. This projection may be compared with others
for future analyses.
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). On this basis, the
predominant fuel both in 1990 and projected for 2010 is coal, at
approximately 81 percent of the total industry fuel usage. Oil
and gas consumption in 1990 were, respectively, 6 and 13 percent
of the total industry fuel usage on a Btu/yr basis. For 2010,
oil consumption was projected to decrease to 2 percent, and gas
consumption was projected to increase to 17 percent on a Btu/yr
basis for the total industry fuel usage.88 Projected coal
consumption in 2010 is expected to be the same percentage of the
total utility fossil fuel usage as in 1990 (81 percent).
Figure 2-7 shows the projected growth of each utility fuel
between 1990 and 2010.88 Between 1990 and 2010, the consumption
of coal, oil, and natural gas is projected to increase by
32 percent, decrease by 50 percent, and increase by 59 percent,
respectively.
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 and oil) 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 cannot estimate the
2-28
-------
0
o
CM
>»
3
4->
to
"55
u_
tD
O
T"-
X
CM
CM
"rt
o
0
Tf
'o
X
CM
CO
5
m
O
X
10
•"fr
V)
as
O
00
S
O
5 •**
£
O r-
* S3
3 C8
— O
. e
00
3
o
•^
>.
3
*->
m
"35
u.
(O
0
X
N-
T"
15
o
0
in
0
X
CM
"r~
6
in
O
X
CO
CM
(I)
CO
O
2-29
-------
0)
s
IT)
W
T3
CO
o
2010
years
Gas
Oil
Coal
Figure 2-7. Projected use of fuels by 2010 for
utility industry.88
2-30
-------
size.nor the location of the new units, the increased consumption
has been allocated to existing units (in 2010) for the analyses.
The decrease in oil consumption could result in units being
retired or in a decrease in capacity factor, or a mix. The
decreased consumption has been allocated among those oil-fired
units believed to still be operating 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.
The Agency projects that 135 units will be retired during
the period between 1990 and 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
Provisions, listed in Title IV, establish a cap on the national,
annual S02 emissions. To achieve compliance with the
requirements, utilities may do 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 not natural gas-fired) replacement,
(4) use of a Phase II compensating unit, or (5) purchase of
emission allowances. Emission allowances are allocated to
existing utility units based upon historical operating
conditions. One allowance equals the right to emit 1 ton S02.
Affected units are required to turn in to the EPA one allowance
for each ton S02 emitted in a calendar year. Unused or "excess"
allowances may be sold on the open market.
The Phase I requirements affect 202 units at 110 plants,
which must comply with the Phase I requirements by January 1996.
To date, only 27 units at 16 plants (total generation capacity of
about 14,058 MWe) have announced plans to install scrubbers to
meet the Phase I requirements. The remaining 175 units
(94 plants) will comply with Phase I requirements by (as
mentioned above) either fuel switching (to low-sulfur coal), by
buying more allowances than allocated, or by otherwise having
enough allowances at the end of 1995 to cover their emissions.85
Nearly 50 percent of the Phase I units plan to fuel switch or to
2-31
-------
blend to obtain low-sulfur coal. Each of the 27 units known to
be installing scrubber units was modeled with the scrubber unit
for the 2010 scenario. In the absence of specific plans for the
remaining 175 units in Phase I, EPA modeled these 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 Provisions, all utility
units will be covered by 2000. 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 using alternate
methods.90 This assumption was based on several factors
including: (1) the increased availability of low-sulfur coal at
favorable prices; (2) the introduction of processes that reduce
sulfur 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.91
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 vs. 2010 analysis. No change in a
unit's burner configuration (i.e., "old" vs. new low-NOx) was
included in the 2010 scenario. The impact of low-NOx burner
installation is discussed in chapter 10.
Under the acid rain program, the rules for NOX control
require that tangential-fired and dry-bottom wall-fired boilers
subject to Phase I S02 reduction requirements also meet annual
average NOX emission limits of 0.45 Ib/MMBtu arid 0.50 Ib/MMBtu,
respectively, by January 1, 1996. Utilities that did not meet
the limits were to comply with the regulation by installing low-
NOX burner technology or by averaging emissions among several
units. This rule was issued as a direct final rule on
April 13, 1995 (60 FR 18751).
Under Phase II of the acid rain program, the EPA will
establish NOX emission limits for all other boilers, including
wet-bottom wall-fired boilers and cyclones, by January 1, 1997,
2-32
-------
and the affected units must be in compliance by 2000.92 Also by
1997, the EPA will determine if more stringent emission limits
should be established based on technology developments for
dry-bottom wall-fired boilers or tangential-fired Phase II units.
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.90 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.93 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.94 It is not known how transient these excess
emissions will be.
2.8 DISCUSSION OF FEDERAL INTERAGENCY REVIEW COMMENTS
Previous drafts of chapters 1 through 10, along with the
appendices, were reviewed by numerous non-EPA scientists
representing industry, environmental groups, academia, and other
Federal Agencies during the summer of 1995. 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. The EPA has revised the report, as
appropriate, based on the reviewers' comments. The EPA revised
the report to incorporate the majority of the comments received.
However, there were several comments that could not be fully
addressed because of limitations in data, methods, and resources.
This section presents comments received by other Federal agencies
that could not be substantially addressed in this interim report.
2-33
-------
2.8.1 Comment
The Department of Energy (DOE) commented that w[p]ther 2010
utility industry scenario forecasts should be considered..." and
the utility industry scenario forecasts used by the EPA should be
compared to those of other groups (e.g., Edison Electric, EIA,
GRI, The WEFA Group, Data Resources, Inc., etc.).95
2.8.2 Response
The EPA added discussions of the uncertainties and
limitations to the approach used to estimate emissions for 2010.
Also, the EPA acknowledges that other methods and other
projections exist. However, the EPA did not evaluate or present
the alternative approaches or projections in this report. To the
extent feasible, the EPA plans to review other industry growth
scenarios and projections of what the industry will look like in
2010 before issuing the final report.
2.8.3 Comments
The DOE also commented that "...EPA has chosen to describe
utility sector emissions in a manner which misrepresents to
Congress both the present and the future emissions...and thus
overstates the argument for regulation." In addition, the DOE
commented that the EPA had not captured "...the effects of fuel
switching, fuel cleaning, combustion controls, and
post-combustion controls implemented since [1991]" which have
reduced air toxics. This omission leads to an overestimate of
air toxic emissions.96
Similarly, the Council of Economic Advisors (CEA) commented
on the EPA's growth estimates and noted that w[n]o justification
is provided...to explain the assumed lack of improvement in
either generating efficiency or abatement technology that would
make it possible to meet the increased demand for electricity
without causing as much pollution. In particular, the report
seems to underestimate the potential of biomass as fuel. The
predicted growth in demand for electricity for the 1990 to 1995
period exceeds actual industry experience, raising doubts about
the accuracy of the predicted growth in emission from 1990 to
2010. Also, the report does not adequately examine the
possibility of substituting demand side management for expansion
of output." 97
2.8.4 Response
As noted in this chapter, the baseline was chosen to
characterize the industry as it stood, and to quantify the risk
to human health, as of 1990, the most recent year for which
facility data were available when the project started. This
2-34
-------
approach is believed to be appropriate and consistent with the
mandate of section 112 (n) (1) (A) .
As to future emissions, the EPA has incorporated into its
2010 analysis those changes announced by utilities for compliance
with Title IV of the Act. The EPA is unaware of any significant
changes in the industry over the past 5 years (or planned for the
next 10 years) that would increase generating efficiency or HAP
emission removal efficiency to any great extent. The EPA notes
that fuel use has increased in the past 5 years which may have
had the effect of negating any potential HAP emission reductions
accruing from increased use of control devices. The use of the
control measures noted by the DOE and CEA have been addressed in
the interim report (see chapter 10) to the extent possible using
current information. Further analyses may be performed for the
final Report to Congress should additional data become available.
No change in a unit's combustion control efficiency was made
to account for increased use of NOX controls in 2010. Existing
data indicate that combustion controls would not significantly
affect HAP emissions. Where new controls for SO2 were known to
be installed, these controls were employed in the 2010 analyses.
Where EPA has estimated the impact of new facilities, the
emission reductions are based on compliance with the new source
performance standards (NSPS) for both PM and S02 control.
Although the EPA believes its approach is reasonable, the
EPA also recognizes that there are uncertainties in the
assessments. The EPA intends to address these uncertainties and
consider the above comments, to the extent feasible and
appropriate, before issuing a final report. In addition, the EPA
intends to address other factors that may impact on emission
projections, including improvements in HAP emission removal by
FGD units and the impact of any actions taken under Title I of
the Act (e.g., any tightening of the NAAQS for PM) before issuing
a final report.
2-35
-------
2.9 .REFERENCES
1. Utility Data Institute. EEI Power Statistics Data Base.
Washington, DC. 1992. (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
1995. pp. 76,78.
6. Utility Data Institute. State Directory of New Electric
Power Plants. Third Edition. Washington, DC. 1992.
pp. 2,13.
7. §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.
8. Ref. 2, p. 4-29.
9. Ref. 2, p. 12-15.
10. Babcox & Wilcox. Steam, Its Generation and Use. New York.
1978. p. 10-3.
11. Ref. 2, p. 12-4.
12. Ref. 2, p. 12-2.
13. Elliot, T. C. Standard Handbook of Powerplant Engineering.
McGraw-Hill, Inc., New York. 1989. p. 4.50.
14. Ref. 2, p. 9-2.
15. Ref. 2, p. 9-4.
16. Ref. 2, p. 9-5.
2-36
-------
17. .Ref. 2, p. 9-30.
18. Ref. 2, p. 7-6.
19. Ref. 2, p. 1-13.
20. Federal Energy Regulatory Commission. Cogeneration.
Washington, DC. 1985. pp. 3, 5.
21. Ref. 2, p. 1-16.
22. Darnle, A. S., D. S. Ensor, and M. B. Ranade. Coal
Combustion Aerosol Formation Mechanisms: A Review. Aerosol
Science and Technology. 1 (1):119-132. 1982.
23. 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.
24. U.S. Environmental Protection Agency. Control Techniques
for Particulate Emissions from Stationary Sources --
Volumes I and 2. EPA-450/3-81-005a, b. Research Triangle
Park, NC. 1982. p. 4.2-23.
25. Leith, D., and D. Mehta. Cyclone Performance and Design.
Atmospheric Environ. 7:527-549. 1973.
26. Ref. 24, Vol. 1, pp. 4.3-20 and 4.3-22.
27. 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.
28. Ref. 24, Vol. 1, pp. 4.3-14 to 4.3-23.
29. White, H. J. Industrial Electrostatic Precipitation.
Addison-Wesley, Reading MA. 1963.
30. Ref. 24, Vol. 1, pp. 4.5-22 to 4.5-29.
31. 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.
32. Ref. 24, Vol. 1, p. 4.4-12.
2-37
-------
33. .McKenna, J. D., and J. H. Turner. Fabric Filter - Baghouses
I, Theory, Design, and Selection (A Reference Text). ETS,
Inc., Roanoke, VA. 1989.
34. Donovan, R. P. Fabric Filtration for Combustion Sources:
Fundamentals and Basic Technology. Marcel Dekker, New York.
1985.
35. Memorandum from J. H. Turner, RTI, to J. D. Cole, RTI.
March 17, 1993. Table presenting a comparison of
particulate matter collection systems.
36. Khoury, D.L. Coal Cleaning Technology. Noyes Data
Corporation, Park Ridge, NJ. 1981. p. 24.
37. 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.
38. 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.
39. 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.
40. Ref. 36, p. 21.
41. Ref. 39, p. 1.
42. 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.
43. Satriana, M. New Developments in Flue Gas Desulfurization
Technology. Noyes Data Corporation, Park Ridge, NJ. 1981.
p. 10.
44. Ref. 43, pp. 13-14.
45. Ref. 43, p. 15.
2-38
-------
46. .Ref. 43, pp. 180-181.
47. Ref. 43, pp. 85-88.
48. Ref. 43, pp. 155-161.
49. Ref. 43, pp. 142-146.
50. Ref. 43, pp. 75-78.
51. Ref. 2, pp. 4-30 to 4-34, 12-6 to 12-8, 15-64 to 15-68.
52. 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.
53. U.S. Environmental Protection Agency. Summary of NOX
Control Technologies and Their Availability and Extent of
Application. EPA-450/3-92-004. February 1992. pp. 1-1 to
3-25.
54. 1991 Joint Symposium on Stationary Combustion NOX Control,
EPA/EPRI, March 25-28, 1991.
55. Ref. 2, p. 4-30.
56. Ref. 53, p. 2-3.
57. 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.
58. Ref. 57, p. 1.
59. Low-N0x Combustion Retrofit Projects, Pittsburgh Energy
Technology Center. Pittsburgh, PA. Review, 6, Summer 1992,
p. 9.
60. Ref. 59, p. 9, Figure 2. Values represent emissions before
the 1993 new source performance standards.
61. Ref. 59, p. 9.
62. Ref. 2, p. 9-4.
63. Ref. 2, p. 4-30.
2-39
-------
64. .Ref. 53, p. 2-1.
65. Ref. 57, p. 5.
66. Ref. 53, p. 2-4.
67. Davis, Wayne T., and Arijit Pakrasi. In: Air Pollution
Engineering Manual, Buonicore, Anthony J., and Wayne T.
Davis (eds). Van Nostrand Reinhold, New York. 1992.
p. 241.
68. Ref. 52, pp. 4-4, 4-25.
69. Ref. 52, p. 4-4.
70. Ref. 53, p. 3-19.
71. Ref. 52, p. 4-4.
72. Ref. 53, p. 3-8.
73. Ref. 53, p. 2-8.
74. Borio, R., R. Lewis, D. Steen, and A. Lookman Long Term NOX
Emissions Results with Natural Gas Reburning 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.
75. LaFlesh, R. C., R. D. Lewis, R. E. Hall, V. R. Kolter, and
Y. M. Mospan. Three-Stage Combustion (Ruburning) 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.
76. Ref. 57, p. 2.
77. Ref. 53, pp. 2-7 to 2-14.
78. Ref. 53, p. 2-10.
79. Ref. 53, p. 2-14.
80. Ref. 53, pp. 2-14, 2-15.
81. Ref. 53, p. 2-16.
82. Ref. 52, p. 4-32.
2-40
-------
83. . Ref. 53, p. 2-16.
84. Ref. 52, p. 4-35.
85. Ref. 52, pp. 4-35, 4-37.
86. Ref. 53, p. 2-17.
87. 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, D.C. February 1994. p. 8.
88. Ref. 87, p. C-l.
89. Fax from S. Terry, EPA/Acid Rain Division, to W. Maxwell,
EPA/ESD/CG. September 15, 1993. "Plants and Units with
Acid Rain Phase I Applications - Explanation of Type."
pp. 2,13.
90. Letter from Zeugin, L. B., Hunton & Williams, to W. H.
Maxwell, EPA. April 11, 1994. UDI Particulate Control
Equipment Survey.
91. J. Makansi. New Supply Lines, Tools Keep Coal Switching in
Forefront Power. May 1994. p. 41.
92. Federal Register, 57(228), 1992, p. 55633.
93. 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.
94. Letter from Schultz, D., EPA (Region 5), to W. H. Maxwell,
EPA. July 18, 1995. Future year (2010) projection of
utility HAP emissions.
95. Letter from Chupka, Marc W., DOE, to Maxwell, William H.,
EPA/CG. May 3, 1996.
96. Letter from Chupka, Marc W., DOE, to Maxwell, William H.,
EPA/CG. May 3, 1996.
97. Memorandum and attachment from Munnell, Alicia, CEA, to
Maxwell, William H., EPA/CG. May 6, 1996.
2-41
-------
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 literature data 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 from 1990 to 1994 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 during 1993 and 1994.
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 tested
emissions at locations of about 25 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 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
test reports from eight of its coal-fired plants for testing
performed from 1990 to 1992.
The DOE, through its Pittsburgh Energy Technology Center
(PETC), 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 the DOE
program.
The EPA also completed the initial development of the
Fourier transform infrared spectrometry (FTIR) field testing
3-2
-------
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.
For inclusion in this report, a total of 52 tests were
conducted at 48 sites using FTIR and conventional sampling and
analysis methods from the EPRI, the DOE, the NSPC, and the EPA.
Although 52 test reports were received by EPA in time for
inclusion in this study, 3 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.
The data reliability and the precision and accuracy of the
sampling techniques were addressed by the individual test
contractors in their test reports. If a contractor had concerns
about the quality of the data or about the precision or accuracy
of a particular test sample, the EPA did not use the data in its
computations.
3.2 POLLUTANTS STUDIED
As many of the 189 HAPs listed in section 112(b) as possible
were included in this study. Table A-l (Appendix A) lists the
HAPs that were detected at least once in the utility test data
(excluding FTIR-detected data), their estimated nationwide
emissions in 1990, and 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
-------
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 U.S. (plus
any additional unit types tested that were below this cutoff).
The matrix was then used 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.
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 U25 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, while 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
-------
USGS coal data
(by State and coal type)
Apply coal cleaning
factor
UDI/EEI plant configuration
information
USGS coal data
(by State and coal type)
No cleaning factor
Trace elements (TE) to boiler
Apply boiler TE emission
modification factor
What is
paniculate
^matter (PM) control,
type?*
Apply the PM control TE
emission modification factor
What is
the SOg control
type?*
Apply SO2control TE emission
modification factor
kg/yr of specific trace element
exiting unit stack
Figure 3-1. Trace elements in coal.
3-5
*Taken from UDI/EEI data.
-------
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
participate
^matter (PM) control,,
type?*
Apply the PM control TE
emission modification factor
What is
the S(>2 control
type?*
Apply SCfe 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
-------
Type of fuel
burning?
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/109 cu ft emission
factor for a specific
HAP3
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/lbcoal)
Individual fuel
consumption x
emission factor x
higher heating value
for subbituminous
coal (9,967 Btu/lb
coal)
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
-------
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 UDI/EEI Power Statistics database (19"91
edition). The UDI/EEI database is 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
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 State of origin in the U.S.
Geological Survey (USGS) database, which analyzed core and
channel samples (3,331 samples) of coal from the top 50 (1990 or
later) economically feasible coal seams in the U.S.
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). 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.
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 ^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.
Coal-fired boiler concentrations were modified for different
heating values, depending on the type of coal, before being
converted to a rate basis (kilograms per year [kg/yr] of
individual HAP). This procedure was necessary because different
coal ranks have different heating values. For example, it would
3-8
-------
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 3-1).5
The 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.6 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.
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.7 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 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 8 to meet customer specifications on
heat, ash, and sulfur content, a coal cleaning factor was applied
to most bituminous coals in the EFP.9
For a unit that burned bituminous coal, the kg/yr feed rate
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 Equations
No. 1 and No. 2 in Table D-2, Appendix D).
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
3-9
-------
Table 3-1. Average Higher Heating Values of Coal
10
Class and group*
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
Agglomerating
character
^.^jff::'
commonly
agglomerating*
*
-
-
-
agglomerating
Fixed carbon
ImJts. % (dry.
free basis)
Equal or
greBter
than
78
69
-
-
-
-
Less
than
86
78
69
—
—
—
% (dry. mineral-
matter-fne basis)
Equator
Qfttter
than
14
22
31
—
—
-
Less
than
22
31
—
~-
—
—
Calorific value limits.
Btu/lb (moist,"
basis)
Equal or
greater
than
\ ;
—
—
14,000"
13,000"
1 1 ,500
10,500
Less
than
—
—
—
14,000
1 3,000
11,500
Average of Averages (Value used in EFP for bituminous coal)
II. Subbftuminou*
1 . Subbituminous A Coal
2. Subbituminous B Coal
3. Subbituminous C Coal
nonagglomerating
-
•
—
—
—
-
—
—
-
—
—
-
—
10,500
9,500
8,300
11,500
10,500
9,500
Average of Averages (Value used in EFP for Subbituminous coal)
III. Ugnhte
1 . Lignite A
2. Lignite B
nonagglomerating
-
-
-
-
-
—
-
—
—
6,300
—
8,300
6,300
Average of Averages (Value used in EFP for lignite coal)
Average
14,000
13,500
12,250
1 1 ,000
12.688
1 1 ,000
10,000
8,900
9,967
7,300
6,300
6,500
• 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.
c 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.
3-10
-------
test.reports 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 HCl and HF Concentration in Fuel
To obtain hydrogen chloride (HCl) 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 HCl or HF
throughout the boiler system.11 For example, for each part per
million of chloride in the feed coal at one of the test sites,
0.61 Ib/h of HCl was found in the gas stream leaving the boiler
and 0.00145 Ib/h in the stack gas. Similarly for HF, the boiler
emissions were 0.56 Ib/h for each part per million of fluoride in
the coal and 0.00448 Ib/h in the stack. For ease of programming,
the HCl 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 0.61 or 0.56,
respectively. The resulting numbers allowed direct conversion
into boiler emissions that could be further modified for systems
with PM control or S02 control.
The chloride concentrations were not available for coals
from the following States: Alaska, Illinois, Indiana, Iowa,
Missouri, Utah, and Washington. Chloride concentrations were
assigned, as shown in Table 3-2, for coals originating from these
States.12
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 emission modification factors (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.13 These EMFs were averaged by
taking the geometric mean of similar devices (e.g., all oil-fired
tangential boilers, all cold-side ESPs). Geometric means were
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 were then applied to the kg/yr feed
rates entering the boiler, the effect of which either reduced or
left unchanged the emissions that passed through them. Those
EMFs calculated as being greater than 1.0 (i.e., more material
exiting a device than entering it) are set to equal 1.0.
3-11
-------
Table 3-2. Assigned Chloride ppmw and HC1 ppmw Concentrations in
coal, by State of Coal Origin 12
State
Alaska
Illinois
Indiana
Iowa
Missouri
Utah
Washington
Conversion of assigned ppmw
chloride to assigned HC1 ppnrw
54 x 0.61 =
1,136 x 0.61 =
1,033 x 0.61 =
1,498 x 0.61 =
1,701 x 0.61 =
220 x 0.61 =
104 x 0.61 =
Assigned ppmw HC1
in coal
32.9
693.0
630.0
914.0
1,038.0
134.0
63.0
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. 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 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. After
accounting for variables such as coal cleaning (bituminous coal
only) and coal type (higher heating value), the emission
concentration of an inorganic HAP was thus converted into 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 kilogram/year
3-12
-------
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.
Appendix C contains all of the EMFs used to develop the unit
emission estimates for inorganic HAPs.
3.4.7 Organic and Mineral Acid HAPs
Organic and mineral acid HAP emissions were handled in one
of two ways. The first method was used only with HC1 or HF
emissions. The numbers resulting from the method allowed direct
conversion into boiler emissions 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 four test reports where contractors
conducted tests individually for HCl and chlorine as well as HF
and fluorine, before and after each control device. The rest of
the available reports showed tests only for chlorine and fluorine
and estimated the fractions that were HCl or HF. In developing
the HCl and HF EMFs for wet FGDs and dry scrubbers, the EPA
decided to address the effect of flue gas bypass. After
analyzing test data and having discussions with industry
representatives, it was decided to assume an industry average
flue gas bypass of 17 percent for wet FGDs and 14 percent for dry-
scrubber systems. This assumption was used only in the
development of HCl and HF EMFs.14 Because each of the four test
sites was different than the others regarding S02 and PM control,
the emission factors for chlorine and fluorine were maintained
separately for the four system types rather than averaging them.
The second method of handling organic and mineral acid HAPs
was for organics. 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.
If stack emission or APCD exit emission data were
unavailable or reported as nondetected, and, if at least
one-third of the data samples at the inlet of the APCD were
detected concentrations, EPA used organic emissions at the inlet
of the APCD and accounted for the effect of the APCD with EMFs.
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
3-13
-------
rate, for all organic HAPs that were observed at least once during
testing.
3.4.8 Model Estimates for the Year 2010
Emission estimates for 2010 were derived from the same basic
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.
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 and 2010
nationwide emissions of this subset of HAPs from utility units
(see also Table A-l, Appendix A).
3.6 COMPARISON OF EFP ESTIMATES WITH TEST DATA
Comparisons were made between test data from 19 utility
boiler stacks and predicted emissions for the same plants using
the EFP.15 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.
For three elements and 19 boilers, the average of predicted
emissions (as represented by pounds emitted per trillion Btu) was
about 1 percent different from the average of reported values.
Averages for estimates of three individual elements were
different from the test values by +38, -28, and -6 percent. The
highest individual difference between predicted and reported
values was represented by a factor of 5,000.
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.
The test data represent only a few hours of operation at each
plant, while the EFP estimates are extrapolated to annual
emissions. Plants 1 through 14 fire coal, plants 18 and 19 fire
3-14
-------
o
H
O
O
en
ot
S
H
• 5 h
•a a >i
e a -C
*1 0 B
O -H P
• a 0
rH B U
• -H
w
4)
H
^
^"
"H
^
"H
4J
3
0
M
•
u
0
l""|
w
CO
9
•d
-H
8
H
A
0
[[Hydrogen
in
0
o
o
0
o
0
.000013
o
S
H
^"*
a
1
•H
•d
|
A
0
0
A
•rl
*o
Q
^|
o
^
0
M
4J
0
43
1
CO
r*
i
i
fa
B'
s
!
.,*
'/- ,,
',;=-
'-
•'
-,
'$•'.
*i.
'•"„
^.
•>"%
Jl'
'-: ^.
.^^
^
V ,
,-,rB .
.-£.£
&'
fc-
s
^v
&•»*
'$jjjt
1
to
•u
§
H
0,
•u
•H
H
•H
4)
3
U
"JJ
•U
o
_l
n
^j
9
|j
*H
^4
«
0
s
H Natural
in
o
vo
rH
0
[[Arsenic
in
ro
ro
H
I
ro
CO
in
in
C
^|
•d
H
1
0
h
•H
•0
43
n
8
I
1
I
u
C
H
•
m
I
0
3-15
-------
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
Average for all
3 elements
Arsenic EFP/ test
0.12
0.26
1.25
1.41
7.67
0.99
1.54
0.16
0.0004
0.72
2.25
0.20
0.01
5.49
0.05
1.92
1.23
0.59
0.30
1.38
Chromium EFP/ test
0.15
0.11
0.35
1.19
1.76
0.25
0.59
1.12
0.0040
0.33
0.18
0.0042
1.09
3.13
0.16
0.54
0.75
1.11
0.89
0.72
1.01
Nickel EFP/ test
0.0030
0.27
0.08
1.30
0.79
0.37
0.79
7.33
0.0002
L 0.11
0.18
0.0008
0.03
0.29
0.17
0.63
0.40
2.44
2.70
0.94
Values presented are the ratio of emission factor program estimates to
test data in terms of lb/1012 Btu.
3-16
-------
a combination of coal and petroleum coke, and plants 16 and 17
fire oil.
Possible reasons were examined for large differences between
projected and actual emissions. In the EFP, only one fuel was
assumed to be burned. However, some of the plants burned
combinations of coal and petroleum coke, but the EFP recognizes
only coal from one State. The petroleum coke used by one plant
had nickel concentrations that may be more than 100 times higher
than that found in the Montana coal used for that plant by the
EFP, and concentrations in 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 5,000
as mentioned above.
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 U.S. Once the specific plant
was chosen, its 1990 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).
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 deyice 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 1990 fuel consumption values, and the emission
testing (on which the EFP is based) was performed under
3-17
-------
o
A
O\
in
H
•8
1990
0)
g
in
cs
ro
s^
i-H
0)
JJ
fti
6
-H
^
O
rl
ft
ft
a
.u
\
3
I
<*>
in
•
in
•
rH
M-l
O
JJ
C
0)
G
O
U
M
rH
rj
CO
^
ra
o
G
•H
JJ
•H
A
J>l
*^
0
3
G
0)
O
O
g
n
H
s
o
ns/year
0
o
o
o
o
o
ijs
^
rH
0)
4J
(8
g
-H
X
0
M
ft
ft
§
•ri
41
Q
g
H
a
§
u
H
1
. — .
0)
73
rl
CO
rH
0
O
rl
0
JJ
(C
4J
H
ft
rl
O
0)
^
ft
U
rl
jj
(C
(0
0
^J]
.p
U
rH
w
0
u
•ri
O
•0
H
Ij
•g
0
M
*J
§
O
4
H
g
•ri
41
h
a
Si"
w
*-^
G
(C
ft
G
O
r(
plementat
E
0)
JJ
(C
w
v^
rH
0)
rj
M-l
0)
U
C
(C
-H
rH
ft
g
0
u
9
O
">
•0
H
8
u
•0
•ri
Q
•H
•0
«H
H
>* ^
.
jf-:
J1
jd
#g
"?
*>/,
»,/**
*
«
^
S
f'S
44
•V*-*
^^
^ 3
*,*~t*
-^?:
-;'';
.-. •;
•*:
ll
^
?
A$i.
4^4
•"C
i
0
IH
jj
0
•C
S
Q
4J
fi
•ri
^^
•0
x>
s (estima
1990
a
0
•ri
n
n
1
fa
3
BC
1
4J
U
0
H
0
(0
H
00
O
o
U
•ri
a
9
rH
O
0
g
•ri
H
H
rH
in
0
o
0
0
•H
C
vc
00
o
o
•ri
1
O
a
ro
rH
O
01
n
|
in
o
o
j?
i
tH
0
•0
•ri
0
rH
^
o
0
I
rH
9
•0
•ri
^
•3
H
§
0)
O
rl
.00000014
o
3
a
•H
X
0
5
1
ft
0
N
Jj
•JS
!
0
H
u
rl
^J
0
JJ
1
00
fO
in
rH
in
o
o
o
9
C
•ri
§
H
I1
4J
G
•H
t
0
tros
•ri
a
a
3-18
-------
o
Ot
Ot
JJ
H
h
1?
•ri
•ri
g
h
jj
u
o
H
H
"8
•rl
O
U
•H
JJ
n
•rl
rl
0
JJ
u
g
I
rl
V4
n
E*
o
Ot
Ot
rH
0)
§
o
VD
rH
^
rH
0)
JJ
IS
g
•H
X
0
rl
ft
ft
(S
a
jj
jj
!
*
in
rH
V
JJ
C
a)
jj
0
u
rl
3
rH
"J
to
^-*
rH
•H
o
rH
(tf
3
T)
•H
01
a)
a
H
IS
*
--»
?•
^\
'"^r*
Si
rl-
v'£
i;p^
i
m"
;pj
rl
0
ximately
o
r<
ft
ft
IS
g
•rl
JJ
i
g
u
H
0
£
^
•X^
2
0
U
•ri
"0
rH
0
rl
JJ
g
U
rl
0
JJ
JJ
g
0
JJ
0
rH
a
o
•rl
5
BI
,,.
^^
s
0
u
1
•e
rH
a
jj
a
0
u
•§
-rl
J
•0
3
rH
P-
;*;
li.
^ff"
v "•
fz
%>.
!>'
g\
.,
^, ,
^s
.4-"
C","
fe
v"
i
if
i
n
0
rl
0
0
a
g
jj
C
•H
"O
0
JJ
!
JJ 0
n ot
0 Ot
^ H
n
g
*H
tt
a
•H
I
jj
u
0
rH
m
n
o
o
o
o
O
•H
S
a
o
0
o
H
O
U
•rl
0
o
0
i
rH
0
h
3-19
-------
essentially steady-state conditions (with little or no variation
from the baseline operating condition). Therefore, the
characteristic emissions 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.
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.
3-20
-------
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. Reasearch 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. Singer, J.G. ed. Combustion Fossil Power, 4th ed.
Combustion Engineering, Incorporated, Windsor, CT. 1991. p
2-3, modified table.
6. 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.
7. Memorandum from Heath, E., RTI to Maxwell, W.H., EPA.
July 14, 1993. State of coal origin used in the computer
emission program.
8. 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.
9. Memorandum from Heath, E., RTI to Maxwell, W.H., EPA.
April 5, 1994. Proposed coal cleaning factors.
10. Ref. 5, p. 2-3.
11. Memorandum from Turner, J.H., RTI, to Cole, J.D., RTI.
April 19, 1994. Methodology for determining HCl and HF
concentrations fom utility boilers.
12. Memorandum from Heath, E., RTI to Maxwell, W.H., EPA. May
27, 1994. USGS data gaps in chloride concentrations for
seven states.
3-21
-------
13. .Memorandum from Cole, J.D., RTI, to Maxwell, W.H., EPA.
March 31, 1994. Emission factor memorandum.
14. 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.
15. Memorandum from Turner, J.H., RTI, to French, C., EPA.
June 19, 1995. Comparison of RTI emission model projections
with test data.
3-22
-------
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 NAS recent report Science and
Judgement in Risk Assessment2 and the EPA Science Policy
Council's (SPC's) Guidance for Risk-Characterization.*
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 to 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).
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 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
4-2
-------
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.
4.1.2 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 (RAC) Guidance
The RAC 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,
presentations of central tendency and worst-case
portions of the range of risk, and, 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.
4-3
-------
.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.
4.1.5 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.
4-4
-------
.To assess the public health concerns due to emissions of
HAPs from utilities, the EPA conducted inhalation and
multipathway exposure and risk analyses. First, to be consistent
with the NAS recommendations, 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
a multipathway analysis of radionuclides, .a long-range transport
modeling analysis for mercury and arsenic, 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
an evaluation for dioxins, lead, cadmium, and arsenic, and
chapter 9 presents the assessment for radionuclides.
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 the
EPA's Integrated Risk Information System (IRIS). The IRIS is an
online database maintained by the EPA, containing 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 utilized. If other data sources
were used, they are indicated by footnotes in tables 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 of these 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
4-5
-------
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 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
4-6
-------
Table 4-1. Weight-of-Evidence (WOE) Classification
Group
A
Bl
B2
C
D
E
Description
Known Human Carcinogen
Probable Human Carcinogen, Limited
Human Data Are Available
Probable Human Carcinogen, Sufficient
Evidence in Animals and Inadequate or
No Evidence in Humans
Possible Human Carcinogen
Not Classifiable as to Human
Carcinogenicity
Evidence of Noncarcinogenicity for
Humans
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.
4.3.3 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
4-7
-------
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 WA", (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 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.
There are factors involved with the human occupational data
that may result in high- or low-biasing effects, including
4-8
-------
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 risks. The RfD is expressed in units of
milligrams/kilograms/day (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
4-9
-------
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 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 for HAPs
listed in chapter 3 and other health effects information.
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 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 has four
major components:
• Emissions characterization
• Environmental fate and transport
4-10
-------
.• 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.
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
arrays, or 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 U.S. and its Territories. The
STAR summaries combine several available years (typically 6
4-11
-------
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; although, 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
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.
4-12
-------
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 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 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.
If the highest modeled ambient air concentration occurs in
an area (e.g., 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 assuming 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
4-13
-------
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
• The percentage of the population, or the number of
persons, exposed above a specified level of lifetime
risk (e.g., 10~6) .
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, 10'6, 10"5, or 10'4.
4-14
-------
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 was 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.16 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, there are differences between how
cancer risks and noncancer risks are estimated. 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
called the inhalation RfC.
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
4-15
-------
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.7-17
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. Should the HI exceed unity, the HI should be
reevaluated with HQ values summed only for noncarcinogens with
similar target organs based on U.S. EPA Risk Assessment
Guidelines7 and the assumption 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 Mixtox18 that contains information
regarding 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.
4.5.9 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
4-16
-------
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 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.
• 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 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.
• Varying exposures as a result of differences such as
age, sex, health status, and activity are not
considered.
The impact of using these default values is evaluated in
sections 6.12 and 6.13 and Appendix G.
4-17
-------
4.6. REFERENCES
1. U.S. Environmental Protection Agency. Draft Dioxin
Reassessment Report. Office of Research and Development.
EPA/600/BP-92/001. 1994.
2. U.S. EPA. Deposition of Air Pollutants to the Great Waters,
First Report to Congress. Office of Air Quality Planning
and Standards. Research Triangle Park, NC 27711.
EPA-453/R-93-055.
3. National Research Council. Science and Judgment in Risk
Assessment. 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, D.C. 1993.
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 Administrator,
Washington, DC. 1992.
9. U.S. Environmental Protection Agency. Integrated Risk
Information System (IRIS) Database. Environmental Criteria
and Assessment Office, Cincinnati, OH. 1994.
4-18
-------
10. .U.S. Environmental Protection Agency. Risk Assessment
Guidance for Super fund - Human Health Evaluation Manual,
Part A. EPA/540/1-89/002. Office of Solid Waste and
Emergency Response, Washington, DC. July 1989.
11. Weinstein, I. B. The Relevance of Tumor Promotion and
Multistage Carcinogenesis to Risk Assessment. In: Banbury
Report 19: Risk Quantitation and Regulatory Policy, D. G.
Hoel, R. A. Merrill, and F. P. Perera (eds). Cold Spring
Harbor Laboratory, Cold Spring Harbor, NY. 1985.
12. National Research Council, 1986. Reference to be completed.
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. Ref. 7.
17. U.S. Environmental Protection Agency. General Quantitative
Risk Assessment Guidelines for Noncancer Health Effects.
ECAO-CIN-538. Research and Development. Cincinnati, OH.
June 1990.
18. MIXTOX (version 1.5). An Information System on
Toxicological Interactions for MS-DOS Personal Computer,
Environmental Criteria and Assessment Office, Cincinnati,
OH., September 1992.
4-19
-------
5.0 SCREENING RISK 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 for 66 of
the 67 HAPs identified in the emissions database. 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 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 a priority 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 of 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 VI)'
Beryllium
Cadmium
Nickel compounds8
Dioxinsh
PAHs1
Naphthalene
Hexaclorobenzene
Carbon tetrachloride
Quinoline1
Vinylidene chloride
Formaldehyde
n-Nitrosodimethylaminek
1 ,1 ,2-Trichloroethane
Acetaldehyde
Benzene
Benzyl chloride
Bis(2-e-h)phthalatel
Bromoform
Chloroform
Ethylene dichloride
Isophorone"
Methyl chloride1
Methylene chloride
Highest MEI
cone.*
(MQ/m3)
0.0014
0.0023
0.00025
0.00009
0.0027
2x 10'9
0.00002
0.00009
9x 10'6
0.00038
0.000006
0.0011
0.00047
0.00008
0.00054
0.00078
0.00029
6x 10'7
0.00047
0.00077
0.00037
0.00036
0.003
0.0007
0.0015
EPA
WOE"
A
A
B2
B2
A"
B2
B2
C
B2
B2
C
C
B1
B1
C
B2
A
B2
B2
B2
B2
B2
C
C
B2
IUREC per
M9/m3
0.0043
0.0016'
0.0024
0.0018
0.00048
30.0h
0.002V
4x 10'6
0.00046
0.00001 5
0.00351
5x 10'5
1 x 1 0'5
0.014
2x 10'6
2x 10-6
8x 10-*
5x 10'6
4x 10"6'
1 x 10*
2x 10'5
3x 10'5
3x 10'7'
2x 10"6'
5x 10'7
MEI
cancer
risk"
6x 10'6
4x 10"*
6x 10'7
2x 10"'
1 x 10'6
7 x 10'8
4x 10'8
4x 10'10
4x 10'9
6x10'9
2 x 10'8
6x 10'8
6x 10'9
1 x 10"6
9x 10'9
2x 10'9
2x 10'9
3x 10'11
2x 10'9
9x 10'10
9x 10'9
9x ID'9
9x 10'10
1 x 10'9
7x 10'10
Primary type
of cancer
assoc. w/
inhalation*
Lung
Lung
Lung
Lung
Lung & nasal
Tongue,
lung, nasal,
liver
Lung (BAP)
NA
Liver
NA
Nasal, lung
Liver & other
NA
Nose&
larynx
Leukemia
NA
NA
NA
Kidney &
liver
Kidney
Liver & lung
MEI cancer
risk > 10"7
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
Yesk
No
No
No
No
No
No
No
No
No
No
No
5-3
-------
Table 5-1. (continued)
Hazardous air pollutant
Trichloroethylene"
Pentachlorophenol'
Tetrachloroethylene1
Highest MEI
cone.'
(M9/m3)
0.00036
1 x 10-6
0.00036
EPA
WOE"
B2/C
B2
B2/C
IUREC per
M9/m3
2x 10"6'
3x10'5i
6x 10'7i
MEI
cancer
risk"
6x 10'10
3x 10'n
2x 10'10
Primary type
of cancer
assoc. w/
Inhalation*
Lung, liver,
& testicular
NA
Liver
MEI cancer
risk > 10 7
No
No
No
NA «= Not available.
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. The MEI * the Maximally Exposed Individual.
b WOE •= Weight of evidence, for carcinogenicity. See section 4.3.1 and Table 4-1. for explanation of WOE.
c IURE = Inhalation Unit Risk Estimate. The IURE is the estimated increased risk of cancer from breathing 1 ^g of pollutant per cubic meter
of air for 70 years. The lUREs were obtained from EPA's Integrated Risk Information System (IRIS),2 unless indicated otherwise by
footnotes.
d This is the estimated increased lifetime cancer risk to the highest MEI due to inhalation exposure.
e 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.
f 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.
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's) by 0.11 (11 percent).
g 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.
h The emissions were estimated using the toxic equivalency (TEQ) approach described in the draft EPA Dioxm Reassessment Report.3
Exposure was estimated by modeling the TEQ emissions with HEM. The IURE is for 2,3,7,8-tetrachlorodibenzodioxin (TCDD) and was
obtained from the draft EPA Dioxin Reassessment Report.
i 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 - 62).4 (These are listed in Appendix H). 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 IURE for benzolalpyrene
(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"4 per /jg/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.
j 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 of De Minimis Emission Rates - Proposed 4O CFR Pan
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 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.
k 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.
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
Cumene"
Ethyl benzene
Ethyl chloride
Hexane
Hydrogen chloride
Hydrogen cyanide
Lead1''
Manganese
Mercury1
Methyl bromide
Methyl chloroform11
Methyl ethyl ketone
MTBE
Styrene
Toluene
1 ,3-Dichloropropene
Vinyl acetate
RfC*-"
(Atg/rm)
0.03
0.02
9.0h
1000
10000
200
20
3.0
1.5'
0.05
-
5.0
1000h
1000
3000
1000
400
20
200
Noncancer health
effect on which
RfC is basedb,c
Hyperplasia of nasal
resp. epith. in rats
Metaplasia and
inflammation rat nasal
epithel.
-
Developmental effects
Delayed fetal
ossification
CNS & nasal epith.
lesions humans
Hyperplasia of nasal
mucosa & larynx in
rats
CNS symptoms and
thyroid effects
CNS & devel. humans
CNS, humans
--
Lesions of olfactory
epithelium
Hepatotoxicity
Decreased fetal birth
weight (mice)
Increased liver &
kidney weight (rat)
CNS in humans
Neurological effects;
degeneration of nasal
epithelium
Hypertrophy/hyperplas
ia of nasal respiratory
epithelium
Nasal epithelium
lesions
Confidence"-
* in RfC
low
med
NA
low
med
med
low
low
NA
med
-
high
NA
low
med
med
med
high
high
Highest
MEI"
cone.
(M9/m3)
3x 10'5
4x TO"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
0.00054
0.00005
Max. HO'
0.001
0.02
3x 10'6
5x 10-"
2x 10'8
5x 10-6
0.115
0.001
0.0057
0.4
-
2x 10'5
4 x 1 0'7
9x 10'7
7x 10'8
4x 10'7
1 x 10*
3 x 10'5
3 x 10'7
Is the
highest
HQ
above
0.1°
No
No
No
No
No
No
Yes
No
No
Yes
No
No
No
No
No
No
No
No
No
[See Footnotes on next page]
5-5
-------
FOOTNOTES for Table 5-2
* RfC = Inhalation reference concentration.
b See chapter 4, Appendix E, and references for more information.
c This is the critical adverse noncancer health effect that was observed in the animal or human studies.2-6
CNS = central nervous system.
" This is the overall confidence in the RfC as reported on IRIS.
• MEI = Maximally Exposed Individual. 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.
' 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).
8 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.
" The RfC was obtained from the 1992 EPA Health Effects Summary Tables. It has not been verified by the
EPA's RfC/RfD workgroup.
i There is no RfC available for lead compounds. Therefore, as a substitute, the lead National Ambient Air
Quality Standard7 (1.5 Mg/m3) was used in this assessment. However, the lead NAAQS is not 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 B2 carcinogen.2-5
1 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 presented 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 disurride
Chlorobenzene
Cobalt compounds
o & p-Cresols
Cumene
Dibutyl phthalate
Hydrogen fluoride
Methyl
methacrylate
MIBK
Phenol
Phthaiic anhydride
Phosphorus
Propion-aldehyde
Selenium
compounds
m,o,p-Xylenes
2,4-Dinitrotoluene
Methyl iodide
NIOSH
REL/420
(M9/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/420
lug/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/420
(uglm3)1
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 cone.
teg/m3)"
0.00008
0.0005
0.0005
0.00037
0.0017
0.0003
0.00003
0.00033
0.365
0.00013
0.00058
7x 10-*
6x 10-4
0.0036
0.0012
0.0056
0.0005
1 x 10-*
0.00005
Maximum
HO.
NA
4x 10"4
7x 10'6
4x 10'7
0.014
1 x 10'6
3x 10"6
3x 10'5
0.06
1 x 10'7
1 x 10'6
2x 10'5
4x 10'6
0.015
NA
0.012
5x 10'7
3x 10'6
2x 10-6
Is the HQ > 0.1
NA
No
No
No
No
No
No
No
No
No
No
No
No
No
NA
No
No
No
No
NIOSH
OSHA
ACGIH
National Institute for Occupational Safety and Health, a U.S. government organization
that focuses on research.
Occupational Safety and Health Agency, a U.S. Government Agency
American Conference of Government Industrial Hygienists, which is a professional
society, not a government agency.
5-7
-------
FOOTNOTES for Table 5-3 (continued)
The N10SH, OSHA, and ACGIH are primarily involved with the safety and health of workers.
REL = Recommended Exposure Levels are developed by NIOSH and are 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.
PEL = Permissible Exposure Levels, legal limits established by OSHA.
The RELs, PELs, and TLVs are relatively similar. Breathing concentrations below these levels are expected to
be reasonably protective of healthy workers, exposed for 8 hours per day, 5 days per week <~40 hours).
However, there are uncertainties and often the data are less than complete. Also, for some of these values
(especially the PELs), measurement techniques and economic factors are sometimes factored in.8-9-10
The use of 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 40 hr/week to 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 (CARB) in the
Air Toxics "Hot Spots" Program." CARB also divides the TLV by 420 to calculate some of their noncancer
reference exposure levels (4.2 to account for exposure time adjustment, 10X to account for sensitive
individuals, and another 10X because health effects are sometimes observed at the TLV level).
" MEI = Maximally Exposed Individual. 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 CARB Noncancer Reference Exposure Levels used in the "Hot Spots
Program".11
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 VI)'
Beryllium
Cadmium
Nickel compounds9
Dioxinsh
PAHs!
Formaldehyde
Acetaldehyde
Benzene
Methylene chloride
Naphthalene
Tetrachl oroethylene'
Highest
MEI cone.
fcg/m3)1
0.0032
0.0025
0.0003
0.0009
0.21
4x 10'9
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
IURE3 per
M8/me
0.0043
0.0028 6
0.0024
0.0018
0.00048
30.0
0.0021
1.3x 10'5
2.2 x 10-6
8.3 x10'6
4.7 x 10'7
4.2 x 10'6
5.8x 10'7
Cancer MEI
Risk"
1 x 10'5
7x 10'6
7x 10'7
2X10-6
1 x 10"*
1 x 10'7
6x 10'8
9 x TO'8
4 x 1 0'9
3x 10'9
4x 10'9
3x 10'10
8x ID'11
Primary Type of
Cancer
associated w/
Inhalation.*
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
[SEE FOOTNOTES FOR Table 5-1]
5-9
-------
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
Lead1
Manganese
Mercury
Methyl
chloroformh
Toluene
Vinyl acetate
RfC*
(Mg/m3)
1000
20
1.5
0.05
-
1000"
400
200
Critical noncancer health
effect that RfC is based on."
Developmental effects
Hyperplasia of nasal mucosa,
larynx, and trachea in rats
Neurotoxicity and
developmental in humans
Neurobehavioral effects in
humans
-
Hepatotoxicityh
Neurological effects
Nasal lesions
Overall
confidence
in RfCc
Low
Low
NA
Medium
-
NA
Medium
High
Highest
MEId
cone.
teg/m3)
1 x 10"4
1.1
0.005
0.002
0.00014
0.0018
0.002
0.0012
Highest
HQ«
1 x 10'7
0.16
0.003
0.04
-
2x TO'6
5x 10'6
6x 10"6
IsHQ
> 0.1
No
Yes
No
No
No
No
No
No
[SEE FOOTNOTES FOR Table 5-2]
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/420
teg/m3)"
0.12
6.0
48
0.24
0.48
1040
OSHA
PEL/420
(Mg/m3)"
0.12
6.0
45
0.24
0.48
1040
ACGIH
TLV/420
(Mg/m3)"
0.12
NA
45
0.24
NA
1040
Highest
MEI cone.
(ftglm3)'
0.0096
0.03
0.006
0.026
0.001
0.0005
Max HQ
0.08
0.005
0.0001
0.1"
0.002
5x 10'7
Is Max HQ
>0.1
No
No
No
No
No
No
[SEE FOOTNOTES FOR Table 5-3]
5-10
-------
Table 5-7. Inhalation Screening Assessment for HAPS Emitted from
Gas-Fired Utilities
Hazardous air
pollutant
Arsenic
Nickel compounds'1
Naphthalene
Toluene
Lead
Formaldehyde
Mercury
Benzene
Phosphorus
Cobalt
Highest MEI
cone. (M9/m3)
2x 10'5
0.0003
0.0001
0.0018
0.00006
0.008
0.0000002
0.0003
0.0002
0.00002
IURE per
Mg/m3
0.0043
0.00048
4x TO*
NA
NA
1.3x 10'6
NA
8.3 x 1Q-6
NA
NA
HEM Cancer
MEI Risk
1 x 10'7
2x 10'7
4x 10'10
NA
NA
1 x 10'7
NA
2x 10'9
NA
NA
RfC
Atg/m3
NA
NA
NA
400
1.5
NA
-
NA
0.24s
0.12'
Highest HQ
NA
NA
NA
4.5 x 10-6
4x10'5
NA
--
NA
0.0008
0.0002
* These values are not RfCs. They are TLV/420. See Tables 5-3 and 5-6.
[SEE FOOTNOTES FOR TABLES 5-1 to 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 fr"om
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.12 Scientists have collected convincing
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.13 Many
studies indicate that deposition of atmospherically emitted
pollutants can result in indirect avenues of exposure for
humans.14 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.13 HAPs that pose a concern for noninhalation exposure
generally have common characteristics. They are persistent in
5-11
-------
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.15 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.15
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 and 8 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.16 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 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
5-12
-------
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 to select five
HAPs 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 (OURE), 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 I 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
5-13
-------
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
PAH = Polycyclic aromatic hydrocarbons.
POM = Polycyclic organic matter.
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 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 particulate 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). 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
utilit-ies are low relative to other anthropogenic sources of POM.
5-14
-------
Table 5-9. Comparison of Cancer and Noncancer Effects Benchmarks
and Emissions Estimates for 13 Selected HAPs
HAP
2,3,7,8-TCDD (dioxinsKTEQ)
Lead compounds
Mercury compounds
Arsenic compounds'
Cadmium compounds
Hexachlorobenzene
Selenium compounds
Beryllium compounds
Cyanide compounds
Manganese compounds
Pentachlorophenol
Cobalt compounds
POM (PAH)C
RfD
(mg/kg/day)
NA
no
threshold b
--
3x Id"4
5x ID"4
8x 10"*
5x 10'3
5x 10'3
5x 10'3
5x 10'3
NA
NA
NA
OURE
(per M9/L)
3x 10*° '
--
-
5.0 x 10'5
-
4.6 x 10'5
-
1.2x 10-4
-
--
-
-
2.1 x ID"4
WOE
B2
B2
--
A
B2
B2
--
B2
--
-
B2
B2
Coal-fired
emissions
estimates"
(ton/yr)
1.5x 10"
7.2 x 10+1
5.1 x 10+1
5.4 x 10+1
1.9
0.7
1.9 x 10+2
6.6
2.4 x 10+2
1.8x 10+z
7.0 x 10'2
2.1 x 10 + 1
1.9
Oil-fired
emissions
estimates"
(ton/yr)
1.0x 10'5
11
0.3
5
2
NA
2
0.5
NA
10
NA
20
< 1
NA = not available.
PAH = Polyarometic hydrocarbons.
POM = Paniculate organic matter.
a This is an unverified oral unit risk estimate.3
b 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.
c POM emissions were estimated by summing the emissions estimates for each individual PAH
listed in Appendix H.
6 This is an estimate of total nationwide emissions from the source category.
8 RfD is for inorganic arsenic. There was not a clear consensus for developing this value. See the
IRIS database for details.
5-15
-------
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 Cranking" scores
they were assigned. Cadmium was selected for further assessment,
rather than beryllium, because of its higher 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.
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 is 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"6 for cancer and RfC for noncancer) . These lower values
were chosen for screening purposes so that it is unlikely that
potentially important HAPs would not be identified by screen.
That is, these conservative levels were chosen to ensure that all
potentially important HAPs would pass 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"6 for MEIs.1
In addition, three HAPs (HCl, 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,2-10'11 emissions estimates, and recommendations from
the peer review panel. Hydrogen chloride, HF, and acrolein were
5-16
-------
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.
5.7 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 less likely
to present significant risks to public health. However, due to
uncertainties and limitations in the data, it is not possible to
conclusively determine that they do not pose a threat to public
health. Although these HAPs are not analyzed further in this
report, it is possible that future data, such as revised
emissions data or new toxicologic data, could warrant further
evaluation of these HAPs in the future.
5-17
-------
Table 5-10. Pollutants Considered Priority for Further Analysis
Based on Results of Screening Assessment
Pollutant
Acrolein*
Arsenic
Beryllium
Cadmium
Chromium
Dioxins/furans
Nickel
Radionuclides"
n-Nitroso-
dimethylamine
Hydrogen chloride
Hydrogen flouride'
Manganese
Lead
Mercury
Formaldehyde
Priority for
coal
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
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 > 10-7
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
multi-pathway
assessment
No
Yes
No
Yes
No
Yes
No
Yes"
No
No
No
No
Yes
Yes
No
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.
Radionuclides were considered priority based on previous risk assessments conducted in the 1980s on
radionuclides from utilities.1
5-18
-------
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. Draft Dioxin
Reassessment Report. Reference to be completed. 1994.
4. U.S. Environmental Protection Agency. Provisional Guidance
for Qualitative Risk Assessment for Polycyclic Aromatic
Hydrocarbons. EPA/600/R-93/089. Washington DC. 1993.
5. 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.
6. 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.
7. Code of Federal Regulations. Title 40-part 50, Section
50.12. Office of the Federal Register, National Archive and
Records Administration. U.S. Government Printing Office.
Washington, DC.
8. U.S. Department of Health and Human Services. NIOSH Pocket
Guide to Chemical Hazards. Public Health Service, Centers
for Disease Control, National Institute for Occupational
Safety and Health, Cincinnati, OH. 1994.
9. American Conference of Governmental Industrial Hygienists.
Threshold Limit Values. Cincinnati, Ohio. 1993-94.
10. Calabrese and Kenyon. Air Toxics and Risk Assessment. Lewis
Publishers, Inc., Chelsea, MI. 1991.
11. California Air Pollution Control Officers Association. Air
Toxics "Hot Spots" Program. Revised 1992. Risk Assessment
Guidelines. California. October 1993.
5-19
-------
12. .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.
13. 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.
14. 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.
15. 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.
16. 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.
5-20
-------
6.0 INHALATION RISK ASSESSMENT
6.1 BASELINE ASSESSMENT OF INHALATION EXPOSURES AND RISKS FOR 14
PRIORITY POLLUTANTS
This chapter presents estimates of risks due to inhalation
exposure to 14 of the 15 priority HAPs identified in the
screening assessment (chapter 5). 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 the baseline assessment,
for the 14 priority HAPs, risks have been calculated using the
HEM for HAP emissions from all 684 utilities utilizing 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 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 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 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). 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. 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 later) suggests that the
baseline risk estimates are generally conservative, but not
overly conservative. 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.
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 MEI risks, the number
of persons exposed above individual cancer risk levels of 1
chance in 1 million (i.e., 1 x 10"6) , 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 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 10'6 for the "highest-risk"
coal-fired plant. The highest estimated MIR at: a single plant
was 2 x 10"6 for arsenic. Table 6-1 shows that arsenic emissions
from two plants resulted in MIRs greater than or equal to 10"6.
The MIRs for the remaining 424 coal-fired plants were lower than
1 x 10"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,
6-2
-------
Table 6-1. Summary of Baseline Risk Estimates from Chronic
Inhalation Exposure by HAP for 424 U.S. Coal-fired Utilities
Pollutant
Arsenic
Beryllium
Cadmium
Chromium*
Dioxin/Furans
Hydrogen Chloride
Lead
Manganese
Mercury
Nickel"
n-Nitrosodi-
methylamine "
Total
Carcinogens
Highest MEI
cancer risk*
3x10-«
3x10'7
2x10'7
2x10-*
SxlO-8
NA
NA
NA
NA
7x10-7
8x10'7
4x10'6
Population
with risk
> 10"6
852
0
0
107
0
NA
NA
NA
0
0
0
NA
# Plants
with MIR
> 10*
2
0
0
1
0
NA
NA
NA
0
0
0
2
Cancer incidence
(cases/yr)c
0.05
0.004
0.0006
0.02
0.001
NA
NA
NA
NA
0.005
0.02
0.1
Noncarcinogen
Max. HQ
NA
NA
NA
NA
NA
0.12
0.001
0.046
--
NA
NA
NA
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 = Maximum 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.
'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.
The nickel emitted is a mixture of various nickel compounds such as soluble nickel. This analysis assumes that all nickel
emitted has the same carcinogenic potency as nickel subsulfide.
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 6.8 for discussion of long-range transport.
The risk assessments for n-nitrosodimethylamine are highly uncertain because of the very limited emissions data available for
n-nitrosodimethylamine. The emissions estimates were based on one measured data value and several nondetects. Therefore,
the risk estimates for n-nitrosodimethylamine are considered conservative and considerably uncertain.
•Of all 424 coal-fired plants modeled with the HEM, this is the estimated increased inhalation cancer risk for a person assumed
to be exposed for 70 years to the highest modeled HAP ambient air concentration around any of the 424 coal-fired plants.
6-3
-------
Rgure 6-1 a. Maximum Individual Risk Posed by HAPs Emitted from All U.S. Coal-Fired Electric Utilities
(Number of coal-fired plants posing various levels fo risk, by HAP)
250
200
ISO
1°°
50
0
<10E-8
Arsenic
1 E-8 to 1E-7 1E-7 to 1E-6 1 E-6 to 1 E-f-
Maximum Individual Risk (MIR)
1E-5 to 1E-4
<10E-8
Chromium
1E-8 to 1E-7 1E-7 to 1E-6 1E-6 to 1E-5
Maximum Individual Risk (MIR)
1E-5 to 1E-4
<10E-8
Nickel
1E-8to1E-7 1E-7to1E-6 1E-6to1E-5
Maximum Individual Risk (MIR)
1E-5 to 1E-4
£ 20D - -
To til: All HAP.
If-t to I 6.7
1E-7 to 1E-6 1 E-» 10 1 E-S
un Ind Ivldm I Run (» IK)
6-4
-------
o
UJ
UJ
D
2
_2
ILJ
D
•o
UJ
UJ
D
LLJ
_g
op
LLJ
v
D
CD
J2
& §
O -Q
Z >
c
E
O
6
I
•o
D
O
S 8
g
CO
LO O
CM CM
pejU-IDoo p
S 8
6-5
-------
(a Class A, human carcinogen). The limited emissions speciation
data (described in Appendix H) found hexavalent chromium between
0.3 and 34 percent of total chromium. The average percent
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"6. The highest total MIR for a single
coal-fired plant is 4 x 10"6. 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 six major carcinogenic HAPs. As with the MIR,
arsenic and chromium are the major contributors. The number of
people estimated to be exposed to risks of 1 x 10"6 or greater
from exposure to arsenic is 850 and from exposure to chromium is
about 107.
The HEM also calculated the annual incidence of cancer
expected for each of the HAPs due to inhalation exposure within
50 km. The total cancer incidence from the carcinogenic HAPs was
estimated, using the HEM, as 0.1 cancer case per year for
coal-fired plants. Arsenic and chromium are again the major
contributors and account for almost 90 percent of the estimated
cancer incidences.
6.1.1.3 Noncancer Risk. The maximum HQ estimated for
noncarcinogenic HAPs emitted from coal-fired power plants was
0.12 for HCl. The next highest was 0.046 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.
6-6
-------
Table 6-2. Summary of Population Exposed at Various Levels of
Inhalation Risk or Greater by HAP: Coal-Fired Utilities
Risk level
5x 10-6
2.5 x 1CT6
1 x 10'6
5x 10'7
2.5 x ID'7
1 x 10'7
Arsenic
0
0
852
5,990
88,800
1,710,000
Chromium
0
0
107
2,160
8,630
80,500
Nickel
0
0
0
0
947
5,100
Beryllium
0
0
0
0
0
1,280
Cadmium
0
0
0
0
0
107
Dioxins/
furans
0
0
0
0
0
0
6.1.2 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 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 9 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, all nickel was assumed to be eguipotent to 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 less than 10
percent of nickel emissions (from oil-fired utilities) are nickel
subsulfide (see appendix H). 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 Baseline Risk Estimates from Inhalation
Exposure for Priority HAPs for 137 U.S. Oil-Fired Utilities
Pollutant
Arsenic
Beryllium
Cadmium
Chromium*
Dioxin/Furans
Hydrogen Chloride
Lead
Manganese
Mercury
Nickel"
Total
Carcinogens
Highest MEI
Cancer Risk
1 x 10'5
7 xlO'7
2X10-6
5x10-*
1 x 10'7
NA
NA
NA
NA
9x 10'5
1 x 10-4
Population
with risk
> 10"6
2,400
0
45
2,300
0
0
0
0
0
1.65M
NA
# Plants
with MIR
> 10"*
2
0
1
1
0
0
0
0
0
20
22
Cancer
Incidence'
(cases/yr)
0.04
0.002
0.005
0.02
0.0007
NA
NA
NA
NA
0.4
0.5
Noncarcinogen
MAX HO
NA
NA
NA
NA
NA
0.06
0.004
0.04
-
NA
NA
MEI = Maximum exposed individual, which is calculated using the highest 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. See chapter 4 for description of MIR and MEI.
NA = Not available.
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.
Total = Total MEI is the sum of the MEI for individual HAPs within a plant. The total HQ (= HI) is the sum of the HQs
within a plant.
* 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.
6 This analysis conservatively assumes that all nickel emitted from utilities has the same carcinogenic potency 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. Emissions tests indicate nickel subsulfide to be present as less than 10
percent of total nickel emitted. 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.
c 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 6.8 for discussion of long-range transport.
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 2 x 10~7 for nickel and 1 x 10"8 for
arsenic. The 90th percentile for MIR among plants is 1 x 1CT6
for nickel (that is, 90 percent of plants pose risks less than
1 x 10"6 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 MIR from the sum
of risks for each carcinogen is 1 in 10,000 (1 x 10'*) at only
one plant. The second and third highest-risk oil-fired plants
pose MEI inhalation risks of 3 x 10"5 and 2 x 10"5, respectively.
The total MIR exceeded 1 x 10"6 as a result of HAP emissions from
22 oil-fired plants. The median total MIR for all plants is
7 x 10~7, and the 90th percentile is 4 x 10'6. 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 as carcinogenic as nickel
subsulfide (see section 6.10).
6.1.2.2 Population 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"6) or more. The number of
people estimated to have a risk from exposure greater than 1 x
10"6 is 1.65 million for nickel, and about 2,400 for arsenic and
chromium.
Incidences from each HAP were summed to estimate total
cancer incidence, which was estimated as 0.5 case per year from
these 137 oil-fired plants. Nickel accounts for over 86 percent
of the total annual incidence and arsenic contributes about
9 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.
6.1.2.3 Alternative Analysis for Estimating Population
Risks. Figure 6-3 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 would be 0.3 case per year
if the potency (IURE) of the mixture of nickel compounds emitted
from oil-fired utilities was about 50 percent nickel subsulfide,
6-9
-------
Figure 6-2a. Maximum Individual Risk Posed by HAPs Emitted from All U.S. Oil-Fired Electric Utilities
(Number of Oil-Fired Utilities Posing Various Levels of Risk, by HAP)
100
90
80
70
60
50
40
30
20
10
0
Arsenic
-------
m
!
«» -w
2 o
2l
O C
2
I
6
TJ
O
O
6-11
-------
Table 6-4. Summary of Population Exposed at Various Levels of
Risk or Greater from Oil-Fired Utilities
Risk Level
5x 10'5
2.5 x 10'5
1 x 10'5
5x 10-6
2.5 x 10-6
1 x 10"6
5x 10'7
2.5 x 10'7
1 x 10'7
Nickel
89
2,240
2,310
9,930
100,000
1,650,000
7,460,000
23,100,000
73,300,000
Arsenic
0
0
45
89
2,280
2,370
32,600
287,000
2,140,000
Chromium
0
0
0
45
89
2,280
2,280
9,490
257,000
Cadmium
0
0
0
0
0
45
89
2,280
3,040
Beryllium
0
0
0
0
0
0
45
89
2,280
Dioxins/
furans
0
0
0
0
0
0
0
0
45
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.
about 0.15 case/yr if the IURE was 20 percent nickel subsulfide,
and about 0.1 case per year if the IURE was 10 percent nickel
subsulfide. Likewise, there would be changes in the risk levels
to which people were exposed. If the nickel mixture IURE were 50
percent as potent as nickel subsulfide about 100,000 people would
be exposed at an MIR > 10'6. If the IURE were 20 percent nickel
subsulfide, about 9,930 persons would be exposed at an MIR >
10"6. Nickel speciation uncertainty is discussed in more detail
in section 6.10.
6.1.2.4 Noncancer Risks Due to Chronic Exposures. The
highest HQ resulting from oil-fired utility emissions was
0.04 for manganese.
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 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.
6-12
-------
a
LU
oc
113
cc
8
•s
LU
CC
(0
ca
til
CC
in
d
m
6
m
CM
c\i
d
in
jad saseo) aouappu) jaoueo ienuuv
6-13
-------
Table 6-5.
Utilities
Summary of Baseline Inhalation Risk for Gas-Fired
Pollutant
Arsenic
Lead
Mercury
Nickel*
Carcinogens
MEI risk
2x 10'7
NA
NA
2x10'7
Population MIR
> 10-*
0
NA
NA
0
# Plants MIR >
io-«
0
NA
NA
0
Noncarcinogen
HO™
NA
1 x 10'7
NA
NA
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
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.
'The nickel emitted is a mixture of various nickel compounds such as soluble nickel. This analysis assumes
that all nickel emitted has the same carcinogenic potency as nickel subsulfide.
6.2 DISTINGUISHING BETWEEN URBAN VERSUS RURAL LOCATIONS
The HEM has two distinct modeling options (urban or rural)
intended to simulate atmospheric dispersion behavior of gases via
different surface roughness.2 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, 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 Quality
Models (40 CFR, Appendix W to Part 51), 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
6-14
-------
people lived within a 3-km radius, then the area was considered
rural and the rural modeling option was chosen.2 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
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 were estimated to be 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 the remainder of the inhalation
exposure modeling analyses 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.
6.3 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. 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. The method
used by EPA is considered reasonable given the available data.
The exposures and risks 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 risks from utility emissions. The
analysis of this subset of priority HAPs provides information
regarding the anticipated potential public health risks for the
year 2010.
6-15
-------
Table 6-6. Comparison of Inhalation Cancer Risk Estimates Based
on (1) HEM Modeling Using Urban Default Assumption and (2) HEM
Modeling Using Urban vs. Rural Distinction
HAP & fuel
As, from Coal
Cr, Coal
(assuming 1 1 %
CrVI)
Be, Coal
Cr, Oil (assuming
18% CrVI)
Be, Oil
Cd, Oil
Ni, Oil
As, Oil
Urban default
MEI risk
6x10-«
3x 10-6
6x 10'7
5x 10-*
7x 10'7
1.6x 10"6
9x 10'5
1 x 10'5
Cancer
incidence
(cases/year)
0.08
0.03
0.006
0.02
0.002
0.007
0.5
0.05
Population'
with cancer
risk> 10*
21,000
890
0.0
2,300
0.0
45
2,300,000
4,600
Rural vs. urban
MEI risk
3x 10"6
2x 10-6
3x 10'7
5x 10-6
7x 10'7
1.6x 10-6
9 x 1 0'5
1 x 1 0'5
Cancer
incidence
(cases/year)
0.05
0.02
0.004
0.02
0.002
0.005
0.4
0.04
Population*
with cancer
risk> 10-*
850
110
0.0
2,300
0
45
1,600,000
2,400
HAP = Hazardous air pollutant.
MEI = Maximum exposed individual.
* The number of people estimated to be exposed to ambient air concentrations causing an estimated
increased risk of cancer of 1 in 1 million or greater.
Table 6-7. Comparison of Inhalation Noncancer Risk Estimates
Based on (1) HEM Modeling Using Urban Default Assumption and
(2) HEM Modeling Using Urban vs. Rural Distinction
HAP & Fuel
HCI from Coal
Mn from Coal
Urban default
MEIHQ
2.3/20 = 0.12
0.02/0.05 = 0.4
# People above an
HQ of 0.01
157,000
104,000
Selection of appropriate setting ( rural vs.
urban)
MEIHQ
2.3/20 = 0.12
0.002/0.05 = 0.04
# people above an
HQ of 0.01
15,100
27,900
HAP = Hazardous air pollutant.
HQ = Hazard quotient.
MEI = Maximum exposed individual.
6-16
-------
.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 RISKS DUE TO SHORT-TERM EXPOSURE
The potential for exceeding short-term reference exposure
levels (RELs)3 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, HCl, 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 1993.3 The CAPCOA
RELs are listed in Table 6-10.
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 utili-ty 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 that considers all reasonable meteorological
conditions (called TSCREEN) to estimate the maximum 1 hour
concentration of the three compounds in the vicinity of selected
coal- and oil-fired utilities. The 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
6-17
-------
Table 6-8. Estimated 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
n-Nitroso-
dimethylamine
from Coal
Cr from Oil (18%
CrVI)
Be from Oil
Cd from Oil
Ni from Oil
Dioxins from Oil
As from Oil
Cancer risk 2010
MEI risk
3x 10"6
3x 10'7
3x 10-8
1 x 10*
2 x 10'9
3x 10'7
8 x 10'7
SxlO-6
4x 10'7
8x 10'7
5x 10*
7x 10*
7x 10*
Cancer
incidence
(cases/year)
0.051
0.004
0.0007
0.021
0.0012
0.006
0.011
0.009
0.0008
0.0026
0.2
0.0004
0.026
Population w/
MIR > 10*
590
0.0
0.0
399
0.0
0.0
0.0
89
0.0
0.0
240,000
0.0
2,300
Cancer risk 1990
MEI risk*
SxlO-6
3x 10'7
2x 10'7
2x10-*
5 x 10-"
7x 10'7
8x 10'7
5x 10*
7x 10'7
2x 10-6
1 x 10"4
1 x 10'7
1 x 10'5
Cancer
incidence
(cases/year)
0.045
0.0035
0.0006
0.02
0.001
0.005
0.016
0.02
0.0017
0.0053
0.40
0.0007
0.042
Population w/
MIR > 10*
852
0.0
0.0
107
0.0
0.0
0.0
2,300
0.0
45
1,600,000
0.0
2,400
Note: The EPA used Urban vs. Rural modeling data distinction in this analysis.
* These MEI risk estimates are for the "highest risk" plant.
b This is the estimated cases of cancer predicted to occur in the U.S. due to emissions of this HAP from all
utilities of that fuel type based on the HEM analysis.
Table 6-9. Estimated Inhalation Noncancer Risks for Coal-fired
Utilities for the Year 2010 Compared to the Year 1990
HAP
HCI
Manganese
RfC (M9/m3)
20
0.05
Highest MEI
Cone, for 2010
2.6 uglm3
0.003 M9/m3
Maximum HQ for
2010
0.1
0.06
Highest MEI Cone.
for 1990
2.3 M9/m3
0.002 //g/m3
Maximum HQ for
1990
0.1
0.05
6-18
-------
Table 6-10. Noncancer Reference Exposure Levels3 (Acute) from
CAPCOA
Pollutant
Acrolein
Hydrochloric acid
Hydrogen fluoride
REL - Hourly average concentration l^g/m3)
2.5
3,000
580
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.
To illustrate this methodology, a sample utility is
presented in Table 6-12. 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 cancer analysis were located in relatively flat terrain. (The
6-19
-------
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
Table 6-12. Stack and Emission Values Input to TSCREEN
Pollutant
Stack height (m)
Stack gas exit
velocity (m/s)
Stack diameter
(m)
Stack gas
temperature (K)
Emission rate
(g/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
uglm3)
2.5
3,000
580
Coal-fired
maximum
predicted
concentration
(hourly avg t^glm3}
0.016
21.5
1.0
Oil-fired maximum
predicted
concentration
(hourly avg uglm3)
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.
6-20
-------
effects of terrain are analyzed in Appendix G.) 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 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.
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 U.S. 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
6-21
-------
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.4 This regional model
simulated monthly S02 and sulfate (SOj") concentrations, 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 SO2 and S0l~
concentrations and wet and dry deposition patterns and generated
matrices of interregional exchanges of sulfur for a user-defined
configuration of regions.4'5, 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 /urn) and coarse (2.5 ^m < diameters < 10.0
A«n) 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, S0l~, 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.6 The next
section discusses modifications made to the original sulfur
version of RELMAP to enable the simulation of atmospheric
arsenic.
6.6.2 RELMAP Modeling Strategy for Atmospheric Arsenic
6.6.2.1 Introduction. Previous versions of RELMAP have
been described by Eder et al.6 and Clark et al.7 The goal of the
current effort was to model the emission, transport, and fate of
airborne arsenic from utilities in the continental U.S. for the
6-22
-------
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
3x 10-6
1 x 1C'5
3x 10"6
7x 10-*
Incidence
0.05
0.04
0.05
0.03
Population w/
risk > 10-*
850
2,200
590
2,200
Double-count runs
MEI risk
3x 10*
1 x 1 0'5
3x 10"6
7x 10*
Incidence
0.05
0.04
0.05
0.03
Population w/
risk > 10"6
850
2,400
590
2,300
year 1989. Modifications to the RELMAP for atmospheric arsenic
simulation were based on the assumption that all arsenic
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 arsenic,
the data accounting task of a source-receptor run for all
electric 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.6
were preserved for the RELMAP arsenic modeling study.
6.6.2.2 Physical Model Structure. Because of the long
atmospheric residence time of fine PM, significant long-range
transport of arsenic 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 (approximately 40 km2) to provide
high-resolution coverage over the entire continental U.S.
Since the descriptive document by Eder et al.6 was produced,
the original three-layer puff structure of the RELMAP has been
replaced by a four-layer structure. The following model layer
definitions were used for the RELMAP arsenic simulations:
6-23
-------
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 arsenic 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 arsenic, as for
most other pollutants, emission rates for each source cell were
defined from input data and a time step of 3 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. Wind field initialization data for a National Weather
6-24
-------
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 arsenic
simulation. 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 is 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.
6.6.3 Model Parameterizations
6.6.3.1 Chemical Transformation. The simplest type of
pollutant to model with RELMAP is the inert type. To model inert
pollutants, one can simply omit chemical transformation
calculations for them and not be concerned with chemical
interactions with other chemical species. Arsenic was treated as
an inert pollutant species.
6.6.3.2 Dry Deposition. All atmospheric arsenic was
assumed to be in particulate form. Since arsenic and its
compounds make up only a small fraction of total PM loading of
the atmosphere, it was treated as a minor component of the
general population of conglomerate aerosol particles. Heavy
metals have been generally associated with fine particle sizes
(<1 /^m diameter) , but there is evidence larger particles may play
a significant role in dry deposition in urban areas.5'6 Therefore,
arsenic particles were modeled in five sizes; 0.1, 0.3, 1.0, 3.0,
6-25
-------
and 10.0 ,um diameter. The results of the RELMAP simulation for
each particle size were then used in a postprocessing operation
to estimate effects of dry deposition on a particle size
distribution appropriate for regional-scale air masses with urban
influences. The following percent fractions of particle mass
were assumed to be in each size class: 20 percent in 0.1 //m, 50
percent in 0.3 /urn, 20 percent in 1.0 //m, 5 percent in 3.0 yum, and
5 percent in 10.0 /urn.
The semi-empirical dry deposition model as described in
Sehmel8 was used to estimate dry deposition velocity for
conglomerate particles in the 5-y.m size class. This model
requires as input the particle density, the particle diameter,
the friction velocity, the Monin-Obhukov length, the surface
roughness length, and the air temperature. Assuming sulfate,
nitrate, and organic compounds make up most of the particulate
mass for particles less than 10 yum in diameter, a density value
of 2 g/cm3 was used to represent all particles containing
arsenic. Although arsenic and most of its compounds have
densities of over 2 g/cm3, it was assumed that they make up only
a small part of the conglomerate aerosol particles in the modeled
size range. Dry deposition velocities for particulate arsenic
were calculated using a FORTRAN subroutine developed by the
CARB.9 Table 6-15 shows the windspeed (/um) (m/s) used for each
Pasquill stability category in the calculation of deposition
velocity from the CARB subroutine, and Table 6-16 shows the
roughness length used for each land-use category.
6.6.3.3 Wet Deposition. Alcamo et al.10 used a scavenging
ratio of 0.5 x 106 for both arsenic and cadmium, noting that
these values are close to the average values reported by Chan et
al.11 They also note that Chan et al. did not address As, but
used values based on measurements in Canada, not Europe.
Schroeder et al.12 show a range of measured values for the
scavenging ratio of arsenic, cadmium, and lead.
Some of the parameters used for dispersion and deposition
modeling are shown in Tables 6-15 and 6-16. These parameters
include windspeed vs. stability category and roughness length vs.
land-use category. As a compromise, a scavenging ratio of 0.25 x
106 was used for the spring and summer seasons and a value of
0.40 x 10s was used for the autumn and winter seasons.
6.6.4 Exposure and Risk Estimates
The RELMAP analysis produced an average annual air arsenic
concentration for each grid cell in the continental U.S. The
maximum annual arsenic RELMAP concentration was 0.28 ng/m3. Of
the 12,600 grid cells in the study area, 33 grid cell
6-26
-------
Table 6-15. Windspeeds Used for Each Pasquill Stability Category
in GARB Subroutine Calculations
Stability category
A
B
C
D
E
F
Windspeed (m/s)
10.0
5.0
5.0
2.5
2.5
1.0
Table 6-16. Roughness Length Used for Each Land-Use Category in
GARB 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)
Autumn-winter
0.5
0.15
0.12
0.5
0.5
0.4
10'6
0.1
0.2
0.135
0.1
Spring-summer
0.5
0.05
0.1
0.5
0.5
0.4
10-6
0.1
0.2
0.075
0.1
6-27
-------
concentrations were greater than 0.100 ng/m3. The 50th
percentile of the grid cell values was 0.0034 ng/m3 and the
average cell value was 0.01 ng/m3. The figure indicates that,
typically, the largest concentrations occur in the eastern part
of the U.S.
Once the grid cell concentrations are known, public exposure
and risks can be calculated by applying the population database
used in the HEM. This population database contains the location
of and number of people living within each census block. By
overlaying the population database onto the grid cells, exposure
can be estimated for each group of people within the continental
U.S. By multiplying the grid cell arsenic concentration by the
number of people within that grid cell and summing these products
over all the U.S. grid cells, one can estimate total population
exposure. To estimate annual population risks (cancer
incidence), the model multiplied the total exposure product by
the arsenic IURE and divided by 70 years. The results are shown
in Figure 6-4 and Table 6-17.
To evaluate potential impacts due to long-range transport,
the coal, oil, and gas emissions were modeled together. By
applying the algorithm described above, 0.6 cancer case/year was
estimated, for all three fuels, for arsenic emissions from
utilities in the continental U.S. This estimate is about seven
times greater than the population risks estimated modeling
arsenic emissions within 50 km of each facility using the HEM
(i.e., 0.05 case/year for coal and 0.04 case/year for oil).
The potential impacts to the MEIs appear to be considerably
less than for population exposures. The maximum RELMAP
concentration of 0.27 ng/m3 is about 40 percent of the highest
HEM arsenic concentration for coal-fired utilities. The modelers
expect that the other metals of potential concern (e.g.,
chromium, nickel, cadmium) would show similar results. These
trace metals are also associated with fine 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, for a
screening exercise, the factor of 7 can be applied to these other
HAPs to roughly estimate the potential impact of long-range
transport of HAPs on the overall cancer incidence. In the HEM
analysis (i.e., within 50 km) the total cancer incidence (not
including radionuclides) was estimated to be up to 0.6 case per
year (0.1 case per year for coal-fired utilities and 0.47 case
per year for oil-fired utilities). Multiplying the 0.6 case per
year by the factor of 7 results in a cancer incidence estimate
6-28
-------
•H
0
§
u
1
•rl
H
•rt
4J
9
5,
rt
n
a
4>
H
6-29
-------
Table 6-17. Exposure and Risk Estimates Based on RELMAP Modeling
of Arsenic Emissions from All Oil- and Coal-fired Utilities in
the U.S.
Exposure Concentration (uglm*)
2.5 x 10"4
1 X10"4
5x 10'6
2.5 x 10'6
1 x 10'5
5x 10*
2.5 x 10*
1 x 10"6
Inhalation Risk*
1 x 10'6
4x10'7
2x 10'7
1 x 10'7
4x 10'8
2x 10"8
1 x 10'8
4x 10'9
Number of People Exposed to
this Level of Risk
42,000
1 1 ,700,000
96,400,000
147,000,000
183,000,000
204,000,000
221,000,000
230,000,000
a Based on standard cancer risk equation, assuming 70-year exposure for people living in exposure grids
(described in chapters 4 and 6).
of roughly four cancer cases/year due to emissions of
nonradionuclide HAPs from coal- and oil-fired utilities
(including radionuclides, which are analyzed in chapter 9, cancer
incidence is estimated to be as high as 6 cases/year) . Assuming
that the factor of 7 applies equally to oil and coal utilities,
the cancer incidence for coal-fired utilities for nonradionuclide
HAPs is estimated to be roughly 0.7 case/year (i.e., 0.1
multiplied by 7) and that the cancer incidence for oil-fired
utilities is roughly three cases/year (i.e., 0.47 x 7).
However, there are numerous uncertainties in the modeling,
the assumptions, the extrapolations, and the resulting cancer
incidence estimates. For example, the long-range transport of
emissions from oil-fired utilities may be different than the
long-range transport of emissions from coal-fired utilities.
Also, 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, since nickel is the HAP
contributing most to the cancer risks in the HEM analysis, the
cancer incidence estimate for oil-fired utilities (i.e., three
cases per year) and the overall cancer estimate (i.e., four cases
6-30
-------
per year) is heavily dependent on the assumption that the mix of
nickel compounds is as carcinogenic as nickel subsulfide.
Because of these and other uncertainties, the cancer incidence
estimates and the extrapolation factor of 7 should be viewed with
caution. The resulting cancer incidence estimates are considered
high-end, conservative estimates. Further evaluation of the
data, models, and methods is needed to reduce the uncertainties
and to fully evaluate the impacts of long-range transport.
6.7 DISCUSSION OF BACKGROUND EXPOSURES
6.7.1 Arsenic
There are over 250 sites that 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.
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.
6-31
-------
.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.
However, 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.
Considering the above information, it is difficult to draw
definitive conclusions from the data comparisons. The largest
concentration from the monitored data set was about 8 ng/m3 and
this site was near two plants. Similarly, over half of the
monitors never detected annual arsenic concentrations above the
MDL of 3 ng/m3, so the highest possible impact at the typical
monitoring site must be below 3 ng/m3. The analysis indicates
that the predicted concentrations from the HEM arsenic air
dispersion analyses were not radical underestimates of actual
plant emission impacts.
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
6-32
-------
below detectable levels and did not provide much insight into the
relative concentration impacts from utility emissions.
Based on the HEM modeling, manganese and HCl 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 HCl from the highest-risk plants were estimated
to be 10 times lower than the RfC. All other HEM-modeled
concentrations for HCl and manganese were even lower. Therefore,
regardless of background exposure levels, the emissions of HCl
and Mn from utilities are not likely to contribute significantly
to an RfC problem. 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 the toxicity of the trivalent chromium (Cr III)
versus 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.13-14 However,
there are uncertainties in the health effects of Cr III. For
more information on chromium toxicity see Appendix E.
Data on speciation of chromium were available from 11 test
sites. The limited emissions speciation data (see Appendix H)
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 6-18 and 6-19 present the results of the assessment.
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. This IURE is the EPA-verified value
6-33
-------
Table 6-18. Chromium Speciation Analysis for Coal-fired
Utilities: Inhalation Risk Estimates due to Chromium Emissions
Based on Various Assumptions of Percent Cr VI.
% Chromium VI assumption*
Assume 1 00% Cr VI
Assume 23% Cr VI
Assume 1 1 % Cr VI
Assume 0.4% Cr VI
Lifetime
MEI risk
2x 10'5
4x 10*
2x 10-*
7x 10-"
Lifetime
MIR
1 x 10'5
2x 10*
1 x 10*
4x 10'8
Population w/ >
10* lifetime cancer
risk
69,000
2,300
no
0.0
Cancer incidence
(cases/year)
0.2
0.04
0.02
0.0007
' Based on speciation data from emissions tests for four coat-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 6-19. Chromium Speciation Analysis for Oil-fired
Utilities: Inhalation Risk Estimates due to Chromium Based on
Various Assumptions of Percent Chromium VI
% Chromium VI
assumption*
100% Cr VI
34% Cr VI
18%CrVI
5% Cr VI
Lifetime MEI risk
3x 10'5
1 x 10'5
5x10-«
1.5x 10"6
Lifetime MIR
3x 10'5
1 x 10'5
5x10*
1.5x 10*
Population w/ > 10*
lifetime cancer risk
40,000
2,300
2,300
45
Cancer incidence
(cases/year)
0.1
0.04
0.02
0.005
' 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.
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. 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.
6-34
-------
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.15
The Canadians have also reviewed the available data recently and
established an IURE of 6 x 10'3. The Canadian IURE also appears
to be within the plausible range of potency for arsenic.15
Table 6-20 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 6-20). 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.16
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, 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
(see Appendix H).
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.17-18 Nickel refinery dust and nickel
subsulfide are classified as human carcinogens (WOE = A). The
lUREs for nickel refinery dust and nickel subsulfide are
6-35
-------
Table 6-20. Arsenic Inhalation Risk Estimates: Comparison of
Results Using the EPRI, EPA-verified, and Canadian IURE
Arsenic from
Oil-fired Utilities
Arsenic from
Coal-fired Utilities
Risk estimates using EPRI IURE1
(lAxlO^perjug/m3)
MEI risk
4X10"6
6x 10'7
# > 10*
2,200
0.0
Incidence
0.014
0.015
Risk estimates using EPA
IURE2 (4.3 x 10"3 per uglm3)
MIR
1 x 10'5
3x 10-6
# > 10*
2,400
850
Incidence
0.042
0.045
Risk estimate w/ Canadian IURE
(6x IQ^perMg/m3)
MIR
2x 10'5
4x ID'6
# > 10"6
3,000
850
Incidence
0.05
0.06
8 The EPRI IURE for arsenic (1.4 x 10-3 per ^fl/m3) is three times lower than the EPA-verified IURE for arsenic (4.3 x 10'3 per
Aig/m3). And, the Canadian value is approximately 35 percent greater than the EPA estimate.15
2.4 x 10~4 and 4.8 x 10~4, respectively. 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 IARC considers nickel
monoxide, nickel hydroxide, and metallic nickel as having
sufficient evidence in experimental animals for
carcinogenicity.18 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.19
The American Conference of Governmental Industrial Hygienists
(ACGIH) has stated that all nickel compounds should be considered
carcinogenic.20 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.17-18 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 is as carcinogenic as
nickel subsulfide) are considered conservative, upper-bound risk
estimates.
To assess the potential impact of the speciation
uncertainty, the EPA conducted an assessment for cancer risks
6-36
-------
utilizing different assumptions for speciation and cancer
potency. The assessment (summarized in Table 6-21) provides a
range of the potential cancer risks due to nickel emissions.
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 6-22) .
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 tested have hot-
side 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.21 Since
this phenomenon was observed at MWCs, the EPA assumes that it is
possible that the same situation may possibly occur at utilities.
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. Therefore, as a
scoping effort for this report, the dioxins were modeled a second
time with the assumption that dioxin emissions are 10 times
greater from all utilities that have hot-side ESPs (145 units).
The results of this scoping effort showed an increase of roughly
double the national total dioxin emissions from utilities, from
1.5 x 10~4 ton/yr to 3.5 x 10"4 ton/yr. The cancer MIR increased
by a factor of 3.5, from 5 x 10~8 to 1.8 x 1CT7.
This was based on a hypothesis. Utilities are different
than MWCs. There are differences in fuel and operations.
Therefore, it is not known whether utility units with hot-side
ESPs are likely to emit more dioxins. More data and analyses are
needed before any conclusions can be made regarding dioxins from
utilities with hot-side ESPs.
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
6-37
-------
Table 6-21. Nickel from Oil-Fired Utilities: Inhalation Cancer
Risk Estimates Based on Various Assumptions of Speciation and
Cancer Potency
Nickel Speciation*
100% NiSubsulfide
20% Ni Subsulfide
10% Nisubsulfide
1 % Ni Subsulfide
Cancer potency
(IURE)b
4.8 x 10-4
9.6 x 10'5
4.8 x 10'5
4.8 x 10-6
MIR
9.6 x 10'6
2x 10-5
9.6 x 10"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 estimates.
MIR = Maximum individual risk
' 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.
" The Inhalation Unit Risk Estimate (IURE) of 4.8 x 10"4 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.
Table 6-22. Comparison of Nickel Exposure to Various Noncancer
Health Benchmarks
Various health benchmarks for nickel
compounds
# People exposed0 above the benchmark
# People exposed above 1/1 Oth the
benchmark
Maximum HQ"
CARB REL1 = 0.24
^g/m3
0.0
2,300
0.82
EPRI" value = 2.4 ^g/m3
0.0
0.0
0.082
LEGEND:
REL = Reference exposure level
CARB = California Air Resources Board
EPRI = Electric Power Research Institute
HQ = Hazard quotient
' This value was obtained from the CARB Hot Spots Program.3 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.
• The EPRI benchmark22 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.
c The exposed population is estimated from the results of the Inhalation Human Exposure Modeling.
" 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.
6-38
-------
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.
6.12.1 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 NEC.23 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.
• 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.
6-39
-------
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.24 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 data24 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 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
6-40
-------
been, altered by either the agent or one of its metabolites.24
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.24 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 particular pharmacologic effects commonly
correlate to surface area. Because the body surface area is
6-41
-------
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 fj.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)"1 when derived from oral data and (rng/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 in3/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
6-42
-------
is defined as a statistical best estimate of the value of a
parameter from a given data set.25 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.
6.12.3 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
6-43
-------
(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 and that all chemicals
produce the same effect). If these assumptions are incorrect,
over- or underestimation of the actual multiple-substance risk
could occur.26 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 subsulf ide) , 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
6-44
-------
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).23
Certain groups of individuals within the population are
inherently more sensitive to carcinogen exposure than others.
Factors 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.
• Children spend substantially more time outdoors than
adults and may be exposed to higher concentrations.
• Young organisms appear to be inherently suspectible
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.
6-45
-------
.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 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 for details on the uncertainty analysis. 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.23 Furthermore, any
procedure that relies on a combination of point values (some
conservative and some not conservative) yields a point estimate
6-46
-------
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.
The uncertainty analysis focused on the three HAPs (nickel,
arsenic, and chromium) that accounted for over 95 percent 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 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. Table 6-23 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.
6-47
-------
w
0)
4-1
-H
,-(
-H
4->
P
U
•H
H
4-)
O
0)
i-H
W
M
O
4-1
4->
c
I
W
W
0)
to
tn
<
X
in
•ri
OS
C
•H
T3
0)
W
D
CO
H
0)
4-1
0)
g
(0
n
RJ
CU
U
•rl
tn
td
W
u-i
o
>
M
(0
W
CO
05
I
10
•s
S
So
3
if
I
§1
< u.
• • '
'S 1 >
W = O)
^
il_
I8H1I
> o s o " E
> I > ° " fe
' f 5 £ £ S
1"?
I I
ft
it
o o
IJ
f ~
af I
3 g
11
! o E
ill!
IB
o .
II
** p
"F .5
§1
* (0
o £
111
HI
.1
11
O
. (0
W 09
.2::
H!
lill
6-48
-------
TJ
0)
C
•H
4-1
c
o
u
ro
OJ
i
vo
rj5f«*
i 3. a i o »
I I s
II s
S |-g
it|!
Ill
ill
*ss
SHi
T3 E
•g £
|s,
•M-s:
O O J
ili
f t£
: £
i -£
II
111 £
£ e
6-49
-------
•H
.U
C
0
u
CN
I
•8
n ~
•&* T5 C
II!
*1B 1*1*
s 5 i. $ q-Jl!
(0 t c £ CO
i. s r I I s
• $ > < a = '
~
ii
i
£
«
ID o> •
l
sill
li&l
01 o &. E
BO
||
l!
b-g | _
i Si i
>. .
£* •«
5 «
II
1}
li
f
O "G
09 V)
II'
i *
2 — *
1 ^ S ri i-
If Sit
Illfl
« » 5 S
= s £ o-
.2 E § iB
-•ill*
siSi2--2'
•= i I « S 5 '
5
Ul
u>
•§.«
6-50
-------
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 for explanation).
In general, numerical methods (i.e., 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.
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 6-5 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.
6-51
-------
LU
w
I
cc
UJ
cc
CO
2
X
UJ
CO
o
CO
cc
UJ
a.
CO
3"*
O q>
CO 0)
2>r
Si
33
0
CD O
CC Z
v eg
z oc
< CO
uj -j
u<
UJ
o
z
z
m
O
o
LL
O
O
Q.
UJ
O
in
<6
£
f
£
a.
Frequency
Jr
CO
c
UJ 2
Q 1
O W
55 8 5 5 I
s
Q.
CO
CD
(0
c
o
»
Q
Frequency
2 £ i
T
111
8.
S 5
x
3.2
O uj
i§
O =
a. to
w t:
S |
M
si
X ^
Frequency
6-52
-------
.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).
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 6-24 and Figure 6-6
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 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 6-24,
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
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
6-53
-------
<
COIN
c
o
CO *-
£ E
« 2
w )=
CD CD
*
(0 «
% §
E '55
3 CO
CO'
LU
eo
CM
•o
i
LU
O
LL
o>
i- m
00 ^
T>
O
C
O
0
to
0
oo
q
o
qqoocosgfc
cc
a.
[£
O)
CO
«S co
0) T3 0)
S en 8
" o «
C CO CD
'o c w
co co o
a
« •£
OB CD
T3 X:
v — O
8 W «
?2f
ID = ^r rt
O "5.5 c 2
r~ r« *-
— O
re
11 i i i
W 2 CO O 3
c Q. — •- ->
2
H—
CO
1 s
0-.2
o
._ CD '55
w —. w
.I 8 I 5 I § .i
CO — C *• t *-
w en CD « *= ^ T=;
co<
CM T-
CD
co
LU
q in q o
c> CM tri ^
ui
s
o
£z r\
CO Q. tO -2
n *± •= «s
CO •— ^^ CO
S cu E £
2 §• o -s
o. -Q
in
«
q in
in N o
en o> o
i= o
.« „
0)
CO P CO
CD o S
CO O £
CO
i
1 r x
CD tf
CJ Q.
CB 8-
D. -5-
6-54
-------
Figure 6-6. Summary of Results of Monte Carlo Simulation of
HAP Emissions from Oil-Fired Plant#29
FCEM Concentration Data
SGS Concentration Data
Cell E11
Forecast Plant 29 As Emissions
Frequency Chart 2,942 Trials Shown
III
Cell E12
.027 •
.020 ..
Forecast: PI 29 As Emission (SGS)
Frequency Chart 2,941 Trials Shown
.013
.007 .. . . .
500.00
CellH11
Forecast: Plant 29 Cr Emissions
Frequency Chart 2,937 Trials Shown
Forecast PI 29 Cr Emissions (SGS)
Frequency Chart
2,444 Trials Shown
. 212
llllllilllliiii.in.i i .1.. i.. ..i..
7500
Cell K11
.036 -f
Forecast: Plain 29 Ni Emissions
Frequency Chart 2,930 Trials Shown
Cell K12
.030 j
Forecast: PI 29 Ni Emissions (SGS)
Frequency Chart 2,940 Trials Shown
15.00000 20.00000
11.25000
16.875.00 22.50000
a. Note that the abscissa scales are not the same from the FCEM to SGS data displays.
b. FCEM = Field Chemical Emissions Monitoring from EPRI program.
c. SGS = Subsequent data
6-55
-------
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 1.0, while averages
for arsenic, chromium, and nickel are 1.4, 0.7, and 0.9,
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 5,000 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 5,000 would not change the overall
results. The EFP tended to underestimate rather than
overestimate emissions about 60 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.
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 for
further discussion.
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 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
6-56
-------
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.
The software program Crystal Ball® (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
6-57
-------
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
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 6-25. 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 (generally around the
90th or 95th percentile). The uncertainty analysis supports the
baseline risk estimates as reasonable high-end estimates of risks
because they are within the range of predicted risks.
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
6-58
-------
Table 6-25. 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%
100.0%
FCEM
1E-07
(96)
2E-12
3E-09
6E-09
1E-08
3E-08
6E-08
8E-08
2E-07
5E-07
1E-06
3E-05
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
2E-05
FCEM
1E-07
(87)
2E-10
2E-09
3E-09
5E-09
1E-08
4E-08
1E-07
3E-07
5E-07
7E-07
6E-06
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
6E-06
FCEM
2E-06
(90)
2E-09
2E-08
5E-08
9E-08
2E-07
6E-07
2E-06
4E-06
7E-06
1E-05
9E-05
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
7E-05
Ratio
95th
95th
95th
mean
median
baseline
0.8
8.7
3.5
4.1
10.6
3.8
Arsenic
2.3
12.8
4.2
0.7
9.6
3.9
1.7
11.1
4.0
Variability
Chromium
2.5
10.4
3.8
Nickel
Mean
Initial Point Estimate
(percentile)
Percentiles:
0.0%
2.5%
5.0%
10%
25%
50%
75%
90%
95.0%
97.5%
100.0%
FCEM
1E-07
(95)
5E-12
1E-09
3E-09
6E-09
1E-08
3E-08
6E-08
2E-07
5E-07
1E-06
1E-05
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
2E-05
FCEM
1E-07
(90)
4E-10
2E-09
3E-09
6E-09
1E-08
4E-08
1E-07
3E-07
4E-07
6E-07
3E-06
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
2E-06
FCEM
2E-06
(90)
5E-09
3E-08
6E-08
1E-07
3E-07
7E-07
2E-06
4E-06
6E-06
9E-06
4E-05
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
5E-05
FCEM = (Field Chemical Emissions Monitoring [EPRI]) Original oil concentration data, distribution defined by probability
SGS = Subsequent data, trace metal analysis from samples collected for radionudide analysis.
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.
6-59
-------
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. 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 U.S. 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 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.
6.15 DISCUSSION OF FEDERAL INTERAGENCY REVIEW COMMENTS
Previous drafts of chapters 1 through 10, along with the
appendices, were reviewed by numerous non-EPA scientists
representing industry, environmental groups, academia, and other
Federal agencies during the summer of 1995. In February, April,
6-60
-------
and September 1996, all sections of the draft report underwent
additional review by EPA, State and local agencies, and other
Federal agencies. The EPA has revised the report, as
appropriate, based on the reviewers' comments. The EPA revised
the report to incorporate the majority of the comments received.
However, there were several comments that could not be fully
addressed because of limitations in data, methods, and resources.
This section presents comments received by other Federal agencies
that could not be substantially addressed in this interim report.
6.15.1 Comment
The Department of Energy (DOE) commented that EPA's
assumption that 100 percent of nickel emissions are nickel
subsulfide w...is not justified..." and "...results in an overly
conservative risk assessment..." rather than in a worst-case
scenario based on the available data.27
6.15.2 Response
Faced with the uncertainties associated with nickel
emissions and health effects, the EPA originally chose the
conservative assumption that all nickel is as carcinogenic as
nickel subsulfide. Several reviewers including State and local
representatives and external scientific peer reviewers supported
this approach. However, several industry commenters agreed with
DOE's position. To address these issues regarding the nickel
risk assessment, the EPA has evaluated various alternative
assumptions and presented ranges of potential risks using these
various assumptions. Also, the EPA has presented in this chapter
a thorough discussion and presentation of the uncertainties
associated with the nickel risk assessment.
6-61
-------
6.16. 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. Guidelines on Air
Quality Models. Code of Federal Regulations, 40, Appendix W
to Part 51, July 1, 1994.
3. California Air Pollution Control Officers Association. Air
Toxics "Hot Spots" Program, Risk Assessment Guidelines.
CITY, STATE. October 1993.
4. Johnson, W.B., Wolf, D.E. and Mancuso, R.L. Long Term
Regional Patterns and Transfrontier Exchanges of Airborne
Sulfur Pollution in Europe. Atmospheric Environment.
12:511-527, 1978.
5. 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.
6. U.S. Environmental Protection Agency. RELMAP: A REgional
Lagrangian Model of Air Pollution - User's Guide.
Atmospheric Sciences Research Laboratory, Research Triangle
Park, NC. March 1986.
7. 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.
8. Sehmel, G.A. Particle and Gas Dry DEposition: A Review.
Atmospheric Environment. 14:983-1011, 1980.
9. California Air Resource Board. Deposition Rate Calculations
for Air Toxics Source Assessments. Air Quality Modeling
Section, Technical Support Division, California. September
16, 1987.
6-62
-------
10. .Alcamo, J., Bartnicki, J., Olendrzynski, K. and Pacyna, J.
Computing Heavy Metals in Europe's Atmosphere - I. Model
Development and Testing. Atmospheric Environment.
26A(18):3355-3369, 1992.
11. 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.
12. Schroeder, W.H., Dobson, M., Kane, D.M., Johnson, N.D.
Toxic Trace Elements Associated with Airborne Particulate
Matter: A Review. Journal of Air Pollution Conttrol
Association. 37(11). 1987.
13. 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 27711. December 1994.
14. U.S. Environmental Protection Agency. Integrated Risk
Information System (IRIS) Database, Environmental Criteria
and Assessment Office, Cincinnati, OH. 1994.
15. Memorandum from Hugh McKinnon, M.D., to Seitz, John, EPA.
April 26, 1995. Advice on unit risk estimate for airborne
arsenic.
16. Handbook of Chemistry and Physics, 60th ed. , CRC, CITY,
STATE. YEAR.
17. U.S. Environmental Protection Agency. Health Effects
Document for Nickel and Mercury Compounds. EPA/60078-
83/012ff, final report. September 1986.
18. 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.
19. California Air Resources Board. Initial Statement of
Reasons for Rulemaking. Proposed Identification of Nickel
as a Toxic Air Contaminant. 1991.
20. American Conference of Governmental Hygienists. Threshold
Limit Values for Chemical Substances and Physical Agents and
Biological Exposure Indices. CITY, STATE? 1995.
6-63
-------
21. .Memorandum from Maxwell, William, EPA to the EPA Air Docket
No. A-92-55. March 8, 1996. Dioxin from Hot-side ESP
units.
22. Letter and enclosure from Peck, Stephen C., Electric Power
Research Institute, to Maxwell, William H., EPA:BSD.
September 15, 1995. Transmittal of unlicensed Electric
Utility Trace Substances Synthesis Report.
23. National Research Council. Science and Judgment in Risk
Assessment. Washington DC. 1994.
24. 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.
25. U.S. Environmental Protection Agency. Guidelines for
Developmental Toxicity Risk Assessment; Office of Health and
Environmental Assessment, 54 FR 6398-63826, 1991.
26. 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.
27. Letter from Chupka, Marc W., DOE, to Maxwell, William H.,
EPA/CG. May 3, 1996.
6-64
-------
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°) . 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"10 and to birds.n"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 is a potential concern in
the U.S. as indicated by numerous fish advisories19 and
mercury-related water quality standards issued by State Agencies.
7.1.1 The Mercury Cycle
Environmental mercury passes through various environmental
compartments and may change physical form and chemical species
during this process; these movements are conceptualized as a
cycle. The mercury cycle has been studied and described in
several recent reports3-20"22 and its understanding continues to
undergo refinement.
The movement and distribution of mercury in the environment
can be confidently described only in general terms. There has
been increasing consensus on many, but not all, of the detailed
behaviors of mercury in the environment.23 The mercury cycle in
Figure 7-1 attempts to illustrate mercury release by both natural
and anthropogenic sources into environmental media. The figure
illustrates the various transport and transformation processes
that are expected to occur. The cycle is global in nature and
includes a number of intercompartmental transfers (e.g., water to
air to water) that result in a series of loops within the cycle.
7.1.2 Sources of Mercury
Contemporary anthropogenic emissions of mercury are only one
component of the global mercury cycle. Releases from human
activities today are adding to the mercury reservoirs that
7-1
-------
-------
already exist in land, water, and air, both naturally and as a
result of previous human activities. Given the present
understanding of the global mercury cycle, the flux of mercury
from the atmosphere to land or water at any one location is
comprised of contributions from the following:
. • The natural global cycle
• The global cycle perturbed by human activities
• Regional sources
• Local sources.
Local sources could also include other routes of pollutant
release such as direct water discharges in addition to air
emissions. Past uses of mercury, such as fungicide application
to crops, are also a component of the present mercury burden in
the environment.
Different techniques have been used to estimate the mercury
concentrations in environmental media that occurred prior to the
onset of the industrial revolution. Not surprisingly, there are
a broad range of estimates and a great deal of uncertainty
associated with each. When these estimates are considered
together, they indicate that between 40 and 75 percent of the
current atmospheric mercury concentrations are the result of
anthropogenic releases.24"27 This overall range appears to be in
agreement with the several-fold increase noted in inferred
mercury deposition rates.28"30 The percentage of current total
atmospheric mercury that is of anthropogenic origin may be 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
reemitted to the atmosphere from existing reservoirs.
Because of the continuous cycling of this element in the
environment, it is impossible to separate current mercury
concentrations in the environment by origin (i.e., anthropogenic
or natural). For example, stack releases of Hg° may be oxidized
and deposit as Hg(II) far from the source; the deposited mercury
may be reduced and re-emitted as Hg° only to be deposited again
continents away.
Anthropogenic mercury releases are believed to be dominated
on the national scale (and global) by industrial processes and by
combustion sources that release mercury into the atmosphere.
Stack emissions may include both vapor and particulate forms with
7-3
-------
both divalent and Hg° species in the various fractions. The
analytic methods needed to speciate mercury in exit gases and
emission plumes are being refined, and, at present, there is no
scientific consensus in this field of study.
Chemical reactions occurring in the emission plume are also
possible. Methods are still being developed to examine the
chemistry of emitted mercury close to the emissions source. The
speciation of mercury emissions is thought to depend on the fuel
used (e.g., coal or oil), flue gas cleaning, and operating
temperature. The exit stream may range from almost entirely
Hg(II) to nearly all Hg°. The Hg° fraction is thought to be
emitted primarily in the gas phase although exit streams
containing soot can bind up some fraction of Hg°. The divalent
fraction is thought to be split between gaseous and
particle-bound phases.31 Much of this Hg(II) is thought to be
HgCl2.32
An emission factor-based approach was used to develop the
nationwide emission estimates for utility boilers. The emission
estimates for utilities are based on ratios of mass mercury
emissions to measures of source activities and nationwide boiler
activity levels. The mercury emissions data are estimates;
uncertainties occur in the measurement techniques, emission
factors, estimates of pollutant control efficiency and nationwide
source class activity levels. Details of the emission factor
approach are described in chapter 3. Oil- and gas-fired electric
utilities are estimated to emit 0.25 ton/yr, arid coal-fired
electric utilities are estimated to emit 51 tori/yr. This
represents approximately 21 percent of the total annual
anthropogenic mercury emissions in the U.S. estimated to be about
250 tons for the years 1990 through 1992. If one assumes that
the estimate that between 40 and 75 percent of the total current
atmospheric emissions are the result of anthropogenic activity is
correct, one can calculate that mercury emissions fron. utilities
represent roughly 8 to 15 percent of total emissions of mercury
in the U.S. (i.e., anthropogenic plus other sources [e.g.,
natural emissions]).
Recent estimates of global anthropogenic mercury emissions
are about 4,400 tons.24 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.24 Given
this estimate U.S. anthropogenic mercury emissions could account
for 6 percent of the global total, and electric utilities in the
U.S. would account for about 1 percent of the global
anthropogenic mercury emissions, using 1990 emission estimates.
7-4
-------
.7.1.3 Deposition of Mercury
Most of the mercury species in the atmosphere are in the
vapor phase, and measurement data indicate that the vapor phase
consists primarily of Hg°.3 Particulate-bound mercury species
have also been measured in the atmosphere.33 Concentrations of
vapor phase Hg(II) have been predicted34 near some anthropogenic
emission sources. Although methods are being developed,35 at
this time there is no generally accepted method for determining
the vapor fraction in the divalent form.
Atmospheric Hg(II) species, in either the vapor or
particulate phase, are thought to be subject to more rapid
removal from the atmosphere than Hg°.34"37 This is the result of
the reactivity and water solubility of the gaseous Hg(II) and the
gravitational forces that act on particles. Vapor phase Hg(II)
species are expected to be scavenged readily by precipitation.34
These mercury species have lower Henry's law constants than Hg°
and are assumed to partition strongly to the aqueous phase. Dry
deposition (i.e., deposition in the absence of precipitation) of
vapor phase Hg(II) is thought to be significant because of its
reactivity with surface material. Because of the anticipated
rapid depletion, these vapor-phase species are anticipated to be
detected in the atmosphere only in the proximity of emission
sources. The dry deposition of particulate-bound Hg(II), i.e.,
Hg(II)(P), is predicted to be dependent on atmospheric conditions
and particle size. Particulate mercury is also assumed to be
subject to wet deposition due to scavenging by precipitation.
In contrast, Hg° vapor is not thought to be susceptible to
any major process resulting in direct deposition to the earth's
surface. Elemental mercury has a strong tendency to remain
airborne34, and, on non-assimilating surfaces, Hg° deposition
appears negligible.38 Elemental mercury can be formed in soil
and water39 through the reduction of Hg(II) species by various
mechanisms; Hg° typically comprises a minor fraction of the total
mercury concentration in these systems. Once formed, the Hg° is
expected to volatilize to the atmosphere.24
There appears to be a potential for deposition of Hg° via
plant-leaf uptake. Lindberg et al.36 indicated that forest
canopies could accumulate Hg° 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 Hg° from the atmosphere, resulting in a deposition velocity.
Recent evidence39 indicates that this process does occur but only
at air concentrations of Hg° well above background for a typical
forest area. At more common mercury levels, the forest appears
to act as a source of Hg° to the atmosphere, with the measured
7-5
-------
mercury flux in the upward direction. Lindberg et al.38 noted
this may be explained by the volatilization of Hg° from the
canopy/soil system, most likely the soil. Hanson et al.40 stated
that "dry foliar surfaces in terrestrial forest landscapes may
not be a net sink for atmospheric Hg°, 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." Mosbaek et al.41 also 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 Hg° released from the
plant-soil system greatly exceeds the amount collected from the
air by the plants. It is also likely that many plants accumulate
airborne mercury to certain concentrations, after which net
deposition of Hg° does not occur.
The average residence time of Hg° in the atmosphere is
approximately 1 year.3 Atmospheric Hg° vapor is thought to
deposit to the earth through a series of atmospheric reactions
that occur in the aqueous phase (cloud droplets). These reactions
may both oxidize Hg° to Hg(II) and reduce the Hg(II) to Hg°. The
most important reactions are the oxidation of Hg° by ozone, the
reduction of Hg(II) by sulfite (S03~2) ions, and the production of
Hg(II)(P) by complexation with soot:
Hg°(g) -> Hg°(aq) Reaction 1
Hg°(aq) + 03(aq) -> Hg(II)(aq) Reaction 2
Hg(II)(aq) + soot/possible evaporation -> Reaction 3
Hg(II)(p)
Hg(II) (aq) + S Hg°(aq) Reaction 4
(g) = gas phase molecule
(aq) = aqueous phase molecule
(p) = particulate phase molecule
Hg(II) = divalent mercury = Hg2*
The divalent species that result from the oxidation of Hg° by
ozone can be reduced back to Hg° by sulfite; however, the
oxidation of Hg° by ozone is a much faster reaction than the
reduction by sulfite. Thus, a steady-state concentration of
aqueous Hg(II) is built up in the atmosphere and can be expressed
as a function of the concentrations of Hg°(g), 03(g), H"
(representing acids) and S02(g).42 Note that IT and S02(g),
although not apparent in the listed atmospheric reactions,
control the formation of sulfite.
7-6
-------
The aqueous-phase Hg(II) is expected to be susceptible to
atmospheric removal via wet deposition (precipitation). The
third reaction, however, may transform most of the aqueous-phase
Hg(II) into the particulate form. Sulfur ions in soot are
expected to bind aqueous-phase Hg(II), forming mercury-bearing
particles, which may be removed from the atmosphere through wet
and dry processes. Concentrations of atmospheric aqueous-phase
Hg(II) are expected to limit the reaction with soot. The mercury
deposition that results from this proposed chemical pathway may
occur far from the mercury emission sources due to the slow
reaction rates of the reactions.25
7.1.4 Mercury in Soil
Mercury species in soil are subject to a wide array of
chemical and biological reactions. Soil conditions (e.g., pH,
temperature, and humic content) are typically favorable for the
formation of inorganic Hg(II) compounds such as HgCl2, Hg(OH)2,
and complexes of these molecules and organic anions.43 Although
inorganic Hg(II) compounds are quite soluble, they form complexes
with soil organic matter (mainly fulvic and humic acids) and
mineral colloids. This complexing behavior 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 particles 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 results of analyses
presented in Appendix L, which detail the calibration of
soil-water partition coefficients in the watershed model, are
consistent with these observations. The affinity of mercury
species for soil results in soil acting as a large reservoir for
mercury emissions.22-44
Elemental mercury can be formed in soil by reduction of
Hg(II) compounds/complexes mediated by humic substances.45 This Hg°
is expected to vaporize to the atmosphere. Methylmercury may also
be formed in soil by various microbiological and abiotic processes
acting on inorganic mercury. Typically 1 to 3 percent of the
total mercury in surface soil is methylmercury.46 Some garden
soils with high organic content and some contaminated sediments
may contain a higher percentage of methylmercury.47"49
7-7
-------
7.1.5 Plant and Animal Uptake of Mercury
The inorganic Hg(II) and methylmercury complexes in soil are
available theoretically for plant uptake and subsequent transfers
through the terrestrial food web. Although scientific
uncertainty and inter-plant variability are noted, it appears
that under typical environmental conditions, plant uptake from
soils (especially to aboveground parts of plants) is quite
low.41-43-50 For example, Mosbaek et al.41 determined that
atmospheric mercury uptake accounts for 90 to 95 percent of the
total mercury content of the leafy parts of lettuce and rye grass
and 40 to 70 percent of the concentration in radish roots.
Concentrations of mercury in leafy vegetables generally exceed
those of legumes and fruits,47'51 where it is not clear whether the
mercury content results from air and/or soil uptake.
Most plant uptake studies do not explicitly measure both the
surrounding soil and air concentrations, and most uptake studies
do not examine the species of mercury present in soil. It is
unclear if some of the measured variability is attributable to
the species of mercury present in the soil or atmosphere, soil
properties, or species-specific differences between plants.
Further study of the mechanisms of mercury uptake by plants
is needed. Identification of the mercury species present in
plants is also needed and may indicate the ability of mercury to
be transferred through the food web. Studies indicate that some
of the Hg° absorbed from the air is readily converted to Hg(ll)
species in plant tissues.47 At least some plants appear to
methylate some of the accumulated mercury.52
Although not well understood, plant uptake of environmental
mercury is typically low. Divalent mercury, which appears to be
the primary form, does not appear to be as bioavailable to
animals as methylmercury. Livestock typically accumulate very
low concentrations of mercury in their tissues from foraging or
silage/grain consumption.
7.1.6 Mercury in the Freshwater Ecosystem
There are a number of pathways by which mercury can enter
the freshwater environment: wet or dry deposition of atmospheric
mercury directly into the waterbody; deposition of atmospheric
mercury to the watershed followed by transport to waterbodies in
runoff (bound to suspended soil/humus or attached to dissolved
organic carbon); and leaching of soil mercury species into the
waterbody through ground water flow in the upper soil layers.
Once in a waterbody, 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
7-8
-------
biota. The movements of mercury through any specific waterbody
may be unique. The same complexation and transformation
processes that occur to mercury species in soil can occur in the
aquatic medium in addition to processes specific to the aqueous
environment. Only mercury in the water column, the sediment, and
other aquatic biota appears to be available to aquatic organisms
for uptake.
Much of the mercury in waterbodies binds to matter suspended
in the water column and bottom sediment (See Appendix L) . The
primary form of dissolved mercury in the water column is
Hg(II)(see Appendix J). Elemental mercury is produced in
freshwater by humic acid reduction of divalent species or
demethylation of methylmercury. Some of the Hg° will remain in
the water column as a dissolved gas, but most will volatilize.
Typically, less than 10 percent of the mercury dissolved in the
water column exists as a methylmercury complex (See Appendix J).
Methylation appears to be a key step in the entrance of mercury
into the food chain.53
Most of the methylmercury in the water body is the result
of methylation of inorganic Hg(II). Biotransformation of
inorganic mercury species to methylated organic species in
waterbodies can occur in the bottom sediment54 and the water
column.55 Abiotic processes (e.g., humic and fulvic acids in
solution) can also methylate the mercuric ion.56 There appears
to be a large degree of variability among waterbodies concerning
the processes that methylate mercury. Bacterial methylation rates
appear to increase under aerobic conditions, high temperatures,57
and low pH.54'55 Increased quantities of the mercuric species, the
proper biologic community, and adequate suspended soil load and
sedimentation rate are also important factors; however, these are
not the only factors affecting methylation.57 Anthropogenic
acidification of lakes appears to increase methylation rates as
well.54 Not all mercury compounds entering an aquatic ecosystem
are methylated, and demethylation reactions55 as well as
volatilization of dimethylmercury decrease the amount of
methylmercury available in the aquatic environment.
Most of the methylmercury in the aquatic ecosystem is
present in biota, particularly fish.22-58 Methylmercury
accumulates in fish through the aquatic food web; nearly 100
percent of the mercury found in fish muscle tissue is
methylated.58 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 waterbody. A relationship exists between methylmercury
7-9
-------
content in fish and lake pH, with higher methylmercury content in
fish tissue typically found in more acidic lakes.54-59 (See
Appendix J for more details.)
7.2 MERCURY CONCENTRATIONS IN BIOTA
This section provides data on measured levels of mercury in
various biota as reported in the scientific literature or as
provided by other Federal agencies. It is not feasible, at this
time, to determine the sources or source categories responsible
for these environmental levels. There are many categories of
natural sources (e.g., volcanoes, forest fires) and anthropogenic
sources (e.g., municipal waste combustors, medical waste
incinerators, lead smelters, chlor-alkali plants, electric
utilities) of mercury emissions in the U.S. and worldwide. The
utilities are one category of sources of mercury emissions.
The data are being presented to provide general information
on estimated levels of mercury in the environment. This
information is useful for understanding where mercury is found in
the environment and the relative magnitudes of mercury in each of
the media or biota. In addition, these data aid in the
understanding of environmental fate of mercury, including the
bioaccumulation potential of mercury in various species.
In the U.S., the major nonoccupational routes of exposure
are expected to be to Hg° released from dental amalgams and to
mercury in foods. Because of bioaccumulation in the aquatic and
marine food webs, the mercury concentrations in fish muscle
tissues show the greatest potential for exposure through foods.
Mercury concentrations in fish vary greatly, often showing little
correlation with proximity to mercury emission sources. Fish in
lakes seemingly far removed from anthropogenic sources have been
found to have mercury levels of potential concern. In a given
waterbody, fish methylmercury concentrations are generally
thought to increase with trophic level, as well as with the age
and size of the fish. Other lake characteristics such as pH have
been found to correlate with methylmercury concentrations in
fish.
Tables 7-1 and 7-2 summarize some of the measured mercury
concentrations in freshwater sport fish and saltwater fish,
respectively. The values in these tables are the average
measured concentrations for the fish samples in these studies.
They may or may not accurately represent the concentrations in
these fish species obtained from other waterbodies or obtained at
other times. Also, fish mercury concentrations can vary
substantially depending on the size, age, and type of fish within
7-10
-------
3
O
M
d)
(0
4J
O
EH
W
-H
fc,
O
a
-U
1
w
0)
c
•H
w
c
0
-H
JJ
(0
M
JJ
G
Q)
U
G
O
U
U
4J
T5 0)
3 tn
W \
(0 tn
0)
U
C
QJ
<+-
QJ
cc
3
£
CO
CO
0.
E £
O 09
11
1/9
Cfl
co
m
09
1
CO
1
«
U
'.a.
"5
i£
TJ
5)
g
"5
0)
Bahnick
^-
d
CM
CM
d
r-
CM
d
en
o
d
co
d
^.
CO
o
co
d
CM
m
d
[Nationwide Survey
(338 locations)
ifi
NJDEPE
£
*""
X
(0
05
CO
in
o
d
CM
00
CM'
§
d
INJDEPE, 1992
^_
ID
*Q
05
C
1
55
in
co co t- co co
CM co TJ- .
09 09 CD 09 i
CM 4 «6 CO O
1 2 Adirondack Lakes
07
C
Sorense
in
T—
CO
d
CM
in
4
d
c
IMean concentrations
65 northern MN lakes
m
to
UJ
fx
•>—
O
O
0
'OJ
09
3
m
in
O
d
Q.
CD
O
in CM
«— t—
d d
Jd i:
4-* *-•
II
-I OT
IAshtabula River, OH;
means
S<
Q.
Ill
O
d
D.
CO
O
CM
y
«*«
LL
Isaginaw River, Ml;
means
in
(C
CO
+•*
CO
s
0
8
*^~
*—
d
CO
6
d
1 Northern Michigan
lakes
to
CO
in
CD
5
in
^*
d
0
CM
d
CO
d
o
4-
in
0
d
V
o
in co
O CN
d d
V
o
4-*
in CM
O f»-
d d
V
co
CM
d
^
d
en
co
O)
d
en
^
o
c
134 Northern Wisconsi
Lakes
0)
rH
7-11
-------
•d
0)
•H
JJ
c
o
u
H
I
f-
(1)
rH
•a
CD
U
C
£
•£
"5
tr
S
o
^~
1
**••
C
CO
a.
E S
O 05
*•* ^
CD >£
CO
CO
CQ
1
To
"5
1
.0
"it
J^
D.
•o
3
V)
S.
"5
CD
DJ
CO
O
•
T
S
*J
d
CO
03
2
CO
o
LL
CO
in
S
LT
LU
Q
CO
O
LL
in co
°°CM
o ,
1 CN
!•> CM
°d
0 ..
.. CO
CO C
if
-1 w
£
2
CO
>
Florida Surface
r*
"3
4-1
CD
C
|
U
8
CO
4
CM
O
"o
3
19 Canadian Shii
Lakes
N
"5
£
CM
CO
O)
o
d
CO
CM
T™
o
6
y Ontario Lakes
&•£ E
03 "5
V* «^- ~J
•2^5
o
7-12
-------
Table 7-2. Mercury Concentrations Measured in Some Marine Fish
and Shellfish
Fish
Mercury
Concentration
U/g/g, wet
weight}'
Comments
Tuna
0.136-0.264
Based on NMFS" data 75, the mean concentrations measured in 3
types of tuna are as follows: albacore tuna (0.264 fjg/g), skipjack
tuna (0.136 fig/g) and yellowfin tuna (0.218 //g/g). The U.S. FDA
measured the methylmercury concentration in 220 samples of
canned tuna in 1991; the average amount of methylmercury
measured in these samples was 0.17 //g/g and the measured
range was <0.1 - 0.75 ^g/g.76 Richardson, et al.77 reported an
average mercury concentration in Canadian tuna as 0.195 ug/g
(wet weight); the range of measured concentrations in tuna was
reported as < 0.01-0.97 ug/g (wet weight).
Shrimp
0.024 - 0.074
Based on NMFS data75, the mean concentrations measured in
seven types of shrimp are as follows: royal red shrimp (0.074
fjg/g), white shrimp (0.054 fjg/g), brown shrimp (0.048 fjg/g),
ocean shrimp (0.053 fjg/g), pink shrimp (0.031 fjg/g), pink
northern shrimp (0.024 fjg/g) and Alaska (sidestripe) shrimp
(0.042 fjg/g).
Pollack
0.04-0.15
The Pesticide and Chemical Contaminant Data Base for U.S. FDA
(1991/1992) reports the methylmercury concentration in pollack
in commerce as 0.04 ,ug/g- The value 0.15 ^g/g came from
NMFS.75
Salmon
0.019-0.063
Based on NMFS data75, the mean concentrations measured in five
types of Salmon are as follows: pink (0.019 fjglg), chum (0.030
fjg/g), coho (0.038 fjg/g), sockeye (0.027 fjg/g), and chinook
(0.063 fjg/g).
Cod
0.114-0.127
Based on NMFS data75, the mercury concentrations Atlantic Cod
is 0.114pg/g and for Pacific Cod is 0.127//g/g.
Clam
0.027 - 0.034
Based on NMFS data75, the mean concentrations measured in four
types of clam are as follows: hard (or quahog) clam (0.034 jjg/g),
Pacific littleneck clam (0 //g/g), soft clam (0.027 /jg/g), and
geoduck clam (0.032 /jg/g).
Flatfish (Flounder)
0.066-0.151
Based on NMFS data75, the mean concentrations measured in nine
types of flounder are as follows: Gulf (0.147 fjg/g), summer
(0.127 fjg/g), southern (0.078 fjglg), four-spot (0.090 j/g/g),
windowpane (0.151 //g/g), arrowtooth (0.020 fjg/g), witch (0.083
fjglg), yellowtail (0.067 fjg/g), and winter (0.066//g/g).
Crab
0.07-0.183
Based on NMFS data75, the mean concentrations measured in five
types of crab are as follows: blue crab (0.140 A/g/g), dungeness
crab (0.183 fjglg), king crab (0.070 fjg/g), tanner crab (C. opilio)
(0.088 /jg/g), and tanner crab (C. baird!) (0.102 fjg/g).
7-13
-------
Table 7-2. Continued
Fish
Mercury
Concentration
U/g/g, wet
weight)*
Comments
Lobster
0.108-0.378
Based on NMFS data75, the mean concentration in 3 types of
lobster are as follows: Spiny (Atlantic) lobster (0.108 fjg/g),
Spiny (Pacific) lobster (0.210 j/g/g), and Northern (Atlantic)
lobster (0.378 fjglg).
Fish Sticks
0.02-0.15"
In an FDA memorandum78 dated September 24, 1996, "The
correct level for fish sticks is 0.02 ppm... These unpublished
findings reflect work under the FDA Total Diet Study for the years
1982 to 1990."
Scallop
0.004-0.101
Based on NMFS data75, the mean concentrations measured in four
types of scallop are as follows: sea (smooth) scallop (0.101
fJQ/9), Atlantic Bay scallop (0.038 fjg/g), calico scallop (0.026
A/g/g), and pink scallop (0.004 fjg/g).
Note: The U.S. EPA has not determined the source(s) of emissions or effluent responsible for the fish mercury
concentrations reported in Tables 7-1 and 7-2. There are many categories of natural and anthropogenic
sources of mercury emissions in the U.S. and worldwide. The fossil-fuel-fired utilities in the U.S., which are
the sources being studied in this report, are just one category of sources of mercury emissions. The U.S. EPA
cannot determine if, or how much, the utilities contribute to the mercury levels reported in Tables 7-1 and 7-2.
There are several other species of saltwater commercial fish (e.g., shark, marlin, swordfish, grouper, croaker)
that have been found to contain mercury at concentrations higher than the 10 species included in Table 7-2.
' Unless otherwise specified the source of the mercury concentration data is the National Marine Fisheries
Service (NMFS).75
b In the USDA description of Food Codes used in coding the USDA Coding File79, Fish Sticks are described as
containing Atlantic Pollack or Atlantic Cod. The concentration of mercury in these fishes as reported in
NMFS75 are 0.15 ppm for raw Atlantic Pollack and 0.12 for raw Atlantic Cod. The recipe file lists five types of
preparation for fish sticks: 1) Cooked, 2) Baked/Broiled, 3) Breaded/Battered Baked, 4) Floured/Breaded, Fried
and, 5) Battered, Fried. After adjusting for moisture loss asociated with cooking and for the addition of
non-fish ingredients (e.g., Batter), the concentration in the fish sticks are expected to be roughly the same
concentrations measured in raw Cod and Pollack.
7-14
-------
a species or among different species. It is also important to
note that the mercury concentrations in fish are in all
likelihood the result of both natural and anthropogenic
activities.
The modeling assessment did not predict levels in saltwater
fish. However, the EPA believes that providing data on measured
levels in saltwater fish (Table 7-2) is useful since it is one of
the biota that contain mercury. The EPA has not determined if,
or how much, mercury emissions from U.S. utilities contribute to
these saltwater fish levels. Such an analysis was not possible
for this report because of the limitations in the available data,
methods, and models and because of the immense complexity of such
an assessment.
The little recent data available on mercury in meat products
(e.g., beef, chicken, pork) show concentrations to be very low
(near the detection limits) for concentrations of both Hg(II) and
methylmercury. Meat consumption does not appear to be a major
route of exposure to mercury for humans, especially in comparison
to mercury concentrations in fish tissues.
Mercury concentrations in green plants are also typically
low. Of the plants that do have detectable-mercury
concentrations, the concentrations tend to be highest in leafy
vegetables. Plants grown in mercury-contaminated areas appear to
accumulate more mercury than plants in areas where background
concentrations are prevalent. However, there are no other
noticeable trends in plant concentrations, with mercury levels
varying widely among plants and studies. A comparison of
measured mercury concentrations in fish with concentrations in
plants shows that fish have higher concentrations.
7.3 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 Smith80 measured mercury levels in
environmental media and biota around a 200-MW coal-fired utility
7-15
-------
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 Kinnison81 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 with 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.
There are two recent reports of atmospheric mercury
measurements in the vicinity of multiple anthropogenic emissions
sources. Both studies are of short duration but show elevated
mercury concentrations in the local atmosphere or locally
collected rain.
Dvonch et al.33 conducted a 20-day mercury sampling study
during August and September of 1993 at four sites in Broward
County, Florida. This county contains the city of Ft. Lauderdale
as well as an oil-fired utility and a municipal waste combustion
(MWC) facility. One of the sample collection sites (site 4) was
located 300 m southwest of the MWC facility. Daily measurements
of atmospheric particulate and vapor-phase mercury were collected
at three of the four sites; daily atmospheric concentrations were
not collected at the site near the MWC (site 4) and daily
precipitation samples were collected at all sites. The average
vapor and particulate phase atmospheric mercury concentrations
were higher at the inland sites than at the site near the
7-16
-------
Atlantic Ocean, which was considered a background site by the
authors. Diurnal variations were also noted; elevated
concentrations were measured at night. For example, at site 2,
an inland site, the average nighttime vapor-phase concentration
was 4.5 ng/m3;; see Table 7-3 for comparison data. This was
attributed to little vertical mixing and lower mixing heights
that occur in this area at night. Particulate mercury made up
less than 5 percent of the total atmospheric mercury. Mercury
concentrations in precipitation samples at the four sites were
variable; the highest mean concentrations were measured at the
inland sites. Given the high levels of precipitation in this
area of the U.S. and the short collection period, it is not
appropriate to extend this analysis beyond the time frame
measured. These mercury concentrations, shown in Table 7-3, are,
nonetheless, elevated when compared with the background site.
Keeler et al.82 and Lamborg et al.83 reported results of a
10-day atmospheric mercury measurement collection at two sites
(labeled as sites A and B) in Detroit, Michigan. There is a
large MWC 9 km from site A and a sludge combustor 5 km from
site B, and there is a coal-fired utility in the city as well.
The vapor-phase mercury concentration encountered at site B
during the first days of the experiment exceeded the capacity of
the measurement device. Subsequent analyses indicated that the
concentrations of mercury encountered were significantly higher
than other reported U.S. observations. Table 7-4 presents the
concentrations measured at each site.
7.4 MODELING THE FATE OF MERCURY EMISSIONS FROM UTILITIES
It was decided that existing measured mercury data alone
were insufficient for an adequate nationwide assessment of the
impact of mercury releases from utility boilers; the U.S. EPA
therefore chose to examine the fate of mercury released from the
stacks of utility boilers using a series of fate and transport
models and hypothetical environmental constructs. The mercury
concentrations that result from utility emissions were estimated
for some environmental media and biota. Significant limitations
and uncertainties are associated with the quantitative
predictions of the modeling analysis. The results of the analysis
provide a qualitative assessment of the environmental fate and
transport as well as plausible estimates of environmental mercury
concentrations that result from utility emissions. In this
analysis, there were three main types of modeling efforts:
(1) modeling of. mercury atmospheric transport on a regional
(i.e., long-range transport) basis (beyond 50 km of
source); (2) modeling of mercury atmospheric transport on a local
scale (within 50 km of source); and (3) estimating environmental
7-17
-------
Table 7-3. Mercury Concentrations in the Atmosphere and Mercury-
Measured in Rainwater Collected in Broward County, Florida 33
Site Description
Background near
Atlantic Ocean
(site 1)
Inland (site 2)
Inland (site 3)
Inland (site 4),
300 m from MWC
Avg. vapor-phase
Hg cone., ng/m3
1.8
3.3
2.8
-
Avg. paniculate Hg
cone, pg/m3
34
51
49
-
Avg. tot. Hg cone.
in rain, ng/L (range)
35(15-56)
40(15-73)
46(14-130)
57 (43-81)
Avg. reactive Hg cone.
in rain, ng/L (range)
1.0(0.5-1.4)
1.9 (0.8-3.3)
2.0 (1.0-3.2)
2.5 (1.7-3.7)
- = not measured
Table 7-4. Mercury Concentrations Measured at Two Sites in the
Atmosphere over Detroit, Michigan82-83
Site
Detroit, Ml, site A
Detroit, Ml, site B
Mean vapor-phase mercury concentrations
in ng/m3 (max. meas. value)
>40.8 O74)
3.7 (8.5)
Mean particulate-phase mercury
concentrations in pg/m3, (max. meas. value)
341 (1,086)
297 (1,230)
concentrations using the indirect exposure model (IEM) along with
the results of the atmospheric modeling. (For a complete
description of the models utilized see Appendix M). The models
used for these aspects of the study are shown in Table 7-5.
Factors important in modeling the fate, transport, and
environmental concentrations are listed in Table 7-6. This table
briefly describes the possible effects of these factors on the
fate and transport of mercury and the estimates of mercury
concentration in various media and the means by which these were
addressed.
7.4.1 Long-range Transport Analysis
The long-range transport analysis modeled site-specific,
utility mercury emissions data to generate mean, annual
atmospheric mercury concentrations and deposition estimates
across the continental U.S. (see Figures 7-2, 7-3, 7-4, and 7-5).
The RELMAP atmospheric model was used to model cumulative mercury
deposition from coal- and oil-fired utility boilers in the U.S.
Assumptions were made concerning the form and species of mercury
emitted from each source class.
7-18
-------
Table 7-5. Models Used to Predict Mercury Air Concentrations,
Deposition Fluxes, and Environmental Concentrations
Model
RELMAP
COMPDEP II
IEM2
Description
Predicts average annual atmospheric mercury concentration and wet and dry deposition flux
for each 40-km2 grid in the United States due to all anthropogenic sources of mercury in the
United States
Predicts average concentration and deposition fluxes within 50 km of emission source
Predicts environmental concentrations based on air concentrations and deposition rates to
watershed and waterbody
Table 7-6. Factors Potentially Important in Multipathway
Modeling of Mercury — How They are Addressed in This Assessment
Factor
Mercury emission
rates from stack
Mercury species
emitted from stack
Form of mercury
emitted from stack
Deposition
differences
between vapor and
particulate-bound
mercury
Transformations of
mercury after
emission from
source
Facility locations
Location relative to
local mercury
source
Contribution from
nonlocal sources of
mercury
Importance and possible effect on mercury
exposure
Increased emissions will result in a greater
chance of adverse impacts on environment.
More soluble species will tend to deposit
closer to the source.
Transport properties can be highly
dependent on form.
Vapor-phase forms may deposit
significantly faster than particulate-bound
forms.
Relatively nontoxic forms emitted from
source may be transformed into more toxic
compounds.
Effects of meteorology and terrain may be
significant.
Receptors located downwind are more
likely to have higher exposures. Influence
of distance depends on source type.
Important to keep predicted impacts of
local sources in perspective.
Means of addressing in this study
Emissions of model plants based on
emissions inventory-
Two species considered to be emitted
from source: elemental and divalent
mercury
Both vapor and particle-bound fractions
considered.
Deposition (wet and dry) of vapor-phase
forms calculated separately from
particulate-bound deposition.
Equilibrium fractions estimated in all
environmental media for three mercury
species: elemental mercury, divalent
species, and methylmercury.
Both a humid and less humid site
considered. Effect of terrain on results
addressed separately.
Three distances in downwind direction
considered.
Results of local mercury source are
combined with estimate of impact from
nonlocal sources from RELMAP.
7-19
-------
Table 7-6. Continued
Factor
Importance and possible effect on mercury
exposure
Means of addressing in this study
Uncertainty
Reduces confidence in ability to estimate
exposure accurately.
Probabilistic capabilities possible for any
combination of sources and scenarios. In
the current study, limited uncertainty
analyses conducted for major aspects of
atmospheric transport modeling.
RELMAP = Regional Lagrangian Model of Air Pollution.
7.4.1.1 RELMAP Model Uncertainty. In general, it is
believed that the RELMAP will tend: (1) to overestimate Hg values
in urban areas; and (2) to underestimate Hg values in rural
areas and in the urban center of larger cities . There were
limited data available from several years to check the model and
confirm those beliefs. The RELMAP analysis was somewhat generic
because the evaluation was based on the meteorology of 1989,
thought to be a reasonably typical year, and used 1990 Hg
emission estimates. The discussion below summarizes the limited
comparison between model predictions and measured data. Please
note that the model predictions described in this section include
other major sources and a natural background concentration.
Overall, the RELMAP seems to predict Hg values within a factor of
2 of measured values and is relatively unbiased in its
predictions.
For the ambient vapor-phase mercury concentrations, the data
collected at three remote sites agreed very closely with the
RELMAP predictions. With limited data in a more urbanized
setting in Florida, the RELMAP underpredicted average
concentrations at two sites by about a factor of 2 and was in
agreement with the data at a third site. For particulate phase
Hg, the RELMAP underpredicted concentrations by a factor of 2 at
Detroit (18-day sample) and was in approximate agreement in
Broward County in Florida (2 weeks of data) . At three other
sites (Ann Arbor and Pellston, MI, and Underbill, VT) , the
modeled concentrations were in approximate agreement with a
tendency for slight overestimates of the measured values.
For wet deposition, there were 10 sites in Minnesota, upper
Michigan, and northeastern North Dakota that recorded averages
from about 4 to 10 ug/m2 (multiple years of data). The RELMAP
concentrations for these 10 sites ranged from 2 to 11 ug/m2, thus
showing reasonable agreement with a slight tendency to
overestimate. At two lake sites in Wisconsin, the RELMAP
7-20
-------
7-21
-------
7-22
-------
7-23
-------
7-24
-------
estimates were very close to recorded values. In the early 1990s,
there were four sites in somewhat less remote areas that recorded
data that were overpredicted by the RELMAP, but the
overprediction was within a factor of 2.
Urban wet deposition comparisons cannot be made due to the
paucity of long-term data. Limited data (20 days) from four
sites in the Fort Lauderdale area indicated that the RELMAP was
approximately on target, with a tendency to underestimate.
7.4.1.2 Priliminary Observations from Long-range Transport
Modeling Analysis. Based on this modeling analysis, the response
from internal and external peer reviewers of the atmospheric
mercury modeling strategy, and a review of recent scientific
literature on atmospheric mercury, the Agency has arrived at
several observations. Tables 7-7 and 7-8 provide specific data
on local and regional impacts at an eastern and western site.
The Agency has a high level of confidence in the following
observations:
1. Spatial Pattern of Emissions of Mercury from Utilities
The pattern of mercury emissions from utility boilers is
fairly uniform over the eastern half of the continental U.S. with
the highest concentration of emissions in the Ohio River Valley
and in other major coal mining areas.
The Agency's confidence in the following observations is
moderate:
2. Chemical and Physical Form of Emissions of Mercury from
Utilities
Engineering estimates of the chemical and physical forms of
the emissions of mercury from utilities in the continental U.S.
have been developed. A number of measurement studies have been
conducted with emission samples taken directly from the interior
of exhaust stacks. These samples of stack exhaust that may have
temperatures of more than 300° C (570° F) probably do not
accurately describe the chemical and physical forms of the
emissions as they manifest themselves in the regional-scale
atmosphere. The Agency has considered many possible effects of
cooling and dilution with ambient air of exhaust plume
constituents immediately after emission from the stack in the
development of the following estimates of mercury emissions to
the regional-scale atmosphere by utilities:
7-25
-------
W
0)
u
(0
c
o
•H
Cn
M-l
O
o
•H
JJ
•H
M
4J
C
O
U
ui
4J
U
Id
H
flj
C
o
•H
Cn
Q)
C
(C
U «
O Q)
J 4J
-H
O 0)
•H 4-)
C
•H
(0
«
I O
0) 0)
.c "3
;J2 *g)
3_
*NT *••*
* 0
gB
J= CO
^,
"i. *
O c
i_ O
•~ o
3
c?
^* *
C
o
CO
is
c
u
o
o
1
5
5
"tT
ct
Q.
5
— J
LU
cc
£
H
" CO
_O 3
'•5
Q.
CL
2
LU
CC
T3
2 g -O
•° £ >
| 1 II
nj i © ^
•fc "5 ~ T
"V O _L =
i s i! i
L, ||1|
CM 5 E » E
^ ^ ^ ^
CM CO CO LO
•i w w pj
b b b b
X X X X
co in «-_ co
LO Tt CO' CM
CM CO CO LO
^- ^- CO CM
CM «- O CO
LO co o) cn
b b b b
x x x x
CM CD CO CM
CM CD in LO
CO r- CM «t
•* LO T^ CO
in co O) cn
b b b °a
X X X X
in oo f» in
Tj- CO CM CM
^ ^ g? ^
oo «t CD cn
r« CD CM CM_
CO O5 CM CO
CO 00 CO CO
N fM (M CM
b b b b
X X X X
<°. ^ *°. *
CM CM CM CM
_£ w
1 2: 1 -°
^ 1 £ I
1 1 H
co !!L £ •-
£ 75 ~ ^r
E^ i! I
si! 11
^ ^ ^ #
CD P* CD O
CD t- rt CD
^t 00 !•» CD
CO 00 CD CD
(M tsj (N W
b b b b
X X X X
CO CM CO CO
CO CO CM CM
CD r^ co o
CD t- it CD
^t oo r^ CD
00 00 CO CD
b b b b
X X X X
Tt CO CM CM
CO (*» ^ O
in co in co
^ oo r^ CD
00 CO CD CO
b °0 °0 °0
X X X X
CD CO LO LO
CM CM CM CM
^ ^ ^ ^
co r^ co co
00 CD «-; LO
•^•' ^ co en
CD CD CO CD
(M (S M
-------
0
4J
Q)
u
(0
c
o
•H
Cn
CD CO
r- CM co en
b b b b
X X X X
t- «* CM in
CM r^ tO ^~
oo co in oo
t q oo CM
en en to co
«- CM co en
b b b b
X X X X
co in - 'g J3
ja .t: -° v
> =5 >• ~
1 1 ^ |
® i 0) *-
•f 1 ••= 1
•= O 7? O
• 5
o c o p
E S 1 - 1
•^ g, -o w -a
O ™ ® c eo
v^ CMC
eP e? eS **
in cn co JJ>
rv co q gj
co CD in •
^t in oo O
o b o o
X X X X
in ^- en co
en fv ^t ^~
in en eo •*
iv eo q in
co co in rv
•ft It) to O)
b b b b
X X X X
rv O) en co
CO OJ r-^ r-^
r- »- r- IV
TV en co rv"
>* in oo en
b b b b
X X X X
rv co cq «N
CD in co co
^ g? ^ g;
O CM T- •*
^ * q en
rv r-^ ^- rv
rv rv oo o>
b b b b
X X X X
in q q «--
cd rv co in
CD
111 |
° > s -°
ja .« -° >,
>- 5 > .*
.ti 3 ~ s=
1 "S 3 =
•§ * « £
•- "<5 -^ T
E 8 | 8 |
-1111
CN 5 E § E
7-27
-------
• About 24 metric ton/yr (50 percent of the total) are
emitted in the form of Hg°
• About 15 metric ton/yr (30 percent of the total) are
emitted in the form of Hg(II)
• About 10 metric ton/yr (20 percent of the total) are
emitted in the form of Hg(P).
3. Range of Total Mercury Deposition Magnitudes
Based on modeling and emission inventory analysis, the
Agency finds that the 10th and 90th percentile range of total
mercury deposition from utilities for the continental U.S. spans
more than 2 orders of magnitude.
4. Factors Contributing to Fate of Mercury Emissions from
Utilities
There are three principal factors contributing to the
modeled deposition patterns:
• Emission source locations
• Amount of divalent and particulate mercury emitted or
formed in the atmosphere
• Climate and meteorology.
Whether these factors are strongly controlling the actual
deposition patterns of utility emissions of mercury cannot be
tested due to the absence of source-specific deposition
observations.
The Agency's confidence in the following statements is low:
5. Fate of Different Forms of Mercury Emissions from Utilities
Modeling estimates of the transport and deposition of
utility air emissions of Hg°, Hg(II), and Hg(P) in the
continental U.S. have revealed the following atmospheric fates.
• Of the total amount of Hg° that is emitted, about
1 percent (0.3 metric ton/yr) may be dissolved in cloud
and rainwater, most of which is chemically transformed
into Hg(II) by dissolved tropospheric ozone and/or
adsorbed to particulate soot in aqueous suspension.
This mercury is subsequently deposited in rainfall and
7-28 '
-------
snowfall to the surface. The vast majority of Hg° does
not readily deposit to the surface and is transported
outside the United States or vertically diffused to the
free atmosphere to become part of the global cycle.
• Of the total amount of Hg(II) that is emitted, about
70 percent (10.2 metric ton/yr) deposits to the surface
through wet or dry processes within the continental
U.S. The remaining 30 percent is transported outside
the U.S. or is vertically diffused to the free
atmosphere to become part of the global cycle.
• Of the total amount of Hg(P) that is emitted, about
39 percent (3.8 metric ton/yr) deposits to the surface
through wet or dry processes within the continental
U.S. The remaining 61 percent is transported outside
the U.S. or is vertically diffused to the free
atmosphere to become part of the global cycle.
The Agency's confidence in the following observation is low
due to compounding scientific uncertainties regarding the
chemical and physical form of mercury emissions from utilities
and the lack of data on atmospheric transformations of mercury
between these different forms.
6. Area Most Impacted by Deposition of Mercury from Utilities
Based on modeling analysis of the wet and dry deposition of
utility air emissions of all forms of mercury within the
continental U.S., 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 U.S. and in numerous isolated locations where
relatively large coal-fired utilities are located.
The Agency's confidence in this observation is low since
there are no observational data with which to evaluate model
performance for utilities alone.
7-29
-------
7. Area 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
U.S., 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 Agency's confidence in this prediction is low since
there are no observational data with which to evaluate model
performance for utilities alone.
7.4.2 Local Analysis
The local impact analysis estimated the impacts of mercury
emissions within 50 km of individual coal- or oil-fired utility
plants. Natural and recycled sources of mercury were not
factored into the analysis. Model plants, which were developed
to represent actual utilities and their emissions, were located
at both a hypothetical western and eastern U.S. site (for
descriptions of the sites, see Appendix K). Mercury
concentrations in environmental media and biota were estimated
through fate and transport modeling of the stack emissions. The
contribution of regional mercury transported from utilities was
also included in the assessment of the impacts of the single
source. The models used in the local-scale analysis include a
modified version of the COMPDEP air dispersion model, the
Methodology for Assessing Health Risks Associated with Indirect
Exposure to Combustor Emissions,8* which is composed of a series
of fate and transport models, and a 1993 addendum to the 1990
methodology document.85 The COMPDEP is an air dispersion model,
based primarily on the COMPLEX I model (the "COMP" part of the
acronym) , which has been modified and updated to use wet and dry
deposition algorithms (the "DEP" part of the acronym). The
COMPLEX I is designed for screening analysis of air pollution
transport in complex terrain, i.e., for locations where terrain
features have elevations above the top of the stack of a
facility. Together these models were used to estimate mercury
concentrations in environmental media.
The three forms of mercury modeled were Hg° vapor, Hg(II)
vapor, and particulate mercury Hg(P). The vapor/particle (V/P)
ratio was assumed to be equal to the V/P ratio as it would exist
in stack emissions. Average ambient U.S. particle size profile
7-30
-------
data were used.86 Approximately 93 percent of the particles were
assumed to be accumulation particles, approximately
0.3-^m-diameter particles, and approximately 7 percent were
assumed to be coarse particles with a diameter of 5.7 ^m. Both
wet and dry deposition of mercury emissions was considered in the
analysis. The assumed deposition parameters for emitted forms
of mercury addressed in this study are compared in Table 7-9.
7.4.2.1 Multipathway Modeling. An updated methodology,
called Indirect Exposure Modeling 2 (IEM2), uses atmospheric
chemical loadings to perform mass balances on a watershed soil
element and a surface water element. The mass balances are
performed for total mercury, which is assumed to speciate into
three components: Hg°, Hg(II), and methylmercury. The fraction
of mercury in each of these components is specified for the soil
and the surface water elements. Loadings and chemical properties
are given for the individual mercury components, and the overall
mercury transport and loss rates are calculated by the
methodology. An overview of the IEM2 watershed modules is shown
in Figure 7-6.(See Appendix M for more details.)
The IEM2 first performs a terrestrial mass balance to obtain
mercury concentrations in watershed soils. Soil concentrations
are used in addition to vapor concentrations and deposition rates
to calculate concentrations in various food plants. These are
used, in turn, to calculate concentrations in terrestrial
agricultural animals (i.e., beef, pork, poultry). The IEM2 next
performs an aquatic mass balance driven by direct atmospheric
deposition as well as the runoff and erosion loads from watershed
soils. Methylmercury concentrations in freshwater fish are
estimated from total dissolved water concentrations using
bioaccumulation factors (BAFs). The modeling did not estimate
concentrations in marine fish.
For the analysis, the IEM2 methodology was expanded to
handle multiple chemical components in a steady-state
relationship. The fraction of each chemical component in the
soil and water column was specified by the user. The methodology
predicts the total chemical concentration in watershed soils and
the waterbody based on loading and dissipation rates specified
for each of the components. The model tracks the buildup of
watershed soil concentrations over the years given a steady
depositional load and long-term average hydrologic behavior.
To predict mercury levels in freshwater fish, a
bioaccumulation factor (BAF) approach was used instead of the
bioconcentration factor (BCF) approach described in the EPA's
Methodology for Assessing Health Risks Associated with Indirect
7-31
-------
Table 7-9. Comparison of Assumed Deposition Parameters for
Emitted Forms of Mercury8
Parameter
Dry deposition velocity (cm/s)
Washout ratio (unitless)
Wet deposition scavenging
coefficient (/s)b
Elemental mercury
vapor
0
1.6x104
SxlO^to 1 x10'5
Divalent
mercury vapor
0.3- 1.0
1.6x106
3x10-* to 1 xlO'3
Divalent
mercury particulate
0.2 - 1 .4C
NA
2.2 xlO"4 (light
precipitation) to 1.5 x10"3
(heavy precipitation)
* See Appendices J through M for the basis for these parameters and other assumptions.
b For elemental and divalent vapor, this is calculated by L = WP/L, where Wis the washout ratio, P is the
representative precipitation intensity for the hour, and L is the predicted mixing height for the hour. Due to
the dependence on mixing height, the upper end values of the ranges shown routinely occur even for light
precipitation.
c Based on particle density of 1.8 g/cm3, particle diameter of 2 urn, surface roughness of 0.3 m, and ambient
air temperature of 295 K.
Figure 7-6. Overview of the IEM2 watershed modules
Water column
benthic
transformation
'yds
Definitions for Figure'7-6
chemical concentration in upper soil
chemical concentration in water body
vapor phase chemical concentration in air
average dry deposition to watershed
average wet deposition to watershed
mg/L
mg/L
mg/yr
mg/yr
7-32
-------
Exposure to Combustor Emissions.84 A BAF measures the total
uptake rate from water, food, and sediments and is generally
derived from field studies. The BAF selected was based on a
modification of the concept described in the 1993 U.S. EPA Great
Lakes Water Quality Initiative,87 which previously developed a
BAF for mercury of 130,440 based on total measured mercury (all
species) in water. The BAFs used for this analysis were 66,200
and 335,000 for trophic level three and trophic level four,
respectively. (See Appendices J through M for background
information and further discussion of the models, BAFs, and other
assumptions, input data, and parameters used in RELMAP, COMPDEP,
and IEM modeling.)
7.4.2.2 Model Plants. Model plants representing four types
of facilities were developed to represent a range of mercury
emissions sources. The facilities selected were these: a large
coal-fired utility, a medium coal-fired utility, a small
coal-fired utility, and a medium oil-fired utility boiler.
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 assessment include mercury
emission rates, mercury speciation, and mercury
transport/deposition rates. Important model plant parameters
include 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 speciation. Values for these parameters are given in
Table 7-10. Emission estimates are thought to represent typical
levels of mercury emitted from existing utility sources.
Two generic sites were considered: a humid site east of
90 degrees west longitude, and an arid site west of 90 degrees
west longitude. The primary differences between the two
settings as parameterized are the assumed erosion characteristics
for the watershed and the amount of dilution flow from the
waterbody. The eastern site has generally steeper terrain in the
watershed than the other site (see Appendix K for details about
the specific site parameters used in the Assessment) .
7.4.2.3 Hypothetical Settings for Estimating Mercury
Concentrations in Various Media and Biota. Three types of model
plant settings were utilized: rural (agricultural), lacustrine
(near lakes), and urban. These three settings were selected
because of the variety they encompass and because each was
expected to provide "high-end" mercury concentrations in
7-33
-------
Table 7-10. Process Parameters for Model Plants
Model
plant
Large
Coal-fired
Medium
Coal-fired
Small
Coal-fired
Medium
Oil-fired
Plant
size
(MW)
975
375
100
285
Capacity
(% of yr)
65%
65%
65%
65%
Stack
height
(ft)
732
465
266
290
Stack
diamet.
(ft)
27
18
12
14
Hg emission
rate (kg/yr)
230
90
10
2
Spec. %
(Hg°/Hg(ll)/
Hg part)
50/30/20
50/30/20
50/30/20
50/30/20
Exit
veloc.
(m/sl
31.1
26.7
6.6
20.7
Exit
temp.
(°F)
273
275
322
environmental media of potential concern (e.g., elevated mercury
concentrations are expected in the waters of the lacustrine
setting).
Table 7-11 shows the predicted methylmercury concentrations
in soil, water, and air for the eastern and western sites. These
media concentrations were used to calculate predicted
methylmercury concentrations in biota for the two sites, and
these values are also shown in Table 7-11.
7.4.2.4 Discussion of COMPDEP and IEM2 Model Limitations,
Uncertainty, Validation, and Supporting Empirical Data. For the
COMPDEP model results predicted in this interim report, there
appear to be no measured data collected around utilities to
adequately assess whether the predicted air concentrations and
deposition rates can be corroborated. Section 7.3 of this report
briefly describes four studies conducted in the vicinity of
utilities: Anderson and Smith,80 Crockett and Kinnison,81 Dvonch
et al.,33 and Keeler et al. ,82 and Lamborg et al.83 (The last two
studies describe different data collected from the same area over
the same period of time.)
Mercury concentrations generated prior to 1980 are
generally suspect because of contamination and changes in
measurement methods. As a result, comparison of predicted values
with the measured data from Anderson and Smith80 and Crockett and
Kinnison,81 would not appear to be scientifically valid.
The data of Dvonch et al.33 and Keeler et al.82 and Lamborg
et al.83 are collected in areas where there are several
anthropogenic sources including a utility. The predicted air
7-34
-------
(0
J-)
o
•H
CQ
•O
C
•H
Ti
0)
S
c
•H
to
c
o
•H
(U
u
c
o
u
u
^
Q)
S
(0
co
1
r-
Q)
rH
•3
Concentrations in Biota (Mg/9)a'b
1 Summary Concentrations for Eastern Site
fll
K - ""
O
X
i
Poultry %MeHg|
Milk % MeHg
I
i
i
1
II
Fruits % MeHg
Surface Water 1
Concentration 1
(ng/l) 1
Untilled Soil
Concentration
(no/til
Mercury Total
Deposition
Rate (ug/m'/yr)
Mercury Air
Concentration
(no/m3)
f- •- CN CM
9999
t- *- * n
9999
CM 1? CN £
CO CM O) tO
*- T- *- 00
Q c— ID |D
O O O O
9999
«- tfi * CM
# # # #
«- ^ CN CO
CN CM CM CM
ID in to to
i i i 1
CN •- r* 10
* * 5 £
• • •*• iX
WO)**
CD ID CO O
* # # *
CO CO . .
<» »> " 2
S S 5 3
35 # # #
r» « co o>
CD r-* en en
S 0 6 S
CO (O ID IA
< O)
O O O O
£ 3 £ 3
CO CO 10 VD
O O O O
i co i i
o> p"» o co
0008
CO CO 0> CO
T- CO O (N
CN «- r* in
5888
5 S «? ,S
~ oi en PM
CN CN N M
9999
S <£ rS 5
2.5 km with RELMAP 50th Percentile
Large Coal-fired Utility Boiler
Medium Coal-fired Utility Boiler
Small Coal-tired Utility Boiler
Medium Oll-flred Utility Boiler
CN CN CM CN
9999
,1 S S S
ID ^ CO CN
CM CO CO CO
0000
Jfi JS * .£
O O> *" CD
*- 00 CO 10
# # # #
r- O O O
* ID CO CD
9999
^ ^ CM «-
CN CM CO CO
CN CN CM CM
9999
CD PV u> 10
m CM co co
CO •* ID 10
*?!*
CD CO O O
CM CO ID 10
9999
CN to O O
O) O) O O
t- r- CM CM
S S S S
.£ rS £ •
10 IO ID ID
00 CO Cl C)
o o o o
i s i *
v to * o
ID 10 ID ID
9999
* i cb co
CM Cfi CO CM
8888
£ $ S $
tO K ID ^
8888
ID CD P* ID
tf CO CM CN
CM CM CM IN
O O O O
i rS ,8 *
£ 60 flQ CD OB
Hill
1 ? ? « «
10 km with RELMAP 50
Large Coal-
Medium Coal-
Small Coal-
Medium Oll-
CM CN CN CN
9999
n o « cs
o o o o
m CN co in
ce co 10 in
r» o •- *-
in co co co
9999
c? n CN ^
n en co n
CN CM CN 01
§ i § g
$ $ z s
ID IA ID ID
*- ID N tO
ID tD ID ID
ssss
^ • c! £
co en ^ ^
« « in in
s??s
r5 S S S
O) O O O
2 S S S
10 tD 10 ID
A n a a
9999
ID ID ID LO
9999
to O CD D
^ CO CM CN
8888
ID ID ^ ^
8888
m co to to
CM CM CM CM
(N CN CM CN
O O O O
9999
10 CO 10 *
Hill
1 1 1 1 1
25 km and RELMAP BOt
Large Coal-
Medium Coal-
Small Coal-
Medium Oil-
a
to
"a>
"ra
CO
o
m
_c
CO
o
Concentre
Western Site
1 Summary Concentrations for
|l
n _
p u.
1
o
1
0.
Milk % MeHa '
o
i
a
I
9
II
1
3
u.
Surface Water 1
Concentration 1
Ina/l) 1
Unfilled Soil
Concentration
(no/Ql
^ f *Q
|Ij
rcury Air
centratlon
ig/m3)
S5
I
9999
CD <• *- ID
S 8 S S
c? cS 1 5
— CO CN i-
in r ". °.
o o o p
# # # #
•- »- p) *•
CM CN CN CM
§co co r*
.999
co in r- r»
„ O (0 CO
co 0 in CD
S o o o
a s s s
Sf S*?
§m 10 ID
O O O
9999
O W 00 T-
CM 10 O ^
£ £ S S
9 9 9 ?
co i co o
CO CO ID v-
n 10 01 o
S S S S
9999
f" r- ID O
^ *- CM CM
9999
O 9 9 9
^ r- oo «-
8885
+ -f + 9
£55^
co in ^ ^
!??5
H^5
r» r> en n
99°9
O co en »-
2.5 km with RELMAP 60th Percentile
Large Coal-fired Utility Boiler
Medium Coal-fired Utility Boiler
Small Coal-fired Utility Boiler
Medium Oll-flred Utility Boiler
9999
» *- CM
CN * p^ fv
g g g g
8 S S iS
to «- P* OD
r^ W CO CO
(D ID ID ID
O O O O
* * s. s
TO ^ tfJ O
# # # #
co r» in cc
co en o o
I- »- CN CN
? 9 9 9
r» 0) o o
S S S S
CO 9 9 9
ID ID V O
M CM CM CM
O O O O
9999
CN us 3 r*
eo ID CM •-
8855
4- + ' •
CN ^
$999
$ c? 5 £
^ oi * co
SCO CO CO
.999
5 c? S 5
10 km with RELMAP 60th Percentile
Large Coal-fired Utility Boiler
Medium Coal-fired Utility Boiler
Small Coal-fired Utility Boiler
Medium Oll-flred Utility Boiler
en en en en
0000
oi K *t «'
S S S S
$ & & s
•- •- 0) 00
* # * *
o 05 in *•
CO CO P* t-
S S g §
3 i s &
# * # #
CM CM CM CM
CO to r^ r^
9999
3 £ |2 «
^ *- r* CD
4D N ^ CO
ID (6 K P*
g S g g
c? S c? S
O CM O O)
ID* to r^ to
ID IA ID 10
o o o o
i 9 9 9
CM CM * 01
O ^ tO CO
S S S 8
9999
en o o o
o o o o
cS 5 c? S
CM CN (M N
9999
i c « i
r- 03 0) to
o o Q o
^ ^2 i1* *o
o o o o
ec o co i
r* co tS CM
CO ID* CO CO
SCO CO CO
o o o
9999
ID O O ^
26 km and RELMAP 50th Percentile
Large Coal-fired Utility Boiler
Medium Coal- fired Utility Boiler
Small Coal-fired Utility Boiler
Medium Oil-fired Utility Boiler
O)
'
tn
CO
I I
-------
concentrations around the utility are approximately 0.2 ng/m3.
(This predicted value includes the contribution from the 50th
percentile RELMAP value.) Measured average air concentrations
over South Florida and near anthropogenic sources are 2.8 and 3 .3
ng/m3 and average concentrations from sites in Detroit are 3.7
and 40 ng/m3. The measured data were collected over relatively
short periods of time and the measured mercury concentrations
probably reflect the input from multiple local as well as
regional sources. As a consequence, comparison to predicted data
are inappropriate.
It must be stressed that COMPDEP predicts annual average
results and that no systematic collection of measured mercury
data around these sources was identified. The measurement data
tend to be collected for short periods of time, and it is
difficult in an urban environment to apportion the sources of
mercury. It must also be emphasized that: (1) the COMPDEP is a
Gausian Plume model, these models have been used for years; (2)
the modifications made to the COMPDEP model pertaining to the
chemistry of atmospheric mercury were based on "state of the art"
assumptions about mercury (see Petersen et al.34); and (3) these
models have not been validated, but then no other mercury fate
and transport models have (at least to the EPA's knowledge) been
validated at this time. The EPA is currently conducting research
to refine some of the areas of greatest uncertainty. Model
validation will take a much longer period of time and will
require significant resources.
Hourly and daily predictions of COMPDEP were not used in
this report. The EPA's confidence in these predictions is low.
Confidence in the longer-term average predictions is greater.
It should also be noted that both the dry deposition of
Hg(II) vapor and the chemistry of mercury emitted in a plume are
poorly understood. The lack of measured data make it impossible
to compare the predicted dry deposition estimates of sources
expected to emit Hg(II). The COMPDEP results agree with those of
the RELMAP model in that both models predict that the bulk of the
emitted mercury will be transported beyond 50 km from the
emission source.
Given the uncertainty in the COMPDEP results and the lack of
measurement data collected around the local emissions sources, a
comparison of the IEM2 models with measured data collected around
sources was impossible.
The EPA was unable to locate systematically collected
measured mercury data in soils, water, or biota around utility
7-36
-------
boilers. The lack of measured data precluded a comparison of
predicted and measured values and, as a result, a validation of
these models is not possible at this time.
7.5 PRELIMINARY OBSERVATIONS FROM LOCAL ANALYSIS
The Agency's confidence in the following observation is high:
1. There is a lack of reported data near the utilities
considered in this report. The lack of such measured data
precludes a comparison of the model results with measured
data around these sources. These data include measured
mercury deposition rates and concentrations in the
atmosphere, soils, and biota (including fish).
The Agency's confidence in the following observations is
moderate:
2. The modeling analysis of utility emissions in flat terrain
indicates that relatively low mercury concentrations will
result in media and biota in the local modeling domain
(i.e., within 50 km of the facility). Depending on the
facility, this is predicted due to a combination of the
following factors: (1) the predicted -plumes are high due
to a high effective stack height, and (2) the assumed
mercury emission rates are low. The high effective stack
heights are a result of high stacks and large stack exit gas
velocities associated with this source class. This high
effective stack height results in the dispersion of mercury
emissions more to the regional scale than to the local
scale.
3. Based on the analysis, the larger impact from utility
mercury emissions is the result of bioaccumulation of
deposited mercury through the aquatic food chain. The
magnitude of the increase from any individual utility as
configured in this analysis (in particular, in flat terrain)
is predicted to be low. There is a great deal of
uncertainty and potential variability due to differences in
mercury bioaccumulation in various bodies of water.
4. Higher total mercury air concentrations and deposition rates
may result in elevated terrain. The magnitude of these
impacts compared to flat terrain is uncertain.
5. From the analysis of deposition and on a comparative basis,
a utility located in a humid climate has a higher annual
rate of deposition than a utility located in an arid
7-37
-------
climate. The critical variables are the solubility
(estimated washout ratios) of Hg(II) and the annual amount
of precipitation.
6. The lack of measurable impacts of the utility model plants
is a direct result of the relatively tall stacks of utility
boilers. The tall stacks result in impacts beyond the scale
of the local-scale analysis (50 km).
7. Based on this analysis, when the results of the local impact
analysis were combined with the 50th percentile estimates
from the long-range transport analysis for all utilities
across the U.S., the regional contribution was generally
higher than the local contribution for air concentration and
deposition, although there are exceptions.
7.6 GENERAL FINDINGS FOR MERCURY FROM UTILITIES
There are many sources of mercury emissions that occur
around the world; these include both natural and anthropogenic
sources. Fish mercury levels are probably due to mercury
emissions from all of these sources over time. The utilities are
one category of anthropogenic sources of mercury, and they are
estimated to emit approximately 51 ton/yr of mercury emissions in
the U.S., which is approximately 21 percent of the total
anthropogenic emissions of mercury in the U.S.
Although the amount of mercury being emitted from any single
source may be small, there are a number of reasons why these
emissions are of potential concern. The first is that mercury is
a persistent element; it is not degraded or removed, but
continually accumulates. Consequently, over time, there is
potential for levels of biologically available mercury to build
up. Second, current scientific evidence demonstrates that most
of the mercury emitted from a source, especially sources like
utilities that have tall stacks, does not deposit within 50 km of
the source but is deposited far away. As a result, even though
the concentration of mercury around a single source may not be
elevated, there are sufficient data to conclude that the
cumulative impact of many small sources leads to accumulation of
mercury in the biosphere and bioaccumulatron in fish.
The multipathway modeling assessment, in conjunction with
available scientific knowledge, supports a plausible link between
mercury emissions from utilities and mercury concentrations in
air, soil, water, and sediments. The mercury modeling
assessment, in conjunction with available scientific knowledge,
7-38
-------
also supports a plausible link between mercury emissions from
utilities and methylmercury concentrations in freshwater fish.
Other studies conducted by various researchers support the
EPA's observation that there is a plausible link between utility
mercury emissions and the methylmercury found in freshwater fish.
For example, in the EPRI's report entitled Electric Utility Trace
Substances Synthesis Report,88 EPRI estimated the impacts of
utility mercury emissions from two utility plants on
methylmercury concentrations in freshwater fish.
The results of the EPRI modeling analysis indicate that, for
two utility plant sites within 50 km of four lakes, the
methylmercury concentrations in freshwater fish were slightly
elevated. The EPRI report notes that these estimated mercury
concentrations in fish are well below the multimedia hazard index
and that the uncertainty analysis indicated that the estimates
were highly conservative. However, the EPRI modeling analysis
does support a credible link between mercury emissions from
electric utility plants and some incremental increase in
methylmercury concentrations in fish located in local
waterbodies.
The EPA's multipathway modeling assessment of mercury
emissions from four utility model plants suggests that mercury
emissions from utilities may contribute, at least to a limited
degree, to mercury concentrations in freshwater fish. Also, the
modeling assessment suggests that emissions of mercury from all
utilities may contribute to the overall mercury loadings to the
environment. However, the EPA has not yet determined, at this
time, whether the mercury contribution from utilities is a
concern for public health.
7.7 DISCUSSION OF FEDERAL INTERAGENCY REVIEW COMMENTS
Previous drafts of chapters 1 through 10, along with the
appendices, were reviewed by numerous non-EPA scientists
representing industry, environmental groups, academia, and other
Federal agencies during the summer of 1995. 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. The EPA has revised the report, as
appropriate, based on the reviewers' comments. The EPA revised
the report to incorporate the majority of the comments received.
However, there were several comments that could not be fully
addressed because of limitations in data, methods, and resources.
This section presents comments received by other Federal agencies
that could not be substantially addressed in this interim report.
7-39
-------
7.7.1 Comments
The Office of Management and Budget (OMB) commented that the
April 1996 draft report included "...mercury-related material
subject to continued interagency debate..." (e.g., discussion of
the health effects of mercury, the level of "safe" exposure to
mercury, the potential food safety issues associated with fish
consumption). Discussion of this material should be deferred
pending completion of further review of the mercury report.89
The National Marine Fisheries Service (NMFS) commented that
the EPA should defer addressing mercury risks from utilities
until completion of the Mercury Report.90
The Food and Drug Administration (FDA) commented that the
EPA may want to consider issuing the Report without a discussion
of human health effects of methylmercury. The FDA also suggested
"...that the discussion of the RfD be rewritten..." and the
uncertainties acknowledged.91
The Council of Economic Advisors (CEA) and the Office of
Science and Technology Policy (OSTP) similarly commented that the
current report should not refer to the current RfD as a
"reasonable estimate" but should state that the current standard
is under review.92
The NMFS encouraged the EPA "...not to release any documents
that characterize [methylmercury] risks..." from exposure through
fish consumption until release of the Mercury Report.93
The FDA commented that the EPA should "... acknowledge that
typical consumers eating fish in moderation from a variety of
sources and a variety of species are not believed to be at
increased risk from methylmercury." 94
The OSTP commented that:
"...quantitative relationships between utility mercury
emissions and fish methylmercury levels are unknown. This
information is critical for estimating human mercury
exposures and risks from utility emissions. Because mercury
emitted from utilities can be transported long distances and
total deposited anthropogenic emissions in the US are only
1/3 of the total US anthropogenic emissions, it is evident
that the majority of US anthropogenic mercury emissions are
transported outside of the US. Thus, it is not clear to
what extent reductions in US emissions of anthropogenic
mercury from utilities will affect methylmercury
concentrations in fish. The purpose of this report is to
7-40
-------
determine if utility HAP emissions present a health hazard
to humans. The report notes that the fate and transport
model predicts that utilities may contribute about 18
percent of the total anthropogenic mercury deposited in the
US. Would a 30 percent increase in mercury emissions, which
is projected for the year 2010 from coal-fired utilities,
lead to greater than a 5 percent increase in fish levels of
methylmercury, and would a 50 percent reduction in mercury
emissions produce more than a 10 percent reduction in total
anthropogenic mercury emissions? More importantly, would
either of these changes produce a significant adverse or
beneficial health effect? Perhaps health-related issues of
anthropogenic mercury emissions need to be examined on a
more global scale." 9S
The Centers for Disease Control (CDC) commented "...that
mercury is the [HAP] of greatest concern among the pollutants
listed." Also, the CDC agreed with one of EPA's draft
conclusions (presented in the April 1996 draft report) ..."that
mercury emissions should be minimized as part of an overall
strategy for reducing power-plant generated pollutants such as
sulfur dioxide and particulate matter." 96
The Department of Energy (DOE) and CEA suggested, as an
option, that the report should be revised to remove all
discussions of mercury, indicating that this material is being
reviewed by the SAB and will be included in the final Mercury
Study.97
7.7.2 Response
As discussed in the preface, the EPA has removed the
assessments of human exposure, health effects, and risk
discussions for mercury from this interim report. The EPA plans
to consider the above comments and make appropriate changes, as
appropriate, and to the extent feasible, before issuing a final
report.
The EPA has decided not to include conclusions or policy
statements regarding the public health impacts due to mercury
emissions from utilities in the interim report because the Agency
is awaiting new data from human epidemiology studies that will
help address the potential public health impacts. Therefore,
several statements that were presented in the April 1996 draft
report have been removed from this interim report. The EPA plans
to include a more complete assessment of exposures and risks and
plans to include conclusions and policy statements, to the extent
feasible and appropriate, in the final report.
7-41
-------
7.7.3 Comment
The OMB commented that the April 1996 draft report "...does
not place in proper context the contribution of anthropogenic
U.S. air emissions to the total loading of mercury in the
environment." The discussion "...should be expanded to provide a
more comprehensive description of the sources of mercury
contamination...", including other air deposition sources
(e.g., global and naturally occurring sources) and non-air
sources (e.g., industrial and municipal water discharges; the
release of mercury from sediments) .98
7.7.4 Response
The EPA has revised the discussion of mercury emission
sources to include and emphasize that utility emissions, and
anthropogenic emissions as a whole, represent a fraction of total
mercury in the environment. However, global emissions were not
quantified because of the significant uncertainties associated
with estimates of global emissions.
7.7.5 Comment
The FDA commented that "[s]ince methylmercury from marine
fish is not considered in the exposure modeling assessment, there
is no point in referring to or including Table 7-2 (Measured
Mercury Concentrations in Saltwater Commercial Fish) in this
discussion. Further, inclusion of information on shark, marlin,
etc., is also irrelevant, especially since some of the
information on mercury concentrations is inaccurate." " The
OSTP commented that Table 7-2 is "intended for the mercury study
report, should not be in this report." The NMFS submitted the
comment: "We recommend that references to marine fish (text in
Section 7.2, page 7-10 and Table 7-2) be deleted. Their
inclusion is not pertinent to the discussion of mercury
concentrations in biota potentially affected by utility emissions
as constrained by the modeling techniques used in the hazardous
air pollutant study. The last sentence of the footnote to the
table is totally irrelevant in the absence of sample sizes, fish
lengths (ages), and mercury ranges." The NMFS continued in the
comment letter that "If Table 7-2 is retained..." then the EPA
should correct the inaccuracies in the table.
7.7.6 Response
The EPA added caveats and discussions of ranges and
uncertainty and discussed the relevance of this information. In
addition, the EPA revised the table substantially by presenting
ranges and explanations of data and by correcting the
inaccuracies. The data are being presented to provide general
information on estimated levels of mercury in the environment.
This information is useful for understanding where mercury is
7-42
-------
found in the environment and the relative magnitudes of mercury
in each of the media or biota. In addition, these data aid in
the understanding of environmental fate and transport of mercury,
including the bioaccumulation potential of mercury in various
species.
The modeling assessment did not predict levels in saltwater
fish. However, the EPA believes that providing data on measured
levels in saltwater fish (Table 7-2) is useful since it is one of
the biota that contain mercury. The EPA has not determined if,
or how much, mercury emissions from U.S. utilities contribute to
these saltwater fish levels. Such an analysis was not possible
for this report because of the limitations in the available data,
methods, and models and because of the immense complexity of such
an assessment.
7.7.7 Comments
The DOE commented that "[t]he model used in this Report for
regional deposition estimation of all pollutants (RELMAP) as well
as the model used for estimation of local mercury deposition were
both significantly criticized during review of the Mercury Report
and are scheduled for review by the EPA's Science Advisory-
Board." In addition, the DOE adds that w[a]ll sections now using
[the RELMAP model] should contain prominent caveats..." because
the model is "...believed to significantly overestimate
deposition and hence risk of all types..." 10°
The OMB commented that the report "...should clearly state
the strengths, limitations, and uncertainties associated with
modeling the predicted levels of exposure and indicate the effect
of these factors on the overall assessment. The Report should
also indicate the extent to which the predicted exposure levels
have been validated by empirical data." lt}1
The OSTP similarly commented that a "...lack of monitoring
data that reflect or validate predictions of mercury exposures
based on the fate and transport model..." is a major shortcoming
of the report.102
7.7.8 Response
The EPA has added discussions of uncertainty and appropriate
caveats to various sections (e.g., section 7.10 and others) of
the report regarding the exposure models. The models are
considered useful for estimating fate and transport of
pollutants. The EPA believes that the results of the modeling
assessment are reasonable. However, the EPA does recognize there
are limitations and uncertainties. The text has been revised in
various locations to reflect these limitations and uncertainties.
7-43
-------
7.7.9 Comment
With regard to the multipathway modeling assessment for
mercury, the FDA commented that the EPA's confidence in the local
impact analysis must be considered low, rather than high, based
on the stated uncertainties.103
7.7.10 Response
Because of scientific uncertainties and variability and a
lack of measured mercury data around utility boilers, the EPA
interpreted the results of the modeling in a qualitative rather
than quantitative manner. The EPA's confidence in the general
trends of the quantitative data is greater than the confidence in
the actual values predicted. This is reflected in the
qualitative nature of the conclusions. As explained in section
7.5, the EPA's confidence in the observations of the local
analysis are generally moderate rather than high.
7.7.11 Comments
The FDA commented that "...the existence of fish consumption
advisories in 35 States based on conservative methodology does
not itself evidence a national mercury fish contamination
problem." Rather, it could relate "...more to risk assessment
methodologies than to mercury vcontamination' of fish..." and
merely be a reflection of the uncertainty surrounding determining
a "safe" exposure level and risk assessment methodologies than to
actual contamination of fish. The FDA stated that w[t]he
discussion of fish advisories and mercury levels in fish should
be deleted because it will mislead the reader on the extent of
mercury contamination of the environment." m
7.7.12 Response
The EPA revised the text in section 7.1 to say that
"...there is a potential concern for mercury contamination in the
U.S. as indicated by numerous fish advisories..." Also, much of
the discussion of fish advisories has been removed from this
interim report.
7.7.13 Comment
The DOE commented that the "EPA uses a series of fate and
transport models and hypothetical environmental constructs to
assess the emission of mercury from utility boilers. EPA extends
a level of moderate confidence to a conclusion of its modeling
effort that (1) 50 percent of utility mercury is emitted as Hg°;
(2) 30 percent is emitted as Hg(II) (i.e., divalent or ionic
form); and (3) 20 percent is emitted as a particulate. This is a
very critical conclusion in that, with a high percentage of Hg°,
the model predicts relatively low local (within 50 km) deposition
7-44
-------
and high long-range deposition due to the fact that Hg° readily
becomes part of the global mercury cycle."105 The DOE noted that
"...recent bench-, pilot-scale, and field characterization
studies...have indicated that coal-fired flue gas can contain
high percentages of ionic mercury...studies also show that far
less than 20 percent of the mercury is in the form of a
particulate (typically less than 5 percent)." It was indicated
that the EPA needs to reconcile the discrepancy noted between the
various mercury speciation values used in the modeling and those
presented by the DOE.
The DOE also commented that "EPA expresses moderate
confidence in the model's prediction that local (within 50 km of
the source) deposition of, and exposure to, mercury will be
low...based on the assumption that most of the utility flue gas
mercury is elemental, which remains airborne and is transported
greater distances than Hg(II). EPA takes this one step further
in concluding that utility mercury impacts are more regional than
local in nature." It was noted that the EPA should revisit these
conclusions in light of the mercury speciation data provided by
the DOE.
7.7.14 Response
The EPA plans to consider these comments and, if feasible
and appropriate, conduct additional analyses of the data in the
future.
7-45
-------
7.8 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. 22:
73-80. 1980.
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.
19:277-291. 1983.
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 Scientifi 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. 181:230-241. 1973.
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, p. 1728-1745.
1981.
8. Borst, H.A. and C.G. Lieshout. Phenylmercuric acetate
intoxication in mink. Tijdschr. Diergeneesk. 102:495-503.
1977.
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. 40:30-33. 1976.
10. Wobeser, G., N.D. Nielsen and B. Schiefer. Mercury and mink
II: Experimental methylmercury intoxication. Can. J. Comp.
Med. 40:34-45. 1976.
7-46
-------
11. Fimreite, N. Effects of me thy liner cury treated feed on the
mortality and growth of leghorn cockerels. Can. J. Anim.
Sci. 50:387-389. 1970.
12. Fimreite, N. Effects of methylmercury on ring-necked
pheasants. Canadian Wildlife Service Occasional Paper
Number 9. Department of the Environment. 39 p. 1971.
13. Fimreite, N. Accumulation and effects of mercury on birds.
In: The Biogeochemistry of Mercury in the Environment, J.O.
Nriagu, Ed. Elsevier, Amsterdam, The Netherlands.
p. 601-628. 1979.
14. Heinz, G.H. Effects of low dietary levels of methylmercury
on mallard reproduction. Bull. Environ. Contain. Toxicol.
11:386-392. 1974.
15. Heinz, G.H. Effects of methylmercury on approach and
avoidance behavior of mallard ducklings. Bull. Environ.
Contain. Toxicol. 13:554-564. 1975.
16. Heinz, G.H. Methylmercury: Second-year feeding effects on
mallard reproduction and duckling behavior. J. Wildl.
Manage. 40:82-90. 1976.
17. Heinz, G.H. Methylmercury: Second-generation reproductive
and behavioral effects on mallard ducks. J. Wildl. Manage.
40(4) :710-715. 1976.
18. Heinz, G.H. Methylmercury: Reproductive and behavioral
effects on three generations of mallard ducks. J. Wildl.
Mgmt. 43:394-401. 1979.
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. Mitra, S. Mercury in the Ecosystem. Trans Tech Publications
Ltd. Switzerland. 1986.
21. Fitzgerald, W. F. and T. W. Clarkson. Mercury and
Monomethylmercury: Present and Future Concerns. Envir.
Health Perspec. 96:159-166. 1991.
22. Swedish EPA. Mercury in the Environment: Problems and
Remedial Measures in Sweden. ISBN 91-620-1105-7. 1991.
7-47
-------
23. Brosset, C. and E. Lord. Mercury in Precipitation and
Ambient Air: A new Scenario. Water, Air and Soil Poll.
56:493-506. 1991.
24. Expert Panel on Mercury Atmospheric Processes. Mercury
Atmospheric Processes: A Synthesis Report. Report No. TR-
104214. 1994.
25. 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. 55:(all chapters) 1991.
26. Porcella, D.B. Mercury in the Environment: Biogeochemistry.
pp. 3-19 in Watras, C.J. and J.W. Huckabee eds. Mercury
Pollution Integration and Synthesis. 1994.
27. Horvat, 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. Analytics Chimica
Acta, 282:153-168. 1993.
28. 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 257:784-
787. 1992.
29. 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. Pp. 33-66 in L.A. Baker (ed). Environmental
Chemistry of Lakes and Reservoirs. American Chemical
Society. 1994.
30. 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. Pp. 187-202 in
Watras, C.J. and J.W. Huckabee eds. Mercury Pollution
Integration and Synthesis. 1994.
31. 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. 55:(all chapters) (see in
particular Chapter 4) 1991.
7-48
-------
32. Michigan Environmental Science Board. Mercury in Michigan's
Environment: Environmental and Human Health Concerns. Report
to Gov. John Engler. 1993.
33. 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. Accepted
for Publication in Water, Air and Soil Pollution. 1995.
34. Petersen, G., A. Iverfeldt and J. Munthe, 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.
Atmospheric Environment 29:47-68. 1995.
35. Stratton, W.J. and Lindberg, S.E. "Use of a refluxing Mist
Chamber for Measurement of Gas-phase Water-soluble Mercury
(II) Species in the Atmosphere." Water Air Soil Pollution.
Vol. 80. pages 1269-1278. 1995.
36. 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. 97(D2):2519-2528. 1992.
37. Shannon, J. D., and E. C. Voldner. Modelling 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.
38. 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. IVater, Air and Soil Poll. 56:577-
594. 1991.
39. 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. 56:745-767. 1991.
40. 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. 3rd International Conference on Mercury as a
Global Pollutant. Whistler, BC, Canada (July 10-14, 1994)
7-49
-------
41. .Mosbaek, H., J. C. Tjell, and T. Sevel. Plant Uptake of
Mercury in Background Areas. Chemosphere 17(6):1227-1236.
1988.
42. Ref. 25, (Please see Chapter 6 of this reference)
43. 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.
56:667-680. 1991.
44. 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 56: 439-453. 1991.
45. Nriagu, J. 0. The Biogeochemistry of Mercury in the
Environment. Elsevier/North Holland. Biomedical Press: New
York. 1979.
46. Revis, N. W. , T. R. Osborne, G. Holdsworth, and C. Hadden.
Mercury in Soil: A Method for Assessing Acceptable Limits.
Arch. Environ. Contain. Toxicol. , 19:221-226. 1990.
47. Cappon, C. J. Uptake and Speciation of Mercury and Selenium
in Vegetable Crops Grown on. Compost-Treated Soil. Water, Air
and Soil Poll. 34:353-361. 'l987.
48. Wilken, R. D. and H. Hintelmann. Mercury and Methylmercury
in Sediments and Suspended particles from the River Elbe,
North Germany. Water, Air and Soil Poll. 56:427-437. 1991.
49. 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. Aq. Sci. 46:2184-2202. 1989.
50. Ref. 25, Chapter 9.
51. Cappon, C.J. Mercury and Selenium Content and Chemical Form
in Vegetable Crops Grown on Sludge-Amended Soil. Arch.
Environm. Contain. Toxicol. 10: 673-689. 1981.
52. Fortmann, L. C., D. D. Gay, and K'. 0. Wirtz. Ethylmercury:
Formation in Plant Tissues and Relation to Methylmercury
Formation. U.S. EPA Ecological Research Series,
EPA-600/3-78-037. 1978.
7-50
-------
53. Sorensen, J., G. Glass, K. Schmidt, J. Huber and G. Rapp.
Airborne Mercury Deposition and Watershed Characteristics in
Relation to Mercury Concentrations in Water, Sediments,
Plankton and Fish of Eighty Northern Minnesota Lakes.
Environ. Sci. Technol. 24:1716-1727. 1990.
54. Winfrey, M. R. and J. W. M. Rudd. Environmental Factors
Affecting the Formation of Methylmercury in Low pH Lakes.
Environ. Toxicol. and Chem., 9:853-869. 1990.
55. Xun, L., N. Campbell and J.W. Rudd. Measurements of
Specific Rates of Net Methyl Mercury Production in the Water
Column and Surface Sediments of Acidified and Circumneutral
Lakes. Can. J. FishAquat. Sci. 44:750-757. 1987.
56. Nagase, H., Y. Ose, T. Sato, and T. Ishikawa. Methylation
of Mercury by Humic Substances in an Aquatic Environment.
Sci. Total Environ. 32:147-156. 1982.
57. New Jersey Department of Environmental Protection and
Energy. Final Report on Municipal Solid Waste Incineration.
Volume II: Environmental and Health Issues. 1993.
58. 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. 56: 477-491. 1991.
59. Driscoll, C. T., C. Yan, C. L. Schofield, R. Munson, and J.
Holsapple. The Mercury Cycle and Fish in the Adirondack
Lakes. Environ. Sci. Technol. 28 (3): 136 A - 143 A. 1994.
60. Bahnick, D., C.Sauer, B. Butterworth and D. Kuehl. A
National Study of Mercury Contamination of Fish.
Chemosphere 29(3) : 537-546. 1994.
61. Simonin, H. A., S. P. Gloss, C. T. Driscoll, C. L.
Schofield, W. A. Kretser, R. W. Karcher, and J. Symula.
Mercury in Yellow Perch from Adirondack Drainage Lakes (New
York, U.S.),in (1994) Watras, C. J. and J. W. Huckabee
[eds], Mercury Pollution: Integration and Synthesis, Lewis
Publishers, Boca Raton, FL. 1994.
62. Mills, E.L., W.H. Gutenmann, and D.J. Lisk. Mercury content
of small pan fish from New York State Waters, Chemosphere,
Vol. 29, No. 6, pp. 1357-1359. 1994.
63. U.S. EPA. Assessment and Remediation of Contaminated
Sediments (ARCS) Program. EPA 905-R92-007. 1992.
7-51
-------
64. U.S. EPA. Assessment and Remediation of Contaminated
Sediments (ARCS) Program. EPA 905-R92-008. 1992.
65. Gloss, S. P., T. M. Grieb, C. T. Driscoll, C. L. Scholfield,
J. P. Baker, D. H. Landers, and D. B. Porcella. Mercury
levels in fish from the Upper Peninsula of Michigan (ELS
Subregion 2B) in Relation to Lake Acidity. USEPA Corvallis
Env. Res. Lab. Corvallis. 1990.
66. Giesy, J., D. Verbrugge, R. Othout, W. Bowerman, M. Mora,
et. al., Contaminants in Fishes from Great Lakes-influenced
Sections and above dams of three Michigan Rivers. I.
Concentrations of organochlorine insecticides,
polychlorinated biphenyls, dioxin equivalents and mercury.
Arch. Environ. Contain. Toxicol. 27:202-212. 1994.
67. Lathrop, R. C., K. C. Noonan, P. M. Guenther, T. L. Grasino,
and P. W. Rasmussen. Mercury Levels in Walleyes form
Wisconsin Lakes of different Water and Sediment Chemistry
Characteristics. Tech. Bull. No. 163. DNR, State of
Wisconsin, Madison. 1989.
68. Gerstsenberger, S., J. Pratt-Shelley, M. Beattie and J.
Bellinger. Mercury Concentrations of Walleye (Stizostedion
vitreum vitreum) in 34 Northern Wisconsin Lakes. Bull.
Environ. Contam. Toxicol. 50:612-617. 1993.
69. Lange, T. R., H. R. Royals, and L. L. Conner. Influence of
Water Chemistry on Mercury Concentrations in Largemouth Bass
from Florida Lakes. Trans. Amer. Fish. Soc. 122:74-84.
1993.
70. Florida Department of Environmental Regulation. Mercury,
Largemouth Bass and Water Quality: A Preliminary Report.
1990.
71. MacCrimmon, H. R. , C. D. Wren, and B. L. Gots. Mercury
Uptake by Lake Trout, Salveiinus namaycush, relative to age,
growth and diet in Tadenac Lake with comparative data from
other Precambrian Sheild lakes. Can. J. Fisher. Ag. Sci.
40:114-120. 1983.
72. Wren, C. D., W. A. Scheinder, D. L. Wales, B. M. Muncaster,
and I. M. Gray. Relation Between Mercury Concentrations in
Walleye (Sitzostedion vitreum vitreum) and Northern Pike
(Esox lucius) in Ontario Lakes and Influence of
Environmental Factors. Can. J. Fisher. Aq. Sci. 44:750-757.
1991.
7-52
-------
73. NMFS (National Marine Fisheries Service). The current
available NMFS database was supplied to the U.S. EPA via fax
from Malcolm Meaburn (Charleston Laboratory/Southeast
Fisheries Service Center/NMFS/NOAA/USDOT to Kathryn Mahaffey
(Environmental Criteria and Assessment Office-Cincinnati,
OH/Office of Health and Environmental Assessment/Office of
Research and Development/U.S. EPA). February 23, 1995.
74. Yess, N.J. U.S. Food and Drug Administration Survey of
Methylmercury in Canned Tuna. Journal of AOAC Intl. Vol.
76(1):36-38.
75. Richardson, M., M. Mitchell, S. Coad and R. Raphael.
Exposure to Mercury in Canada: A Multipathwy Analysis.
Water, Air and Soil Pol. 80:21-30. 1995.
76. Memorandum from G. Cramer, FDA, to R. Lake, EPA, dated
September 24, 1996.
77. United States Department of Agriculture. Recipe File
Database. 1995. (Available on the internet)
78. Anderson, W. L. and K.E. Smith. Dynamic of mercury at
coal-fired utility power plant and adjacent cooling lake.
Environ. Sci and Technol. 11:75. 1977.
79. Crockett, A. and R. Kinnison. Mercury residues in soil
around a coal-fired power-plant. Envir. Sci. Technol.
13:712-715. 1979.
80. Keeler, G., M. Hoyer, and C. Lamborg. Measurements of
Atmospheric Mercury in the Great Lakes Basin. In Watras,
C. J., and J. W. Huckabee (eds). Mercury Pollution
Integration and Synthesis, pp. 231-241. 1994.
81. 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
Watras, C. J. and J. W. Huckabee (eds). Mercury Pollution
Integration and Synthesis, pp. 251-259. 1994.
82. 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.
7-53
-------
83. 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 Research and
Development, Washington, DC. 1993.
84. Whitby, K. The physical characteristics of sulfur aerosols.
Atmosph. Env. 12:135-159. 1978.
85. U.S. EPA. Wildlife Criteria Portions of the Proposed Water
Quality Guidance for the Great Lakes System. (Proposed)
Office of Water and Office of Science and Technology. EPA-
822-R-93-006. 1993.
86. 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.
87. Memorandum and attachment from Fraas, Art, and Hillier,
Troy, OMB, to Maxwell, Bill, EPA/CG. May 3, 1996.
88. Letter from Schmitten, Rolland A., NMFS, to Nichols, Mary,
EPA/OAR. April 26, 1996.
89. Letter and enclosures from Lake, L.R., FDA, to Maxwell,
William H., EPA/CG. April'30, 1996.
90. Memorandum and attachment from Munnell, Alicia, CEA, to
Maxwell, William H., EPA/CG. May 6, 1996; Memorandum and
attachment from Gibbons, Jack, OSTP, to Maxwell, William H.,
EPA/CG. May 6, 1996.
91. Letter from Schmitten, Rolland A., NMFS, to Nichols, Mary,
EPA/OAR. April 26, 1996.
92. Letter and enclosures from Lake, L.R., FDA, to Maxwell,
William H., EPA/CG. April 30, 1996.
93. Memorandum and attachment from Gibbons, Jack, OSTP, to
Maxwell, William H., EPA/CG. May 6, 1996.
94. Letter from Mannino, David M., CDC, to Maxwell, William H.,
EPA/CG. May 2, 1996.
95. Memorandum and attachment from Munnell, Alicia, CEA, to
Maxwell, William H., EPA/CG. May 6, 1996; and letter from
Chupka, Marc W., DOE, to Maxwell, William H., EPA/CG. May
3, 1996.
7-54
-------
96. Memorandum and attachment from Fraas, Art, and Hillier,
Troy, OMB, to Maxwell, Bill, EPA/CG. May 3, 1996.
97. Letter and enclosures from Lake, L.R., FDA, to Maxwell,
William H., EPA/CG. April 30, 1996.
98. Letter from Chupka, Marc W., DOE, to Maxwell, William H.,
EPA/CG. May 3, 1996.
99. Memorandum and attachment from Fraas, Art, and Hillier,
Troy, OMB, to Maxwell, Bill, EPA/CG. May 3, 1996.
100. Memorandum and attachment from Gibbons, Jack, OSTP, to
Maxwell, William H., EPA/CG. May 6, 1996.
101. Letter and enclosures from Lake, L.R., FDA, to Maxwell,
William H., EPA/CG. April 30, 1996.
102. Letter and enclosures from Lake, L.R., FDA, to Maxwell,
William H., EPA/CG. April 30, 1996.
103. Letter and enclosure from Chupka, Marc C., Department of
Energy, to Maxwell, William H., EPA:ESD. September 25,
1996.
7-55
-------
8.0 QUALITATIVE MULTIPATHWAY ASSESSMENT FOR ARSENIC, DIOXINS,
LEAD, AND CADMIUM
8.1 BACKGROUND
A multipathway exposure analysis, 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
HAPs of interest (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. Quantitative
multipathway assessments were performed for only one class of
HAPs (radionuclides). A general assessment of the fate and
transport of mercury is presented in chapter 7. For the other
four HAPs — arsenic, dioxins, lead, and cadmium, only qualitative
assessments of the potential concerns to human health from non-
inhalation exposure were performed.
The completion of quantitative assessments of inhalation
exposure risks for all HAPs and of multipathway exposures for
only one class of HAPs (radionuclides) is not a reflection of 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 due to 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
arsenic, dioxins, 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, potentially,
a more significant route of exposure than inhalation exposure.
The mercury assessment suggests there is a need for further
analysis of noninhalation exposures.
Efforts are underway to collect 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 of noninhalation exposure to arsenic, dioxins, lead, and
cadmium is presented in the following sections.
8.2 ARSENIC COMPOUNDS
Arsenic is highly persistent in the environment, has a
tendency to bioaccumulate, and is toxic from both the inhalation
and oral exposure routes. Arsenic is also a known human
carcinogen.
For 1990, the estimated arsenic emissions from all coal-,
oil-, and gas-fired units were 54 ton/yr, 11 ton/yr, and
0.16 ton/yr, respectively. To put these estimates into
perspective, if the total amount of arsenic emitted from these
units in 1990 were added to those reported in the toxic release
inventory (TRI) database for 1 year (data from 1988) (i.e.,
270,000 pounds1), the TRI estimate would be increased by
50 percent.8
Because arsenic is a naturally occurring compound, it has
been found in low levels in all media, including air, soil,
water, sediment, fish and shellfish, and other food products.1
Arsenic is released from anthropogenic activities (e.g., metal
smelting, chemical production and use, coal combustion, and waste
disposal), and these emissions of arsenic can contribute
significantly to environmental contamination.1
Once released from stack sources, such as utilities, arsenic
can be deposited to various environmental media and undergo
complex transformations. In the atmosphere, arsenic exists as PM
in the form of trivalent and pentavalent states.1 Studies
indicate that the trivalent state is more toxic to the biological
system than the pentavalent species.2 A typical residence time
of particulate arsenic is 9 days. The primary removal mechanism
from the atmosphere is wet or dry deposition to soil,, water, and
plants.1 Once deposited onto soil, arsenic tends to sorb to
soils.1 However, precipitation may leach soluble forms of the
compound into surface waters and ground water. The predominant
species of arsenic in soils varies, with pentavalent arsenic
The emissions reported in the TRI database represent atmospheric releases
of arsenic from all facilities included in the Standard Industrial
Classification (SIC) 20-39 that manufactured, processed, or otherwise used
arsenic above the established threshold. As a consequence, the TRI
emissions estimate does not include emissions from the electric utility
industry (i.e., SIC 49). Furthermore, the ATSDR Toxicological Profile for
Arsenic1 points out that the TRI data do not include emissions data from
coal combustion facilities and pesticide spraying operations, two major
sources of arsenic.
8-2
-------
dominating in aerobic soils, trivalent arsenic dominating in
slightly reduced soils, and arsine, methylated arsenic, and
elemental arsenic dominating in extremely reduced conditions.1
Terrestrial plants accumulate arsenic by root uptake from soil or
by absorption of airborne arsenic deposition onto leaves.1 In
the aquatic environment, arsenic will partition to sediments or
will remain dissolved in the water. In water, arsenic can
undergo complex transformations, but the predominant form of
arsenic is usually pentavalent arsenic, arsenate.1 The pH of the
water and the redox potential of the aquatic system greatly
affect the chemical forms of the inorganic arsenic species.2
The highest levels of arsenic are found in seafoods, meats,
and grains. U.S. dietary levels of arsenic in these food
products typically range from 0.02 ppm in grains and cereals to
0.14 ppm in meat, fish, and poultry.1 Fish and shellfish are
typically associated with the highest levels of arsenic
concentration, although, arsenic accumulates in fish and
shellfish in a form thought to be relatively nontoxic.
Background levels of arsenic reported for milk are 0.5 to
70 /ug/kg3. With the exception of significantly contaminated
areas, exposures occurring through the intake of air, soil, and
water are usually lower than those occurring through food
products. Background levels of arsenic for -these media have been
reported as follows: soil is 5,000 ppb, water 2 ppb, and air
0.02 to 0.10 yug/m3. Furthermore, surveys of drinking water in
the U.S. indicate that greater than 99 percent of the drinking
water supplies have concentrations below the EPA maximum
contaminant level (MCL) of 50 ppb.1 The ATSDR1 states that, for
the general population, the highest levels of exposure to arsenic
occur through their diet, with an average intake of approximately
50 /ug/d. Based on an EPA reference dose of 3 x 10"4 mg/kg/d,
this intake value corresponds to a hazard quotient of slightly
less than l.b
A study that examines the transfer of metals to bovine milk
indicates human exposure to arsenic through consumption of milk
may be of concern. Because the contribution to human exposure
through the food chain has not been thoroughly examined, a study3
was undertaken to estimate the steady-state bovine milk
biotransfer factors (i.e., the rate at which the compounds are
transferred to milk) for six metals: arsenic, cadmium, chromium,
lead, mercury, and nickel. Results from this study indicate
b This risk estimate was calculated based on the following exposure
assumptions: exposure duration of 30 years, exposure frequency of
350 d/y,; body weight of 70 kg, and an average life expectancy of 70 years.
8-3
-------
that, of the metals studied, lead and arsenic transfer to milk to
the greatest extent.3 The BCF estimated for inorganic arsenic is
3.0 x 10"5 d/L. To put this value into perspective, Stevens
points out the estimated biotransfer factor of tetra
chlorodibenzo-p-dioxin (TCDD) is 2.6 x 10"2 d/L and that the
particular food chain pathway may be less important for these
metals than it is for TCDD.
A study conducted on children in a community located less
than 1 mile from a copper smelter indicated that exposure to
arsenic through the ingestion of soil may be of concern.
Specifically, Polissar et al. conducted a study that attempts to
establish a link between direct soil ingestion and arsenic
exposure in a community located near a copper smelter (i.e.,
median distance from the smelter is 0.5 miles). The
contamination in the community was predominantly in the form of
arsenic trioxide. Analyses of samples obtained from 17 children,
ages 6 or under, revealed elevated levels of urinary arsenic
(i.e., a relatively high median urinary arsenic concentration of
43 .6 ppb) . The children of the community appeared to be
receiving their arsenic dose from hand-to-mouth activities (i.e.,
soil ingestion). Calculations showed that inhalation could have
contributed only a small fraction of the arsenic, and dermal
absorption could be ruled out on the basis of poor dermal
absorption of the inorganic arsenic in particulate or bulk soil.
Thus, the remaining exposure route is ingestion. Also, arsenic
concentrations in homegrown fruits and vegetables in the study
area were low, and the arsenic concentrations in drinking water
were far below the EPA standard of 50 ppb.
Furthermore, a study conducted by Chen et al. on ground
water samples obtained from southwest Taiwan indicated that
arsenic consumption through drinking water may be of concern if
arsenic is present in the trivalent state.0 As part of this
study, ground water samples obtained from the Blackfoot Disease
(BFD) area of Southwest Taiwan and well water samples from a city
where no BFD has ever been reported were collected and analyzed.
Analyses of these samples revealed that the total dissolved
arsenic levels detected in the BFD area samples were
approximately 1,000 times higher than in the samples obtained
from the control site. In the samples obtained from the BFD
area, the ratio of the trivalent species to the less toxic
pentavalent species was approximately 3 to I.2
This study states that Blackfoot Disease (BFD) is a peripheral vascular
disease found in a limited area of Taiwan. The cause of this disease is
still unknown, but it is generally attributed to the high concentrations of
trivalent arsenic found in deep well waters.
8-4
-------
.Volume 1 of the EPRI's document Electric Utility Trace
Substances Synthesis Report (Synthesis Report) presents results
of a multimedia risk assessment of utility emissions. The
assessment was performed by EPRI, using their "Total Risk of
Utility Emissions" model, for HAPs emitted by four selected
utilities. For arsenic, results of the assessment indicated that
the dominant pathway or exposure to arsenic emissions can be
ingestion (perhaps as much as 90 percent of total exposure).
Further discussion of the assessment is provided in Volumes 1 and
2 of the EPRI Report.4
Because utility emission sources have the potential to
contribute significantly to the total amount of arsenic emitted
annually in the U.S. and because arsenic compounds tend to
accumulate in the environment, the risks posed by noninhalation
exposures to arsenic compounds in locally and regionally impacted
areas may be of concern. Based on the studies discussed above
and the physical and chemical properties of arsenic, it can be
hypothesized that the primary exposure routes of concern for
adults are those that are related to the ingestion of food
products and that, for children, the primary route of exposure is
ingestion of soil. Exposure through these routes is most likely
to occur through the consumption of vegetation or soil
contaminated by atmospheric deposition or the consumption of
animals or fish contaminated through the ingestion of
contaminated media or organisms.
Since inhalation risks are estimated to be above 1 x 10"6
for arsenic emissions from four utilities, since utilities emit
approximately 59 ton/yr of arsenic nationwide, and since
ingestion of arsenic can pose a cancer risk, the Agency believes
that further evaluation of multipathway exposure is necessary to
fully understand the risks posed by arsenic emissions from
utilities.
8.3 DIOXIN AND DIOXIN-LIKE COMPOUNDS
Polychlorinated dibenzo-p-dioxins and polychlorinated
dibenzofurans, which will be referred to collectively as dioxins,
are ubiquitous in the environment.5 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 are probable human
carcinogens.5 In addition, EPA has concluded that there is
adequate evidence to support the inference that humans are likely
8-5
-------
to respond with a broad spectrum of noncancer effects from
exposure to dioxins, if exposures are high enough.5
It has been hypothesized that the primary mechanism by which
dioxins enter the terrestrial food chain is through atmospheric
deposition.5 Therefore, a qualitative assessment of the
potential risk to human health attributable to noninhalation
exposure to dioxins emitted from utilities is presented here. In
general, this assessment focuses on chlorinated dibenzodioxins
(CDDs) and chlorinated dibenzofurans (CDFs) as a group and not on
individual congeners. A multipathway exposure analysis was not
performed for dioxin emissions from utilities. Therefore, data
primarily from the external review draft dioxin reassessment
report are used here for a qualitative assessment of potential
concerns from multimedia exposures to dioxin. The primary
sources used for this assessment are the external review draft
dioxin reassessment documents (Health Assessment Document for
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) and Related Compounds5
and Estimating Exposure to Dioxin-Like Compounds.6''') The Final
dioxin report is expected to be released by the summer of 1997.
The occurrence of dioxin and dioxin-like compounds in the
environment appears to be primarily the result of human
activities.5-6-7 The national estimated loading of these compounds
from identified sources into the environment is approximately
12,000 g toxicity equivalents (TEQ)/yr. 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 150
g TEQ/yr, dioxin emissions from utilities represent about 1.3
percent of total dioxin emissions.
When this loading of 12,000 g TEQ/yr is compared to annual
loadings of other HAPs, it appears to be relatively low.
However, dioxin emissions are of concern because these compounds
are extremely toxic to humans and wildlife, are persistent in the
environment and tend to bioaccumulate.d
The CDDs and CDFs 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
The draft Health Assessment Document for 2,3,7,8-Tetrachlorodibenzo-
p-Dioxin (TCDD)S and Related Compounds (Volume III of III) states that
these compounds are extremely potent in producing a variety of effects in
experimental animals based on traditional toxicology studies at levels
hundreds or thousands of times lower than most chemicals of environmental
interest.
8-6
-------
be attributed to their stability under most environmental
conditions and also to the great number of sources located
throughout the U.S. 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, CDDs and CDFs remain
adsorbed to PM, 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 through sedimentation and, ultimately, burial of
sediments.
Once CDDs and CDFs are deposited and make their way into
various media and into biota, they are available for human uptake
through ingestion. Consumption of dioxin-contaminated food is
considered the primary route of exposure in the general
population. Table 8-1 presents background media concentration
data and TEQ daily intake rates associated with each medium. The
EPA estimates that the highest daily TEQ intakes for humans occur
through the consumption of animal fats, particularly fat of beef
and veal, dairy products, milk, pork, and chicken. The EPA
estimates that the lowest daily TEQ intakes occur through water
ingestion, soil ingestion, and inhalation. Based on the data in
Table 8-1, exposure occurring through ingestion of beef and veal
is estimated to be approximately 16 times higher than exposure
through inhalation.6
Volume III of the draft Dioxin Reassessment report5 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/d. This value is more than 500-fold
higher than the EPA's 1985 risk-specific dose of 0.006 pg TEQ/kg
body weight/d associated with an upper-bound risk of 1 in a
million (1 x 10"6) and is several hundredfold higher than the
revised risk-specific dose estimates presented in the draft
Dioxin Reassessment report.5
It should be noted that the dioxin reassessment documents state that only a
small number of samples were available for analysis, particularly for food,
and this fact should be considered when evaluating the data.
8-7
-------
CO
(U
4J
(0
Tl
(U
Q)
X!
-U
C
•H
(0
0)
w
o
-H
X
O
•H
P
•O
c
u
(C
CQ
O
H
EH
TJ
Q)
4->
(0
E
•H
4J
(Q
I
CO
(U
rH
•8
EH
•
ID
co
CM
CO
It)
CO
o
d
I Inhalation
o
d
00
8
d
CM
6
X
00
*
cr
Q.
(O
in
8
d
H Water Ingestion
CO
CD
CO
O
X
CO
•a
O>
T—
in
CM
Q.
a
O
d
H Milk Ingestion
CO
d
CM
•*'
CM
00
o
X
CM
ID
•o
o
CO
d
H Dairy Ingestion
CO
T-
cn
o
X
t
10
TJ
CD
0)
CM
a
a
d
| Egg Ingestion
CM
CO
q
CO
00
6
X
CO
ID
•o
D>
a
a
CO
d
c
o
in
a
f
8
ID
1
ro
d
CM
CM
00
6
X
CM
ID
•o
O5
a
0.
(O
CM
d
I Pork tngestion
0)
d
CO
cvi
00
O
X
CO
10
•o
CD
00
<0
a.
a
CO
d
1 Chicken Ingestion
8
^~
^
^
o
r—
X
00
o
"ID
8
'c
'6
•12.
o
1
8
"E
u
o
u
u
a
*•»
O
u
ID
*••
Q
8-8
-------
As part of the draft dioxin reassessment, methodologies for
conducting site-specific indirect exposure modeling are
presented. 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.
The results from this 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 dioxin-like compounds 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.6 This is of
significance since the uptake of plants by foraging animals,
including cows, 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
8-9
-------
in the lipid portion of the breast milk. Based on the estimated
adult intake of dioxin discussed above, an exposure duration of I
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-d, 20 to
60 times higher than the estimated range for background exposure
to adults (1 to 3 pg of TEQ/kg-d).
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., 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. Since these exposure pathways are likely to be the
most significant, they should be considered when assessing the
risk from exposure to dioxins. Though some emissions, health
effects, and exposure data are available, at this time it is
unclear whether utility emissions of dioxins pose a significant
risk. Further evaluation may be needed to more comprehensively
evaluate the risks posed by emissions of dioxins from utilities.
8.4 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 oral and inhalation route. For these reasons,
lead emissions from utilities are a potential concern from
noninhalation exposure.
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 of lead
were emitted into the atmosphere from anthropogenic point and
nonpoint sources during 1989.9 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), solid waste management (2.3 x 103 metric
8-10
-------
tons), transportation (2.2 x 103 metric tons), and fuel
combustion (0.5 x 103 metric tons).f
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.10 Evidence
supports 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 ground water or
surface water is not likely to occur. With the exception of
highly acidic environmental conditions, leaching of lead into
ground water 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 that have low water solubilities and will
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. However, the bioavailability of lead to
plants from soil is limited due to the strong absorption of lead
to soil.9
The highest background levels of lead are found in soils
(<10 to 30 //g/g) and in sediments (i.e., the average
concentration of lead in river sediments is 20,000 (ig/g) .10 In
1988, the average ambient air concentration for 139 sites
f Industrial processes include nonferrous smelters, battery plants, and
chemical plants.
8-11
-------
monitored by the National Air Surveillance Network (NASN) was
determined to be 0.085 //g/m3.9 This value is well below the
NAAQS of 1.5 yug/m3.10
Concentrations of lead found in foods are given in
Table 8-2. These concentrations range from a low of about
0.002 /ug/g of food to a high of more than 0.8 /ug/g (found in
milk). Background levels of lead in milk can range from 23 to 79
,ug/kg.3 The ATSDR9 states that, for the general population, the
highest levels of exposure to lead are most likely to occur
through the ingestion 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 8-2, the average adult dietary intake of lead for the years
1980-82 was estimated to be 56.5 /ug/day.9-11 However, recent data
(1992) indicate that average dietary intakes have reduced
significantly over the past decade to approximately 2 to 4
//g/day.12 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
emissions from utilities do not contribute substantially to the
total amount of lead released annually for 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.
8.5 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/rn3) . 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 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, 3.5 ton/yr, and
8-12
-------
Table 8-2. Concentration of Lead in Various Food Products9
,10
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 Ug/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
0.086 ton/yr, respectively. To put these estimates into
perspective, it was estimated that a total of 700 ton/yr of
cadmium were emitted into the atmosphere from anthropogenic point
and nonpoint sources in the early 1980s.13 In the early 1980s,
the major contributors of atmospheric cadmium included fossil
fuel combustion, smelting operations, manufacturing plants, and
incinerators. The total amount of atmospheric releases of
cadmium reported in the 1988 TRI database was approximately 60
tons.13 g
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
The emissions reported in the TRI database represent atmospheric releases
of arsenic from all facilities included in SIC 20 to 39 that manufactured,
processed, or otherwise used cadmium above the established threshold. As a
consequence, the TRI emissions estimate does not include emissions from the
utility industry (i.e., SIC 49). Furthermore, the ATSDR Toxicological
Profile For Cadmium13 points out that the TRI data do not include emissions
data from the combustion of fossil fuel and incineration of municipal or
industrial wastes, two major sources of cadmium.
8-13
-------
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 and, as a
result, 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 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 cadmium bioaccumulates in all levels of the food
chain.13 Table 8-3 presents concentrations of cadmium in various
foods.
The highest background levels of cadmium are found in
soils.13 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, ground, water, and
drinking water are typically less than 1 /ug/L. In a study
conducted in 27 U.S. cities, 12 food groups were tested and
cadmium was detected in nearly all samples. As seen in
Table 8-3, 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 up to thousands
of times higher than the levels of cadmium found in the water.
In the U.S., the adult intake of cadmium attributable to diet is
estimated to be approximately 30 /ug/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 /ug/cigarette. A cadmium exposure
8-14
-------
Table 8-3. Concentration of Cadmium in Various Food Products13
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)
0.01 6 to 0.1 42
0.01 6 to 0.061
0.002 to 0.033
trace-0.028
trace-0.093
trace-0.033
trace-0.053
trace-0.014
trace-0.01 6
trace-0.016
trace-0.01 2
trace
trace-0.142
and absorption level of 1 to 3 //g/d can result from smoking one
pack of cigarettes per day. Based on these data, the ATSDR13
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 study3 was undertaken to estimate the steady-state bovine milk
biotransfer factors (i.e., the rate at which the compounds are
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.3 The
bioconcentration factor estimated for cadmium is 1.3 x 10"6 L/kg.
To put this value into perspective, Stevens points out the
estimated biotransfer factor of TCDD is 2.6 x 10'2 L/kg and that
8-15
-------
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 for 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.
8.6 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 needed. Due, in part, to low
emissions of these HAPs from utilities relative to other
anthropogenic sources, a quantitative assessment of noninhalation
exposure to dioxins, lead, and/or cadmium has not been given as
high a priority as arsenic for further multipathway assessment.
However, dioxins, 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
needed to more comprehensively evaluate the impacts of emissions
of dioxins, lead, and cadmium from utilities.
8-16
-------
8.7 REFERENCES
1. U.S. Public Health Service. Toxicological Profile for
Arsenic. TP-92/02. Agency for Toxic Substances and Disease
Registry, Atlanta Ga. 1993.
2. Chen, Shun-Long, S. R. Dzeng, M. Yang, K. Chiu, G. Shieh,
and C. M. Wai. Arsenic Species in Groundwaters of the
Blackfoot Disease Area, Taiwan. Environmental Science and
Technology. 28:877-881. 1994.
3. Stevens, J. B. Disposition of Toxic Metals in the
Agricultural Food Chain. 1. Steady-State Bovine Milk
Biotransfer Factors. Environmental Science and Technology.
25(7):1289-1294. 1991.
4. 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.
5. U.S. Environmental Protection Agency. Health Assessment
Document for 2,3, 7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and
Related Compounds—Volume III of III: Site-Specific
Assessment Procedures—(External Review Draft).
EPA/600/BP-92/001c. Office of Research and Development,
Washington, D.C. 1994.
6. U.S. Environmental Protection Agency. Estimating Exposure
to Dioxin-Like Compounds—Volume I: Executive
Summary-External Review Draft. EPA/600/6-88/005Ca. Office
of Research and Development, Washington, D.C. 1994.
7. U.S. Environmental Protection Agency. Estimating Exposure
to Dioxin-Like Compounds—Volume II: Properties, Sources,
Occurrence, and Background Exposures—External Review Draft.
EPA/600/6-88/005Cb. Office of Research and Development,
Washington, D.C.
8. Ref. 7, Table 5-5, Chapter 5, (p. 5-9)
9. U.S. Public Health Service. Toxicological Profile for Lead.
TP-92/12. Agency for Toxic Substances and Disease Registry,
Atlanta GA. 1993.
10. U.S. Environmental Protection Agency. Air Quality Criteria
for Lead. EPA 600/8-83-028f. Washington, D.C. 1983.
8-17
-------
11. .Gartrell et al. Reference to be completed. 1986.
12. 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, 1996, Vol. 13, No. 1, 53-60.
13. U.S. Public Health Service. Toxicological Profile for
Cadmium. TP-92/06. Agency for Toxic Substances and Disease
Registry, Atlanta GA. 1993.
8-18
-------
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 Generation 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, which exploded as a super-nova.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 to 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
man's activities, in particular 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, the 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 reports,
-------
including the Background Information Document supporting the
decision not to regulate radionuclide emissions from coal-fired
boilers issued in 1989.4 This report updates previously
published data and estimates with more recently available
information 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
denotes 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, with an average
content of 1.24 ppm uranium and 2.18 ppm thorium in bituminous
coal, there is a corresponding activity of 0.41 pCi/g for each
member of the U-238 series and 0.24 pGi/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.
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
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 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. Analyses 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, 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 denotes 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
9-3
-------
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.
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 study, estimates of radionuclide emissions and
associated human health risks are based on fossil-fired boiler
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
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 report. 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.
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.
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°
1.2x 10°
1.2x 10°
1.2x 10°
2.3 x 10°
1.2x 10°
1.7x 10°
3.0 x 102
5.6x 10°
5.6 x 10°
1.2x 10°
5.6 x 10°
5.6 x 10°
1.2x 10°
5.6 x 10°
7.1 x 10'1
1.0x 10°
7.1 x 10'1
7.1 x 10'1
1.0x 10°
1.6x 102
3.5 x 10°
3.5 x 10°
7.1 x 10'1
2.1 x 10'1
7.8 x 10°
Gas
1.3x102
1.3x 10'2
1.3x TO'2
1.3x10'2
2.5 x 10'2
4.9 x10'2
2.9 x 10'2
2.5x 103
3.1 x 10'2
3.1 x 10'2
3.1 x 10'2
3.1 x 10-2
3.1 x 10°
3.1 x 10°
3.1 x 10°
2.1 x 10-2
4.7 x 10-1
4.7 x 10'1
4.7 x 10"1
4.7 x 10'1
5.7 x 10'1
4.7 x 10'1
4.7 xlO'1
4.7 x 10'1
1.4x 10'1
6.2 x 10 3
Oil
1.1 X10'1
1.1 x 10'1
1.1 x 10'1
1.1 x 10'1
2.1 x 10'1
4.1 x 10'1
2.6 x 10'1
3.8 x 102
2.7 x 10'1
2.7 x 10'1
2.7 x 10'1
2.7 x 10'1
2.7 x 101
2.7 x 101
2.7 x 101
1.8x 101
4.1 x 10°
4.1 x 10°
4.1 x 10°
4.1 x 10°
8.4 x 10°
4.1 x 10°
4.1 x 10°
4.1 x 10°
1.2x 10°
5.2 x 10'3
Per Billion kWe-h Generated (mCi/y)
Coal
1.5x 10°
7.7 x10'1
7.7 x 10'1
7.7 x 10'1
1.5x 10°
7.7 x10'1
1.2x 10°
2.0 x 102
3.8 x 10°
3.8 x 10°
7.7 x 10'1
3.8x 10°
3.8 x 10°
7.7 x 10'1
3.8 x 10°
4.7 x 10°
7.1 x 10'1
4.7 x 10'1
4.7 x 10'1
7.1 x 10-'
1.1 x 102
2.4 x 10°
2.4 x 10°
4.7 x 10"'
1.4x 10°
5.3 x 10°
Gas
2.6 x 10'2
2.6 x 10'2
2.6 x TO'2
2.6 x 10-2
4.9 x 10'2
9.5 x 10'2
5.7 x TO'2
4.9 x 103
6.0 x 10'2
6.0 x 10'2
6.0 x 10'2
6.0 x 10'2
6.0 x 10°
6.0 x 10°
6.0 x 10°
4.1 x 10'2
9.1 x 10'1
9.1 x 10-1
9.1 x 10'1
9.1 x 10'1
1.1 x 10°
9.1 x 10'1
9.1 x 10'1
9.1 x 10'1
2.7 x 10'1
1.2x 10'2
Oil
1.8x 10'1
1.8x 10'1
1.8x 10'1
1.8x 10'1
3.4 x 10'1
6.7 x 10'1
4.3 x 10"'
6.2 x 102
4.4 x 10'1
4.4 x 10'1
4.4 x 10'1
4.4 x 10'1
4.4 x 101
4.4 x 101
4.4 x 101
3.0 x 10'1
6.7 x 10°
6.7 x 10°
6.7x 10°
6.7 x 10°
1.4 x 101
6.7 x 10°
6.7 x 10°
6.7 x 10°
1.9x 10°
8.5 x 10'3
9-6
-------
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 gas-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
For a given facility, 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. Based on availability, primary
model parameters for plume dispersion and depletion are based on
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.)
From plume dispersion and plume depletion calculations, the
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.
9-7
-------
.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 (NAS 1988, NAS 1991) exposed to radon. 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 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.
a 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.5 The revised estimates yield a nominal value of 5.1 x 10~*
fatal cancer per rad for uniform whole body exposure to low-LET
radiation and 2.2 x 10~* 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
-------
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 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.
9-9
-------
.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 (i.e., 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 as to 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 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.
9-10
-------
The CAP-93 assesses risk for a circular grid that is defined
by 16 sectors and a radial distance of 50 km around a facility.
Risk to the population is determined by summing individual risks
by distance and sector 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 a separate addendum. These data are summarized in
section 9.2.
9.2 RADIONUCLIDE UNCERTAINTY ANALYSIS6
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. Since, as in almost all assessments
of environmental health risk, the risk estimates were based on
modeling rather than direct measurements of -exposure and risk,
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. And 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, and
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
9-11
-------
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.
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
9-12
-------
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 that were 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 1CT6 to 10'4.
9.2.1 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 10'5) .
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
9-13
-------
Table 9-4. Frequency Distribution of Lifetime Fatal Cancer Risks
for All Plants
Lifetime cancer
risk range
1 x 10° to 1 x 10'1
1 x 10'1 to 1 x 10'2
1 x 10'2to 1 x 10'3
1 x 10'3to 1 x 10"4
1 x 10-* to 1 x 10'5
1 x 10'5to 1 x 10"6
Less Than 1x10"*
Number of
people
0
0
0
0
1,027
95,745
196,000,000
Average
individual lifetime
risk
0
0
0
0
1.3x 10*
2.2 x 10-6
1.2x 10'7
Deaths per year
in this risk range
0
0
0
0
1.92x 10"*
3.06 x 10'3
3.32 x 10'1
Death per year in this
risk range or higher
0
0
0
0
1.92x 10-*
3.26 x 10'3
3.36 x 10'1
risk attributable to radionuclide emissions from electric utility
SGUs (includes coal-, oil-, and gas-fired utilities )is less than
1 cancer death per year (i.e., 3.36 x 10"1 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 arsenic, which was modeled with
the RELMAP (see chapter 6), the 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
two facilities are identified as exclusively oil-fired plants.
9-14
-------
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*
3x 10'5
3x 10'5
2x 10'5
2x 10'5
2x 10'5
2x 10'5
2x 10'5
2x 10'5
2x 10'5
1 x 10'6
1 x 10'5
1 x 10's
1 x 10'5
1 x 10'6
1 x TO'6
1 x 10'6
1 x 10'5
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 is the maximum individual risk expressed as lifetime fatal cancer risk.
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 10"6. The other 135 oil-fired utilities and
all coal-fired utilities are estimated to pose cancer risks less
than 1 x 10"6 due to inhalation exposure to radionuclides.
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
-------
9.3 REFERENCES
1. U.S. EPA. Estimates of the Health Risks Associated with
Radionuclide Emissions from Fossil-fueled Steam-Electric
Generating Plants., August 1995, Office of Radiation and
Indoor Air, Washington, B.C. EPA 402/R-95-16.
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. 174:424. 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. Estimating Radiogenic
Cancer Risks. EPA-402-R-93-076. Office of Radiation and
Indoor Air. Washington, D.C. 1994.
6. 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. pp. 54.
9-17
-------
10.0 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 a potential health risk. These consist of arsenic,
cadmium, chromium, lead, manganese, mercury, and nickel; dioxins
and furans (due to the toxicity of the organic chemical); and HCl
and HF (due to the estimated emissions of the compounds).
10.1 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 is
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.
10.1.1 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 concern with fuel switching is
whether or not it will increase HAP emissions, due to potentially
increased concentrations of trace elements in the fuel.
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 10-1 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 10-1, 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.
-------
Table 10-1. Comparison of Average Concentrations of Trace
Elements in Utility Fuelsa-b
- / «;»'»* J < '• -•;* ;
Sulfur
average*
SD (mean)1
No. averages
Coal'
(Ib/MMBtu)
1.24
0.19
26
Residual
(tb/MMBtu)
0.31
0.07
13
Natural gas"
(Ib/MMBtu)
0.00006
0.00006
2
, , '; ,, ^'.iafc*': • •
.;•>-.. •%-^*r-
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
Mercury1'2
overage
SD (mean)
No. (averages
Nickel
average
SD (mean)
No. averages
Coal'
(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
(IbAritlion)
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 gas*
(Ib/trillion)
0.19
0.06
2
-
-
-
-
-
0.001
1
-
• The coal data listed in Table 10-1 were not weighted for coal production by State of coal origin.
b There were only two sets of data lor concentrations of trace elements in natural gas in Table 10-1.
c 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 data 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.
a Natural gas values were determined from the preliminary EPRI test reports for Sites 120 and 121. The listed values are
detected concentrations.
e Averages of averaged data sets.
f This is the standard deviation of the number of averages directly below.
10-2
-------
10.1.1.1 Switching to Natural Gas Combustion. As shown in
Table 10-1, natural gas has the lowest average concentrations, on
a lb/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.
10.1.1.2 Switching from Coal to Residual Oil Combustion.
As shown in Table 10-1, 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.
10.1.1.3 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 10-1 shows the relationship between the
concentrations of mercury and sulfur in 153 samples of coal
shipments.2 As shown in Figure 10-1, 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
10-3
-------
g a
I8
•5. W
5 UJ
• D
in
CO
en
jj
(!)
I
•H
.c
to
tO
o
u
U-l
0
CO
D
D
in
cvi
D
D n
CD
£
3
D
a
in
in
ci
in
in
CM
o
CM
in
-H
i-l
3
03
i-i
3
U
H
0)
14-1
o
03
C
O
-H
JJ
(8
in
4J
Q)
U
O
C
0)
(U
4J
JS
C
0
•H
JJ
(0
I
O
0)
>H
3
cn
•H
(n;g uoi||uyqi) Ajnojaw
10-4
-------
concentrations (above 2.5 Ib/MMBtu) and lower sulfur
concentrations (below 1.5 Ib/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 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 is the
possibility of blending coals that contain different species of
mercury, and changes in the amount of vapor- and PM-phase species
of mercury would affect mercury control with PM control devices.
Another factor is that the blending of two different coals might
change the higher heating value of the resulting mixture.
The qualitative effects of switching to
lower-sulfur-containing coals on other metallic HAPs are examined
in Figure 10-2(a-g) through plots of the average concentrations
of each HAP, excluding mercury, with sulfur content in coal. As
shown in Figure 10-2(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.
10-5
-------
E
•o
03
O
If)
CO
c\
3 3
I 1
| |
3 U
in
d
CO
CNJ
.2
V)
8 8 § § 5
(nig UOJIIUVQI) iun;ujpeo
*O
CM"
CM
(nig
U
•mm
C
0
I •§
-
CM
CM
I
CO
8
ID
o
o
o
§
n CM
8 8
(njg uotiiw/qi) apuomo
10-6
-------
CO
TJ
(0
0)
ffi
CN
w
o
8
n CM
(nig uoinuvqi) pean
0
O
3
CO
CO
2
£D
s.
J
•-
S-1
eo
8
o
§
o
o~
o
o
o
co"
§
o
CO
10-7
T3
0)
D
C
•H
O
u
CT
f
(C
csj
I
O
0)
M
-------
10.1.2 Coal Cleaning
Approximately 77 percent of eastern and midwestern3
bituminous coal shipments are cleaned to meet customer
specifications on heat, ash, and sulfur content.9 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) .5 In the process of
removing the mineral matter, coal cleaning generates solid refuse
that contains trace elements; the solid refuse must be handled
carefully and disposed of properly. Any coal cleaning liquid
wastes will also contain trace elements, but the liquid wastes
may be properly clarified and then recycled.
Table 10-2 lists the limited amount of available data on
trace element reductions achieved through conventional coal
cleaning. In Table 10-2, 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 10-2, 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 10-2, 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
Research has only recently begun to look at coal cleaning specifically for
trace metal removal.
10-8
-------
Cn
C
-H
ti
(0
u
(0
O
U
(0
c
O
•H
.u
c
(U
c
O
u
01
d
O
n
EH
Ti
Q)
>
0)
•H
c
O
-H
4J
U
Q)
Q)
6
a>
rH
H
Q)
u
(d
r-l
EH
CN
I
O
"3
|I
as '3
»S
"* >.
|i
% u
E ^
U U
Removal
fluoride
^
o .=
E E
« J
1 S
O 3
" 1
6? °
1 «
i S
™ «
SS
U
1
u
a
C/5
1
o
in
~
£
JC
3
ON
1
u
en
00
NO
S
en
cs
1
o
3
E
00
fs
%
^
in
_J
0
en
en
P
t?
f^
Tt
J
S
eN
9
t-
rN
es
eN
en
J
J
00
-
oo
NO
5:
Tt
C3
J
in
eN
p-
en
es
$
§
TT
00
•t
in
J
oo
00
oo
a
S
,
S,
a.
R
o
S
£
NO
Tt
m
NO
en
n!
| Sewickley
JO
1
R
T?
~
00
en
(V
^-
£
D*
in
3
en
oo'
o
00
5?
0
m
t^
iS
^,
{
o
S
NO
in
I—
in
Tt
T?
^f
r~
ft
[Upper Freepor
S
S
S
in
NO
£
0
en
es
5>
I
1
in
1
en
00
ON
in
^
o
Tt
en
B
E
cn
r-
s
X
B
JQ
£
r-
i1^
(9
M
S
18
o
Tt
OO
en
in
Tt
| average
.Q
(0
O)
V
10-9
-------
from the report:
and Clifford Raleigh, The Effect of Coal Cleaning on Trace Elements, Draft Report, Development of
_
(continued
•
rq
i
o
T-{
Q)
rH
ft
(0
EH
o>
09
O
C
£
.£
09
CC
£
09
.*
CO
•*-»
£
09
•a
t
CO
ences
09
^
0)
^
o
H—
CO
ij
00
O
>
1, Robert Dospo
u
CO
O
w
03
^
<
u
_c
a
U
£>
E
0.
LU
1
•o
£
CO
Q.
09
a.
CO
O5
O9
December 16, 1
09
E
j*j
0
O)
<
the report:
rey, Rodney R. Ruch, Heinz H. Damberger, Chusak Chaven, John D. Steele, Wayne T. Frankie, Ken
vailable Coals from Illinois Mines," draft report, December 28, 1993, Illinois State Geological Survey
E
o
H—
09
.*
CO
+J
£
09
09
09
U
£
03
»+-
09
W
^
O
•*-
CO
+J
to
o
c <
\, Richard D. Ha
racterization of
fc CO
CO J=
J= U
^ S
>k *,
15
09
0 ^
T3
09
O9
'«
M
CO
C
XJ
0
03
&
E
3
c
_®
•^
jz
O9
'»
CO
CL
•o
CO
to
•o
09
*-•
09
U)
£
Q.
ation of the Effects of Coal Cleaning on Fugitive Elements: Final Report, Phase III," DOE/EV/04427-
3
CO
LU
s
CD
0
£
< oi
_ 00
12
"f
12
LL. (D
U
10-10
-------
The average trace element reductions, listed in Table 10-2
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 10-2, 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.
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.
10.1.3 Coal Gasification
Coal gasification converts coal to a 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
technology (IGCC) and conventional cold-gas cleanup. In this
process, gas from coal is used to generate electricity from both
10-11
-------
a steam turbine and a gas turbine. Steps in the process, shown
in Figure 10-3, 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.
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 10-3.10
The hot gas from the gasifier was treated in a moving bed with
zinc titanate sorbent, followed by a cyclone. Emissions from the
hot-gas cleanup were vented to a flare.
Hot-gas cleaning may fail to condense volatile trace species
(e.g., mercury) and may produce a high concentration of trace
elements in the vapor phase. This could lead to higher trace
element emissions compared to the concentrations normally
encountered in conventional coal-fired power plants.11
10-12
-------
OXIDANT
COAL
PREPARATION
GASIFICATION
COAL
GASIFICATION
SLAG/ASH
RECOVERY
SULFUR
RECOVERY
WATER
DISPOSAL
GENERATOR -*
GENERATOR
COMBINED CYCLE
Figure 10-3 . Coal gasification combined cycle technology-
Table 10-3. Emissions from an Air-Blown, Fixed-Bed Gasifier
Trace metal
Arsenic
Cadmium
Chromium
Mercury
Nickel
Selenium
Emissions to flare, /yg/Nm3
639
16
155
20
1,530
68
Emissions from turbine
simulator, //g/Nm3
8
0.19
20
2
26
0.56
Total air, //g/Nm3
647
16.2
175
22
1,556
68.6
10-13
-------
10.2. 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 most
of the recent emission testing on utility units dealt with the
generation and control of trace metals and not organic HAPs,
trace metals are used 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.12 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.13 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 lb/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 10-4, there appears to be a trend toward
reductions in HAP emissions through the addition of NOX control.
However, this trend is neither uniform (see arsenic, beryllium,
10-14
-------
z ?
o
co
§
90.
I
1
*
1 1
— o>
8 *
10-15
|
£ E
1 i
£ 8
S i
S £
a §
! I
I I
.2 I
o
-------
and cadmium percentage change between Sites 110 and 114) nor
universal (see Site 16 vs. 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 10-5) 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 10-5. 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, beryllium, lead,
and mercury seemed to be emitted in higher amounts by wet bottom
units while cadmium, chromium, and manganese seemed to be emitted
in higher amounts by dry bottom units (see Figure 10-4a) . When
units firing only bituminous coal were analyzed, the same effect
was observed (see Figure 10-5a). 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 10-6a). 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
10-16
-------
CM 0-
III
Si*
> to
1 !
0 8 «
in ^
2
UJ
0
o
•iB
2 £ CO
£i
11
ni
S -5 5
U) <0 -0
c
£ s
111 g
« x
£ O
£1°
UJ —
!
b
*
i
s
*
UJ
s
1-
uf
*
t
Bottom Typt
o
o
M
o
c
o
u
5
i
(J
o
c
o
s
0
1
o
0
o
I
m-
3
w
%
;
ffl "
3
CO
'
t
_
t"
-
t s % C
;w 1
i" §
i- 1
-
A
*
in
CM
CO
CO
CO
in
CO
CO
CM
t
CM
Arsenic
,_
q
in
CO
*
CO
in
CM
CO
O)
1
cn
*~
CM'
to
00
r>
Cadmium
o
to
1O
CO
$
q
to
00
CM
Chromium
q
CM
CO
CO
in
o
CO
in
in
.
CM
__
in
CO
I
0
CO
to
CD
to
in
CO
6
CM
in
CO
CM
CO
i
„
;
q.-.
CO.
CM
in
^s
I
'"' 1
-- I
« UJ
1
CO
o
in
6
00
in
0
in
00
o
in
CM
O
in
O
Arsenic
S
o
00
o
6
CD
fM
O
o
CO
o
in
o
in
o
1
CM
O
o
CO
6
,_
o
o
CM
O
5
0
o
o
E
5
in
CM
O
CM
CM
ci
00
CM
O
in
CO
0
to
CN
O
o
CO
o
Chromium
CO
o
00
CO
6
CD
in
0
S
o
00
o
o
in
o
I
CO
CO
o
CO
6
in
O
cn
o
CO
o
o
CM
O
1
i
>.-•: '•
8^
*-:•.*..
... . . ..
VV.S
§;•:•-•••
£ !;! i| '
!.':=
o|
£:
:--:- :•
"i
£: i
(;> 1
in :' ;• c?
6;- :; 5
If 1
y'$y g
y"^'< £
m*. ? -"•'•' ' **
C*3%.-H>!" 0
°l I
8 1
few*1 S
:;•••••:. n
K:i
•IF-
5 i'-vly
?-'«:" ': '
un
CM
CO
CO
CO
cn
in
CM
S
CM
1C
in
00
00
o>
.
1
CM
CO
Px
in
i
CO
CD
in
CO
00
CD
CO
CM
CO
CM
CM
I
cn
00
in
CO
CO
CM
CO
CM
IX
CM
q
CO
CM
CD
in
CO
Cadmium
r*.
CO
§*
in
O)
O)
CM
in
CM
CM
CO
CO
6
CM
CM
CO
s
0
Chromium
CO
ps.
Si
148.82
q
IX
s
CO
in
°
N
S"
CM
1
CO
q
00
o
PS.
CO
g
o
CO
ci
o
CM
10
1O
CM
CM
S'
O>
CO
CO
CM
i
1
r-
co
"*
O)
rs.
CM
CM
CO
CM
O
CO
oa
s
00
Mercury
o
1
.If
II1
C -— *-
10-17
-------
•» *
1
•H
1
U
in
i
o
H
I
1
09
t- cfl
ss
UJ
CD
is
-. O
UJ
in
1
UJ
o
5
5
a.
UJ
CM
o
v>
£
0-
UJ
"Z
£
,«
UJ
O
0
10
E
UJ
CO
O
S
o
o
*!
o_
Ul
EPRISite 110
(w/LNB and
offset air)
1
1
i
„
Q
CC
a
,.
Q
EC
f?
D
fr
a
>•
or
a
cc
a
!
I
i
a
0
I
O
o
1
TJ
*
1
S
S
U.
1
5
I
c
CD
5
1
C
fC
1
c
i
1
K"
Si"
ft?
x %
tl;
? s
b
l-s
s
K
%
,_"
t1
snjg uoniin
•. -, -,.. .-.;.-:«-•; ;.fi >.
^tlx:lx*x:xi?»^
1IB
Finltfpf
"%*3%M~-'- ''¥'?'•
' : ' '' ?'< f^ : '' :' ?
Emissions in IbJ
Arsenic
O)
to
09
o
cn
CM
d
CM
CO
*
CM
CO
0
cn
09
to
CO
CO
r^
i
1 Arsenic
E
I
6
CO
00
o
CO
to
09
10
O
CO
^
09
S'
01
f.
1
1
i
_
CM
O>
o
CO
CO
CO
CO
CO
CO
^.
00
CO
CM
00
?'
0>
c
o
1 Cadmium
i
CO
in*
.
CO
•*
cn
CM
o
10
cc
10
!g
CM
O
id
n
CO
S
09
09
A
CN
10
^
O
pi
1 Chromium
'
CM
CO
CO
q
to
CM
10
^*
CM
CO
O
CO
(0
CO
00
to
09
CD
00
CO
09
e
I
1
cn
d
CD
in
in
CM
CO
CO
CN
•-
CO
n
CO
cn
in
in
O
CN
GO
0
cri
en
cn
a
1
i
s
1
*«
&
•v
CO >»
^h
00 ..
»--.';4x:s.i':is
.;-: ^.y"-5:x;5. i-'%-->:- ••. . ^-Jl" ;, .•
Arsenic
8
8
D
O
g
O
8
n
o
S
o
8
cn
CO
o
at
CO
<:
1
I
g
o
8
D
o
CM
cn
o
8
CD
O)
0
in
o
8
n
o
CO
a
6
i
j
Cadmium
,_
o
8
09
10
O
g
O
*
o
8
0
o
8
o
0
8
1 Cadmium
Chromium
S
O
09
10
O
n
o
D
O
8
D
O
00
to
o
8
8
8
1 Chromium
•
s
o
8
CD
CO
O
D
O
8
§
8
8
09
CO
O
g
1
Manganese
S
o
S
o
09
IO
o
CO
cn
o
8
r.
0
D
O
8
S
o
R
a
s
•
*:
T?
Jl
lili
B&iiX:
gf
~!
D & 'i •
°|;
II
°l
°K
£.
>••' ' '
:• -
5 v'/
O:-- •
8p:.
Bf
&
'<••'•• to
>••" |
10 6^; a
•s, •:•.;• •. O
?!:••• 1
1.00
S«;g^?|:|t:;?:
ncentratlons In ci
fe K
i
Arsenic
0
to"
2
09
00
O
CM
O
CM
CO
10
o
in
5
09
in
CD
<
n
m
cn
cn
tn
Ul
pi
cc
I
3
=
5
10
at
o
E
""
M
en
A
M
£
n
CO
00
CO
;
*
106.24
1
Cadmium
5
8
09
09
cn
CO
CM
CM
09
to
CM
-
*~
CO
cn
CD
CM
CO
cn
CD
rri
CN
CC
i
i
Chromium
00
00
M
«
ID
£
3
a
cn
in
0
cn
CD
CN
O)
CO
cn
ID
8
M
CN
cn
CO
en
(N
?
CM
Ifl
to
r.
\ Chromium
1
cn
cn
CM
in
to
CM
0
CO
(0
5
D
•-
CM
to
D
2
o
CO
cn
to
CD
09
ri
CM
*
n
•
Manganese
o>
•*
o
CO
CO
6
CM
to
CM
CO
CO
8
CD
CM
6
CO
cn
CM
S
in
m
CO
CM
in
es
CO
CO
cn
6
cn
1C
1
«
c
c
t
>
3
j
I
«
u
in
cn
0
in
**.
*~
en
8
*
f
D
J)
in
V.
CO
I
o
I-
CO
S
§
'•5
*
ill
O 3 10
I 2 »
I = 5
s « ii
n " x
CD U3
-------
•0
u
in
i
o
H
I
.9
I
s.
• I I
g I g
C *-" +j
CD 03 C/J
10-19
-------
Figure 10-4a. Average emissions
2500 T
5 10
Arsenic
Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 10-4b. Average trace metal removal by boiler
92%
82%
Arsenic Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 10-4c. Average trace metal concentration in feed coal
Arsenic Beryllium Cadmium Chromium
Lead
Manganese Mercury
Figure 10-4 (a-c). Average coal-fired emissions, average trace metal removal, and average trace
element concentration in feed coal vs. bottom type (bituminous and subbituminous coal-fired)
* Denotes a boiler emission average higher than the concentration found in the feed coal. This is caused by figures (a) and (b) having been computed from
data with EMFs limited to a maximum of 1.0. 1 f) — ? 0
-------
Figure 10-5a. Average emissions
1189
Arsenic
Beryllium Cadmium Chromium
Lead
Manganese
Mercury
Figure 10-5b. Average trace metal removal by boiler
92%
80%
Arsenic
Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 10-5c. Average trace metal concentration in feed coal
3500
3000 -
Arsenic
Beryllium Cadmium Chromium
Lead
Manganese Mercury
Figure 10-5 (a-c). Average coal-fired emissions, average trace metal removal, and average trace
element concentration in feed coal vs. bottom type (bituminous coal-fired only)
Denotes a boiler emission average higher than the concentration found in the feed coal. This is caused by figures (a) and (b) having been computed from
data with EMFs limited to a maximum of 1.0.
10-21
-------
Figure 10-6a. Average emissions
4000
3500
3000
£2500
.22000"
{1500 -
1000 --
500--
0--
3754
70 42
29 4
52 2
Arsenic
Beryllium
Cadmium
Manganese
Mercury
Figure 10-6b. Average trace metal removal by boiler
94%
91%
77%
Arsenic
Beryllium Cadmium Chromium Lead Manganese Mercury
Figure 10-6c. Average trace metal concentration in feed coal
<- to
o t
t 00
10 •*
Arsenic
Beryllium Cadmium Chromium
Lead
Manganese Mercury
Figure 10-6 (a-c). Average coal-fired emissions, average trace metal removal, and average trace
element concentration in feed coal vs. bottom type (subbituminous coal-fired only)
* Denotes a boiler emission average higher than the concentration found in the feed coal. This is caused by figures (a) and (b) having been computed from
data with EMFs limited to a maximum of 1.0.
J- U ~" & £
-------
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.
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 10-4(a and
b), 10-5(a and b), and 10-6(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 did not occur in Figures 10-4(c), 10-5 (c), or 10-6(c)
because these data were taken directly from the coal feed without
modification. The result of this 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.
10.3 POSTCOMBUSTION CONTROL
The following sections assess how different APCDs affect
removal of selected HAPs from fossil-fuel-fired electric utility
flue gas.
10.3.1 Particulate Phase Controls
Figures 10-7 through 10-14 and Tables 10-6 through 10-9 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). 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
10-23
-------
S
S
3
o
o
00
o
£
t
§
O
o
8
a.
CM
CO
o
o
8
IBAOOI8J %
O CM
^- S
»•
r%
i i
• •
— cS
in
g
in
CM
0)
'o
£
0)
o
1
S
1
:i
1-i
s i
l«
O c
f-I
S'.
o E
it
c
o
m
c
01
Q
cn
c
m
•O
«
C
CD
f"»
cn
cn
A
cn
CM
cn
en
cn
'E
(0
h-
r*
en
cn
A
O
cn
CO
CM
03
cn
CO
cn
E
I
r*.
00
cn
CO
CM
cn
in
CO
E
E
•a
CO
CJ
!••
0)
O)
A
CO
CO
CO
CM
cn
^
CO
1
1
6
r-
cn
CO
A
CO
in
cn
CO
CO
•o
CD
Is*
cn
O)
A
in
cn
in
CO
cn
CO
cn
CO
w
CO
co
cn
CD
!*•
CN
CO
O
CO
CM
CO
in
CM
>.
o
i
CO
2
3
00
c
• ^N
c
o
J3
O
c
'o
S
o
S
o
'•4—>
c/5
'-t->
•4—»
cn
>
O
o
cd
|BAOW9J
10-24
-------
o
o
10
s
|EAOLU9J %
in
CM
o
'o
o
=
8
«N
Si
II
'5 "2
!> S
&~*
II
11
2*
I 6
2 E
•M
3
O
O
X
CD
C
i
0)
Q
V)
r.
CO
c
CO
CD
CM
m
cn
in
00
O
CM
CD
CM
CD
[Arsenic
CM
CD
CD
CD
CD
O
CD
OS
Cn
en
Beryllium
CM
en
A
en
cn
6
a>
cn
| Cadmium
CM
CO
en
CD
cn
CO
IV
0
cn
| Chromium
CM
CO
CD
CD
CM
CD
CD
•o
(0
CD
_J
CM
cn
cn
CO
cn
in
cn
cn
®
| Manganes
CM
O
o
o
o
0
| Mercury
I
o
*o
c
CO
o
o
C/3
o
cd
O
I
S-l
fc.
tw
o
•4— »
tW
CO
o
W
u
Q
o
o
CM
|BAOUJ9J o/0
10-25
-------
r
CO
0)
o
o
in
tn
CM
|BAOLU9J
I
o n
O
O
o
S
:§
8
fe.
5-S
•2 =
c
o
o
X
m
C
s
O
V)
c
m
•o
V
ID
m
CM
v
in
in
*
CO
CO
CO
o
'r
02
in
<
CM
^
in
CO
in
en
CM
in
in
in
in
•o
to
V
CM
CO
CO
CM
*
CD
CM
CM
(O
CM
CO
2-
3
S£
a)
CM
CO
CO
^
2
CO
CO
CO
CO
"S
o
z
00
c
o
Cfl
00
4>
'3
_u
'o
e
o
s
,o
BO
u
03
•*•*
on
u
a <*
o
O
m
o
c^
g
o
oo c
(O
£•§
£ !
II
•- p
00
o
•^ I
O ,E-
8
|BAOLU9J
10-26
-------
IBAOLU9J o/0
O
I
O
0)
'o
o
0)
in
g
IBAOLU9J
m
CM
o
** i 2
• • I
*-+.
*A 1 Jr j
W \ "«— ' "
^^
- o
» •
•» i
1 !
0)
'o
_.«
o> c
o
=8
<1>
r 8
a.
CO
35
1 a
A i
t E
s -a
oS
•E-D
.0 a
»2 «S
il
talic HAPs b
ium, chromii
11
•8 3
1.2
If
m y
11
li
•3
c
o
o
CO
c
I
V
Q
(0
'•s
c
n
CO
en
en
A
en
CO
*~
en
en
en
en
[Arsenic
in
en
en
A
en
^
CM
en
en
A
en
en
E
3
f
m
*
en
en
O
CO
en
CM
*^
[Cadmium
in
en
en
A
in
0
en
en
^j-
en
Chromium
CO
en
en
A
en
CN
*~
en
en
en
en
S
CO
en
en
A
in
00
CO
in
en
en
CO
en
0}
[Manganes
in
CO
^
0
CO
co
CO
00
CM
Mercury
en
I
o
O
en
o
c
u
.2
e«
O
en
>
t-l
o
Os
o
03
H
in in
s
10-27
-------
particulate form are readily controlled by PM control devices.
These HAPs include arsenic, beryllium, cadmium, chromium, lead,
and manganese. Table 10-10 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.
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.
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, dioxins may be formed in
particulate control equipment. Studies of MWC systems have shown
that dioxins and furans can be formed in particulate-laden flue
10-28
-------
Table 10-10. Particulate Metallic HAP Removal Percentage from
ESPs and FFs (Excluding Mercury)
Particulate Control Device (Coal)
ESP (cold-side)
ESP (hot-side)
Fabric Filter
Number of data
points
102
12
32
Percentage of data with a
HAP removal efficiency
greater than 90 percent
90
92
91
gases in a temperature range of 480° to 1,020° F (250° to
550° C) .41'43 Units that are equipped with hot-side ESPs (ESPs
upstream of an FGD unit) are of 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.
10.3.2 Vapor Phase Controls
Figures 10-15 through 10-18 and Tables 10-11 and 10-12 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
five data sites at which two of the five 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.
10-29
-------
I
I
I
Q
oO
1 ' 1 ' 1
3 O O
"1 ^-« rrt j— . rvi i-n
emoval
*
o
3
.£?
E
CO
o
o
Q
O
u.
• I*
1
|BAOUiej
r
-. °
JT |I\
^ <'E"£
^ a|s
.H u
jo =". a
2 E §
1= E
- -sFS
il o 5 cs
^ S ,- "2
^> «^ o ro
— o •-
anganes
^
CO
en
in
o
CO
CM
fx
in
C
&0
o
en
o
IEAOLU3J %
10-30
-------
00
l—<
O
^•M
•o
8
0
CO
CO
CM
o
(O
CM
%
CO
IS
b~.
i
o
(0
k
L.
I
0)
O
l/J
Median
c
(0
ID
^
O)
0)
A
(O
0)
00
,_
O)
A
0)
0)
|Arsenic
CO
OS
O)
A
OS
CO
o
O)
O)
A
0)
0)
A
| Beryllium
^
o>
O)
A
00
04
00
O
O
m
Cadmium
CO
O)
O)
A
o>
in
,_
£
00
O)
| Chromium
CO
O)
O)
A
o>
O)
^
o
o>
A
O)
0)
•o
fa
V
CO
0)
ff>
A
0>
A
r-
O
os
05
A
0)
en
A
to
JManganes
*
in
in
0
O)
CM
CM
CD
CM
(Mercury
o
o
in
8
|BAOUJ9J
in
2
£
E
*-
SO)
— .
CM C
• •~>
^"
Q
CO
II
" E"
11
a 3
11
« =
E t
"8-2
-..2
Ii
« „•
(N
i
O
u
1
10-31
-------
.The HAP metal removal by SDA/FF-eguipped 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.
10.3.3 Acid Gas Control
There was a limited amount of data (using EPA Method 26a)
available on the removal efficiencies for HCl 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, respectively, 17 percent and
14 percent bypass. The test report data indicated that: ESPs
removed less than 6 percent of the acid gases; FFs removed
approximately 44 percent of the HCl and essentially none of the
HF; an FGD with 17 percent bypass was estimated to remove
approximately 79 percent of the HCl and approximately 28 percent
of the HF; and an SDA/FF with 14 percent bypass was estimated to
remove approximately 82 percent of the acid gases.45 Despite the
inconsistencies in removal efficiencies achieved for HCl and HF
with FFs and FGDs, the data indicate that the S02 control devices
remove more of the acid gases than do PM controls.
10.3.4 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 sectipn 10.6.1.2), 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.
10-32
-------
10.4 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.47 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.48
Future potential electrical transmission equipment could include
the development and use of superconductive power lines, which
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) .49 A final method would be to switch to a source of
renewable energy (e.g., wind, solar, biomass-firing), 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.)
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
2030.50
10-33
-------
10.5. POLLUTANT TRADEOFFS
10.5.1 HAP Increase/Decrease
The various strategies for limiting HAP emissions, discussed
in sections 10.1 through 10.4, have different effects in
controlling air emissions of all HAPs. Table 10-13 presents the
qualitative effects of the different control strategies on air
emissions.
Table 10-13 provides a comparison of HAP removal
effectiveness of different existing and alternative control
strategies. As shown in Table 10-13, 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 10.3.1, 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.
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.
10.5.2 Water/Solid Waste Considerations
Coal cleaning can produce a variety of waste problems.51
The process creates a liquid waste containing fine coal particles
and inorganic elements and compounds dissolved from the ash in
10-34
-------
«w
«M
H
s
•H
I
o
H
10-35
•g
«
fl)
M
3
JJ
«
i-i
0)
0)
X
JJ
o
u
0)
•g
to
cu
w
w
0)
T3
•H
CO
I
4-)
I
c
•H
T3
0)
U
3
-O
0
^
a
0
a
d) en
J-> 4J
o u
a
-------
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 ground water can occur from
coal wastepiles or storage piles if water is allowed to
infiltrate them. Contaminants such as iron and manganese and
heavy metals such as cadmium and silver may leach from the
wastes. Because processing wastes are higher in ash (material)
content than in the cleaned coal, leachate from waste poses a
greater threat.51
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 the ratio of bottom ash to fly ash. Increasing the
10-36
-------
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.52 In addition to
removing sulfur oxides, regenerable processes generate a usable
product from the sludge, such as gypsum, that 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.53 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
10-37
-------
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.54
All elements of the alternative controls, from conservation
to technology improvements, can reduce the amount of waste
produced by the utility industry.
10.6 AVAILABLE CONTROL TECHNOLOGY AND STRATEGIES FOR MERCURY
CONTROL
Typical mercury removal efficiencies for conventional
emission controls are discussed in section 10.3. Conventional
controls are generally inconsistent in their effectiveness and
range from 0 to more than 80 percent removal.
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. This hypothesis comes from data on MWCs,
particularly from units burning refuse-derived fuel, which tend
to have greater carbon carryover and a larger drop in mercury
concentrations across particulate control devices than do mass
burn systems, which tend to have lower carbon levels in the fly
ash. Because combustion modification NOX controls can lead to
increased carbon in the fly ash of coal combustion units, some
level of mercury capture may take place in units with combustion
control for NOX through a similar mechanism. This has not been
verified for the conditions occurring in utility units, however.
Strategies for reducing mercury emissions from electric
power generation include demand reduction to decrease overall
fossil fuel use, use of other forms of generation (such as
nuclear power), switching to fuels having less mercury, 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 (/zg/dscm) to
21 /ug/dscm at 7 percent oxygen and standard conditions of 20° C
and 1 atmosphere, while utility flue gas flow rates may range
from 400,000 to 4,000,000 dscm/h.55 Thus, at utility plants, any
strategy for mercury control must consider adequately treating
large volumes of gas in order to remove relatively small
10-38
-------
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.
The rest of this section describes the impact of process
variables on mercury removal in existing systems and then
describes a potential new control, mercury removal by adsorption
onto AC by various methods. Other potential technologies are
also briefly described.
10.6.1 Impact of Fuels and Temperature on Mercury Emissions
Fuels, their sources and precombustion treatment, and the
way they are burned can have a significant impact on the quantity
of mercury emissions in the flue gas from a boiler.
10.6.1.1 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 gas56 in
one of three forms: (1) elemental, (2) ionic, or (3) organic.
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., Kg**) is water soluble and
less volatile than elemental mercury (i.e., Hg°) . Thus, reducing
the temperature of the flue gas or wet scrubbing of the flue gas
may result in increased ionic mercury removal. 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.
The distribution of ionic mercury, most likely mercuric
chloride (HgCl2) , in coal-fired utility flue gas ranged from
12 to 99 percent of the total mercury content and averaged
79 percent, while the distribution of elemental mercury in
coal-fired utility flue gas ranged from 0.8 to 87.5 percent of
the total mercury content and averaged 21 percent. Analysis of
two samples suggests that mercury is predominantly in the
elemental form when the fuel is oil. 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;57 however, validated methods are needed
to establish the speciation of mercury before any relationship
between mercury speciation and control device performance can be
inferred.
10-39
-------
It has been observed that higher concentrations of ionic
mercury are obtained in utility flue gas when the combusted coal
has a high chloride concentration (0.1 to 0.3 weight percent),
but more data are needed to verify this association.58"60 The
distribution of mercury species in coal-fired flue gas also
appears to vary with the type of coal (e.g., bituminous,
subbituminous, or lignite) .61>62 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.
The speciation of mercury is important in planning control
strategies but is still in the early stages of investigation.
Studies of a pilot-scale wet FGD system treating coal-fired flue
gas indicate that more than 90 percent of the ionic mercury was
removed while hardly any of the elemental merciiry was removed.60
Similarly, studies at a pilot-scale SDA/ESP system treating
coal-fired flue gas suggest that 95 percent of the ionic mercury
and essentially none of the elemental mercury were removed.59
10.6.1.1.1 Coal cleaning. As mentioned in section 10.2.2,
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.56
Advanced coal cleaning techniques are being investigated for
improved trace metal removal potential.
10.6.1.1.2 Coal switching. As mentioned in section
10.1.1.3, switching to certain higher sulfur coals (above
2.5 Ib/MMBtu) and selected lower sulfur coals (below
1.5 Ib/MMBtu) , which contain less than 15 Ib mercury/trillion
Btu, could reduce mercury emissions from utility units. However,
coal blending for mercury control is not a proven control
strategy. Changes in the electrical resistivity and amount of
fly ash resulting from coal blending could reduce PM capture
efficiencies by ESPs and subsequently lead to increased emissions
of PM and HAP metals. Blending for mercury control could also
increase levels of other HAPs or sulfur. Another uncertainty
with coal blending for mercury control is the possibility of
blending coals that would result in different species of mercury,
and changes in the amount of vapor- and particulate-phase species
of mercury would affect mercury control with PM control devices.
10-40
-------
10.6.1.2 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.55 Mercury is found
predominantly in the vapor phase in utility flue gas.56 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
PM56 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 study suggests that mercury removal
efficiencies apparently increase as the temperature of the flue
gas decreases. Specifically, as the flue gas temperature
decreased from 107° C (225° F) to 99° C (210° F) to 96° C
(205° F), the mercury removal percentages for a pilot-scale FF
correspondingly increased from 27 percent to 33 percent to 51
percent.63 However, mercury control is not entirely temperature
dependent — FGD systems have exhibited no mercury removal (see
Figure 10-17) with outlet temperatures below 107° C (225° F).
10.6.2 Developing Technologies: Activated Carbon
Research has been performed on AC injection at a pilot-scale
SDA/FF plant59 at a pilot-scale pulse-jet FF (by EPRI)63 and at a
pilot combustor and FF (by the University of North Dakota Energy
and Environmental Research Center [UNDEERC] and EPRI).61
10.6.2.1 Factors Affecting Mercury Removal Efficiency.
Preliminary data from two studies suggest that factors besides
the optimum amount of AC that is injected may affect mercury
removal. These factors are temperature, the form of the
vapor-phase mercury, and the type of activated carbon61 injected
into the flue gas.
10.6.2.1.1 Temperature. Mercury removal efficiencies and
the required amount of AC injection were apparently temperature
dependent within a range of 88° to 121° C (190° to 250° F) in a
pilot-scale study conducted on reducing mercury levels in utility
flue gas through AC injection upstream of an FF.63 At lower
temperatures, 88° to 96° C (190° to 205° F) , mercury
concentrations were reduced by 97.7 percent with an AC injection
rate of approximately 155 /^g carbon//ug of inlet mercury. At
higher temperatures, 110° to 121° C (230° to 250° F) mercury
concentrations were reduced by only 75 to 87 percent with an AC
10-41
-------
injection rate of approximately 3,500 ^g carbon///g of inlet
mercury.
These studies suggest that more mercury is removed and less
carbon 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 96° C (205° F) and still
avoid acid condensation, provided low-sulfur coals (less than
about 1 weight percent sulfur) are burned.64 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.
10.6.2.1.2 Speciation of Mercury. The effectiveness of AC
injection in recovering different forms of mercury is still in
the early stages of investigation. In testing at a pilot-scale
SDA/ESP system in Denmark,59 the flue gas contained from 66.6
percent to 83.4 percent ionic mercury with an average of 75.2
percent ionic mercury, and elemental mercury constituted the
remainder of the total mercury concentration. Without AC
injection, the pilot-scale SDA/ESP system removed 96.8 percent of
the ionic mercury and essentially none of the elemental mercury
from coal-fired flue gas (i.e., the system removed 72.5 percent
of the total mercury). During testing with AC injection, the
flue gas contained from 58.4 to 77.7 percent ionic mercury with
an average of 69.5 percent ionic mercury, and elemental mercury
made up the remainder of the total mercury concentration.
Activated carbon injection ahead of the SDA/ESP system removed
46.4 percent of the elemental mercury and 84.3 percent of the
total mercury.59 In testing at another facility that had a full-
and pilot-scale SDA/FF system, the flue gas contained 85 to 90
percent elemental mercury. Without AC injection, the full- and
pilot-scale SDA/FF systems removed 10 to 20 percent of the total
mercury from the coal-fired flue gas,59 and the low removal of
total mercury may be attributed to essentially complete removal
of the ionic mercury and little removal of the elemental mercury.
Activated carbon injection ahead of the pilot-scale SDA/FF system
increased the removal of total mercury to approximately 55
percent, and injection of iodide- and sulfur-impregnated AC
10-42
-------
increased the removal of total mercury to approximately
90 percent.59 Thus, the studies at this SDA/FF system suggest
that sulfur- and iodide-impregnated ACs are needed for total
mercury removals of 90 percent when elemental mercury is the
predominant mercury species. Furthermore, the studies suggest
that total mercury removal efficiencies are dependent upon
mercury speciation.
Finally, recent tests conducted at the pilot-scale combustor
and FF at UNDEERC also suggest that mercury removal is dependent
upon mercury speciation and the type of carbon used.61
Since mercury speciation affects total mercury removal from
utility flue gas with AC injection and because the speciation of
mercury is not understood at this time, more data are needed to
establish the factors that affect, as well as to characterize,
mercury speciation in utility flue gas.
10.6.2.2 Comparison of Characteristics of MWC/utility Flue
Gas. The level of mercury control achieved in 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 10-14, 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.
10.6.2.3 Results of Pilot-scale Carbon Injection.
Figure 10-19 shows mercury removal efficiency ranges for existing
control devices with AC injection at different temperatures based
on limited available information.59-63 Figure 10-19 shows that the
tested SDA/FF systems with AC injection had a median mercury
10-43
-------
Table 10-14. Comparison of Typical Uncontrolled Flue Gas
Parameters at Utilities and MWCsa
Uncontrolled flue gas
• parameters
Temperature (°C)
Mercury content (^g/dscm)
Chloride content (Atg/dscm)
Flow rate (dscm/min)
Coal-fired utility boiler65
121 - 177
1 -25
1,000- 140,000
1 1 ,000 - 4,000,000
Oil-fired utility boiler
64,65.66
121 - 177
0.2-2
1 ,000 - 3,000
10,000-2,000,000
MWC" •»•"
177-299
400 - 1 ,400
200,000 - 400,000
80,000 - 200,000
' Standard conditions are 0° C and 1 atmosphere.
" Moisture content in the MWC flue gas was assumed to be 13.2 percent.
removal efficiency of 60 percent with a range from 50 percent to
99 percent removal. (The 99 percent removal may have been
obtained with an impregnated carbon, but the test report did not
specify the type of carbon. The use of impregnated carbon is
discussed in section 10.6.1.2.4.) Figure 10-19 also shows that
the tested SDA/ESP systems with AC injection had a median mercury
removal efficiency of 85.9 percent, with a range from 74.5
percent to 90.9 percent removal.
Figure 10-19 suggests that FFs with AC injection ahead of
the FF have a median mercury removal efficiency that varies with
temperature and AC injection rate.63 It should be noted that the
higher temperature test data, at flue gas temperatures between
107° and 121° C (225° and 250° F) , had prior PM control while the
lower temperature test data, at flue gas temperatures between 88°
and 107° C (190° and 225° F) , were obtained with and without
prior PM control. The effect of the presence and absence of the
prior PM control on mercury removal efficiency was not apparent
in the lower temperature test data. With a low AC injection rate
(<1,000 wt C/wt inlet Hg) and an average flue gas temperature
between 107° and 121° C (225° and 250° F) , the median mercury
removal efficiency was 29 percent with a range from 14 to 47
percent removal. With a low AC injection rate (same as above)
and an average flue gas temperature between 88° C (190° F) and
107° C (225° F) , the median mercury removal efficiency was 97
percent with a range from 76 to 99 percent removal. A high AC
injection rate (>1,000 wt C/wt inlet Hg) and an average flue gas
temperature between 107° and 121° C (225° and 250° F) produced a
10-44
-------
g
•rl
4J
0
-------
median mercury removal efficiency of 81.5 percent with a range
from 69 to 91 percent removal. A high AC injection rate (same as
above) and an average flue gas temperature between 88° and 107° C
(190 and 225°F) produced a median mercury removal efficiency of
98 percent with a range from 95 to 99 percent removal.
Any strategy for add-on mercury control at utility plants
must consider adequately treating large volumes of gas in order
to remove relatively small concentrations of mercury, as well as
any resulting impacts on power plant equipment operations (such
as particulate control devices) and on waste disposal issues.
Later testing done on FF-equipped units using AC injection showed
much more variability and generally lower collection efficiencies
than those shown in Figure 10-19.68 Thus, while AC injection
shows promise as a mercury control technology, the limited
results to date are inconsistent and more data and research are
needed.
10.6.2.4 Emerging Technologies for Controlling Mercury
Emissions from Utilities. Research continues on developing
potential technologies for mercury emission reduction from
utility plants. 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.
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.59
With iodide-impregnated AC injection, the carbon-bound
iodide reacts with mercury to form mercuric iodide (HgI2) on the
carbon and the carbon is removed by a PM control device. In a
pilot-scale study, iodide-impregnated carbon increased mercury
removal to nearly 99 percent, an increase of nearly 45 percent
over results achieved with an equal amount of nonimpregnated
AC.59 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 (certain coal types at temperatures of
177° C [350° F] and higher), a portion of the captured mercury
10-46
-------
(postulated to be mercuric iodide) may be revolatilized as
oxidized mercury.69
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.67 Although the amount is small, chloride-impregnated
AC injection would introduce additional chlorine (a HAP) into the
flue gas stream.
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.66 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.70
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.71 There are several potential limitations to Na2S
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.72 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.
Another potential process for the reduction of mercury
emissions is the use of AC in a CFB.56 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
10-47
-------
FF, and recycled to the reactor. A small part of the used AC is
withdrawn from the process and replaced by fresh material.67
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.
As noted earlier, another potential technique, advanced
coal cleaning, may reduce the concentrations of mercury contained
in the mineral and organic phases of coal. The reliability and
feasibility of advanced coal cleaning techniques in reducing the
concentration of mercury in coal are unknown at this time, but
are being investigated.
Additional advanced processes for controlling mercury
emissions include the condensing heat exchanger, gold sorbent
technology, other sorbent injection processes, advanced fine
particulate control technologies, and enhancement of wet
scrubbers.73 However, data on these processes were not available
for this report; data will be incorporated into the final Report
to Congress as they become available.
10.7 DISCUSSION OF FEDERAL INTERAGENCY REVIEW COMMENTS
Previous drafts of chapters 1 through 10,.along with the
appendices, were reviewed by numerous non-EPA scientists
representing industry, environmental groups, academia, and other
Federal agencies during the summer of 1995. 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. The EPA has revised the report, as
appropriate, based on the reviewers' comments. The EPA revised
the report to incorporate the majority of the comments received.
However, there were several comments that could not be fully
addressed because of limitations in data, methods,, and resources.
This section presents comments received by other Federal agencies
that could not be substantially addressed in this interim report.
10.7.1 Comment
The DOE commented that the report "...understates the
possibilities for air toxics control through fuel switching and
coal cleaning." The DOE "...has found that coal cleaning holds
10-48
-------
significant promise for many forms of air quality control.
Efforts in 2010 should be significantly more successful in
emission reductions than indicated in the Report." 74
10.7,2 Response
There is no way to tell which plants will use advanced coal
cleaning technologies or which coals are going to be cleaned in
the future. The current report considers HAP reductions through
fuel switching. In addition, it was assumed that 100 percent of
bituminous coal is currently washed using conventional coal
cleaning techniques (even though only 90 percent was washed in
1993) . At the time the report was being developed, there were no
data available to characterize advanced coal cleaning. For
subbituminous and lignite coals, to the Agency's knowledge there
is currently no coal cleaning being performed and no projections
for cleaning these coals in the future. Therefore, no coal
washing factors were added to HAP concentrations for lignite and
subbituminous coals. Data on advanced coal cleaning will be
incorporated into the final Report to Congress as they become
available.
10.7.3 Comment
The Office of Science and Technology Policy (OSTP) commented
that w[t]he report does not adequately address the impact of
demand-side management strategies though a number of utilities
are pursuing this aggressively rather than building out
generation capacity. The report implies that investments in
alternative energy technologies may occur at the expense of
investments in more efficient coal-burning technologies (a
theoretical possibility but no data are presented). The report
maintains that biomass will not become a viable option until the
year 2030. With realistic investments in Research & Development
(R&D) in both crop production and generation techniques, biomass
can be competitive (without subsidies) within 5 years in niche
markets (with high energy prices) and within 10 years in larger
markets. [DOE and IPCC studies]." Also, OSTP adds that
"...control strategies dominate pollution avoidance strategies,
both in discussion and analysis." 7S
10.7.4 Response
Demand-side management's potential for reduction of HAPs is
mentioned in this chapter, based on the limited data available.
The EPA will review the data as they become available and provide
10-49
-------
additional discussion, as appropriate, of other alternative
control strategies in the final report.
10.7.5 Comment
The DOE and the Office of Management and Budget (OMB)
commented that the discussion of control options should include
cost estimates.76'77
10.7.6 Response
As noted in Chapter 1, the EPA has not included any cost
estimates in this report. However, as a part of the mercury
study, the EPA did evaluate general costs of mercury control.78
This analysis was based on a model-plant approach and
assumed that all plants within a source category would achieve
the same emission reduction and incur the same costs as the model
plants. The cost estimates derived, therefore, were only for
relative comparisons among source categories. The cost of
mercury control incurred by any specific facility may be
underestimated because of the variability inherent in the
assumptions that were made (e.g., mercury reduction efficiency of
the various control techniques; flue gas stream mercury content;
site-specific factors such as down-time and labor costs). Costs
for monitoring and record keeping were not included. The costs
represent retrofit application of controls; installation of
controls at new facilities can be significantly less expensive
than retrofitting. These costs also do not include the benefit
of co-control of other pollutants in addition to mercury. In
addition, the cost estimates represent analysis of relatively new
applications of mercury control technologies to utility units.
New or emerging control technologies will undoubtedly have lower
costs. (Cost analyses were not performed for such emerging
technologies as coal gasification or for other HAPs.)
Based on this model plant analysis, the EPA estimated the
national cost of mercury control (using activated carbon) to be
approximately $3 billion. The EPRI has estimated the cost to the
industry for mercury control to range from $1 to $10 billion.79
Should the EPA embark on a regulatory program, detailed cost
analyses of the various control options and strategies considered
would be undertaken."
10-50
-------
10.7.7 Comment
The OMB commented that it was "... concerned about the
experimental nature of the technologies for controlling mercury
emissions and the uncertainties associated with costs and
effectiveness of those technologies." 80
10.7.8 Response
The EPA shares this concern and would like to see the
technology tested on full-scale units to establish their actual
emissions reduction capacity and costs across a range of
operating conditions.
10.7.9 Comment
The OMB commented that "...the discussion of the
effectiveness of potential control strategies for HAPs should
include a full discussion of the strengths, limitations, and
uncertainties of the information provided for each option." 81
10.7.10 Response
The EPA will look in more depth at these strengths,
limitations, and uncertainties when more data on these
alternative controls become available. This material will be
presented in the final Report to Congress.
10.7.11 Comment
The DOE commented that the "EPA characterizes the mercury
removal effectiveness of the...FGD as poor. Based on...1992-1994
utility emissions characterization studies...this statement is
generally valid. However, recent pilot-scale testing has shown
relatively high removal of divalent mercury (Hg++) across a wet
scrubber...similar results were found in tests recently performed
at a utility boiler...due in part to developments in the methods
used to measure (speciate) mercury. It is also more directly the
result of a concerted effort to better understand the role of
conventional flue gas cleanup technology in controlling mercury
emissions. Continued research may show that existing FGD
technology can have a significant impact on mercury emissions,
particularly Hg++. This will become more relevant if additional
FGD i[s] installed in compliance with Phase II, and will have an
impact in terms of lowering EPA's mercury emission projections
through the year 2010."
10.7.12 Response
10-51
-------
.The EPA was not able to obtain the necessary data, or
conduct appropriate analyses and review of the data, from the
recent pilot-scale studies referred to by the DOE before
publication of this interim report. However, the EPA plans to
consider these data and, if appropriate, conduct additional
analyses of the data from the pilot-scale studies before
completing the final utility study Report to Congress.
10-52
-------
10.8 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.
10-53
-------
10. Baker, B.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.
11. Thambimuthu, K.V. Gas Cleaning for Advanced Coal-based
Power Generation. IEACR/53. IEA Coal Research, London.
March 1993.
12. 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.
13. 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.
14. 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.
15. 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.
16. 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.
17. 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 93-209-061-0.1. November
1993.
10-54
-------
18. .Battelle. Preliminary draft emissions report for Miles
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.
19. 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.
20. 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.
21. 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.
22. 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.
23. 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.
24. 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.
10-55
-------
25. 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.
26. 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.
27. 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.
28. 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.
29. 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.
30. Carnot. Preliminary draft emissions report for EPRI Site
115, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. Carnot report No.
EPRIE-10106/R022C855.T.
31. 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.
32. 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.
10-56
-------
33. 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.
34. 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.
35. 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.
36. 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.
37. 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.
38. Carnot. Preliminary draft emissions report for EPRI Site
112, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. Carnot report No.
EPRIE-10106/R016C374.T. March 1994.
39. Carnot. Preliminary draft emissions report for EPRI Site
118, Field Chemical Emissions Monitoring Project. Prepared
for Electric Power Research Institute. Report No.
EPRIE-10106/R140C928.T. January 1994.
40. 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.
10-57
-------
41. 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.
42. Goldfarb, T.D. Evidence for Post-Furnace formation of PCDDs
and PCDFs -- Implications for Control. Chemosphere.
18:1051-1055. 1989.
43. 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.
44. 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.
45. 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.
46. Memorandum from Cole, Jeffrey, RTI, to William Maxwell, EPA.
May 11, 1994. Emission factor output for Systems
Application International (SAI).
47. Department of Energy. National Energy Strategy, Powerful
Ideas for America. First Edition. Washington, DC.
February 1991. p. 44.
48. Ref. 47, p. 8.
49. Ref. 47, p. 39.
50. Ref. 47, p. 127.
51. Kilgroe, J. D. Cleaned Coal (Chapter 19). In: Handbook on
Air Pollution Control, John Wiley & Sons, Inc.. April 1983.
pp. 38-39.
52. 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.
53. Ref. 52, p. 8-23.
10-58
-------
54. 58 FR 42466. August 9, 1993. Final Regulatory
Determination on Four Large-Volume Wastes from the
Combustion of Coal by Electric Utility Power Plants.
55. 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.
56. 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.
57. 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.
58. 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.
59. Felsvang, Karsten, Rick Gleiser, Gary Juip, and Kirsten
Kragh Nielsen. Air Toxics Control Jby Spray Dryer Absorption
Systems. Second International Conference on Managing
Hazardous Air Pollutants, Washington, DC. July 1993.
60. 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 .
61. 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.
62. Letter from Boyce, P. L., Northern States Power Company, to
Martha Keating, EPA. January 19, 1994. Comments on the
draft report Afercury Control Technologies and Costing of
Activated Carbon Injection for the Electric Utility
Industry, prepared by RTI, September 1993.
10-59
-------
63. 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.
64. 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.
65. Peterson, J., D. Seeger, R. Skarupa, M. Stohs, B. Hargrove,
Mercury Removal by Wet Limestone FGD Systems: EPRI HSTC Test
Results. Radian, Corporation. Paper No. 94-RP114B.01.
AWMA annual meeting, 1994.
66. 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. 1991. p. 3-14.
67. 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.
68. Miller, S. J., D. L. Laudal, and G. E. Durham, Evaluation of
Activated Carbon for Control of Mercury from Coal-Fired
Boilers. Eleventh Annual Coal Preparation, Utilization and
Environmental Control Contractors Conference, Pittsburgh,
Pennsylvania. July 1995.
69. Ref. 68, p. 3.
70. Memo from Maxwell, William H., EPA, to Kenneth R. Durkee,
EPA. August 3, 1993. Second International Conference on
Managing Hazardous Air Pollutants.
71. Ref. 66. p. 3-1. (equations given in this reference were
balanced)
72. Ref. 66, pp. 3-6 and 3-7.
73. Letter and attachment from Chupka, Marc W., Department of
Energy, to Maxwell, William H., EPA:ESD. September 25,
1996. Comments on revised OMB review draft interim final
report.
10-60
-------
74. Letter from Chupka, Marc W., DOE, to Maxwell, William H.,
EPA/CG. May 3, 1996.
75. Memorandum and attachment from Gibbons, Jack, OSTP, to
Maxwell, William H., EPA/CG. May 6, 1996.
76. Letter from Chupka, Marc W., DOE, to Maxwell, William H.,
EPA/CG. May 3, 1996; Memorandum and attachment from Fraas,
Art, and Hillier, Troy, OMB, to Mcixwell, Bill, EPA/CG.
May 3, 1996.
77. Memorandum from Fraas, Art, and Hillier, Troy, Office of
Management and Budget, to Maxwell, Bill, EPA:ESD. September
24, 1996. Comments on EPA's interim utility report.
78. Mercury Study Report to Congress. Volume VII: An Evaluation
of Mercury Control Technologies and Costs. SAB Review
Draft. EPA-452/R-96-001g. June 1996.
79. Letter and enclosure from Peck, Stephen C., to Maxwell,
William H., EPA-.ESD. September 15, 1995. Transmittal of
unlicensed Electric Utility Trace Substances Synthesis
Report.
80. Memorandum and attachment from Fraas, Art, and Hillier,
Troy, OMB, to Maxwell, Bill, EPA/CG. May 3, 1996.
81. Memorandum and attachment from Fraas, Art, and Hillier,
Troy, OMB, to Maxwell, Bill, EPA/CG. May 3, 1996.
10-61
-------
11.0 PRELIMINARY OBSERVATIONS
The following observations are based on this study and the
currently available scientific data.
11.1 INDUSTRY GROWTH AND HAP EMISSIONS
1. Utility units emit a significant number of the 189 HAPs
included on the section 112(b) list, with coal-fired
units emitting the largest number of 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 21 percent (between
18 and 23 percent) of the U.S. anthropogenic emissions
of mercury.
2. Estimated growth in the number of utility units during
the period 1990-2010 is predicted to resultin an
overall increase in HAP emissions. Over this period,
utility coal consumption is estimated to increase by
approximately 32 percent, oil consumption is estimated
to decrease by approximately 50 percent; and natural
gas consumption is expected to increase by about 59
percent. Because coal- combustion is the primary source
of HAP emissions from utility units, its relative
increased usage accounts for the increase in HAP
emissions, even though much of the growth will be at
units subject to emission regulations. This results
from the fact that the emission controls used to comply
with these regulations are not 100 percent efficient in
removing HAPs. (The year 2010 was selected as the end
year for analysis because it is projected that all
measures to effect compliance with other provisions of
the Act will be completed by then.)
3. Anticipated future actions taken by the utility
industry to comply with other provisions of the Act
through installation of add-on controls are not
expected to significantly impact HAP emissions.
Similarly, actions taken to enhance existing controls
or to utilize emerging technologies may also impact HAP
emissions. Unknown actions that, for example, take the
form of pollution prevention or fuel switching may
significantly impact HAP emissions. Planned
installations of S02 scrubbers are running below
estimates made at the time of enactment of the 1990
Amendments to the Act, with compliance with Title IV
' being effected 'by switching from relatively higher to
-------
relatively lower sulfur coals, allowance trading,
switching from coal to oil or natural gas, etc. rather
than by installation of add-on controls. In addition,
these SO2 scrubbers do not exhibit significant control
over the majority of HAPs, including mercury. Upgrades
and modifications planned for PM control devices appear
to be related to accounting for changes in coal ash
properties when switching from relatively higher to
relatively lower sulfur coals, rather than to any
improvement in overall PM (and associated trace metal)
control efficiency. The limited data available do not
suggest that the various NOX controls being installed
to comply with Titles I and IV have any impact on HAP
emissions. However, the impact on HAP emissions of
advanced processes (e.g., the condensing heat
exchanger, gold sorbent technology, other sorbent
injection processes, advanced fine particulate control
technologies, enhanced wet scrubbers) cannot be
estimated at this time but depend on the extent to
which they are utilized by the industry. Switching
from one coal to another does not impact overall HAP
emissions (i.e., there are no "low-HAP" coals as there
are low-sulfur coals). The extent of the impact of
pollution prevention (e.g., repowering for efficiency,
DSM) on HAP emissions cannot be estimated but depends
on the extent to which these practices are utilized by
the industry. They would at least serve to reduce the
rate of increase in HAP emissions. Switching from coal
or oil to natural gas will effectively reduce HAP
emissions, but the extent of this practice over the
period 1990-2010 cannot be estimated.
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 had a hand in establishing these
protocols. The test sites covered the range of types
of facilities and configurations to the extent that the
EPA finds that the units tested are representative of
the industry.
11-2
-------
11.2 INHALATION RISK ASSESSMENT
5. The incremental increased cancer risk due to inhalation
exposure to HAP emissions (based on 1990 emission
levels and calculated in the baseline assessment) from
the "highest risk" coal-fired unit for the MEI for the
year 1990 is estimated to be approximately 4 in
1,000,000 (i.e., 4 x IP'6) . This does not include
risks due to either multipathway exposure or background
exposures. Arsenic and chromium VI are the HAPs that
contribute most to this cancer risk. The MEI
inhalation risk for arsenic emissions is estimated to
be 3 x 10"6, and the MEI inhalation risk for chromium
VI is estimated to be 2 x 10~6. The cancer risks from
other HAPs are estimated to be lower than those from
arsenic and chromium VI.
6. For the year 2010, the MEI cancer risks from coal-fired
utilities are not expected to be significantly
different. However, due to unknown actions and
uncertainties in the projections, the EPA has low
confidence in this observation.
7. The incremental increased cancer risk due to inhalation
exposure to HAP emissions from the "highest-risk" oil-
fired utility for the MEI for the year 1990 is
estimated to be between 3 in 100,000 (i.e., 3 x IP'5)
and 1 in 10,000 (1 x IP'4) . Nickel, arsenic, chromium,
and radionuclides are the HAPs contributing most to
this estimated increased risk.
8. For the year 2010, the MEI cancer risks from oil-fired
utilities are predicted to be approximately 30 to 50
percent lower than the 1990 estimates.
9. Based on the quantitative parameter uncertainty
analysis conducted for the inhalation risk assessment,
the EPA estimates that the baseline MEI risk estimates
due to inhalation exposure presented above are
reasonable, high-end estimates. The quantitative
variability and uncertainty of many 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") MEI inhalation risk estimate is
roughly 5 to 10 times lower than the MEI baseline risk
11-3
-------
estimates presented above. That is, these MEI risk
estimates are more likely to be overestimating rather
than underestimating the true risks due to inhalation
exposure for the MEI. However, there are limitations
to the uncertainty analysis (e.g., it did not consider
multipathway exposures) and limitations in available
data and the range of uncertainty is, most likely,
larger than estimated by this study.
10. 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)
and other uncertainties and variabilities, that were
not quantitatively assessed.
11. Based on the baseline assessment, 2 of the 426
coal-fired utilities are estimated to pose an MIR of
cancer greater than 1 x 10'6 due to inhalation exposure
to HAP emissions. The exposure model predicts that
approximately 2,477 people are exposed to air
concentrations from the two coal plants that result in
an estimated inhalation cancer risk of greater than 1 x
10'6. The other 424 coal-fired utilities are estimated
to present inhalation cancer risks less than 1 x 10~6.
12, The risk estimates from the "highest-risk" oil plant
(MEI cancer risk of I x 10'4 assuming all nickel is as
carcinogenic as nickel subsulfide) are estimated to be
3 to 4 times higher than the second "highest-risk" oil
plant. The MEI risk from the second "highest-risk"
oil-fired utility is estimated to be 3 x 10~5 (assuming
all nickel is as carcinogenic as nickel subsulfide).
The third "highest-risk" plant is estimated to present
a risk of 1 x 10"5. All remaining plants are estimated
to be below 10"5.
13. Using the assumption that the mix of nickel compounds
are as carcinogenic as nickel subsulfide, 20 oil-fired
utilities present an estimated MEI risk of greater than
10"6. If the mix of nickel compounds is assumed to be
20 percent as carcinogenic as nickel subsulfide, then
approximately four oil-fired utilities present an MEI
risk greater than 1 x 10"6.
11-4
-------
14. The population cancer incidence due to inhalation
exposure to HAP emissions from all coal-fired
utilities, considering only local (within 50 km of each
plant) impacts, is estimated to be 0.1 case per year.
15. The population cancer incidence due to inhalation
exposure to HAP emissions from all oil-fired utilities,
considering only local impacts, is estimated to be up
to 0.5 case per year.
16. The test-case evaluation of long-range transport of
particle-bound arsenic indicates that the population
inhalation exposures could increase the estimated
population cancer incidence by a factor of 7 (0.6 case
per year for all fuels) in comparison to the local
dispersion within 50 km of the source (less than 0.1
case per year. Thus, the cancer incidence may be
dominated by exposures on the regional, rather than on
the local, scale. Arsenic, as well as other
carcinogenic heavy metals of concern (e.g., chromium
and nickel), are mostly sorbed to fine particulate in
utility stack emissions, will experience long residence
times in the atmosphere, and are mostly subject to
long-range transport over large geographical areas.
17. With regard to noncancer effects, the highest HQ for
any HAP, considering both short- and long-term
exposures, is approximately 0.1 for HCl from coal-fired
utilities. This estimate represents the exposure due
to utilities only and does not consider background
exposures or long-range transport.
11.3 MERCURY
I
I
18. Mercury emitted by utility units has the potential to
be transported long distances. The modeling analysis
indicates that the deposition of mercury emitted from
such units is dominated on the regional rather than the
local scale. Predictions of the COMPDEP and RELMAP
models indicate that most of the mercury emitted by
utilities is transported further than 50 km from the
emission source."
19. Based on the modeling conducted for this study, along
with evidence from related studies, the EPA finds that
there is a plausible link between mercury emissions
from anthropogenic combustion and industrial sources,
11-5
-------
including utility boilers, and mercury concentrations
in air, soil, water, and sediments.
20. Based on the modeling conducted for this study, along
with evidence from related studies, the EPA finds that
there is a plausible link between methylmercury
concentrations in freshwater fish and mercury emissions
from utility units.
11.4 QUALITATIVE ASSESSMENT OF DIOXINS, ARSENIC, CADMIUM, AND LEAD
21. Dioxins, arsenic, cadmium, and lead are highly
persistent in the environment, have the potential for
bioaccumulation, and exhibit toxicity. These
characteristics make it likely that noninhalation
exposure pathways are a significant route of exposure
to these HAPs. Therefore, further evaluation of
multipathway exposure may be needed to more
comprehensively assess the risks.
22. Exposure to arsenic through noninhalation exposure
routes may pose additional risk, beyond the risk posed
by inhalation exposure to arsenic; however, no
quantitative assessment of noninhalation exposure to
arsenic emissions from utility units has been performed
at this time. Available data indicate that consumption
of fruits, grains, and vegetables; soil, ingestion; and
ingestion of milk may be significant routes of exposure
to arsenic. These noninhalation pathways are of
potential concern due to the emissions of arsenic by
utility units, because arsenic persists in the
environment, bioaccumulates, and can cause cancer
through the oral route of exposure.
11.5 RADIONUCLIDE ANALYSIS
23 . The highest MIR risk to any individual, within a 50-km
radius, resulting from multipathway exposure to
radionuclide emissions from utility units is estimated
to be 3 x IP'5, and 17 of the 684 plants have an MIR
greater than 10'5. Based on the limited analyses of
multiplant exposure, neither of these values would be
changed if the analysis had systematically addressed
multiplant exposure.
24. For the vast majority of the 196+ million persons
living within 50 km of any utility unit, the lifetime
fatal cancer risk is less than 1 x 10'6.
11-6
-------
I
I
25. 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 per
year. Had the exposures and risks due to overlapping
plumes been explicitly addressed, this value would not
increase by more than a factor of 3.
26. The quantitative uncertainty analysis performed for the
radionuclide analysis indicates that these risk
estimates are central values of the true probability
distribution.
11.6 ALTERNATIVE CONTROL STRATEGIES
27. There are a number of alternative control strategies
that are effective in controlling some of the HAPs
emitted from utility units. These strategies are
summarized below.
A. Conversion of coal- and oil-fired units to natural
gas firing effectively eliminates emission of
HAPs.
B. 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).
C. 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.
D. Current methods of coal cleaning are able to
remove portions of. the trace metals contained
within the coal. These emission reductions range
from approximately 20 percent for mercury to
approximately 50 percent for lead. Further
research is needed in methods of effecting greater
trace metal removals during coal cleaning.
E. 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.
11-7
-------
F. 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.
G. 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. Organic HAPs
are not well controlled by PM control devices.
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.
H. 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. Flue
gas desulfurization units (as operated on utility
units) generally effect good control (i.e.,
greater than 80 percent) of HCl, but control of HF
is not uniform.
I. Add-on technologies for the control of mercury
have not been demonstrated on utility units in the
U.S. 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.
J. Pollution prevention methods (e.g., 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.
11-8
-------
I
I
I
11.7 AREAS FOR FURTHER RESEARCH AND ANALYSES
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 are needed.
11.7.1 Emissions Data for Dioxins
Dioxins emissions data 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).
11.7.2 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 of the health effects associated
with those different forms would be of value for this study.
Further evaluation of chromium speciation is also needed.
11.7.3 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. Mercury, arsenic, dioxins, cadmium, and lead
are five HAPs identified as priority for further multipathway
assessment.
11.7.4 Long-Range Transport Exposures
Uncertainties remain regarding long-range transport of HAPs.
Further modeling and evaluation is needed to assess the impacts
of long-range transport of HAPs from utilities.
11.7.5 Mercury Issues
There are numerous areas regarding mercury that need further
research and evaluation, which are discussed in the draft mercury
study being reviewed by the Science Advisory Board (SAB). A few
areas relevant to the content of this interim report are the
following: (1) what percent of mercury emissions is elemental
versus divalent mercury; (2) how much mercury is removed during
coal cleaning; and (3) what control technologies or pollution
prevention options are available that could potentially reduce
mercury emissions, and what are the feasibilities and economic
impacts of such options.
11-9
-------
11.7,6 Projections to the Year 2010
There are significant uncertainties and unknowns in the
emissions and risk projections made to the year 2010. Further
research and evaluation in this area is needed.
11.7.7 Ecological Risks
Other than the ecological assessment for mercury, the
effects of HAPs on wildlife, endangered species, and terrestrial
and aquatic ecosystems were not evaluated. Although not mandated
by section 112(n)(1)(A), further evaluation of ecological risks
due to HAP emissions is needed to fully evaluate the impacts of
utility HAP emissions.
11.7.8 Criteria Pollutant and Acid Rain Programs
Further evaluation is needed to assess the impacts of the
Acid Rain and Criteria Pollutant programs on HAP emissions.
11.7.9 Short-Term Emissions
A limited assessment of. short-term exposures was completed.
However, further evaluation of short-term releases, especially
high-end, peak releases, is needed to fully assess the potential
impacts to public health due to emissions of HAPs (particularly
HC1 and HF) from utilities.
11.8 FEDERAL INTERAGENCY REVIEW COMMENTS
11.8.1 Comments
The Food and Drug Administration (FDA) commented that it was
concerned that "...this report reaches conclusions about the
possible public health impact of mercury emissions that are
inconsistent with actions involving the Mercury Study report."
The FDA also commented that "...it seems inappropriate to
conclude that mercury emissions do in fact pose a potential
concern for public health." x
The Office of Science Technology and Policy (OSTP)
recommended that "... the utility report...[should] be delayed
until completion of the SAB and stakeholder review [of the
Mercury Report] and revision." 2
The Department of Energy (DOE) commented that "...it is
premature to conclude that it is necessary 'to pursue an
evaluation of a range of options to minimize mercury emissions
from various sources, including utilities'..." The DOE also
commented that the language of the conclusions should be revised
to say that "... it is not appropriate or necessary to regulate
[HAPs] from [utilities] under section 112 at this time." The EPA
11-10
-------
should indicate that "...it will revisit this determination if
information to the contrary becomes available." 3
The Centers for Disease Control (CDC) commented "...that
mercury emissions should be minimized as part of an overall
strategy for reducing power-plant generated pollutants such as
sulfur dioxide and particulate matter." 4
The Council of Economic Advisors (CEA) commented that the
April 1996 draft report "...is not clear on the health related
benefits of reducing mercury emissions from utilities...the
relationship between reductions in utility generated mercury
emissions and deposition of mercury in the [U.S.] from
anthropogenic sources is not clear." 5
11.8.2 Response
Conclusions regarding the impacts to public health due to
mercury emissions have not been included in this interim report.
Also, policy decisions and regulatory decisions are not included
in this interim report. The EPA plans to include conclusions
regarding the impacts to public health, policy statements, and
regulatory determination in the final report, as appropriate.
Also, the EPA has not conducted a cost/benefit analysis for this
interim report.
11-11
-------
11.9 REFERENCES
1. Letter and enclosures from Lake, L.R., FDA, to Maxwell,
William H., EPA/CG. April 30, 1996.
2. Memorandum and attachment from Gibbons, Jack, OSTP, to
Maxwell, William H., EPA/CG. May 6, 1996.,
3. Letter from Chupka, Marc W., DOE, to Maxwell, William H.,
EPA/CG. May 3, 1996.
4. Letter from Mannino, David M., CDC, to Maxwell, William H. ,
EPA/CG. May 2, 1996.
5. Memorandum and attachment from Munnell, Alicia, CEA, to
Maxwell, William H., EPA/CG. May 6, 1996.
11-12
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
1. REPORT NO.
EPA-453/R-96-013a
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Study of Hazardous Air Pollutant Emissions from Electric
Utility Steam Generating Units — Interim Final Report
Volume 1.
5. REPORT DATE
October 1996
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)(l)(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 numerous HAPs (excluding
mercury); (4) an assessment of risks due to multipathway (inhalation plus non-inhalation) exposure to
radionuclides; and (5) a discussion of alternative control strategies. The assessment for mercury includes a
description of emissions, deposition estimates, control technologies, and a dispersion and fate modeling
assessment which includes predicted levels in various media based on modeling from four representative
utility plants using hypothetical scenarios. The EPA plans to publish a final report at a later date which
will include (1) a more complete assessment of the exposures, hazards, and risks; (2) conclusions, as
appropriate and feasible, regarding the significance of the risks and impacts to public health; and (3) a
determination as to whether regulation of utility HAPs is appropriate and necessary.
17
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c COSATI Field/Group
Air Pollution
Atmospheric Dispersion Modeling
Electric Utility Steam Generating Units
Hazardous Air Pollutants/Air Toxics
Air Pollution Control
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
384
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
-------
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590
------- |