-------
DRAFT EPA/600/6-88/005A
DO NOT QUOTE OR CITE March 1988
External Review Draft
ESTIMATING EXPOSURES TO 2,3,7,8-TCDD
NOTICE
THIS DOCUMENT IS AN EXTERNAL REVIEW DRAFT. It has not been formally released
by the U.S. Environmental Protection Agency and should not at this stage be
construed to represent Agency policy. It is being circulated for comment on its
technical accuracy and policy implications.
Exposure Assessment Group
Office of Health and Environmental Assessment
Office of Research and Development
Washington, DC 20460
-------
-------
CONTENTS
Page
Tables viii
Figures xii
Foreword xiii
Preface xiv
Document Development xv
1. EXECUTIVE SUMMARY 1
A. Introduction 2
B. Conclusions 5
^- C. Recommendations 8
^ PART ONE: AN UPDATE OF CURRENT KNOWLEDGE AND METHODS FOR
r PERFORMING EXPOSURE ASSESSMENTS FOR 2,3,7,8-TCDD 10
00
r^ 2. PHYSICAL/CHEMICAL PROPERTIES AND GENERAL EXPOSURE
PARAMETERS 11
A. Physical/Chemical Properties 11
B. Body Weights and Pulmonary Ventilation Rates in Exposure
Assessments 18
1. Body Weights 18
2. Pulmonary Ventilation Rates 20
3. FATE 24
A. Fate of 2,3,7,8-TCDD in Soil 25
1. Transient Behavior of TCDD Profile in Soil Layer 25
2. Transient Profile in Soil Column 27
3. Degradation of 2,3,7,8-TCDD in Soil 29
4. Biodegradation of 2,3,7,8-TCDD in Soil 30
B. Fate of 2,3,7,8-TCDD in Sediments 34
1. Aquatic Sediments 34
2. Sediment-to-Water Transport Process 35
3. Estimation of TCDD Concentration in Water Body 36
4. Example Calculation 38
in
-------
CONTENTS, continued
Page
C. Bioaccumulation of 2,3,7,8-TCDD in Fish and Cattle 39
1. Bioaccumulation in Fish 39
2. Bioaccumulation in Cattle 42
D. Plant Uptake 43
4. EXPOSURE 46
A. Inhalation — Indoor Dust Levels Versus Outdoor Levels 46
B. Inhalation — Vapors 48
1. Emission Potential SO
2. Dilution of Emissions in Ambient Air 54
3. Exposure Estimation 57
4. Effect of Photodegradation on Exposure Estimation 59
C. Inhalation -- Particulates 62
1. Vehicular Traffic 63
2. Loading and Unloading Operations 64
3. Spreading Operations 65
4. Transportation in Trucks 65
5. Wind Erosion 66
D. Dermal — Soil Contact Rates and Dermal Absorption 68
1. Contact Rates 68
2. Absorption 71
3. Summary 73
E. Ingestion — Soil 73
1. Available Studies 74
2. Evaluation 78
F. Ingestion ~ Beef and Dairy Products 80
G. Ingestion — Fish Consumption Data 83
1. Available Studies 83
2. Evaluation 88
IV
-------
CONTENTS, continued
Page
5. POST-EXPOSURE 90
A. Absorption from Environmental Matrices (Unavailability) 91
;
1. General Considerations 91
2. Review of Data on Unavailability 92
3. Summary 120
B. Pharmacokinetics and Body Burden of Dioxins 126
1. Body Burden: Estimate of Exposure 128
2. Physiologically Based Pharmacokinetic Modeling 130
3. Calculation of Daily Intake . 139
a. Data 139
b. Data Use 140
c. Findings 140
d. Conclusions Regarding Body Burden Data 141
e. Calculation, Assumptions, Uncertainties, and Actual
Parameter Values 142
f. Parameters Chosen 143
g. Daily Intakes Calculated 144
h. Impact of Daily Background on Risk 144
i. Recommendations for Future Activities 145
PART TWO: APPLICATION OF EXPOSURE ASSESSMENT METHODS IN EVALUATING
2,3,7,8-TCDD EXPOSURES FROM SELECTED SITUATIONS 149
6. USE OF METHODOLOGIES TO ESTIMATE EXPOSURE TO 2,3,7,8-TCDD 150
A. Description of the Exposure Scenarios for Contaminated
Soil and Landfills 151
B. Exposure Pathways 160
1. General 160
2. Exposure Factors Common to More than One Pathway 161
a. Degradation and Dilution 161
b. Sediment Dilution Factor 171
c. Body Weight 172
d. Lifetime 172
e. Pharmacokinetics 172
-------
CONTENTS, continued
3. Specific Factors by Pathway 173
a. Dust Inhalation from Wind Erosion 173
b. Dust Inhalation from Vehicular Traffic 183
c. Vapor Inhalation 184
d. Dermal Exposure 190
e. Soil Ingestion 191
f. Ingestion of Beef and Dairy Products 193
g. Ingestion of Fish 194
h. Water Ingestion—Surface Water 197
i. Ground Water Contamination 199
j., Fruit and Vegetable Ingestion 203
C. Description of Exposure Scenarios for Incineration 206
1. Summary of Incineration Scenarios 208
2. General Calculations and Factors Used 212
a. Summary of Emissions Data and Vapor/Particulate
Distribution 212
b. Municipal Waste Incierator Capacities in the U. S. 221
c. Air Dispersion Modeling 222
d. Surface Water Contamination 227
(1) Vapor Absorption 229
(2) Particulate Deposition 231
e. Land-Disposed Ash 234
f. Soil Contamination 237
D. Incinerator Exposure Pathway Calculations 241
1. Inhalation of Ambient Air 241
2. Ingestion of Contaminated Soil 247
3. Dermal Contact with Contaminated Soil 247
4. Ingestion of Contaminated Drinking Water 248
5. Ingestion of Contaminated Fish 248
6. Ingestion of Contaminated Beef and Dairy Products 249
7. Ingestion of Dairy Products Due to Particulate
Deposition on Fodder 250
E. Risks Associated with Total CDDs vs. 2,3,7,8-TCDD 253
VI
-------
CONTENTS, continued
7. UNCERTAINTY EVALUATION 257
A. Contaminated Soil and Landfills 259
1. Summary of Uncertainties 259
2. Uncertainties in Specific Methods Applied 262
a. Soil Dilution Factor 262
b. Sediment Dilution Factor 265
c. Degradation 267
d. Dust Inhalation 271
e. Vapor Inhalation 275
f. Dermal Exposure 278
g. Soil Ingestion 280
h. Beef and Milk Fat Ingestion 283
i. Fish Ingestion 287
j. Water Ingestion - Surface Water 291
k. Ground Water Contamination 296
1. Plant Uptake 297
B. Incineration Scenarios 297
1. Emissions Data 299
2. Selection of Model Incinerator and Exposure Scenario 300
3. Inhalation and Surface Deposition 300
4. Surface Water Contamination 304
5. Soil Contamination from Emissions 306
6. Dairy Product Exposure Following Deposition on Plants 309
7. Land-Disposed Ash 311
8. REFERENCES 315
APPENDIX. RISK ESTIMATES FOR CHAPTER 6 SCENARIOS 333
vn
-------
TABLES
Page
2-1 Properties of 2,3,7,8-TCDD 19
2-2 Estimated Minute Ventilation Associated with Activity Level
for Average Male Adult 23
4-1 Estimates of Soil Ingestion from Dermal Contact 76
4-2 Rates of Ingestion of Beef and Dairy Products 82
5-1 Guinea Pigs Receiving a Single Gavage Dose of Materials
Containing 2,3,7,8-TCDD 93
5-2 Toxicity of TCDD Contaminated Soil 98
5-3 Comparison of Mortality in Guinea Pigs Following a Single
Dose of Contaminated Soils 103
5-4 Guinea Pig Mortality and Weight Changes Following Treatment
with Contaminated Soot or 2,3,7,8-TCDD 106
5-5 Concentrations of PCDD and PCDF Isomer-Groups in Fly Ash
Extracts (Diluted with Acetone) and Fly Ash 110
5-6 Concentrations of PCDD and PCDF Isomer Groups in Livers
of Rats Fed Fly Ash Materials 111
5-7 Percentage of Cumulative Dose of Dioxins and Furans Present
in Rat Liver Following 19 Days Exposure in Diet 112
5-8 Percentage of Tritium-Labeled 2,3,7,8-TCDD in Rat Liver
Following Administration of 14.7 mg Dose in Ethanol 116
5-9 Percentage of Tritiated 2,3,7,8-TCDD Dose in the Liver
24 Hours After Oral Administration of 0.5 ml, of Various Media 117
5-10 Gut Absorption of 2,3,7,8-TCDD in Rabbits after 7-Day
Treatment 119
5-11 Summary of Data on the Bioavailability of 2,3,7,8-TCDD
Following Ingestion of Environmental Matrices 122
5-12 Animal vs Human Clearance and Half-Lives of TCDD 138
5-13 Calculated Average Daily Intake 144
5-14 Risks Associated with Background Daily Intake of 2,3,7,8-TCDD
Compared with Annual Cancer Incidence in U.S. Population 146
Vlll
-------
TABLES, continued
6-1 Assumptions for Contaminated Soil Scenarios 158
6-2 Assumptions for Landfill Scenarios 159
6-3 Erosion Parameters 167
6-4 Dilution Factors 170
6-5 Factors Used in Exposure Calculations 174
6-6 Exposure Levels Associated with Various Exposure
Pathways/Scenarios - Contaminated Soil . 176
6-7 Simulated Concentrations at Wells 202
6-8 Parameter Values for Incinerator Facilities in Scenarios 16-19 209
6-9 Parameter Values for Populations in Scenarios 16-19 210
6-10 Stack Emission of PCDD's from Municipal Incinerators 214
6-11 PCDD's in Fly Ash of Combustion 216
6-12 Emissions of CDD's from Combustion Devices Other Than MWI 218
6-13 Air Dispersion Modeling of Particulate-Form 2,3,7,8-TCDD
Emissions 224
6-14 Summary of Ambient Air Concentration and Deposition Rate
in the Vicinity of the Facilities in Scenarios 16-19 ,^ 228
6-15 Comparison of Particle Size in Settling Velocity 229
6-16 Exposures Associated with Incinerator Exposure Pathway/
Scenarios A. Exposures Associated with Stack Emissions
and B. Exposures Associated with Fly Ash Disposal 242
6-17 Parameter Values for Calculating Exposures Associated with
Incinerators 244
6-18 Ambient Air Concentrations and Exposures at 0.8 km 246
6-19 Inhalation Exposures at 200 m and 100 km 246
6-20 Exposure from Ingestion of Contaminated Soil 247
6-21 Exposure from Dermal Contact of Contaminated Soil 247
6-22 Exposure from Ingestion of Contaminated Drinking Water 248
IX
-------
TABLES, continued
6-23 Exposure from Ingestion of Contaminated Fish 249
6-24 Exposure from Ingestion of Contaminated Beef and
Dairy Products 250
6-25 Parameter Values Needed in Equation 6-53 252
6-26 Daily Exposure from Ingestion of Milk Resulting from
Particulate Deposition on Fodder 253
6-27 Weight Percent Distribution of CDD Congener Equivalents in
MWI Stack Emissions 254
6-28 Weight Percent Distribution of CDD Congener Equivalents in
MWI Fly Ash 256
7-1 Landfill Assessment - Soil Dilution Factor 266
7-2 Landfill Assessment - Sediment Dilution Factor - Ponds 268
7-3 Landfill Assessment - Sediment Dilution Factor - Streams 269
7-4 Landfill Assessment - Degradation 272
7-5 Landfill Assessment - Dust Inhalation 276
i
7-6 Landfill Assessment - Vapor Inhalation 279
7-7 Landfill Assessment - Dermal Contact 281
7-8 Landfill Assessment - Soil Ingestion 284
7-9 Landfill Assessment - Beef and Milk Ingestion 288
7-10 Landfill Assessment - Fish Ingestion 292
7-11 Landfill Assessment - Surface Water Contamination 295
7-12 Landfill Assessment - Ground Water Contamination 298
7-13 Incinerator Assessment - Air Emissions Estimate 302
7-14 Incinerator Assessment - Inhalation and Surface Deposition 303
7-15 Incinerator Assessment - Surface Water Contamination by
Incinerator Stack Emissions 307
-------
TABLES, continued
7-16 Incinerator Assessment - Soil Contamination Levels
Resulting from Incinerator Emissions 310
7-17 Land-Disposed Ash 314
A-l Upper-Bound Incremental Cancer Risks Associated
With Various Exposure Pathways/Scenarios: Contaminated
Soil 336
A-2 Upper-Bound Incremental Cancer Risks Associated
With Incinerator Exposure Pathways/Scenarios: A. Risks
Associated with Stack Emissions. B. Risks Associated
with Fly Ash Disposal 339
XI
-------
FIGURES
6-1 Landfill Scenario 157
6-2 Ambient Air Concentration and Exposure with Distance
for 1 Acre Site . 187
6-3 Ambient Air Concentration and Exposure with Distance
for 10 Acre Site 187
6-4 Incinerator Scenarios 211
XII
-------
FOREWORD
The Exposure Assessment Group (EAG) of EPA's Office of Health and Environmental
Assessment has three main functions: (1) to conduct exposure assessments, (2) to review
assessments and related documents, and (3) to develop guidelines for Agency exposure
assessments. The activities under each of these functions are supported by and respond
to the needs of the various EPA program offices. In relation to the third function, EAG
sponsors projects aimed at developing or refining techniques used in exposure
assessments.
2,3,7,8-TCDD problems first surfaced in the United States in the early 1970s with
Agent Orange and the Missouri Horse Arenas. Since then, 2,3,7,8-TCDD contamination
has been found elsewhere in Missouri, Arkansas, Michigan, New York, and New Jersey.
The EPA has become increasingly involved in the discovery, assessment, and cleanup of
these sites. A previous EAG document (Schaum, 1984) featured the use of nomographs to
provide quick and appropriate estimates of risks for five exposure pathways. For each
pathway, factors such as contact rates, absorption fractions, and exposure duration were
developed, together with equations to be used in calculating exposure levels. The
purpose of this document is to provide the most recent exposure and risk estimation
methodology for application to 2,3,7,8-TCDD-contaminated sites. This methodology will
help us set priorities and make decisions required to address this important problem.
Michael A. Callahan
Director
Exposure Assessment Group
Xlll
-------
PREFACE
The Exposure Assessment Group of the Office of Health and Environmental
Assessment has prepared this exposure assessment document at the request of the Office
of Solid Waste. This document presents an update of previous work and an analysis of
key issues related primarily to the assessment of exposure of 2,3,7,8-TCDD. Estimates of
exposure and risk for a number of exposure pathways are presented. Current thinking in
the area of bioavailability and pharinacokinetic modeling are presented. The literature
search supporting this document is current to March 1988.
xiv
-------
DOCUMENT DEVELOPMENT
Richard V. Moraski, Ph.D.
Document Manager
Exposure Assessment Group
U.S. Environmental Protection Agency
Washington, D.C. 20460
Chapter Titles and Authors
1 EXECUTIVE SUMMARY - Michael A. Callahan; Richard V. Moraski, Ph.D.;
and Charles H. Nauman, Ph.D.
2 PHYSICAL/CHEMICAL PROPERTIES AND GENERAL EXPOSURE PARAMETERS
Gregory Kew, Ph.D. and Richard Walentowicz
3 FATE - Seong T. Hwang, Ph.D.; Charles H. Nauman, Ph.D.;
John L. Schaum, P.E.; and Gregory Kew, Ph.D.
4 EXPOSURE - Seong T. Hwang, Ph.D.; Gregory Kew, Ph.D.;
Richard Walentowicz; and Paul White
5 POST-EXPOSURE - Jerry N. Blancato, Ph.D. and Paul White
6 USE OF METHODOLOGIES TO ESTIMATE EXPOSURE TO 2,3,7,8-TCDD
Michael A. Callahan; Seong T. Hwang, Ph.D.;
John J. Segna, P.E.; Gregory Kew, Ph.D.; and
John L. Schaum, P.E.
7 UNCERTAINTY EVALUATION - Paul White
Reviewers
Donald Barnes
Office of Pesticides and Toxic Substances
U.S. Environmental Protection Agency
Washington, DC 20460
Judith S. Bellin
Risk Assessment Forum
U.S. Environmental Protection Agency
Washington, DC 20460
xv
-------
Philip M. Cook
Hazardous Waste Branch
Environmental Research Laboratory
U.S. Environmental Protection Agency
Duluth, MN 55804
William Ellis
Science Applications International Corp.
8400 West Park Drive
McClean, VA 22102
James W. Falco
Director
Environmental Monitoring Systems Laboratory
Research Triangle Park, NC 27711
Michael Firestone
Exposure Assessment Branch
Hazard Evaluation Division
Office of Pesticide Programs
U.S. Environmental Protection Agency
Washington, DC 20460
George Fries
Pesticides Degradation Laboratory, Bldg. 050
U.S. Department of Agriculture
Beltsville, MD 20205
Michael Gallo
Department of Environmental and Community Medicine
UMDNJ-Robert Wood Johnson Medical School
675 Hoes Lane
Piscataway, NJ 08854-5635
Lester D. Grant
Director
Environmental Criteria and Assessment Office
U.S. Environmental Agency
Research Triangle Park, NC 27711
Karen Hammerstrom
Exposure Assessment Branch
Exposure Evaluation Division
Office of Toxics Substances
U. S. Environmental Protection Agency
Washington, DC 20460
xvi
-------
Renata D. Kimbrough
Office of Regional Operations
Office of the Administrator
U.S. Environmental Protection Agency
Washington, DC 20460
Steven D. Lutkenhoff
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Cincinnati, OH 45268
Debdas Mukerjee
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Cincinnati, OH 45268
Jerry Schroy
Monsanto Company
800 N. Lindbergh Boulevard
St. Louis, MO 63166
Marcia Williams
Director
Office of Solid Waste
U.S. Environmental Protection Agency
Washington, DC 20460
Howard Zar
Chairman, Dioxin Task Force
Region 5, U.S. Environmental Protection Agency
230 South Dearborn Street
Chicago, IL 60604
xvn
-------
1. EXECUTIVE SUMMARY
This report was prepared by scientists and engineers from the Exposure Assessment
Group, Office of Health and Environmental Assessment. The primary purpose of the
report is to provide a review and update of information related to exposure to
2,3,7,8-TCDD that has come to light since 1984. In addition, this report provides an
illustration of the application of this information in performing exposure assessments for
2,3,7,8-TCDD. This is accomplished by using the information to construct several
scenarios where contaminated material may result in exposure to 2,3,7,8-TCDD, and
estimating what the exposure (and risk in the Appendix) would be for various pathways
from source to humans exposed. Sources used as examples in this report include
contaminated soil, various land disposal situations, and municipal waste incinerators. It
must be emphasized that these scenarios are not to be interpreted as an exposure/risk
assessment for all sources of these types. The assumptions used to construct the
scenarios may be quite different from the situations encountered in a specific assessment
in an actual case. It should be emphasized that when assessing actual sites, monitoring
data may be available, or measurements may be made, that would preclude the necessity
of part or all of the estimation methods described here. Good monitoring or
measurement data will add reliability to the appropriate parts of the assessment. In
many cases, however, where measured data are unavailable and estimation techniques are
required, the assumptions used in certain scenarios in this report may be reasonable
approximations of the actual situations being assessed. For those cases, the scenarios
provided as examples here may provide valuable insight into the general magnitude of
exposures to be expected from similar sources. It is hoped that this report will provide
a sound starting point for many exposure assessments of 2,3,7,8-TCDD contamination.
The scope of this report is outlined in the Introduction, followed by the major
conclusions and recommendations.
1
-------
A. INTRODUCTION
2,3,7,8-Tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) is a substance that is of major
concern to public health because of its extreme toxicity. In the family of chemicals
known as chlorinated dibenzo-p-dioxins (CDDs), 2,3,7,8-TCDD has been shown to be the
most toxic compound to animals that has been isolated and tested. Humans and animals
exposed to 2,3,7,8-TCDD have shown acute, subchronic, and chronic effects.
2,3,7,8-TCDD adversely affects the skin, the liver, the nervous system, and the immune
r
system of humans and animals. As evidenced by its high cancer potency slope factor
(Q!*), 2,3,7,8-TCDD is quite potent compared to other known carcinogens, and is
classified a probable carcinogen in humans on the basis of animal carcinogenicity studies
which were positive in multiple species and organs.
Although CDDs have never been intentionally produced as an industrial product for
wide distribution, they are generated as a by-product by many different sources (U.S.
EPA, 1980a, 1986d), including: processes for production of chlorinated phenol; processes
for production of chemicals using chlorophenols; combustion processes in industrial and
municipal incinerators; automobile exhaust; processes resulting in wastes from
pentachlorophenol wood treating operations; and processes involved in production of
bleached kraft paper products.
Additionally, CDDs have been detected in human mothers' milk (Gross, 1982), and it
has been suggested that CDDs may be formed in forest fires (Sheffield, 1985). It is
beyond the scope of this paper to address the ways in which all of these various sources
contribute to human exposure. Many of these sources of CDDs can result in air, water,
and soil contamination; some of the exposure pathways that might be associated with
some of the sources are addressed in detail in Chapter 6.
-------
It has been concluded by the Centers for Disease Control that 2,3,7,8-TCDD soji
concentrations above 1 ppb in residential areas raise concerns about health risks
(Kimbrough et al., 1984).
Human exposure is likely to result from ingestion of contaminated fish, beef, dairy
products, and other foods; ingestion of contaminated soil, especially by children with pica
tendencies; from dermal contact with 2,3,7,8-TCDD-contaminated soil, dust, and sediment;
and from inhalation of contaminated dust and 2,3,7,8-TCDD vapors.
This report is divided into two parts. Part One' (Chapters 2 to 5) presents an
update of previous work and an analysis of key issues related to exposure assessment for
CDDs, with emphasis on 2,3,7,8-TCDD. The updated information builds upon exposure
assessment methods and general concepts developed as part of the Dioxin Strategy (U.S.
EPA, 1983) and published by the Exposure Assessment Group in November of 1984
(Schaum, 1984). The 1984 document includes standard factors, assumptions, computation
methods, and nomographs for estimating exposure under various exposure scenarios
involving 2,3,7,8-TCDD-contaminated soil. Schaum (1984) addressed five pathways of
exposure: dust inhalation, fish ingestion, dermal absorption from soil contact, soil
ingestion, and ingestion of beef and dairy products. In the interim, results of new
studies have provided information and an understanding of scientific issues which now
enable us to make more informed judgments regarding qualitative and quantitative inputs
to exposure assessments for 2,3,7,8-TCDD and related compounds, including the ability to
make estimates of exposure through a number of additional exposure pathways. This new
information has enabled us to construct more realistic scenarios, narrow ranges of
estimates for parameter input values, broaden the scope of assessments by including
municipal waste incinerators and disposal of fly ash generated from the incinerators, and
thus to lessen the degree of uncertainty in resulting exposure assessments.
-------
Several of the items discussed in this update report were originally identified in the
1984 document. An expanded and updated analysis is provided relative to the physical
and chemical properties of 2,3,7,8-TCDD, the behavior of 2,3,7,8-TCDD in soil and
sediment, the concentrations of 2,3,7,8-TCDD in indoor versus outdoor dust and soil, and
inclusion of additional pathways. Also discussed are biologically related issues such as
bioavailability from soil; body weight and respiration rates; absorption rates of soil-sorbed
2,3,7,8-TCDD through human lung, gut, and skin; consumption rates for fish, beef, and
dairy products; soil ingestion rates; and bioaccumulation in fish, beef, and dairy products.
Methods for considering exposure due to inhalation of vapor-phase 2,3,7,8-TCDD are
presented; exposures due to inhalation of dust from vehicular traffic, and loading and
unloading of ash are considered; food chain contamination by direct and indirect routes is
incorporated; and the analysis of pharmacokinetics for estimating exposure to
2,3,7,8-TCDD is discussed. Available information on the plant uptake of 2,3,7,8-TCDD is
incorporated in estimating human exposures to make the inclusion of pertinent exposure
pathways as complete as possible.
Part Two (Chapters 6 and 7) of this report shows how the refinements discussed in
Part One can be put to use in a new set of calculated exposure scenarios. Twenty
scenarios are calculated, each one considering many different exposure pathways, and
these exposures are converted into risk estimates. For scenarios 1-15, which cover
contaminated soil, open dumps, abandoned dumps, and a capped landfill, contamination
levels of 2,3,7,8-TCDD of one part per trillion show relatively low exposures in most
pathways of these scenarios. Future developments in the areas of bioavailability and
pharmacokinetic modeling may substantially improve our confidence in exposure
assessment for 2,3,7,8-TCDD.
As part of assessing exposure associated with releases of the emissions from
combustion devices and the disposal of fly ash collected in control equipment, two
-------
municipal waste incinerators with different capacities and levels of control are considered
for evaluation. The exposure assessment scenarios presented here (scenarios 16-20) differ
from the traditional approach in that exposure from multimedia pathways is evaluated in
addition to the direct inhalation pathway, and exposure associated with disposal of fly
ash is considered.
The risk numbers calculated in the Appendix relate to the individual scenarios
chosen, and the reader is again cautioned that application of these numbers to specific
(actual) sites is not appropriate without a careful examination of the circumstances that
exist at the site relative to the scenario assumptions.
B. CONCLUSIONS
The authors are in consensus on the following conclusions. Where risk is discussed,
it refers to the upper bound incremental lifetime cancer risk only.
(1) For the scenarios considered in the report (contaminated soil, dumps, municipal
waste incinerators) where 2,3,7,8-TCDD was openly available to the
environment (i.e., not capped in a landfill), the highest exposures and
consequent risks were associated with the food chain (e.g., plants, beef, fish,
dairy products). The apparent reasons for this include the tendency of
2,3,7,8-TCDD to partition and accumulate in organic substances, the higher
access humans have to contamination through these pathways, and the
activities of the populations which may make it more likely to be exposed
through these pathways than through other pathways such as soil ingestion,
dust inhalation, and dermal contact with soil.
(2) Reasonable worst case scenarios illustrate that in the absence of any controls
on disposal of 2,3,7,8-TCDD-contaminated material at the 1 ppb level, exposure
>j
may result in increased cancer risks as high as 10"^ as shown in the Appendix.
However, careful handling and disposal of contaminated material in both the
-------
landfill and incinerator ash scenarios (e.g., capping properly designed landfills
or monofills, proper incinerator design, runoff control, etc.) results in
predicted exposures correlated to substantially reduced risks (10~^ or below).
Exposures calculated for contaminated materials of 1 ppt or below in land-
related scenarios were low in all cases regardless of controls, except for the
reasonable worst case soil contamination scenarios, where the pathway with the
highest exposure was correlated with a risk in the range of 10"^.
(3) For the two incinerator capacities considered, poor operating practice and
improper control of stack emissions resulted in higher exposures and risks with
the smaller capacity unit. In all scenarios dealing with stack emissions,
highest exposures and risks were associated with indirect routes of exposure
(e.g., dairy products, beef) in contrast to the direct inhalation pathway. The
apparent reasons for this would be similar to the ones described in earlier
conclusions coupled with the deposition of particulate and vapor 2,3,7,8-TCDD
on land, water and fodder over the lifetime of incinerator operation.
(4) Fly ash disposal scenarios illustrate that in the absence of containment
measures such as landfill cap, exposure may be significant especially for the
population that derives some of its food from a farm operation located in the
vicinity of the disposal site (see Appendix for risk estimates).
;
(5) Recent literature is divided and seemingly contradictory on the issue of
whether, and how much, 2,3,7,8-TCDD is taken up into plants from
contaminated soil. The authors of this report conclude that there is evidence
that 2,3,7,8-TCDD is taken up by plants growing in contaminated soils, but the
amount taken up, or subsequent transport within the plant itself (say, to edible
portions) is very uncertain. The worst-case calculations (using the highest
plant-to-soil ratio from the literature) result in very high exposures, at least
-------
as high as all other pathways. On the other hand, using other values from the
literature would result in exposures of little concern.
(6) The properties of 2,3,7,8-TCDD make widespread ground water contamination
from landfills unlikely, provided uncontaminated water filters through the
capped landfill. Preliminary calculations of the effects of codisposed solvents
indicate a slight increase in solubility of 2,3,7,8-TCDD with one-phase solutions
(i.e., saturated solutions or moderate mixtures of miscible solvents with water),
but the effect of this solubility increase on mobility has not been fully
investigated. For systems where two distinct liquid phases exist (water and a
relatively nonpolar organic solvent), the authors believe much greater mobility
of 2,3,7,8-TCDD is possible. For these cases, and in cases where physical
transport of soil particles to ground water can occur, there may be an
associated threat to ground water.
(7) The weight of evidence indicates that 2,3,7,8-TCDD is often bioavailable from
contaminated soils, although certain soils may bind 2,3,7,8-TCDD very tightly,
decreasing the bioavailability by an order of magnitude or more. The reasons
for this difference in bioavailability from one material to another is not well
understood at this time. The data base upon which this conclusion is drawn is
very slim. The implications of this conclusion are that after additional data
are collected, sufficient to draw more firm conclusions, bioavailability may be
an important factor in site-specific assessments.
(8) Pharmacokinetics have been considered in order to calculate, from body burden
data, "background" daily intake levels in the U.S. population. " While the data
do not allow an estimation of an average body burden in the U.S. population
from which to calculate an average daily intake, an upper limit of 6.72 ppt in
the adipose tissue has been estimated. From this upper bound estimate of
-------
body burden, the upper bound daily intake ranges from 0.04 pg/kg to 0.51
pg/kg. The upper limit of risk which results from such estimates is then
compared with the annual cancer incidence in the U.S. Further aims, future
research and application goals regarding physiologically based pharmacokinetic
models are discussed.
(9) The authors believe that a significant amount of uncertainty in the exposure
assessment could be reduced by a focused, limited research program addressing
the areas where critical information is needed. These areas are outlined in
the section for recommendations, below.
C. RECOMMENDATIONS
The authors believe that, should additional work be necessary to reduce the
uncertainty in exposure estimates for 2,3,7,8-TCDD, the following areas are important.
They are listed in priority order (highest first) in the ability to address major areas of
uncertainty having an impact on the overall assessment.
(1) Because of the uncertainty of many orders of magnitude, and the possibility of
the exposures being significant, it is the authors1 consensus that the area of
plant uptake and transport of 2,3,7,8-TCDD within the plant, especially to
edible portions for both humans and animals, is a major area needing further
work.
(2) In addition to being an aid for the calculation of background levels,
pharmacokinetics may also assist in determining background levels in soils,
target organ dose, absorbed dose, lactational/placental transfers, and effects on
offspring. For this reason, the authors feel that pharmacokinetics is an
important area for future work.
(3) The bioavailability data base is very slim, yet for site-specific situations,
bioavailability may be a very important issue. The authors recommend
8
-------
additional work to determine whether a simple test for bioavailability can be
developed, and if so, to develop guidance on how the results of such a test
could be interpreted for exposure assessments.
(4) The potential effect of solvents on 2,3,7,8-TCDD mobility in soils is another
area where the authors feel that additional work may help fill a significant
gap in the assessment.
(5) The authors feel that additional work in the areas of erosion factors,
distribution factors for fish, degradation rates of 2,3,7,8-TCDD in various
media, and in understanding the vapor/particulate behavior of 2,3,7,8-TCDD in
stack emissions would all lead to significant reductions of uncertainty in
important parts of the assessment.
-------
PART ONE
AN UPDATE OF CURRENT KNOWLEDGE AND METHODS FOR PERFORMING
EXPOSURE ASSESSMENTS FOR 2,3,7,8-TCDD
10
-------
2. PHYSICAL/CHEMICAL PROPERTIES AND GENERAL EXPOSURE PARAMETERS
In the past few years, there have been several refinements in what is known about
general exposure parameters (e.g., pulmonary ventilation rates and human body weights as
a function of age and sex) and also in what is known of the properties of 2,3,7,8-TCDD.
Most of these refinements have little effect on exposure assessments, but they add to
our confidence in such assessments by making the input data more precise. Updated
input data include revised Henry's law constant which would be a factor in predicting the
volatilization potential of 2,3,7,8-TCDD, and recent work on increased solubility of
2,3,7,8-TCDD when other organic materials are present. In the past the volatilization
potential of high molecular weight organics such as 2,3,7,8-TCDD had been neglected by
many exposure assessors. This chapter will discuss the refinements to our knowledge in
the areas of general exposure parameters and physical and chemical properties including
solubility enhancement. Implications of the increased importance of volatilization from
soil and water are developed more fully in Chapter 4.
A. PHYSICAL/CHEMICAL PROPERTIES
Knowledge of the basic physical and chemical properties of 2,3,7,8-TCDD is essential
to understanding or modeling its environmental transport and fate as well as its
pharmacokinetic or toxicologic behavior. The most essential parameters appear to be
vapor pressure (Pv), octanol/water partition coefficient (Kow), and water solubility (Sw).
Numerous other important but less frequently investigated parameters are available by
derivation or through published correlations, e.g., to soil or sediment partition
coefficients and bioconcentration factors (Lyman et al., 1982, Chapters 4 and 5
respectively). The ratio of Pv to Sw (PV/SW) yields Henry's law constant (Hc) for
low-solubility organic compounds, an index of partitioning for a compound between the
atmosphere and the water phase (Mackay et al., 1982).
11
-------
Brief summaries of recent scientific articles on the physical-chemical properties of
2,3,7,8-TCDD and some related compounds are provided in the following paragraphs.
These have been chosen based on credibility of experimental methods and results. No
discussion relative to earlier findings is included here, although such discussion usually is
offered in the original papers and in reviews such as Webster et al. (1985), Schroy et al.
(1985a), and Mackay et al. (1985). [The most recent comprehensive list of the physical
properties of 2,3,7,8-TCDD is found in Schroy et al. (1985a).]
Podoll et al. (1986) recently measured the vapor pressure of 2,3,7,8-TCDD using
l^C-labeled 2,3,7,8-TCDD and a gas saturation technique followed by combustion to
^CO2- The mean value and standard error of five determinations were 7.4 ± 0.4 x
10"'0 mm Hg at 25°C. Henry's law constant was then calculated as 12 mm Hg M~^ (or
1.6 x 10"5 atm m^ mol"1), using a water solubility reported by Marple et al. (1986a).
Based on this Henry's law constant, Lyman et al. (1982) offers guidelines, though not
specific to 2,3,7,8-TCDD, to compare organic compounds that may or may not be
volatilized from water at a significant amount, and provides the ranges of Henry's law
constant at which volatilization represents "significant transfer mechanisms" and at which
volatilization would be insignificant. The compound listed as having the potential for
volatilization include polycyclic aromatic hydrocarbons and other halogenated aromatics
such as PCBs. Further discussion can be found in Section B of Chapter 4.
Marple et al. (1986b) reported the octanol/water partition coefficient of
2,3,7,8-TCDD as (4.24 ± 2.73) x 106 at 22 ± 1°C. Two similar experimental techniques
were used, but the more expeditious and reliable one involved equilibration of octanol
presaturated with water and containing the 2,3,7,8-TCDD, with water, presaturated with
octanol, over 6 to 31 days.
Burkhard and Kuehl (1986) used reverse-phase HPLC and LRMS detection to
determine octanol/water partition coefficients for 2,3,7,8-TCDD and a series of seven
12
-------
other tetrachlorinated planar molecules, including three more TCDD isomers and
2,3,7,8-tetrachlorodibenzofuran (TCDF). Log Kow = 7.02 ± 0.50 was reported for
2,3,7,8-TCDD. These authors also reevaluated data on 13 chlorinated dioxins and
dibenzofurans previously obtained by Sarna et al. (1984) by very similar experimental
techniques. In the reevaluation, Burkhard and Kuehl (1986) used experimental rather
than estimated log Kow values in correlations with gas chromatographic retention time.
This approach yielded log octanol-water partition coefficients ranging from about 4.0 for
the non-chlorinated parent molecules to about 8.6 for the octa-chlorinated compounds,
much lower than previously reported. Coefficients in this range usually mean that the
substance tends to adsorb strongly to organic components in the soil.
The water solubility of 2,3,7,8-TCDD recently was reported as 19.3 ± 3.7 parts per
trillion at 22°C (or 5.99 ± 1.15 x 10'11 M) by Marple et al. (1986a) after equilibrating
thin films of resublimed 2,3,7,8-TCDD with a small volume of water followed by gas
chromatography (GC) analysis with "•'Ni electron capture detection. A corresponding
experiment using radiolabeled 2,3,7,8-TCDD and GC plus scintillation counting was found
to be less reliable due to only 80% purity of radio-labeled versus 98% purity of
non-radiolabeled material, low scintillation count rates (three to four times background)
and unexplained losses after equilibration.
The very low solubility of 2,3,7,8-TCDD would imply a low likelihood of leaching to
ground water from the soil surface if pure water were the only transporting medium.
However, some hazardous waste landfill leachates may be viewed as mixed solvent
systems, with an aqueous phase containing a variety of man-made organic chemicals and
possibly a second, organic phase which could dissolve and transport significant amounts
of chemicals like 2,3,7,8-TCDD having very low solubilities in pure water or
predominantly aqueous mixed solvents (demonstrated below). [Solubility of 2,3,7,8-TCDD
has been reported for benzene (570 mg/L), methanol (10 mg/L), and acetone (110 mg/L)
13
-------
(U.S. EPA, 1980a)]. Naturally, partition coefficients could be determined for
2,3,7,8-TCDD between water and any other solvent not miscible with water (i.e., any
solvent which forms a second phase; one not soluble in all proportions). More details on
mixed solvent effects on solvent solubility follow shortly.
Recent work by Kapila et al. (1987) focused on the role of a non-polar dispersing
medium on the future fate and transport of 2,3,7,8-TCDD in contact with soil. The
medium involved in a pollution incident which occurred at Times Beach, MO (used motor
oil mixed with 2,4,5-T process still bottom materials) likely would exist as a second phase
if in contact with ground water. No contact with an experimental equivalent of ground
water was involved in the work of Kapila et al. (1987), although contact with an aqueous
phase in an unsaturated zone was introduced through simulated rainfall. Their results
suggested that 2,3,7,8-TCDD migration through soil and losses due to volatilization and
photolysis or due to surface photolysis alone appear much lower than previously reported
in the literature.
In addition, the effects of dissolved organic matter and very finely divided
suspended solids on measurement of partition coefficients have been discussed in
numerous publications reviewed by Lyman and Loreti (1987). The dissolved organic
matter of surface waters and ground water both before and after pollution shows wide
variability in sorption capacity, especially for hydrophobic chemicals, although the
potential effect of its presence must be increased transport of sorbed pollutants. Details
and examples of this and other factors affecting sorption, such as pH and ionic strength,
are available in Lyman and Loreti (1987).
A "solids concentration effect" connected with the presence of very fine suspended
solids is characterized by a decrease in sorption constant with increasing soil or sediment
oncentration and is believed to be due to the presence of microparticles not removed by
standard filtration or centrifuging procedures. Lyman and Loreti (1987) note that this
14
-------
phenomenon often was not acknowledged in studies published before 1980 and also cite
studies demonstrating that levels of microparticles (or "non-settleable solids") can
increase as a result of chemical, mechanical or thermal modifications to soils or
sediments which tend to break down the soil matrix in non-reproducible ways. Examples
of techniques which can yield test materials much different from the original soil or
sediment include pH adjustment, ionic-strength adjustments, drying and/or sterilization,
grinding, sieving and extended mixing times. In-depth discussion of the solids
concentration effect also is presented in Lyman and Loreti (1987).
Because little information currently is available on the frequency of occurrence of a
second liquid phase in contact with ground water, description of pollutant transport in a
second phase is viewed as a topic for future consideration.
The potential effect on solubility of organics in mixed water/organic solvent systems
is well known. A variety of thermodynamic approaches have been developed for
estimating solute solubility in mixed solvent systems (reviewed in U.S. EPA, 1985d).
More recently, Webster et al. (1986) reported on adsorption and transport of
1,3,6,8-TCDD on dissolved and suspended organics (humic and fulvic acids) in natural
waters. Investigators also have reported on the solubility and sorption properties of a
number of aromatic organic compounds in methanol/water and acetone/water systems
(Nkedi-Kizza et al., 1985; U.S. EPA, 1985d; Fu and Luthy, 1986). (Because methanol and
acetone are miscible with water no second phase would form.)
Earlier work by Yalkowsky et al. (1972) related solubility of organics in water and
binary solvents by showing a semi-logarithmic increase in solubility with increasing
volume fraction of organic solvent,
In Sm = In Sw + of0 (2-1)
15
-------
where Sm = solute solubility in mixed solvent (moles/L), Sw = solute solubility in water,
a = a system-specific empirical parameter related to surface area and surface free energy
of solute, and f° = volume fraction of organic cosolvent, 0 < f° < 1.
Nkedi-Kizza et al. (1985) noted an exponential decrease in sorption coefficient with
increasing fraction of cosolvent for both methanol/water and acetone/water systems,
In (Km/Kw) - -aof0 (2-2)
where Km = sorption coefficient in mixed solvent systems, Kw = sorption coefficient in
water, and a = an empirical constant. These investigators also found that a was unique
to each sorbate (organic solute)/mixed solvent combination and was independent of the
soil (sorbent) used in different experiments, suggesting that this might be an unusual
site-independent phenomenon.
A similar exponential or semi-logarithmic relationship between solubility and fraction
of organic cosolvent was reported in U.S. EPA (1985d), where it was also noted that
solubility increases upon adding a fixed amount of cosolvent are most pronounced for
very hydrophobic compounds (such as 2,3,7,8-TCDD). It was also noted that increases in
solubility with increasing fraction of organic cosolvent did not result in directly
proportional decreases in the sorption coefficient. (This effect is attributed to solvent
swelling of the soil organic carbon material and a corresponding increase in accessibility
of the latter for sorption.)
Equation 2-1 may be applied to the question of 2,3,7,8-TCDD solubility in
methanol/water mixed solvents, since the Hildebrand solubility parameters (5) of methanol
(15.15 calories1/2 cm~3/2) and water (23.50 calories1/2 cm"3/2) are more than three units
larger than that of 2,3,7,8-TCDD (approximately 10 calories1/2 cm~3/2) (Martin et al.,
1982). A technique described in U.S. EPA (1985d, p. 107) may be used to approximate
16
-------
the solubility parameter, a, for 2,3,7,8-TCDD in a mixed solvent system if the solute
solubility in both pure solvents is known. The solubility parameter estimate is obtained
by taking the difference between the log of solute solubility in pure solvent and the log
of solute solubility in pure water.
Substituting into Equation 2-1, the calculated solubility of 2,3,7,8-TCDD in 1%
methanol/water (by volume) is 14% greater than in pure water (22.0 ng/L versus 19.3
ng/L, a 1.14-fold increase). In 10% methanol/water, 2,3,7,8-TCDD solubility is predicted
to be 7.19 ng/L, or a 3.73-fold increase, while in the environmentally unrealistic case of
50% methanol/water, solubility should increase to 13,800 ng/L (a 715-fold increase). If
similar calculations are performed for a saturated solution of benzene in water (1.78 g/L),
2,3,7,8-TCDD solubility is predicted to be 20.0 ng/L (1.036-fold, or less than a 4%
increase). The application criteria for the use of Equation 2-1, referred to earlier, are
not fully satisfied for 2,3,7,8-TCDD in benzene/water, but generally this only poses
problems at high-volume fractions.
Increases in solubility due to the presence of cosolvents thus are predicted as
relatively small for low percentages of cosolvent (typical in landfills), as a consequence
of the logarithmic variation of solubility with linear variation in volume fraction of
cosolvent. Walters et al. (1987) recently reported measuring sorption isotherms for
2,3,7,8-TCDD in soil/water and soil/water/methanol systems. Isotherms were linear up to
0.5 of solubility and logarithmic variation of solubility and partition coefficients with
linear variation of fraction cosolvent was confirmed. (Further details await publication.)
In summary, prediction of increased solubility for hydrophobic solutes is feasible for
binary systems of related solvents, either through UNIFAC calculations (from activity
coefficients) or through existing information in the literature, which might allow
calculation of a and/or o. for the desired cosolvent/solute combination. UNIFAC
calculations apparently have not yet been published.
17
-------
For ease of reference and comparison, the results of at least two experimental
determinations for Pv, Sw, and Kow have been included in Table 2-1.
Changes in the vapor pressure, pure water solubility, and octanol/water partition
coefficient of the properties of 2,3,7,8-TCDD between Schaum (1984) and the values cited
above will have negligible effect on exposures calculated for any of the five routes
covered in Schaum (1984). The major influence of changes in vapor pressure and water
solubility (hence Henry's law constant) concerns volatilization from water and soils. The
effect of volatilization on potential exposure to 2,3,7,8-TCDD is discussed in greater
detail in Section B of Chapter 4.
B. BODY WEIGHTS AND PULMONARY VENTILATION RATES IN EXPOSURE
ASSESSMENTS
1. Body Weights
The performance of representative exposure assessments requires the use of
appropriate physiological parameters. Traditionally, the assumption was made that human
body weight was equal to 70 kg, which represented the typical U.S. male. When
calculations are made for large populations, this is a valid and acceptable number. When
considering smaller numbers of individuals or subpopulations of concern (e.g., children), a
different value might be more appropriate.
In a report on risks from 2,3,7,8-TCDD-contaminated soil (Schaum, 1984), a
procedure was presented which allowed average body weight to be considered as a
function of age, in particular for individuals under 18 years old (those 18 years of age or
older were considered to be 70 kg). One of the concerns in that report was to obtain
representative values for small children who had a tendency to ingest soil, so that more
appropriate exposure values could be calculated.
A more recent report, entitled "Development of Statistical Distributions or Ranges
of Standard Factors Used in Exposure Assessments" (U.S. EPA, 1985a), refines the
18
-------
TABLE 2-1. PROPERTIES OF 2,3,7,8-TCDD
Structure of 2,3,7,8-TCDD:
Molecular Weight: 322
Cl
Cl
O
O
Cl
Cl
Property
Value Temp.(°C) Method
Reference
Vapor pressure
(x 10'9 mm Hg)
3.49 ± 0.55 30.1
1.52
Water solubility
(ppt = ng/L)
19.3 ± 3.7
Octanol/water partition
coefficient (x 106)
6.9 ± 1.6
14.5 ± 1.6
25
0.74 ± 0.04 25
22
7.91 ± 2.7 25
25
4.24 ± 2.73 22
10.5 ± 1.1
25
Gas saturation,
GC/MS
Extrapolation
Gas saturation,
combustion
Schroy et al.
(1985a)
Schroy et al.
(1985a)
Podoll et al.
(1986)
Thin film Marple et al.
equil., GC/LRMS (1986a)
Adams and Elaine
(1986)
Fragment
additivity
Diffusion,
GC/LRMS
HPLC
Fragment
additivity
U.S. EPA (1981a)
Marple et al.
(1986b)
Burkhard and Kuehl
(1986)
U.S. EPA (1984a)
19
-------
treatment of body weight. It provides weight data on age-sex distributions (with
percentiles) for tailoring exposure assessments to whatever level of detail is necessary.
The source of the data for the above report was the second National Health and
Nutrition Examination Survey, conducted from 1976 to 1980 (National Center for Health
Statistics, as referenced in U.S. EPA, 1985a). The survey was a probability sampling of
approximately 28,000 people from the ages of 6 months to 74 years. Being concerned
with nutrition, the survey oversampled subpopulations thought to be at high risk of
malnutrition.
2. Pulmonary Ventilation Rates
One of the most critical and variable factors in considering the inhalation route in
exposure assessments is the pulmonary ventilation rate. Respiration is usually presented
as a flow rate and denoted as minute volume (L/min or m^/d). Minute volume is the
product of the amount of air moved during each cycle (tidal volume = 0.5 L) and the rate
of respiration.
The resting ventilation rate is related to an individual's basal metabolic rate for
oxygen consumption, and is reported as 7.5 L/min (0.5 L x 15 breaths/minute). As the
activity of individuals increases, so does their metabolism and, hence, the ventilation
rate. To obtain a representative description of exposure, information on the activity, its
duration, and the associated ventilation rates must be integrated. A report by the U.S.
EPA (1985a) on standard factors provides tables on distributions of minute volumes and
activity levels and patterns for various age-sex groups. These data are especially useful
in situations when subpopulations are being considered for their particular exposure
levels.
Some factors to consider in addressing ventilation rates relate to individuals in
compromised states, such as emphysema, fibrosis, or those changes that occur with age.
The U.S. EPA (1985a) report contains a table listing various formulae for calculating
20
-------
minute volume (empirically) based on variables including surface area, weight, height, age,
and sex. It should be noted that little information is available for preschool children,
due mainly to the problems involved in clinically studying this age group.
Activity levels can be categorized as light, moderate, or heavy according to criteria
developed by the Environmental Criteria and Assessment Office, RTF [Air Quality Criteria
for Ozone and Other Photochemical Oxidants, June 1984; as cited in U.S. EPA (1985)].
The estimated ventilation rates for these different activity levels, which were originally
presented in Table 4-3 of the U.S. EPA (1985a) report on standard factors, are
reproduced here for the convenience of the reader (Table 2-2).
The activity patterns, when used with the ventilation data, will provide a
time-weighted average ventilation rate for use in assessment. The data on activity
patterns indicate the time which individuals might spend at various activities or in
different microenvironments. This type of data is especially useful when focusing on the
specific locations of individuals' exposure. These activity data are averaged for both
sexes and all age groups. Appendix D in the above-referenced report contains detailed
information on activity patterns for 56 population subgroups.
When performing assessments for general populations, the ventilation rates (U.S.
EPA, 1985a) and activity pattern data [93% light activity, 6% moderate, and 1% heavy
(U.S. EPA, 1985a)] can be combined as follows to provide an approximate overall daily
ventilation rate:
Ventilation rate = [(22.4 hr/d x 13.8 L/min) (2-3)
+ (1.4 hr/d x 40.9 L/min)
+ (0.2 hr/d x 80 L/min)] / (1,000 L/m3 x hr/60 min) = 23 m3/d
21
-------
This value is the same as that reported in a previous EAG report (Schaum, 1984), and is
similar to values used traditionally that range from 20 to 23 m3/d as a daily ventilation
rate.
22
-------
TABLE 2-2. ESTIMATED MINUTE VENTILATION ASSOCIATED WITH
ACTIVITY LEVEL FOR AVERAGE MALE ADULT
Level
of
work
Light
Light
Light
Watts kg-m/mina
25 150
50 300
75 450
L/min
13
19
25
Representative activities
Level walking at 2 mph;
washing clothes
Level walking at 3 mph;
bowling; scrubbing floors
Dancing; pushing
Moderate
Severe
100
300
600
1,800
30
Moderate
Moderate
Heavy
Heavy
Very heavy
Very heavy
125
150
175
200
225
250
750
900
1,050
1,200
1,350
1,500
35
40
55
63
72
85
100+
wheelbarrow with 15-kg
load; simple construction;
stacking firewood
Easy cycling; pushing
wheelbarrow with 75-kg
load; using sledge hammer
Climbing stairs; playing
tennis; digging with spade
Cycling at 13 mph; walking
on snow; digging trenches
Cross-country skiing; rock
climbing; stair climbing
with load; playing squash
and handball
chopping with axe
Level running at 10 mph;
competitive cycling
Competitive long distance
running; cross-country
skiing
akg-m/min = work performed each minute to move a mass of 1 kg through a vertical
distance of 1 m against the force of gravity.
SOURCE: Adapted from U.S. EPA, 1985a.
23
-------
3. FATE
Recent information on the fate of 2,3,7,8-TCDD in soil and sediments tends to
confirm previous findings on this subject. Once 2,3,7,8-TCDD is in the soil.the chemical
and biological degradation processes are very slow, with half-lives estimated in tens of
years or longer. Although 2,3,7,8-TCDD has a low solubility and low vapor pressure, the
long-term (years-to-decades) soil-moisture-interstitial air partitioning system will lead to
slow movement of the chemical. In systems where the contamination is relatively near
the surface, it is likely that the 2,3,7,8-TCDD will migrate to the surface over time and
volatize rather than leach into ground water. Except in unusual cases involving mobile,
organic-contaminants, large-scale leaching of 2,3,7,8-TCDD to ground water from soil is
thought to be unlikely (EPA, 1985g). (Note, however, that some landfills may have these
very conditions.) There is little evidence to support the suggestion that photolysis at
the soil surface plays a significant role in reducing contaminant concentrations. An
estimated biodegradation half-life of several decades would make some difference to an
exposure assessment on a site for a 70-year lifetime, compared to the assumption that
the compound does not degrade at all, perhaps lowering exposure in some scenarios by a
factor of two to four.
Contaminated soil may also be ingested by grazing animals, potentially contaminating
foodstuffs such as meat and dairy products. Recent studies tend to support earlier
estimates of transfer factors from soil to food.
Recent attempts to determine the effect contaminated sediments may have on
aquatic biota have led to estimates that differ by several orders of magnitude for average
bulk water concentrations and those concentrations near the contaminated sediments.
The following sections discuss these areas in more detail.
24
-------
A. FATE OF 2,3,7,8-TCDD IN SOIL
1. Transient Behavior of TCDD Profile in Soil Layer
In this section, a brief description will given of the manner in which concentrations
of 2,3,7,8-TCDD in the soil media can change over a period of time. The method of
incorporating these changes into an exposure assessment will also be discussed.
2,3,7,8-TCDD, like any other organic contaminant, can undergo transport and
transformation processes in soil. These processes may be chemical, biological or physical
in nature. The chemical processes may include hydrolysis, photolysis (discussed in more
depth in Chapter 4), or breakdown of chemical bonds in the molecules, resulting in the
change of the contaminant to products that could be more harmful or less harmful than
the original contaminant. The biological processes occur in the presence of soil microbes
to enhance the breakdown of the contaminant to the other products. Finally, the
physical processes can be thought of as transport processes in which the contaminant
retains its chemical identity, but is transported from one location to another by diffusion
or advection mechanisms, and may be transferred to different media. Examples of
transport include diffusion of 2,3,7,8-TCDD vapor through soil pores and ultimate
volatilization into the air, or the leaching of the contaminant into soil by precipitation
or floods. The severity of the migration will be dependent on the mobility of
2,3,7,8-TCDD in the soil. As a result, the initial concentration distribution will change
as the contaminant is subjected to these transport and transformation processes over a
long period of time.
As discussed above, a variety of physical and chemical processes may affect the fate
of 2,3,7,8-TCDD in soil. However, the most important process appears to be vapor phase
diffusion and photolysis at the surface (Freeman and Schroy, 1986). Diffusion is
discussed further in this chapter. Photolysis is discussed in Chapter 4. Although
25
-------
biodegradation appears to be a relatively unimportant process, it is also discussed in this
chapter.
Freeman and Schroy (1986), EPA (1985b), and Tung et al. (1985) simulated the
concentration and thermal profile of 2,3,7,8-TCDD in soil with initially uniform
contamination. As time progressed, the simulated concentration profile tended to be bell
shaped, with a maximum concentration somewhere in the core of the soil column. The
bell-shaped concentration profiles, calculated as a function of depth in soil, were
compared with the results of analyses for 2,3,7,8-TCDD in soil core samples taken at
various depths. Samples were taken from plots at Times Beach, Missouri (EPA 1985c) and
Eglin Air Force Base (Tung et al., 1986). They showed good agreement between the
model simulation and the measured data in both cases. EPA 1985c noted that "the floods
at Times Beach, Missouri, have not redistributed the TCDD over a large area," and
concluded that based on a simulation of the measured concentration profile at some time
periods, the volatilization process is a major mechanism by which 2,3,7,8-TCDD is
depleted from the soil. EPA (1985g) used field soils to measure the soil/water partition '
coefficients, which ranged from 3 x 104 L/kg to 1.3 x 107 L/kg, and evaluated the
teachability of 2,3,7,8-TCDD from the soils. Based on these partition coefficients and the
use of solute transport models, they concluded that the worst-case movement of
2,3,7,8-TCDD in leaching from soil media is so slow that leaching by water is
unimportant compared with other transport mechanisms, such as volatilization and
erosion. Note that in other situations concentration profiles may differ due to
differences in mode of application and weather conditions.
According to EPA (1985g), EPA (1985b) and Freeman and Schroy (1986), the rate of
movement of 2,3,7,8-TCDD in soil during the leaching process is insignificant compared to
the depletion of 2,3,7,8-TCDD by volatilization. Over a long period of time, this
depletion will create a new profile of 2,3,7,8-TCDD concentration in soil. This
26
-------
redistribution should be an extremely slow process in view of the very low vapor
pressure of 2,3,7,8-TCDD. The change of concentration profile in the soil column, along
with the effect of the soil concentration profile on the exposure evaluation, is discussed
below.
2. Transient Profile in Soil Column
For assessment of exposures over extended periods, it is appropriate to obtain the
average concentration of 2,3,7,8-TCDD in soil over the depth of concern and the
exposure duration. For some pathways, this average concentration is important; for other
pathways, surface concentrations are important. Although the surface concentration may
theoretically appear to be relevant in some cases, the soil surface is not always
quiescent, and could be subject to disturbances due to construction activities, erosion, or
digging. These activities will expose the subsurface soil and make these soils available
for human exposure.
Tung et al. (1985) experimentally measured the variation of the soil temperature
profile as a function of depth and time of a day. The temperature variation along the
soil column in a typical day can affect the volatilization rate of 2,3,7,8-TCDD vapor
because certain properties of 2,3,7,8-TCDD vapor influencing the volatilization rate is
temperature-dependent. These properties are particularly important when dealing with
low volatility organics in a soil matrix, and include vapor pressure and diffusivity. The
experimental data showed that the diurnal temperature variations are noticeable at soil
depths 2 and 10 cm from the surface, and diverge to an essentially constant temperature
at a depth of 48 cm regardless of times of day.
The significant diurnal temperature change on the surface of soil would be an
important factor in dealing with volatilization of 2,3,7,8-TCDD vapors from the shallow
surface of soil. This phenomena would be significant during the initial stage of
contamination. As time progresses, the bulk concentration of contaminant will remain at
27
-------
a depth of the soil column as demonstrated by Freeman and Schroy (1986). Since the
2,3,7,8-TCDD vapor molecules in the soil pore in the bulk of the soil column should
diffuse out to the air-soil interface, the bulk of the transport process occurs at the
depth where the bulk of the contaminant is contained. The transport process will be
affected by the vapor pressure and diffusivity corresponding to the temperature in the
bulk of the soil column. The bulk soil temperature deep inside the soil should remain
fairly constant without being affected by the diurnal temperature variation. In
considering landfills with depth of 10 feet or more, the bulk phenomena would be more
important that the surface phenomena when dealing with long-term exposure.
The methods of estimating the average concentration along a soil depth are
available for two cases in which initially the contaminant is uniformly distributed to a
specific depth in soil. In one case, it is assumed that contamination occurred from the
soil surface to a certain depth. In the other case, it is assumed that the contaminated
surface is covered with a soil material of known thickness initially free of the
contaminant. Methods for estimating the average concentrations in these cases involve
Fourier series solutions to partial differential equations (U.S EPA, 1986a). Although it
may be a tedious process to correct for the time- and depth-dependent concentration
variation, such a correction is particularly important for sites that were contaminated a
long time ago or are being evaluated for long-term exposures. However, the use of the
initial concentration in uniformly contaminated soil will provide an upper-limit value. If
the site-specific concentration profile data are available, the maximum value for the
bell-shaped concentration profile can approximate the initial value.
The use of cover soil initially free from 2,3,7,8-TCDD will retard exposure for a
period of time. The time required to contaminate a clean cover will depend on a number
of site-specific properties. An estimate of this time was made assuming a 25-cm cover
thickness, 1% organic carbon content, and other typical soil properties. This estimate
28
-------
also assumes that the 2,3,7,8-TCDD is infinitely available under the cover, no convective
transport occurs, no degradation occurs, however, equilibrium partitioning occurs between
the 2,3,7,8-TCDD adsorbed to soil and in the air space of the soil. On this basis, it was
estimated that the 2,3,7,8-TCDD concentration adsorbed to soil at the surface would
reach 1% of the concentration under the cover in 1,000 years and 90% of the
concentration under the cover in one million years. The uncertainties in these estimates
could be significant, since the estimates are based on a mathematical model rather than
actual data. However, topsoils frequently contain more than 1% organic carbon and the
2,3,7,8-TCDD source will not be infinite, which suggests that the model estimates are
more likely to be low than high.
Another site-specific property affecting volatilization rates from soil is the
temperature of the soil. Freeman and Schroy (1986) noted that soil temperature
fluctuations were important considerations in the estimates of vapor flux rates.
Vapor-phase diffusion can occur downward and laterally as well as upwards. For
near-surface contamination scenarios, the upward movement is more important, since the
chemical will probably reach the surface before it reaches the ground water. Thus,
although downward diffusion may occur at rates similar to upward diffusion, a much
longer time elapses before the chemical becomes avail able for exposure.
3. Degradation of 2.3.7.8-TCDD in Soil
If the concentration of 2,3,7,8-TCDD in soil changes significantly over a period of
time, the exposure evaluation should reflect this change in concentration. Such changes
can be incrementally accounted for in the exposure computations requiring the summation
of all exposures over the lifetime period, or, alternatively, a representative concentration
of 2,3,7,8-TCDD in soil averaged over the exposure period can be used.
In addition to volatilization, leaching, and atmospheric photolysis, which are
addressed in Chapter 4, another possible mechanism of reducing the concentration of
29
-------
2,3,7,8-TCDD in soil is biodegradation by action of microorganisms. Biodegradation in
soil is discussed in Section A.4 of this chapter, and biodegradation in sediment is
discussed in Section B.I of this chapter.
Many researchers claim that photodegradation is the primary means by which
degradation of 2,3,7,8-TCDD occurs at the soil surface. Czuczwa and Hites (1986) studied
lake sediments and concluded that little photolysis occurred during the long-range
transport of atmospheric 2,3,7,8-TCDD on particulates. However, final conclusions cannot
be drawn, since the sources of lake sediments are not known with certainty.
Additionally, this work focused on incinerator fly ash so it may not be applicable to
soils. Chemical degradation via hydrolysis and oxidation in soil is very unlikely in view
of the insignificant rate of these reactions in aquatic media (U.S. EPA, 1985b). Recent
evidence indicates that photolytic reactivity on fly ash behaves differently from the soil
surface: SRI found experimentally that photolytic degradation of 2,3,7,8-TCDD on the
two types of fly ash tested is negligible (Mill, 1987). Investigators have shown that
2,3,7,8-TCDD on the soil surface photolyzes at a rate slower than observed in organic
solutions (Zepp et al. 1988).
The concentration of 2,3,7,8-TCDD in environmental media may depend on the
degradation of related congeners as well as that of 2,3,7,8-TCDD itself. The process of
reductive dechlorination; more highly chlorinated congeners degrading to 2,3,7,8-TCDD,
may not occur at a significant rate except in anaerobic microbes, where the rate would
be extremely slow. In fact, the available data strongly suggests that more highly
chlorinated PCDDs will not degrade to the 2,3,7,8-TCDD isomers (U.S. EPA, 1985d).
4. Biodegradation of 2.3.7.8-TCDD in Soil
The environmental persistence of 2,3,7,8-TCDD has created concern on a national
level (EPA 1985g). It is thought that one reason for this persistence is that soil
microorganisms cannot degrade 2,3,7,8-TCDD, or that they do so very slowly (Bumpus et
30
-------
al., 1985). However, very few studies have been done on the biodegradation of
2,3,7,8-TCDD in soil. A biodegradation rate constant and a half-life will be derived
below on the basis of the limited available information.
Bacteria, while the most abundant of soil microorganisms, constitute less than half
of the total microbiological cell population (Alexander, 1977). Fungi, because of their
extensive network of mycelium, usually account for a significant portion of the soil
biomass (Alexander, 1977). One of the major activities of fungi in the mycelial state is
the degradation of complex molecules. Fungi require aerobic conditions in order to
•
degrade chemicals, and are affected by the availability of oxidizable organic substrates.
The actinomycetes become predominant in dry and cultivated areas (Alexander, 1977).
Algae, being phototrophic and having a lower population in soil than other
microorganisms, generally play an insignificant role in soil biodegradation.
Young (1983) estimated the half-life of 2,3,7,8-TCDD on his test grid (grass field
located at Eglin Air Force Base in Florida) to be 10 to 12 years. He cited several
mechanisms to account for the disappearance of 2,3,7,8-TCDD in the herbicide applied
within the area of his study. Crosby and Wong (1977) had observed that trace amounts
of 2,3,7,8-TCDD in Herbicide Orange exposed to sunlight on leaves, soil, or grass, were
apparently photodegraded during dissemination of the herbicide. On the basis of their
data, Young calculated that less than 1% of the original 2,3,7,8-TCDD that was sprayed
actually remained in the ecosystem of the test site, and that the majority of the
remaining 1% was retained in the soil. He reasoned that the processes of water and
wind transport of contaminated particles and biomass removal would be unimportant
mechanisms in removing the 2,3,7,8-TCDD retained in the soil. Although he stated that
the role of volatilization and microbial degradation in removing 2,3,7,8-TCDD from soil is
not clear, he estimated the half-life as 10 to 12 years, based on observed changes in soil
concentrations.
31
-------
Ward and Matsumura (1978) summarized the conclusions of other investigators to the
effect that microbial degradation of 2,3,7,8-TCDD was found to be very low or
nonexistent. Freeman and Schroy (1986) also concluded that biodegradation of 2,3,7,8-
TCDD does not occur on the basis of tests they conducted with soil from Eglin Air Force
Base. Bumpus et al. (1985) conducted experiments on the degradation of 2,3,7,8-TCDD
mixed in cultures containing the white rot fungus, and measured the rate of CC>2
evolution as an indication of the progression of biodegradation.
It is difficult to analyze the Bumpus et al. (1985) data in a strict kinetics sense,
since the progress of the reaction was monitored by analyzing the reaction product (CC«2)
being evolved, and exact stoichiometric relations between the reaction product and
2,3,7,8-TCDD are not known. For example, the CC>2 may have evolved from oxidation of
compounds other than 2,3,7,8-TCDD. Additionally, the applicability of experimental
conditions to actual soil conditions is rather speculative because the distribution of
microorganisms in actual versus experimental soil is different, and because fungus favors
aerobic conditions for biodegradation. If it is assumed that aerobic conditions are poorly
maintained in soil, except perhaps in surface layers, the biodegradation constant derived
from these data would tend to overestimate biodegradation.
For these reasons and others, it is widely believed that the application of such
experiments, particularly those based on pure cultures, to actual field conditions may be
inappropriate. Accordingly, the results of this study were not applied anywhere in this
report. However, for illustrative purposes only, a rate constant was derived from the
Bumpus et al. (1985) data. Assuming first-order kinetics, the rate constant derived was
6.6 x 10"^ day~*. This corresponds to a half-life of about 29 years. It should be noted
that these data are applicable for one fungus type only. The effects of the presence of
other microorganisms, including bacteria, on the rate constant cannot be evaluated at
32
-------
present. It appears reasonable to expect a longer half-life than that observed by Young
(1983), since his observation includes the effect of photolytic degradation.
Once a half-life has been estimated, the exposure can be calculated under the
assumption that the concentration of 2,3,7,8-TCDD in soil varies according to first-order
kinetics.
= Csoe-kt (3-1)
where Cs = the concentration of 2,3,7,8-TCDD in soil at any time (days) from initial
contamination at concentration Cso, and k = the rate constant (day~^). The half-life,
tj/2, can be given as
t1/2 = 0.693/k (3-2)
The average concentration in soil under the influence of biodegradation 'can be
obtained by integrating Cs over the exposure time, and can be substituted for use in
exposure evaluation. When the concentration is evaluated considering volatilization, and
based on the depth-averaging process, the initial concentration, Cso, should be replaced
by the time- and depth-averaged concentration in obtaining the concentration corrected
for biodegradation. Alternately, exposure can be estimated by solving for Cs at frequent
intervals, computing exposure, and summing exposure values. Such calculations can be
conveniently done on a computer. The effects of degradation on exposure can be shown
as follows:
(Degradation exposure) =/Csdt/Csot = (l-e"kt)/kt (3-3)
(Non-degradation exposure)
33
-------
The non-degradation exposure is multiplied by this ratio to reflect the effects of
degradation. The ratio of 1 assumes that biodegradation will not occur (i.e., half-life
equals infinity). The lower limit of this ratio will provide the maximum degradation.
Microbial degradation reduces the concentration of 2,3,7,8-TCDD available for human
exposure. Although the biodegradation rate for 2,3,7,8-TCDD has not been established, it
appears that its half-life in soil is in the range of several decades. For purposes of
illustration, if we assume a 30-year half-life for biodegradation in soil, approximately 80%
of the original 2,3,7,8-TCDD would have transformed to other products at the end of a
70-year (life-time) period. For a lifetime exposure evaluation, it is appropriate to take
into account the gradually decreasing 2,3,7,8-TCDD concentration in soil from which the
contaminant is released for human exposure. For a 70-year exposure period, Equation
3-3 indicates that at 30-year half-life causes a 50% reduction in exposure relative to an
infinite half-life (i.e., no degradation).
B. FATE OF 2,3,7,8-TCDD IN SEDIMENTS
1. Aquatic Sediments
Low concentrations of various CDDs have been reported to exist in sediments in the
Great Lakes and other water bodies (Czuczwa and Kites, 1986). Czuczwa and Kites
attribute the existence of CDDs in aquatic sediments to the long-range transport through
the atmosphere of dioxin-related material in emissions of combustion products. Erosion
of contaminated soil may also result in accumulation of 2,3,7,8-TCDD in sediments.
Contaminated sediments will slowly release the contaminant to the water body in
dissolved form or as suspended sediment. The impacts of contaminated sediments must be
considered in exposure estimation.
Recent theoretical discussions (e.g., Thibodeaux e't al., 1986) of the manner in which
2,3,7,8-TCDD may partition from sediments to water to fish and other aquatic biota,
when compared to empirical observations (e.g., Kuehl et al, 1987b), allow some feeling for
34
-------
the importance of the water-to-fish route for bioaccumulation as compared to the sum of
all routes (including ingestion of suspended sediment and other biota) as implied by
empirical data. The following paragraphs discuss the sediment-to-water partitioning of
2,3,7,8-TCDD, and the implications this partitioning may have for the water-to-biota
(alone) route, as well as a way to calculate potential surface water-drinking water
concentrations due to contaminated sediments. The section that follows (Section C)
discusses the use of empirical data for estimating impacts of contaminated sediment on
fish.
2. Sediment-to-Water Transport Process
In estimating the 2,3,7,8-TCDD concentration in a water body above a sediment of
known contamination level, it is useful to picture a system where, initially, the water is
free of 2,3,7,8-TCDD, and therefore a concentration gradient is set up between sediment
and water, providing a driving force for the contaminant to enter the water. In the
process, dilution will occur as a result of mixing with the moving water and diffusion
through it. If the 2,3,7,8-TCDD molecules reach the interface of the water body and
atmosphere, volatilization will occur. In the process of transporting from bottom layer
to top surface, 2,3,7,8-TCDD in water may be subject to photolysis, biodegradation and
other inter-related reactions resulting in some disappearance of the contaminant. It is a
complex process to track the movement of 2,3,7,8-TCDD from the sediments to the
atmosphere across the water body.
Although the process of the transport may be inherently transient, the average
concentration in the water body may approximate steady-state; that is, the concentration
in the bulk water remains constant by assuming that the amount of 2,3,7,8-TCDD leached
from the sediments into the water is equal to that lost to the atmosphere. The
steady-state (non-equilibrium) process can be modeled using steady-state transport models
(Thibodeaux et al., 1986). The results of such a model should represent an average value
35
-------
of 2,3,7,8-TCDD concentration across the water body. Impacts on the organisms
inhabiting near the sediments may be greater than the exposure estimated by this average
value. For such organisms, the use of the equilibrium partitioning relationship will
provide an upper-bound limit of exposure (for the sediment-water-organism route alone).
The more appropriate parameter would be the sediment to organism transfer coefficient
obtained from field data.
3. Estimation of TCDD Concentration In Water Body
Transient aspects of diffusional processes through the sediments and water body
boundary layer will require applications of Fickian-type models. Thibodeaux et al. (1986)
presented a simplified model based on the consideration of two-phase resistances in the
sediment and water sides, and pointed out important parameters controlling the rate of
the contaminant transfer from sediment to water body. These parameters include the
effect of surface winds creating mixing and moving of lake water, thermal stratification,
flow-through, and lake geometry. The effects of these parameters are combined into
mass transfer coefficients based on experimental data (Thibodeaux and Becker, 1982).
Although the model is based on the widely applied two-phase resistance theory and
uses experimentally derived mass transfer coefficients, it has not been validated via field
measurement.
The steady-state model relating the sediment and water concentrations in the
absence of chemical reactions in the water body is given by Thibodeaux et al. (1986) as:
Cw = ((kwke)/(kw+KLA)(kw+ke)-(kw)2Xl/Kd Ce) (3-4)
where Ce = the concentration in sediment (mg/kg), Cw = the concentration in the bulk
water body (mg/L), kw = the water-side mass transfer coefficient above sediment
(cm/hr), ke = the sediment-side mass transfer coefficient (cm/hr), KL& = the overall
36
-------
water/air mass transfer coefficient based on water side for air-water interface (cm/hr),
and K(j = the partition coefficient between sediment and water (mg/kg sediment/mg/L
water). It can be noted that the water body concentration, Cw, refers to the
concentration in the bulk of water, and should be distinguished from the concentration in
the vicinity of sediments which may be close to the equilibrium partitioning
concentration. This model is derived assuming that the sediment is the source of the
contaminant and volatilization losses to the atmosphere are the sinks, that resuspension
of sediments is negligible, and that there is no significant inflow and outflow of water.
Thibodeaux and Becker (1982) presented correlations for individual mass transfer
coefficients for sediment and water sides. These correlations are as follows:
o Water-side mass transfer coefficient
kw = (0.06(CDV2h5/4)/FMl/2)(pa/pw) (3_5)
where CD = the drag coefficient (0.00166 for wind speed 1 to 7 m/s and 0.00237 for wind
speed 4 to 12 m/s), V = the wind speed at 10 meters above the surface of the water
body (cm/min), h = the average depth of the water body (cm), F = the average wind
fetch (cm), M = the molecular weight of 2,3,7,8-TCDD (322), pa = the air density (g/L),
and pw = the water density (g/L).
o Sediment-side mass transfer coefficient
ke = 3,600((DwE4/3)/r) (3-6)
where Dw = the diffusivity of 2,3,7,8-TCDD in water (cm2/s) [5.6 x 10"6 cm2/s)], E =
the sediment porosity, and r = the thickness of contaminated sediments (cm).
37
-------
Equations 3-5 and 3-6 provide the coefficients necessary to calculate the average
2,3,7,8-TCDD concentration in the water body using Equation 3-4. The concentration at
the water body interface above the sediments should be closer to the equilibrium
partitioning concentration than the average water body concentration.
4. Example Calculation
To facilitate the use of Equation 3-4 in estimating the concentration in the water
body contaminated by the presence of 2,3,7,8-TCDD in the sediments, an example
calculation is provided below. This example is tailored for a particular site. For other
conditions, calculations should be performed with appropriate parameter values and
conditions applicable to the sites.
Suppose that a small pond is contaminated with 2,3,7,8-TCDD. For the water body,
it is assumed that the temperature is 15°C - 25°C, depth is 500 cm with a surface area
of 100 m x 100 m, and the lake is unstratified. For the sediment layer, it is assumed
that the temperature is also 15°C - 25°C, the thickness of contaminated sediments is 10
cm, the surface area of contaminated sediments is also 100 m x 100 m, and the sediments
have a porosity of 50% and contain 1% organic matter.
Other parameters assumed are that the wind speed is 6 miles per hour, the
diffusivity of 2,3,7,8-TCDD in water is Dw = 5.6 x 10~6 cm2/s, and the sediment/water
partition coefficient (at 1% organic matter) at the Koc value of 468,000 is Kj = 4,680
L/kg (Schroy et al., 1985b). Using these values, one can calculate the value for kw from
Equation 3-5 as kw = 0.4 cm/hr, the value for ke from Equation 3-6 as ke = 8 x 10~^*
cm/hr, and the value for K-La from the two-resistance theory between the water body
and air as KL& = 0.725 cm/hr (Thibodeaux, 1979). Substituting these values into Equation
3-4, the average water body concentration (Cw) is 2.4 x 10"° ug/L when the
concentration in sediments is 10 ug/kg. As an alternative value, an equilibrium
partitioning model predicts the concentration to be 2.1 x 10"^ ug/L when the
38
-------
2,3,7,8-TCDD concentration in sediment is 10 ug/kg. For large water bodies, an
equilibrium approach for calculating average water body concentrations is unrealistic,
since equilibrium may never be attained.
This example shows the method of using Equations 3-4, 3-5, and 3-6 to calculate
the concentration of 2,3,7,8-TCDD in the water underlain by the sediments using the
steady-state and equilibrium partitioning models.
C. BIOACCUMULATION OF 2,3,7,8-TCDD IN FISH AND CATTLE
1. Bioaccumulation in Fish
2,3,7,8-TCDD has been shown to be bioavailable to fish from sediments and fly-ash.
Many aquatic organisms, including fish, selectively accumulate polychlorinated
dibenzodioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs), which are substituted
at the 2, 3, 7, and 8 positions (Rappe et al., 1981; Kuehl et al., 1985, 1986a, b).
Furthermore, a marked preferential affinity is demonstrated by some fish for
2,3,7,8-TCDD over all other 2,3,7,8-substituted dioxin congeners in the tetra through octa
homologous groups (Kuehl et al., 1987b). This is believed to be due to more rapid
elimination of other TCDD isomers or, for higher chlorinated PCDDs and 1,3,6,8-TCDD, it
may be due to a combination of elimination rate differences and kinetic effects involving
decreasing amounts of each isomer with respect to 2,3,7,8-TCDD in each step along the
food chain, as well as decreased uptake rates across the gills with increasing degree of
chlorination (Cook, 1987).
The exposure assessment procedures detailed in this document, as well as the
exposure assessment methods outlined in a previous document (Schaum, 1984), consider
bioaccumulation in fish as a function of a fish/sediment distribution factor. This
approach avoids difficulties inherent in attempts to use recently updated bioconcentration
factors (BCFs) for 2,3,7,8-TCDD [66,000 for carp, 97,000 and 159,000 for fathead minnows
at two exposure concentrations (Cook, 1987a)]. The principal difficulty arises from the
39
-------
inability to detect environmental water concentrations with the most sensitive techniques
now available. This approach also avoids the difficulty of only using the water-to-fish
route to estimate bioconcentration, as one might do with the concentrations derived in
Section B.4 of this chapter. By using empirical data, all routes are "lumped," making it
somewhat harder to understand and manipulate this information using theoretical models.
The fish/sediment distribution factor approach assumes that the fish and sediment are at
some "steady-state", rather than true thermodynamic equilibrium, over the exposure
period. This means that 2,3,7,8-TCDD levels in the sediment remain essentially constant
at a particular location and that fish have achieved a balance with the environment. A
further assumption is made that the levels of 2,3,7,8-TCDD in fish from a particular
location will remain constant over time and with a constant relationship to the
2,3,7,8-TCDD level in the sediment. Some fish species, such as bottom feeders, will move
toward steady-state conditions faster than others. Many species may never reach
equilibrium due to the fact that they do not spend enough time in one location, and
some species will bioaccumulate more 2,3,7,8-TCDD than others due to greater lipid
content.
As a first approximation, the ratio of 2,3,7,8-TCDD in fish to 2,3,7,8-TCDD in
sediment may be assumed to range from 1 to 10, as has been suggested in a variety of
studies. For example, laboratory-derived fish/sediment ratios for 2,3,7,8-TCDD for the
catfish, Ictalurus (after only 6 days of exposure), ranged from 0.2 to 2.0, with the higher
ratios corresponding to the less-contaminated sediment. Fish/sediment ratios for the
mosquito fish, Gambusia (after only 3 days of exposure), ranged from 0.2 to 12.0, again
with higher ratios being calculated for the less-contaminated sediment (Isensee and Jones,
1975). A fish/sediment ratio of 0.44 for 2,3,7,8-TCDD can be calculated for the northern
brook silverside, Laludesthes. from the laboratory data of Matsumura and Benezet (1973);
exposure in this case was 4 to 7 days. A field study cited by Kenaga and Morris (1983)
40
-------
associates fish (species unspecified) 2,3,7,8-TCDD residues of 4 to 85 ppt with sediment
concentrations of 10 to 35 ppt; the means of these ranges (44.5/22.5) would indicate a
fish/sediment ratio of about 2.0.
Newer data support the range of fish/sediment ratios presented above. Kuehl et al.
(1987b) presented field data which indicate that carp (Cvorinus) fish (70 pg/g)/sediment
(30 to 200 pg/g) ratios are in the range of 0.4 to 2.3 for 2,3,7,8-TCDD. Laboratory data
presented by the same authors for 2,3,7,8-TCDD yield a fish (7.5 pg/g)/sediment (39 pg/g)
ratio of 0.2 after 55 days of exposure; however, this ratio probably does not represent a
steady state. In some Missouri streams, bottom fish such as sculpins appear to have
2,3,7,8-TCDD concentrations exceeding 10 times the sediment concentrations (Cook, 1986).
Thus, fish/sediment ratios for 2,3,7,8-TCDD can be variable depending on a series of
interdependent factors, including species, lipid content, weight, ratio of surface area to
weight, organic carbon content of the sediment, food intake rate, density of suspended
particulate matter, and concentration of 2,3,7,8-TCDD in the sediment. In addition, fish
in a stream may approach a steady-state with higher sediment concentrations upstream
due to food drift and/or suspended sediment with a higher 2,3,7,8-TCDD concentration
than the bottom sediment concentration.
Some of the variability in the values derived for fish/sediment ratios could be
eliminated if they were derived based on 2,3,7,8-TCDD in the lipid of fish and organic
carbon of the sediment. Lake et al. (1984) reported partitioning (preference) factors for
polychlorinated biphenyls (PCBs) in field studies with aquatic invertebrates of 0.1 to 0.5
(two- to 10-fold greater concentration of PCB per gram of lipid than per gram of organic
carbon of the sediment). These authors also report 3.3- to 5.9-fold greater
concentrations of chlordane, DDD, and tetrachlorodiphenyl in the lipid of these organisms
than in organic carbon of sediments. Kuehl et al. (1987b) also used this approach in
order to facilitate comparison of data for laboratory (10-g carp) and field (1.5-kg carp)
41
-------
experiments, and to determine the congener-dependent bioavailability of select
2,3,7,8-substituted PCDDs and PCDFs. The Bioavailibility Indices derived for these
congeners show clearly that 2,3,7,8-TCDD is preferentially bioaccumulated, with other
2,3,7,8-substituted congeners being accumulated to a lesser degree and non-2,3,7,8-
substituted cogeners being accumulated to a much lesser degree.
2. Bioaccumulation in Cattle
Beef and dairy cattle have been shown to accumulate significant levels of
2,3,7,8-TCDD and compounds with generally related structures such as PCBs, DDT, and
PBBs following administration in the diet or ingestion of contaminated soil. The
potential for human exposure through consumption of beef and dairy products is greatest
where the cattle have contact with the soil; soil ingestion by cattle is the major pathway
for the transmission of 2,3,7,8-TCDD residue from soil to these animals. The amount of
soil ingested by grazing cattle can vary between 2% and 15% of dry matter intake,
depending on whether vegetation is lush or sparse (Healy, 1968).
A number of studies have been conducted using compounds having structures
generally related to 2,3,7,8-TCDD, such as PCB, PBB, and DDT, which relate the resulting
level in body fat or milk fat to the level of the contaminant in the diet. Fries (1982)
reported that under constant feeding these compounds reach an upper estimate,
steady-state milk fat/diet ratio of approximately 5, with body fat levels being slightly
lower. Jensen et al. (1981) conducted similar studies using 2,3,7,8-TCDD, and found the
beef fat/diet ratio (after 28 days of feeding) to be about 4, suggesting that 2,3,7,8-TCDD
behaves similarly to PCB, PBB, and DDT. Jensen and Hummel (1982) developed data from
dairy cows given 2,3,7,8-TCDD in their feed which indicated a cream/diet ratio of 1.6 to
2.2 (cream containing 18% to about 40% butterfat).
Fries (1985) analyzed data from a study in which cattle were kept in feed-lots on
four Michigan farms where soils in the confinement areas were the sole source of PBB.
42
-------
Under these conditions, the beef fat/soil ratio was in the range of 0.27 to 0.39 for beef
cows, beef calves, and dairy heifers (never lactated), and the milk fat/soil ratio ranged
from 0.02 to 0.06 for multiparous and primiparous dairy cows. Assuming, again, that
2,3,7,8-TCDD behaves in a manner similar to PBB, and that the conditions on the
Michigan farms represent the typical situation on U.S. farms, a beef fat/soil
bioconcentration ratio of 0.3 to 0.4 and a milk fat/soil bioconcentration ratio of 0.04 are
suggested for use in the procedures described for exposure assessment in this document
and in Schaum (1984).
However, if more specific information concerning the farm management system is
known, such information should be used to adjust these values. It should be recognized
that the significance of soil ingestion as a pathway for animal exposure, and ultimately
for human exposure, is greatly reduced under U.S. agricultural conditions (Fries, 1986).
Lactating dairy cows are rarely pastured. Beef cattle that may have been on pasture are
often fattened for as long as 150 days in feed lots before slaughter, thus giving
considerable opportunity for elimination and dilution of tissue residues.
D. PLANT UPTAKE
The degree to which 2,3,7,8-TCDD can be taken up into plants has not been well
established. The available literature on this issue is somewhat contradictory.
In one study (Wipf et al., 1982; Wipf and Schmid, 1983), investigators collected soil
and vegetation samples in Seveso, Italy, over the period 1976 through 1979, following the
runaway reaction incident at the IMESCO plant. They found 2,3,7,8-TCDD concentrations
on the order of 1 ppm in plant material in 1976. However, the levels dropped by several
orders of magnitude over the following years. The authors suspect the contamination
was due to 2,3,7,8-TCDD absorption through leaves deposited from local dusts. They also
conducted greenhouse tests using carrots grown in highly contaminated soil collected
from the Seveso area. The 2,3,7,8-TCDD levels in the peeled edible portions of the
43
-------
carrots were approximately 3% of the levels in the soil. On. this basis, the authors
concluded that plant uptake was minimal.
Coc.ucci.et al. (1979) conducted very similar tests using Seveso soils and various
types of vegetation, including carrots,, potatoes,, and onions. In contrast to Wipf et al.
(1982),, they found significant uptake levels [2,3,7,8-TCDD levels (sources unspecified) in
inner parts of the vegetables were approximately equal to levels in the soil].
Young-(1983) studied uptake in perennial grasses and small broadleaf plants located
at a. field in. Eglin Air Force Base, Florida. The field was sprayed with 2,4,5-T during
1962-1970. In 1978 and 1979 the levels of T.CDD (isomers unspecified) in roots (-700 ppt)
were found at levels similar to those in the soil (~500 ppt). Young concluded that this
result suggested a "passive" uptake process, in. which soil particles are incorporated into
the epidermis of the root tissue. The upper portions of the plants were found to have
10 to 75 ppt of TCDD (unspecified isomers).
A study by Facchetti et. al. (1986) looked at the potential for maize and beans to
absorb, translocate, and accumulate 2,3,7,8-TCDD from contaminated soils. Although the
study is somewhat difficult to interpret due to poor translation, the authors appear to
have concluded the following:
(1) Roots had higher levels of 2,3,7,8-TCDD than the surrounding soil.
(2) 2,3,7,8-TCDD levels in above-ground parts of plants did not increase
significantly over time or with, increasing levels of soil contamination (1 to 752
PPt).
(3) 2,3,7,8-TCDD contamination of above-ground parts of plants was due primarily
to volatilization from soil.
(4) 2,3,7,8-TCDD was: lost from the soils over time due to volatilization.
(5) High absorption of 2,3,7,8-TCDD by the roots warrants precautions to be taken
in the consumption of root vegetables such as carrots and potatoes.
44
-------
Isensee and Jones (1971) studied uptake of 2,3,7,8-TCDD into oats and soybeans
from treated soils. They observed a time dependency in which uptake peaked soon after
planting and declined to very low levels by maturity. They concluded that accumulation
of 2,3,7,8-TCDD in plants from soil uptake is highly unlikely. Note, however, that these
are above-ground crops, as opposed to the root crops discussed above.
Sacchi et al. (1986) measured levels of 2,3,7,8-TCDD in bean and maize plants grown
inside green houses using soil which was dosed with various concentrations of
2,3,7,8-TCDD. They found that 2,3,7,8-TCDD accumulated in the aerial parts of the
plants. The accumulation levels generally increased with plant age and soil concentration
levels. The ratio between levels in soil and in aerial parts ranged from about 0.3% - 30%
and varied inversely with soil level. At 1 ppb soil-plant ratio was 1% for beans and 3%
for maize. Substantially lower uptake levels were found when peat was added to the
soil.
Because of the contradictory nature of the current literature, it is not possible at
this time to establish an equilibrium partition ratio between the plant and soil, nor the
rate of plant uptake from soil or other overlying media (air or dust on leaves). If
uptakes such as those observed by Cocucci actually occur in a situation where an
individual obtains a significant portion of his root vegetable diet from a contaminated
home garden, very high risks would result. The findings of Young (1983) and Sacchi et
al. (1986) may indicate that ingestion of certain above-ground plant parts would pose a
smaller, but still potentially significant risk.
45
-------
4. EXPOSURE
This chapter looks at some of the assumptions made in the parameters used to
calculate exposure from various pathways. Specifically, the parameters discussed in this
chapter relate to exposure through inhalation of dust, inhalation of vapor, dermal
contact, ingestion of soil, ingestion of beef and dairy products, and ingestion of fish. As
discussed in Chapter 2, some of the refinements of the past few years (e.g., in treatment
of indoor dust levels, soil contact rates, and consumption rates for fish, meat, and dairy
products) have added to our confidence in the exposure estimates, without substantially
changing the estimates themselves. The refinements in the data on soil ingestion have
resulted in the reduction, by approximately a factor of five, in the estimate of the high
end of the range of the "normal" amount of soil ingested by children while playing. It
should be noted that incidental ingestion of soil is common in children through mouthing
of hands with soil on them or through ingestion of airborne soil. The estimates in this
report are not for the so-called "pica child," who intentionally ingests non-food material
("eating mud pies," etc.).
The exposure routes discussed in this chapter include inhalation of particulates, dust
and vapors, dermal contact, and ingestion of contaminated soil, beef, dairy products, and
fish. A recent evaluation of the significance of inhaling volatilized 2,3,7,8-TCDD in the
vicinity of a contaminated site indicates that this pathway cannot always be treated as
negligible.
A. INHALATION — INDOOR DUST LEVELS VERSUS OUTDOOR LEVELS
For exposure assessments near 2,3,7,8-TCDD-contaminated areas, the level of
suspended particulate matter, and its contaminant content, are major variables in the
calculation of exposure or risk. Usually, when attempting to estimate exposure through a
pathway involving dust intake or dermal contact, not all the factors have been analyzed,
46
-------
and measurements might only be available for 2,3,7,8-TCDD levels in the soil. To gain
information on other variables, such as the suspended particulate matter or dust content,
assumptions are made in order to provide estimates.
In a report on 2,3,7,8-TCDD transport from contaminated sites (U.S. EPA, 1985b), a
technique is presented for calculating conversion factors between soil, air, and sediment.
This report (U.S. EPA, 1985b) provides a methodology for using conversion factors in
calculating air concentrations of 2,3,7,8-TCDD:
cair Gig/m3) = 1(T7 kg/m3 (csoil 0*g/kg)) (4-1)
The authors cite a study by Fred C. Hart Associates, Inc. (EPA 1984b) which indicates
that 2,3,7,8-TCDD levels might increase by a factor of 2 to 12 on small particles as
compared to larger ones. However, U.S. EPA (1985c) report uses the assumption that the
concentration of 2,3,7,8-TCDD adsorbed on suspended particles will be the same as that
in the soil.
Hawley (1985) assumed, based on several other studies in which measurements were
made, that the concentration of suspended particulate matter in indoor air is equal to
75% of that outside. Also, his report stated that most household dust is outdoor dust ,
that is transported into the house, and that only a small percentage is developed from
sources within. He then concluded that 80% of the indoor dust is identical in
contaminant content to outdoor soil. Previous reports (e.g., Schaum, 1984) have usually
assumed that 2,3,7,8-TCDD levels in the soil and dust are equal. This refinement (i.e.,
using Hawley's (1985) method rather than Schaum's (1984) method) should have a minor
effect on the overall exposure estimates. The way this refinement enters into the
exposure assessment is further elucidated in Section C.I. of this chapter.
47
-------
B. INHALATION — VAPORS
The compound 2,3,7,8-TCDD exerts a very low vapor pressure (Freeman and Schroy,
1985a, b; Schroy et al., 1985a; Podoll et al., 1986). However, compounds with low vapor
pressure and low solubility exhibit some properties in the environment that are not
readily apparent from looking at pure-state properties individually. Based on the vapor
pressure consideration, Paustenbach et al. (1986) discounted the importance of
2,3,7,8-TCDD uptake via vapor inhalation in risk assessment evaluation, and assumed that
the human intake via inhalation is related to the intake of airborne, respirable
particulates only. On the other hand, Freeman and Schroy (1985a, b,) and EPA (1985c)
considered the vaporization process to be the most important transport process for CDDs
present in soils, and compared the results of their modeling with the concentration data
obtained at different depths of the soil column and at different times. In addition to
concluding that 2,3,7,8-TCDD can be relatively volatile in an environmental setting, they
presented a set of conditions that will influence volatilization from contaminated spill
sites such as Times Beach, Missouri, and Eglin Air Force Base. Thibodeaux (1983)
compared estimated environmental 2,3,7,8-TCDD exposures from vapor and dust inhalation,
and concluded that vapor inhalation is a significant exposure pathway. Eitzer and Hites
(1986), based on a limited experimental study, found that CDD in the ambient air was
present primarily in the vapor phase (this study is discussed in more detail below).
Despite its low vapor pressure, 2,3,7,8-TCDD can volatilize from spill and disposal
sites, and can be emitted into the air from a variety of other sources, including
incineration and combustion processes, facilities manufacturing PCBs, paper products and
pentachlorophenol, and pyrolysis of PCBs and other chlorinated benzene derivatives
(Radian Corp., 1983; Freeman and Schroy, I985a; Commoner et al., 1985; Czuczwa and
Hites, 1986). For example, Nash and Beall (1980) found ambient air concentrations of
2,3,7,8-TCDD when silvex spiked with 2,3,7,8-TCDD was applied to turf and field sites.
48
-------
The data show that the ambient air concentrations decreased as a function of time after
the application. Initially, concentrations of vaporized 2,3,7,8-TCDD as high as 7.98
x 10" H and 9 x 10"^ g/m^ were measured when the spray material spiked with
2,3,7,8-TCDD at concentrations of 7.5 and 15 ppm was applied to turf in a microcosm
experiment and at the field site. The measurements for the volatilized material were
made immediately after the application. These ambient concentrations were due to
volatilization from the field, which the authors noted was a major pathway of
2,3,7,8-TCDD dissipation.
Czuczwa and Kites (1986) showed that particulates collected from the ambient air in
Washington, D.C., and St. Louis, Missouri, were enriched in octachlorodibenzo-p-dioxin
compared with the congener profiles in the combustion source effluent, which were
reported to be the major source of dioxin constituents in the urban air particulates.
Similar enrichment was noted in congener profiles in surface sediments collected from the
Great Lakes. The authors noted that particulates containing PCDD emitted from
combustion sources would travel through the atmosphere to ultimate environmental sinks,
such as lake sediments.
This section of the report, however, concentrates on risk analysis of contaminated
soils, and therefore does not address the procedure for estimating the amount of
2,3,7,8-TCDD emissions from combustion or pyrolysis sources. Since exposure to
2,3,7,8-TCDD via vapor inhalation is directly proportional to the 2,3,7,8-TCDD
concentration in the ambient air that a person breathes and the air respiration rate
(contact rate), the critical parameter in estimating exposure from vapor inhalation is the
2,3,7,8-TCDD concentration in the ambient air. Although the measured concentration
values can be used directly in exposure evaluation, the analytical difficulty and the cost
associated with obtaining properly quality-controlled concentration data from
measurements for low vapor pressure compounds are extremely disadvantageous.
49
-------
Because of the problems in measurement noted above, models are very convenient
tools for estimating the ambient air concentrations at the site where volatilization
occurs, or at a distance away from the site. The more distant the exposure location is
from the area source where the emissions occur, the less concentrated the contaminant
will be in the ambient air. The major cause for this dilution is the mixing of the
contaminant with the winds, dispersion in the air, and possibly some degradation in the
atmosphere by action of sunlight and free radicals. Since some models describing the
emission rate from soil and the dispersion of volatilized CDD are complex and lengthy,
the reader who is interested in details of the model derivation and different applications
is encouraged to consult the relevant references (U.S. EPA, 1986a; U.S. EPA, 1981c; U.S.
EPA, 1979; Turner, 1970). This part of the report will concentrate on presenting the
model results and pertinent applications.
1. Emission Potential
In estimating the on-site or off-site ambient air concentrations of volatilized
2,3,7,8-TCDD to which people would be exposed, the first task should be to estimate the
emission rate from the contaminated area. As a result of changing concentrations of
2,3,7,8-TCDD in the soil column as the emission proceeds (see Section A of Chapter 3),
no matter how small the vapor pressure is, the emission rate estimation involves
consideration of an nonsteady-state process. Because of this, the emission rate can be
presented either on an instantaneous basis or on an average basis; in the latter case, the
emission rate should be averaged over the time period of interest. For evaluation of
long-term exposure, it is most often appropriate to make use of the average values.
As can be seen in the model presented below, the data requirements for model
estimation of the transient emission rate include the concentration of 2,3,7,8-TCDD in
soil, Henry's law constant (which can also be calculated from vapor pressure and aqueous
solubility), the effective porosity of the contaminated soil media, the diffusivity of
50
-------
2,3,7,8-TCDD vapor in air, the area of contamination, and the depth of contamination.
The data that must be determined on a site-specific basis are the concentration of
2,3,7,8-TCDD in soil, and the area and depth of contamination. The other parameters
can be estimated from data on physical properties, or can be found in the listing of
properties for 2,3,7,8-TCDD in Chapter 2.
Other factors affecting the emission rate are the cover layer of contaminated soil,
2,3,7,8-TCDD concentration profile in soil, the moisture content of the soil, the
biodegradation rate in the soil, and the existence of gases generated by decomposing
material. While cover material can reduce the emission rate, once the cover material is
saturated it loses the capability for adsorption, and any enhancement in retardation is
due to the increased path length of vapor diffusion. (See discussion in Section A.2. of
Chapter 3). Vaporization of moisture and generation of other gases from decomposing
material will lead to convective transport as well as increased diffusion rate and thus
increase the emission rate.
When it can be assumed that 2,3,7,8-TCDD initially contaminates soil from the soil
surface down to a specified depth, with a uniform concentration distribution along the
soil column, the time-averaged emission rate, ND, in grams per cm^ per second, can be
estimated from U.S. EPA (1986a) and Hwang and Falco (1986) as follows:
ND = 2Dj E4/3 Kas Cso/>/3.14aT (4-2)
The parameters in Equation 4-2 can be defined as follows: D] = the molecular diffusivity
of dioxin vapor in air (= 4.7 x 10 cm~/s), E = the porosity of soil, T = the exposure
duration (seconds), Kas = the air/soil partition coefficient (mg/cm3 air/mg/g soil), Cso =
the initial 2,3,7,8-TCDD concentration in soil (g/g), and a (cm^/s) is
51
-------
o = Di E4/3/[E + ps(l - E)/Kas] (4-3)
where ps = the true density of soil (g/cnP).
For wet soils, 2,3,7,8-TCDD is partitioned between the soil, water, and air phases.
Henry's law constant and the soil/water partition coefficient can be used to describe
partitioning of 2,3,7,8-TCDD between the soil and air phases, or
Kas = 41 Hc/Kd (4-4)
where Hc = Henry's law constant (1.6 x 10~^ - 4.6 x 10^ atm m^/g mol) (Podoll et al.,
1986; Schroy, 1985a); K
-------
It should be noted that the vapor emission rate changes over time due to changes
in the concentration profile of 2,3,7,8-TCDD along the soil column. Equation 4-2 is used
to estimate the emission rate averaged over the time period of exposure. Freeman and
Schroy (1985a, b,) and EPA (1985c) showed, by modeling a heat balance around the soil
element, that the temperature at the soil surface fluctuates appreciably in a period of 1
day, while the fluctuation diminishes at greater depths. They concluded that these
temperature fluctuations may cause important changes in the volatilization rate. They
also showed that the concentration profile in the soil column changes with time.
Assuming that these changes in the soil concentration profile are due to volatilization
and diffusion, the initial concentration (Cso at t = 0, in Equation 4-2) can be represented
by the maximum value of the concentration profile derived from sampling done a long
time after contamination occurred. The total rate of emission, Q (g/s), over an area, A
(cm^), of contamination can be obtained from
Q = A . Nd (4-5)
It is preferable not to use the temperature at the soil surface in choosing physical
parameters, because the emission results from a diffusion process occurring below the
surface. In reality, the 2,3,7,8-TCDD concentration at the soil surface will rapidly
approach zero following initial deposition (assuming that vaporization/photolysis losses
occur more rapidly than erosion), although the 2,3,7,8-TCDD in the bulk soil may have
volatilized very little. The temperature to be used in choosing values for physical
parameters to calculate the effect on emission rate should be representative of the
contamination depth.
53
-------
2. Dilution of Emissions in Ambient Air
The 2,3,7,8-TCDD emanating from the soil surface in vapor or particulate form are
diluted by winds at the site, and are transported through the atmosphere until they are
ultimately reduced in concentration, possibly destroyed through photolytic reactions, or
deposited on land or in water bodies. The atmospheric phenomena which influence the
dilution, transport, and transformation of 2,3,7,8-TCDD emissions can change extensively
as a function of time and location. Although model simulations may be sufficiently
representative, such variations in atmospheric phenomena can lead to uncertainties of
more concern than those associated with the models themselves. These uncertainties will
arise from assumptions of estimates regarding (1) mixing height, (2) meteorological
conditions (wind speed and direction), (3) stability conditions, and (4) atmospheric
reactions.
The maximum ground-level concentrations resulting from point-source stack
emissions occur some distance from the source. In contrast, ground-source emissions
occurring over a large area cause higher ground-level concentrations at the site of
emission, and become more diluted as they are transported downwind of the source.
Because of the higher concentration of the contaminant on-site than off-site, the
population residing on-site will normally have higher levels of exposure from vapor
inhalation than those off-site. If concerns relating to emission impacts arise for people
living on or having access to the site, it may then also be necessary to conduct an
exposure evaluation for the population located downwind of the site.
Gaussian dispersion modeling cannot be used in estimating the on-site concentrations
or the receptor concentrations within 100 meters of the center of the facility. Despite
the importance of the contaminant dispersion phenomena at the short-range or on-site
locations, experimental data are lacking with which to calibrate the dispersion
coefficients needed in the short-range dispersion equation derived by integrating point
54
-------
source modeling over the area of emission (Hwang, 1987). The on-site or short-range
concentrations along the center of an area source can be estimated from
9
C= v/2/3.14 ((q)a//«rz) EXP (-i(z/az)2) erf (b/2v/2ay) (4-6)
where a, b = dimensions of the site parallel and perpendicular to the wind direction,
respectively; q = emission rate per unit area; u = average wind speed; Sy and Sz are the
horizontal and vertical dispersion coefficients in air, respectively; z = receptor height,
and it is assumed that the source is at ground level.
Under the assumption that a contaminant present at the site has traveled only a
short distance before it is inhaled by the population, the box-model approach considers
mixing of the emitted contaminant with the winds, ignoring the dispersion effects. Based
on the use of an average value for wind speeds that varies logarithmically with respect
to height, and for a mixing height, this approach provides an estimate of the on-site
concentrations of 2,3,7,8-TCDD in ambient air, as follows:
On-site Ca = Q/[(LS)V(MH)] (4-7)
where Ca = the ambient air concentration of 2,3,7,8-TCDD at the exposure location
(g/m^), Q = total emission rate (g/s), LS = an equivalent side length of the site
perpendicular to the direction of the winds (m), V = the average wind speed at the
inhalation height (about 2.2 m/s for 10-mph winds), and MH = the mixing height before
being inhaled by an individual (m).
Two approaches are possible for estimating the ambient air concentrations of
2,3,7,8-TCDD at off-site locations. In the current state of model development, both
approaches make use of virtual source approximations for area sources. The first
t
approach requires extensive site-specific meteorological data, which are employed in using
55
-------
refined computer models. These data are used to sum the frequency distribution of wind
speed for each radial/concentric sector of exposure around the source, and to evaluate
the vertical standard deviation of the plume for each observed stability class on an
annual basis. The details of the equation can be found elsewhere (U.S. EPA, 1986a).
The second approach, which is applicable for estimating concentrations at distances
of 100 meters or more from the source, uses an approximation of the model for a quick
assessment. The simplified form of the equation is given in Turner (1970) as:
Off-site Ca = 2.03 (Q/LvazV) (4-8)
where Lv = the total virtual downwind distance to the receptor (m), obtained from Lv =
L + 2.5S; L = the distance from the center of the facility to the receptor; S = the width
of the facility perpendicular to the wind direction, (m); Ca, Q, and V are as defined for
Equation 4-7, and oz = the vertical standard deviation (m). The values for sz can be
found in most standard air pollution textbooks in graphic or formula form (Turner, 1970;
U.S. EPA, 1986a). In estimating the annual average values, the concentration obtained by
Equation 4-8 is often corrected by multiplying by a frequency factor with which the wind
blows toward a particular sector of interest.
Introduction of the dilution factor, essentially a conversion factor, makes the
estimation of on-site ambient air concentrations very convenient. The procedures
involved in calculating the emission rate and the ambient air mixing can be combined
into a single step. The dilution factor, D, can be defined as
ambient air concentration at exposure location (4-9)
air-phase concentration of TCDD at the soil surface
from which emissions originate
The vapor-phase concentration, which is assumed to be in equilibrium with the
56
-------
contaminated soil, will still reach the soil surface by the diffusion process along the soil
column. Consequently, the vapor-phase 2,3,7,8-TCDD at the soil surface can be given as
C = KasCso (4-10)
Once the dilution factor is evaluated, the ambient air concentration at the exposure point
can be obtained by multiplying it by the 2,3,7,8-TCDD concentration at the soil surface
corresponding to the initial 2,3,7,8-TCDD concentration in soil. In order to do this, one
can combine Equation 4-2 for emission rates in the case of surface contamination, and
Equation 4-8 for on-site mixing, to yield
n_ Xa_ _ 0,02 Dj.E4/3 A (4-11)
CsoKas (LS)V(MHV3.14aT
where A is the area of contamination (m2), and all other units are as given before (i.e.,
Dj in cm2/s, LS in m, V in m/s, MH in m, a in cm2/s, and T in seconds; note mixed
units for the length).
For example, suppose the average 2,3,7,8-TCDD concentration on a contaminated site
is initially 0.05 Mg/g- At the values for T = 2.2 x 109 s, A = 2,024 m2, LS = 45 m, Dj
= 4.7 x 10~2 cm2/s, V = 2.25 m/s, MH = 2 m, and Kas = 4 x 10'7 g soil/cm3 air, the
air-phase 2,3,7,8-TCDD concentration at the soil-air interface, C, and the dilution factor,
D, are calculated to be 2 x 10~2 Mg/m3 and 5.4 x 10~4, respectively. is 2 x 10~9 as
calculated from Equation 4-3. In this calculation, the values E = 0.35 and ps = 2.65 are
used. From these two results, the 2,3,7,8-TCDD concentration in the ambient air at the
exposure point is found to be 1.1 x 10"^ g/m3.
3. Exposure Estimation
The estimation of average lifetime exposure via vapor inhalation requires
information on the contact rate, the duration of exposure, and the fraction of
2,3,7,8-TCDD vapor absorbed through the lungs upon inhalation, as well as the ambient
57
-------
air concentrations being inhaled. The contact rate is the air respiration rate for
children and adults, and should be consistent with the air respiration rate discussed for
dust inhalation. The exposure lifetime period is normally assumed to be 70 years, as
discussed in the section dealing with dust inhalation (Section A of Chapter 4).
The rate of absorption of vapor-phase 2,3,7,8-TCDD through the membranes of
human lungs into the bloodstream has not been adequately addressed in the literature.
Such a toxicokinetic analysis may require consideration of the diffusional process through
the cell membranes of the trace amount of 2,3,7,8-TCDD vapor that entered the alveoli.
The diffusion processes for lung absorption include both passive and facilitated transport
as well as active transport, while the diffusion process through the gut (small intestine)
mostly involves osmosis.
In view of the data showing that 2,3,7,8-TCDD tightly bound on the soil is absorbed
through the gastrointestinal (GI) tract at the rate of 20% to 26% (Poiger and Schlatter,
1980), the vapor-phase 2,3,7,8-TCDD, which is not bound on adsorptive solids, should find
its way through the single-layer membrane between the air space and blood capillaries in
excess of the rate reported for the GI tract. Until rigorous kinetic models or
experimental data are available, it appears reasonable to assume 50% to 100% absorption
of inhaled 2,3,7,8-TCDD vapor through the thin membrane layer, which at the alveoli and
blood capillaries contains both hydrophilic and hydrophobic components.
The average lifetime exposure for the vapor inhalation pathway can be estimated as
follows:
Average Ambient air concentration x respiration rate x exposure
lifetime duration x absorption fraction (4-12)
exposure ~ Body weight x lifetime
The body weight and respiration rate vary from childhood to adulthood. An
approximation for these two variables is possible based on the assumption that changes in
the body weight and respiration rate are proportional. In this case, the body weight and
58
-------
respiration rate for an average adult male can be used for lifetime exposure estimation.
4. Effect of Photodegradation on Exposure Estimation
Once emitted, 2,3,7,8-TCDD vapors or particulates can be destroyed in the
atmosphere through photodegradation. This may occur by direct absorption of sunlight
energy, resulting in the breakdown of 2,3,7,8-TCDD by photolysis, or through
photo-induced oxidation by the reaction of 2,3,7,8-TCDD with free radicals present in the
atmosphere. Photolysis requires sunlight and is facilitated by availability of a hydrogen
donor (Crosby and Wong, 1977). Also, the presence of a solvent on soil appears to make
adsorbed compounds more available for photolysis.
Podoll et al. (1986) compared the half-lives of 2,3,7,8-TCDD vapors under the
conditions of photolysis and hydroxyl (OH) radical oxidation in the atmosphere. In
estimating the half-lives for photolysis of 2,3,7,8-TCDD vapor, the authors used the
quantum yield observed in hexane and estimated an upper-bound photolysis rate of t^ =
58 minutes. At an average concentration of OH radical of 3 x 10"^ M in the
atmosphere, they also estimated the half-life for oxidation of 2,3,7,8-TCDD vapor by OH
radicals as t^ = 200 hours. The authors presented these results with reservations,
suggesting the need for an accurate measurement of the vapor-phase quantum yield for
2,3,7,8-TCDD.
Crosby and Wong (1977) conducted experimental photolysis studies for 2,3,7,8-TCDD
to evaluate its photolysis rates in different herbicide formulations and on different
surfaces. The half-lives for loss of 2,3,7,8-TCDD from herbicide formulations on glass
surfaces ranged from 2 to 6 hours. Crosby and Wong (1977) reported that thin films of
2,3,7,8-TCDD on glass plates were found to be stable in sunlight for at least 14 days.
Note that this test did not use herbicide formulations. They determined the half-lives
on leaves as about 1 to 2 hours, and those on soil as longer than 7 hours. All
experiments were conducted under natural sunlight without using organic solvents.
59
-------
Accounting for daily and annual fractions of sunlight, Thibodeaux and Lipsky (1985)
-adjusted the 6-hour half-life derived by Crosby and Wong to obtain an effective photo-
degradation half-life of 7.2 days.
These findings appear to indicate that photolysis plays a role in the degradation of
2,3,7,8-TCDD only under certain conditions. The use of solvents, such as olive oil or
hexane, can enhance the photolysis of 2,3,7,8-TCDD on solid surfaces. The surface itself,
and perhaps the organic films present on that surface, are factors influencing the
photolysis rate; for example, 2,3,7,8-TCDD was stable on glass surfaces without being
affected by photolysis, but underwent photolytic degradation when organic solvent film
was present on solid surfaces. This suggests that in the environment, 2,3,7,8-TCDD on a
clean surface, such as glass, may react quite differently from 2,3,7,8-TCDD on soil or
leaf surfaces with regard to photolysis.
A preliminary photolysis experiment using 2,3,7,8-TCDD adsorbed on fly ash
particulates suspended in recirculating air indicated that the photolysis of 2,3,7,8-TCDD
in particulate form underwent virtually no photolytic reactions after 30 hours of
illumination (Mill et al 1987). Further experiments will be needed to compare the
relative magnitude of the photolysis rates for 2,3,7,8-TCDD vapors and particulates.
Eitzer and Kites (1986) reported that 2,3,7,8-TCDD in the atmosphere is all in vapor
form. The vapor was captured by adsorption on polyurethane foam. They collected
ambient air particulates using a high-volume sampler and 0.1-um pore size filters, and
could not detect 2,3,7,8-TCDD in the particulates. The study did not determine how
much 2,3,7,8-TCDD may have been present on the particles measuring less than 0.1 um.
However, if it is true that 2,3,7,8-TCDD is present primarily in the vapor phase, then the
disappearance of 2,3,7,8-TCDD in the atmosphere would be controlled by the vapor-phase
60
-------
photolysis. Further study is needed to confirm these rather unexpected results before
final conclusions can be drawn.
Photolysis is generally assumed to decrease the ambient air concentrations that are
subject to human inhalation. However, higher chlorinated CDDs could degrade to 2,3,7,8-
TCDD and the degradation products of 2,3,7,8-TCDD could still be toxic. In the absence
of further experimental data under sunlight conditions, it appears that a reasonable value
for 2,3,7,8-TCDD vapor-phase half-life is in the range of 2 to 6 hours. This range of
half-life is supported by recent experimental results (Mill et al 1987) on the vapor phase
photolysis of 2,3,7,8-TCDD under simulated sunlight conditions, where results show the
vapor-phase half-life being several hours rather than the fly ash particulate half-life of
several hundred hours as noted above. Even with this relatively short vapor-phase half-
life, the on-site ambient air concentrations, or nearby off-site ambient air concentrations
will not have degraded significantly before the population breathes the air. This is
because the time required for the 2,3,7,8-TCDD vapor emissions from the soil surface to
reach a human or structure is short compared with the half-life.
For ambient air concentrations at many off-site locations, however, degradation by
photolysis could be significant. The distance at which the original 2,3,7,8-TCDD
emissions would have degraded will mostly depend on wind speed and direction, particle
size, and the intensity of light. For example, with winds at 10 miles per hour, the
original amount of 2,3,7,8-TCDD vapor would have been degraded by one-half at a
distance of about 20 miles from the site, if it is assumed that the half-life is 2 hours.
In estimating exposure to 2,3,7,8-TCDD vapor of the population residing at a
distance from the contaminated site, the ambient air concentration estimated by Equation
61
-------
4-8 should be corrected by the amount of photodegradation. For first-order
photodegradation kinetics, this correction is exponential with respect to the travel time:
^corrected = C* ''* (4-'3>
The corrected concentration, as given by Equation 4-13, should be used in place of
Equation 4-8 in subsequent exposure evaluations for the off-site population for which
photodegradation is significant. In Equation 4-13, k is the first-order rate constant (=
0.12 hours"', based on a half-life of 6 hours), and t is the time-of-travel for
2,3,7,8-TCDD vapor to reach the receptor by blowing winds. The effect of
photodegradation could be used in evaluating risk associated with the long-range
transport of 2,3,7,8-TCDD in particulate or vapor form in the atmosphere. For the
short-range transport of the contaminant in the atmosphere (e.g., within a S-mile radius
from the site), the effect of photodegradation on the results of exposure evaluation can
be discounted.
C. INHALATION — PARTICULATES
Dust emissions and the resulting TCDD emissions occur as a result of vehicular
traffic, loading and unloading operations, spreading operations, transportation in trucks,
and wind erosion. Emissions from vehicular traffic is pertinent to the contaminated soil
scenarios where the possibility of traffic over the site can be considered. The emissions
due to wind erosion are possible from contaminated soil or ash disposal sites where the
contaminant surface is exposed to the wind. Emissions from loading and unloading
operations, spreading operations, and during transportation in trucks would be more
appropriate in the scenarios dealing with fly ash disposal which typically requires these
operations on a frequent basis. Such operations may also occur at contaminated soil and
landfill sites.
62
-------
If dust emissions from all the operations and phenomena occur simultaneously, the
total emissions for each landfill scenario are the sum of the vehicular traffic generated
emissions, the emissions resulting from unloading, the emissions from spreading, and the
emissions from wind erosion. Emissions during transportation in trucks and loading
operations may affect populations far from the landfill site as well as those nearby.
Assumptions pertaining to each scenario are further described later.
1. Vehicular Traffic
The emissions from a dump or landfill site due to vehicular traffic can be estimated
from an emission factor. This factor can be found in U.S. EPA (1985), and takes the
form of
Ev = K(1.7 (s/12)(s/48) (W/2.7)0-7 (w/4)°-5 (365-p/365) (4-14)
where: Ev = emission factor (kg/VKT)
k = particle size multiplier (0.8 for particulates < 30 um)
s = silt content of road surface material (%)
S = mean vehicle speed, km/hr
W = mean vehicle weight, MG
w = mean number of wheels
p = number of days with at least 0.254 mm
(0.01 in) of precipitation per year
This emission factor is provided in units of kilogram of particulate emitted per
vehicle kilometer traveled (kg/VKT). Once the emission factor is estimated, the vehicle
kilometers traveled in a day for each of the scenarios need to be estimated to get this
emission rate. VKT equals the distance (km) traveled by all the vehicles which will pass
through the landfill or contaminated site. Equation 4-14 can be used to solve for the
63
-------
quantity of particulate emitted if the VKT is known and the other parameter values can
be defined. U.S. EPA (1985) provides 4.3 - 20% as the range of road silt content.
2. Loading and Unloading Operations
Emissions during loading and unloading of contaminated soil or fly ash can be
estimated using an equation given in U.S. EPA (1985). The emission factor equation is
E = k(0.0009)[(s/5)(u/2.2)(H/1.5)/(M/2)2 (y/4.6)0-33] (4-15)
where: E = emission factor (kg/Mg)
k = particle size multiplier (dimensionless)
s = material silt content (%)
U = mean wind speed, m/s
H = drop height, m
M = material moisture content (%)
Y = dumping device capacity, m3
Equation (4-15) provides an emission factor for kilograms of particulate emitted per
megagrams (Mg) of soil unloaded or loaded. For fly ash disposal, the amount of soil that
is transported to the landfill can be considered the same as the fly ash generation rates.
The particle size multiplier varies with aerodynamic particle size and is given numerical
values of 0.73 and 0.77 for batch drop and continuous drop operations for particle sizes
less than 30 um. The other parameters in Equation 4-15 can be defined as part of the
scenario.
64
-------
3. Spreading Operations
t
It is assumed that spreading operations cause emissions on the same order as
agricultural tilling. The emission factor derived from U.S. EPA (1985) is used to estimate
emissions:
ES = k(5.38)(s)°-6 (4-16)
where Eg = emission factor (kg/ha)
k = particle size multiplier
s = silt content
The particle size multiplier (k) varies with aerodynamic particle size, and is given as 1
for total particulate and 0.33 for particulates less than 30 um.
4. Transportation in Trucks
After the ash is discharged from the storage bin to trucks, emissions can be
controlled by hauling the fly ash in closed trucks or open trucks that use water or other
wetting agents for dust control. It is assumed that dry ash collection is used, ash was
transported to landfills in open trucks, and the ash is wetted to control emissions prior
to being transported to the landfill.
Currently, no emission factors are available for estimating emissions from open
trucks. It is assumed that each truckload of ash can be treated as an individual
aggregate pile. The emission factor for estimating emissions from open trucks is
ET = 1.9 (s/1.5)[(365-p)/235](f/15) (4-17)
where: E = total suspended particulate emission factor (kg/d/ha)
s = silt content of aggregate(%)
65
-------
p = number of days with > 0.25 mm (0.01 in) of precipitation
per year
f = percentage of time that the unobstructed wind speed exceeds
5.4 m/s at the mean pile height
For wetted fly ash, the value for p can be assumed to be 364 days based on the
assumption that the surface of the ash is wetted prior to hauling to the landfill. Since
the truck is normally moving at a speed faster than 12 mph, the value for f is 100%.
The emission rate from transporting soil or fly ash in trucks can be estimated using
the following steps:
a. Estimate the daily number of truckloads of ash transported by first
converting the daily quantity of ash loaded at each facility from kg to cubic
meters. Next divide this quantity by the capacity of each truckload (typically
12 m^). This provides the daily number of truckloads of ash transported to
the landfill.
b. Estimate the surface area of fly ash in each truckload capacity.
c. Estimate the quantity of ash emitted per minute of travel time by multiplying
the surface area of each truckload by the emission factor.
d. Estimate the emission rate of particulate matter by multiplying the quantity
of particulate emitted per minute from each truck by the travel time by the
daily number of loads.
5. Wind Erosion
A method for estimating dust emissions generated by wind is described below. This
method assumes that the uncrusted contaminated surface is exposed to the wind and
consists of finely divided particles. This creates a condition defined by U.S. EPA (1985a)
as an "unlimited reservoir" and results in maximum wind-caused dust emissions. Surface
66
-------
activities such as described above would increase the emissions, and these should be
computed separately if they occur.
The flux of dust particles less than 10 um from surfaces with an "unlimited
reservoir" of erodible particles can be estimated as shown below (U.S. EPA, 1985a):
E = 0.036 (1 - V) (Um/Ut)3 F((x) (4-18)
where: E = total dust flux rate of < 10 um particle (g/m^ . hr)
V = fraction of vegetation cover
Um = mean annual wind speed (m/s)
Ut = threshold wind speed (m/s)
F(x) = a function specific to this model.
Threshold wind speed (Uj) is the wind velocity at a height of 7 meters above the
ground needed to initiate soil erosion. It depends on nature of surface crust, moisture
content, size distribution of particles, and presence of non-erodible elements. It can be
estimated on the basis of the following procedure (U.S. EPA, 1985a):
a. Determine the threshold friction velocity. This is the wind speed measured at
the surface needed to initiate soil erosion. For "unlimited reservoir" surfaces,
U.S. EPA (1985a) suggests that this velocity is less than 75 cm/s.
b. Estimate the "roughness height." This is a measure of the roughness of the
surface.
c. Estimate ratio of threshold wind speed at 7 m to friction velocity, by using a
chart provided by U.S. EPA (1985a).
d. Estimate threshold wind speed by multiplying the friction velocity by the ratios
described in step (c.)
67
-------
Finally, F(x) is determined by first calculating the dimensionless ratio, x = 0.886
Ut/Um and finding F(x) from a chart of F(x) versus x, as provided in U.S. EPA (I985a).
The dust flux is converted to an emission rate as follows (U.S. EPA, I985a):
Q = CSEA (1 hr/3,600 s) (4-19)
where:
Q = TCDD emission rate (ng/s)
Cs = TCDD concentration in soil (ng/g)
/
A = site area (m2).
D. DERMAL — SOIL CONTACT RATES AND DERMAL ABSORPTION
1. Contact Rates
Determining the amount of soil with which individuals might come in contact is a
critical first step in obtaining estimates of dermal absorption or soil ingestion. Among
the factors that must be considered for a relevant estimate of exposure by the dermal
route are type of soil (which determines the bioavailability of a compound), exposed skin
areas, soil contact period, soil and/or skin moisture, age of individuals (which influences
the permeability of the skin), temperature, and location on the body of the soil contact.
Surface area data can be readily obtained from various sources, for example, Snyder
(1975) or U.S. EPA (1985a). The latter reference is an investigation and review of
literature to provide information on statistical distributions and ranges of standard
factors used to support EPA's Guidelines for Estimating Exposures. It discusses
measurement techniques and contains formulae along with summary tables on body surface
areas on the basis of age, sex, and body parts or areas.
Measurements of the amount of soil which adheres or accumulates on skin surfaces
have been conducted by Lepow et al. (1975) and Roels et al. (1980). Lepow et al. (1975),
68
-------
using an adhesive, found an average of 11 mg of dirt per 21.5 cm2 on the hands of
children. Roels et al. (1980) obtained similar results from a study with older children
(11-year-olds) using a technique involving the lead content of the dirt. In the above
studies, the amount of adhered soil was 0.5 mg/cm2 and 0.6 mg/cm2, respectively.
Hawley (1985) used the results of Lepow et al. (1975) and Roels et al. (1980) in his
assessment of risk from exposure to contaminated soil, and used a value of 0.51 mg/cm2.
This value was taken as the soil covering for estimating exposure to children playing
outdoors. For adults, Hawley (1985) assumed a value of 3.5 mg/cm2 from doing yard
work. Schaum (1984), after considering Snyder (1975), Lepow et al. (1975) and Roels et
al. (1980), assumed a contact range of 0.5 to 1.5 mg/cm2 and that this range also
represents an average for the entire exposed area of the human body for both adults and
children.
The duration (that is, contact time per event [hr/d] times frequency of the event
[d/yr]) of the exposure to the contaminated soil is an important determination. Schaum
(1984) adopted a range of 247 to 365 d/yr as the exposure frequency. Paustenbach et al.
(1986), in their examination of assumptions used for risk calculations, stated that it
appeared that the Centers for Disease Control (CDC) assumed exposure to soil for about
180 d/yr. Hawley (1985) presented a reasonable description and separation of the
exposure duration. He assumed outdoor exposure of 5 days/week for 6 months for young
children (2i years of age), with a 12-hour contact time (since children often retain soil
on their skin after coming indoors). This corresponds to about 130 d/yr. For older
children, Hawley (1985) assumed an average outdoor playtime of 5 hr/d from May to
September, or 150 d/yr. Adults were assumed to have about 43 d/yr of outdoor soil
exposure, based on 8 hours of yard work 2 d/wk for 5 months.
A similar issue is the length of time the soil is in contact with the skin surface.
This is an important factor, since it will help determine the amounts absorbed. The CDC
69
-------
(Kimbrough et al., 1984) assumed that the contact period was sufficient to cause 1%
absorption. This assumption was based on work by Poiger and Schlatter (1980), who
found 0.05 to 2.2% absorption in rats from a soil and water paste after 24 hours.
Schaum (1984) assumed 24 hours of contact for the days of exposure. In contrast,
Hawley (1985) assumed 12 hr/d of contact for children and 8 hr/d for adults.
Most of the above determinations and assumptions have recognized that contact
with soil is dependent on various factors, including weather, age, and activity patterns of
subgroups or individuals.
The skin surface area which is exposed to soil may vary because of age, type of
activity, or outside conditions such as temperature. Schaum (1984) used values developed
by Sendroy and Cecchini (1954); that is, 2,940 cm2 for an adult wearing a short-sleeved
shirt and no gloves, and 910 cm2 for an adult wearing a long-sleeved shirt and gloves.
For children, Schaum (1984) computed the exposed surface area by multiplying the above
values by the ratio of a child/adult total surface area. Hawley (1985) derived surface
areas of various body parts from Diem and Lentner (1973) and Berkow (1924), and then
made assumptions about the body parts which might be exposed. For young children, he
assumed that both hands and the legs and feet would come into contact with soil, giving
a surface area of 0.21 m2. For older children, he assumed soil contact over both hands,
the forearms, and the legs from the knees down (0.16 m2). For adults, Hawley (1985)
assumed contact on both hands and the forearms (0.17 m2), estimating 3.5 mg/cm2 of soil
on the skin for adults.
Obviously, these surface area designations will change according to the attire of the
individuals or assumptions made as to which body areas come in contact with soil.
So far, only outdoor exposure has been considered in this section. If the soil in
the surrounding area is contaminated, then it can be assumed that soil is carried into the
house, and that household dust is contaminated to significant levels approaching those
70
-------
occurring outdoors. Hawley (1985) assumed that about 80% of indoor dust is identical in
contaminant content to outdoor soil, and that the concentration of suspended particulate
matter indoors is three-quarters of that outdoors. He also assumed that the dust
covering indoor surfaces was 560 mg/m , based on the assumption that the indoor
dustfall rate was 20% of that outdoors, or 80 mg/m^/d and that cleaning took place once
every 2 weeks. Dermal contact, therefore, to children indoors, according to Hawley
(1985), was 560 mg/m^ on a surface area of 0.05 m^. The exposed area was taken as
half that of the feet, hands, and forearms. The contact area for older children was
assumed to cover both hands (0.04 m^) and to continue for 4 hours/day. Adults were
assumed to be in contact with an area equivalent to the surface area of their hands (0.09
m2).
2. Absorption
The skin constitutes a major interface between humans and the environment, and
influences percutaneous absorption by physical, physiological, and biochemical means.
The factors that affect transfer across this membrane, and which can modify the
cutaneous penetration rate, include the intrinsic properties of the skin itself (such as
age, location on the body, and any modifications from trauma or disease), in addition to
the environmental conditions of exposure (temperature, humidity, concentration gradient,
and duration of exposure). All of these factors can influence the permeability of the
skin to various contaminants. In particular, an increase in any of these factors would
cause an increased rate of contaminant transfer.
Suskind (1977) states that there are two major routes of penetration: the epidermis
itself, and the hair follicles and sebaceous glands. According to Suskind, the latter are
particularly important in initial or transient exposures, and also in cases where
chlorinated hydrocarbons and dioxins seem to impact the skin (for example, producing
chloracne at those sites). Therefore, the skin may need to be considered as a target
71
-------
organ. Tschirley (1986) reported that in an experiment with human volunteers where
2,3,7,8-TCDD was applied in a dose of 3 to 114 ng/kg to their skin, no chloracne
developed. However, at a dose of 107,000 oig/kg, ,80% of the subjects developed
.chloracne.
Many investigators believe that the skin may be the most sensitive index of
exposure to CDDs, especially in cases of accidental or occupational exposure (Brodkin and
Schwartz, .1984; Tschirley, 1986; Dunagin, 1984). In fact, there are studies investigating
'the use of human epidermal cell cultures as indicators and models of toxicity to various
compounds (Greenlee et al., 1985; Osborne and Greenlee, 1985). Kao et al. (1985) studied
the metabolic processes in the skin itself, and the ways in which these processes may
influence percutaneous absorption.
With regard to the absorption of 2,3,7,8-TCDD, Poiger and Schlatter (1980) looked
at the radiolabeled amounts of the compound in the liver of rats as an indication of its
uptake. When 2,3,7,8-TCDD was dermally applied as a pure compound (in methanol) for 24
hours, they found that the highest liver content was 14.8% of the dose applied to the
skin. When 2,3,7,8-TCDD was applied in a mixture of soil, Poiger and Schlatter (1980)
found the liver content to be significantly less than half of that measured in the
previous test. When 2,3,7,8-TCDD was applied in a soil-water paste, the liver content
varied with dose from 0.05% to 2.2%. Schaum (1984), using the information reported by
Poiger and Schlatter (1980) to the effect that 70% of total body burden is found in the
liver, modified the range for dermal absorption to 0.07% to 3%.
Hawley (1985) first considered studies by Bartek et al. (1972) and Feldman and
Maibach (1970) on dermal uptake of various compounds in humans when applied as pure
compounds or in acetone for 24 hours. On the basis of these studies, he assumed the
percutaneous absorption rate to be 11% in 24 hours for adults. Therefore for a 12-hour
contact time, the rate would be about 6%. For children, Hawley (1985) assumed the
72
-------
absorption rate to be twice that of adults. He then considered the results of Poiger and
Schlatter (1980) and modified this rate of 1.8% for children and 0.9% for adults on the
basis of a soil matrix effect (i.e., bioavailability) of 15% and a contact time of 12 hours.
3. Summary
There are many variables and assumptions that need to be considered in addressing
dermal exposures. For example, absorption is probably not a linear process, and the
assumption of a fractional contact time directly proportional to absorption in 24 hours
may not be completely valid. However, there is little evidence to argue otherwise for
that assumption and others discussed here. Most of the above-referenced studies have
presented logical approaches to dealing with percutaneous absorption, and their
conclusions should be of considerable assistance.
Absorption from soil contact can therefore be estimated as 0.9% for adults and 1.8%
for children; or as a range of 0.07% to 3%, as given by Schaum (1984), with no
distinction as to age. The duration of contact, both in terms of physical contact and
yearly exposure, can be estimated as 12 hr/d for children and 8 hr/d for adults; and this
occurs for about 140 days for children and about 45 days for adults. These yearly values
can be modified in the case of rural populations. The amount of soil which accumulates
or adheres on the skin can be considered as 0.5 to 0.6 mg/cm^ for children and 1.8 to
3.5 mg/crcr for adults doing yard work.
E. INGESTION — SOIL
Until recently, only speculative information on the levels of soil ingestion by young
children has been available. During the past 2 years, studies have been conducted using a
new methodology, that of measurement of trace elements present in soil and believed to
be poorly absorbed in the gut. While the initial trace element studies must still be
considered preliminary, a more quantitative basis is being developed for assessing soil
ingestion.
73
-------
1. Available Studies
Binder et al. (1986) studied the ingestion of soil in children 1 to 3 years of age
wearing diapers. The children studied were part of a larger study of residents living
near a lead smelter in East Helena, Montana. Soiled diapers were collected over a 3-day
period from 65 children (42 males and 23 females), and composited samples of soil were
obtained from the children's yards. Both excreta and soil were analyzed for aluminum,
silicon, and titanium, elements thought to be poorly absorbed in the gut and to have
been present in diet only in limited quantities, making them reasonable to use as tracers
in a mass-balance calculation. Both soil and excreta measurements were obtained for 59
children. Using a standard assumed fecal dry weight of 15 g/d, soil ingestion by each
child was estimated using each of the three tracer elements (assuming no absorption or
non-soil source of these elements). The average quantity of soil ingested by the children
was estimated at 121 mg/d [range 25 to 1,324] (aluminum tracer); 184 mg/d [range 31 to
799] (silicon tracer); and 1,830 mg/d [range 4 to 17,000] (titanium). The overall soil
ingestion estimate based on the minimum of the three individual element ingestion
estimates for each child was 108 mg/d [range 4 to 708].
The authors were not able to explain the difference between the results for
titanium and for the other two elements. The frequency distribution graph of soil
ingestion estimates based on titanium shows that a group of 21 children had particularly
high titanium values, > 1,000 mg/d; the remainder of the children showed titanium
ingestion estimates at lower levels, with a distribution more comparable to that for the
other elements.
Clausing et al. (1986) conducted a soil ingestion study with Dutch children, using a
tracer element methodology similar to that of Binder et al. (1986). Aluminum, titanium,
and acid-insoluble residue (AIR) contents were determined for fecal samples from
children, ages 2 to 4, attending a nursery school, and for samples of playground dirt at
74
-------
that school. Fecal samples were obtained daily over a 5-day period for the 18 children
examined. Using the average soil concentrations present at the school, and assuming a
standard fecal dry weight of 10 g/d, the authors calculated mass-balance estimates of soil
ingestion for each material: aluminum, average 230 mg/d [range 21 to 878]; AIR, average
127 mg/d [range 48 to 362]; titanium, average 1,080 mg/d [range 53 to 9,588]. As in the
Binder et al. study, a fraction of the children (5/19) showed titanium values of well
above 1,000 mg/d, with most of the remaining children showing substantially lower values.
Based on the minimum of the three chemical measurements for each child, an estimate of
100 mg/d, with a range of 21 to 362, was obtained.
In a second sample, Clausing et al. (1986) collected fecal samples for six
hospitalized, bedridden children. A mass-balance calculation for these children, who
presumably had very limited access to soil, yielded estimates of 46 mg/d based on
aluminum. For titanium, two of the children had estimates well in excess of 1,000 mg/d,
with the remaining four children in the range of 23 to 48 mg/d. The data on
hospitalized children suggest a major non-soil source of titanium for some children, and
may suggest a background non-soil source of aluminum. However, conditions specific to
hospitalization, e.g., medications, need to be considered. AIR measurements were not
reported for the hospitalized children. Speculation on the source of titanium includes
diet, the white coloring in (disposable) diapers and several other items, but this has not
as yet been resolved.
Hawley (1985) developed scenarios for estimating exposure of young children, older
children, and adults to contaminated soil. Hawley addressed exposure by soil ingestion,
inhalation, and dermal contact. His approach to estimating levels of ingestion is
presented here (see Table 4-1). Each year was divided into two activity periods, May
through October, when individuals were assumed to spend much time outdoors, and
75
-------
TABLE 4-1. ESTIMATES OF SOIL INGESTION FROM DERMAL CONTACT
Scenarios
Young child (2.5 years old)
Outdoor activities (summer)
Indoor activities (summer)
Indoor activities (winter)
Older child (6 years old)
Outdoor activities (summer)
Indoor activities (year-round)
Adult
Work in attic (year-round)
Living space (year-round)
Outdoor work (summer)
Exposure
(mg/d)
250
50
100
50
3
110
0.56
480
Days/year
activity
130
182
182
152
365
12
365
43
Annual
average
(mg/d)
90
25
50
165
21
3
24
3.7
0.56
57
61
SOURCE: Hawley, 1985.
76
-------
November through April, when weather conditions were assumed to eliminate outdoor
exposure to soil. The following estimates were made by Hawley (1985).
(1) Young children (2.5 years old. 13.2 kg)
Outdoor activity, May through October, 5 d/wk: 250 mg/d. Estimate
based on analysis of data from Lepow et al. (1974, 1975), who
hypothesized 250 mg ingestion after experiments showed that dirt from a
21.5 cm^ area of a child's hand typically had a mass of 11 mg.
Additionally, Roels et al. (1980) measured contamination of children's
hands by metal contaminants of soil while the children were active on
playgrounds. These data led to an estimate of 40 to 180 mg dirt being
present on the dominant hand of an 11 -year-old (said to be equivalent in
area to both hands of a 2.5-year-old).
Indoor activity, May through October: Child assumed to ingest 50 mg of
household dust each day. Reference was made to the previously cited
experimental data.
Indoor activity, November through April: 100 mg/d ingestion assumed due
to the longer period of indoor activity.
(2) Older Child (6 years old. 20.8 kg)
Outdoor activity, May through October: 50 mg/d. Using the surface dust
value cited from Lepow et al. (1974) of 0.51 mg/cm^ on skin, a child is
assumed to ingest dirt from an area equal to the area of the fingers on
one hand.
Indoor activity, year-round: 3 mg/d. Indoors, the child is assumed to
have dermal dirt present at the reduced level of 0.056 mg/cm^, which is
the quantity of dirt estimated by the authors to be present on surfaces
77
-------
within the home. Dirt from inside surfaces of hands is assumed to be
ingested.
(3) Adult (70 ke)
Work in attics or other uncleaned areas of a house, 12 d/yr: 110 mg/d.
Estimate based on ingestion of a 50-um-thick dust layer from the inside
surfaces of the fingers and thumb of one hand while eating food or
handling cigarettes. Data from Wolfe et al. (1974) are cited to support
dust intake while smoking cigarettes.
Living space activities: 0.56 mg/d. Adults' hands are assumed to have
dust contamination equal to that on indoor surfaces (0.056 mg/cm^), and
dust is ingested from a lO-cm^ area of skin while eating or smoking.
Outdoor activities, May through October: 480 mg/active day. The adult is
assumed to be engaged in yard work or other outdoor physical activity
for 8 hr/d, 2 d/wk. The estimate is based on ingesting a 50 um-thick
layer of soil from the inside surfaces of the fingers and thumb of one
hand twice daily. These estimates are summarized in Table 4-1.
2. Evaluation
The data from the tracer element studies, Binder et al. (1986) and Clausing et al.
(1986), provide support for a preliminary estimate of average soil ingestion by children
on the order of 100 to 200 mg/d, consistent with the "low" estimate used by Schaum
(1984). These estimates are based on findings with silicon or AIR and aluminum.
Estimates based on a titanium tracer are higher by a factor of 5 to 10. This discrepancy
has not been explained, but may be due to dietary and other sources. Estimates based
on the minimum quantity ingested as calculated using the three tracers are not utilized
here, because use of a minimum will tend to bias the estimated ingestion downwards.
Hawley (1985), who estimated quantities of soil likely to be present on skin and
78
-------
subsequently ingested, also arrived at an estimate in the above range. It should be noted
that Hawley's approach would not address children who deliberately ingest dirt or mouth
soiled objects.
Binder et al. (1986) and Clausing et al. (1986) also provided some limited information
on the upper range of soil ingestion in children. With the exception of the titanium
data, the two studies provide evidence of an upper range of soil ingestion in children on
the order of 1,000 mg/d or more. It should be noted that both studies had limited
sample sizes and that neither specifically included (or excluded) children with pica (the
tendency to eat non-food materials). Again, estimates based on titanium would be
substantially higher, on the order of 20 g/d.
The Exposure Assessment Group is sponsoring a systematic study of soil ingestion
by children, using the tracer element methodology. Preliminary research has included a
study in miniature swine to assess the assumption that the tracer elements are poorly
absorbed, and to provide an experimental check on mass-balance calculations.
The study in children includes a pre-pilot study to test field methods and to address
possible non-soil sources of titanium. A randomly-selected sampling of 100 children will
provide an evaluation of soil ingestion in one location (Richland, Washington and
vicinity). Dietary contributions to tracer element intake will be measured in this study.
The results will be available in early 1988.
Based on this review of the limited data now available, particularly the studies of
Binder et al. (1986) and Clausing et al. (1986), the following values for soil ingestion are
suggested: Average soil ingestion in the population of young children (under the age of
7) is estimated at approximately 0.1 to 0.2 g/d. For calculation purposes, an estimate of
0.2 g/d is suggested as an average value. An upper-range ingestion estimate among
children with a higher tendency to ingest soil materials is 1 g/d. These estimates are
based on data using silicon and aluminum as trace elements. The reason for the higher
79
-------
estimates for titanium are not understood, but the increase seems likely to be due to
non-soil factors.
F. INGESTION — BEEF AND DAIRY PRODUCTS
An indirect route of human exposure to 2,3,7,8-TCDD is that of soil contamination
and subsequent uptake in the food chain. The foods subject to direct soil contamination
are those derived directly from plants and those derived from animals.
While some general information is available on plant uptake of inorganic chemicals
and elements from soil, less appears to be known about uptake of large organic molecules
from soil. The data on 2,3,7,8-TCDD uptake is contradictory, as discussed in Section D
of Chapter 3. The few studies available on other halogenated hydrocarbons suggest low
absorption by plants (Fries and Marrow, 1981; Jacobs et al., 1976). Further, most crops
consumed directly by humans are subjected to effective cleaning procedures that remove
most of the adherent soil or dust. Consequently, the discussion that follows focuses on
foods derived from domestic cattle, grazing animals where commercial production often
involves potential direct contact, and ingestion of soil and deposited dust.
Lengthy accounts of the factors involved in calculating human exposure to
2,3,7,8-TCDD from beef and dairy products are available (Schaum, 1984; U.S. EPA, 1985b;
Fries, 1985, 1986). Schaum (1984) and Fries (1985, 1986) cited studies of soil
consumption by cattle and studies examining ratios of contaminant levels in the diet to
resulting levels in body fat and milk fat for chemicals similar to 2,3,7,8-TCDD, such as
PCBs, PBBs, and DDT (under various production scenarios) (see Section C of Chapter 3).
Exposure duration effects are also discussed in Schaum (1984).
The potential effects of "market dilution" of beef and dairy products on human
exposure are discussed briefly in Schaum (1984), at more length by Fries (1986), and at
much greater length in U.S. EPA (1985b) for the particular case of cattle production in
Missouri. Aspects of the beef industry in this region specifically noted in U.S. EPA
80
-------
(1985b) as important to estimating exposure were type of activity within the industry
[e.g. cow-calf production, "backgrounding" (preparing calves for feedlots), feeding (for
slaughter)], replacement rates as a function of activity, fractions of cattle fed to
maturity outside contaminated areas before slaughter, and slaughter categories and rates
relative to national figures. Both Schaum (1984) and U.S. EPA (1985b) concluded that
dilution will vary widely between different marketing areas. Schaum (1984), U.S. EPA
(1985b), and Fries (1986) noted that the subpopulations most likely to receive high
exposures are beef producers, dairy farmers, and their direct consumers and, further, that
exposure evaluations should be very location-specific.
Average consumption rates and fat content data for beef and dairy products are
presented in Table 4-2, which has been adapted from Schaum (1984) by addition of
information from U.S. EPA (1984b, c) and Fries (1986). Much greater "resolution"
actually is available in U.S. EPA (1984b, c) than is found in Table 4-2, since both (U.S.
EPA, 1984b, c) are based on a U.S. Department of Agriculture (USDA) Nationwide Food
Consumption Survey (NFCS) conducted in 1977-1978. The NFCS covered intake of 3,735
possible food items by 30,770 individuals characterized by age, sex, geographic location,
and season of the year. Further description of the survey design is given in U.S. EPA
(1984d).
The average beef fat consumption noted in Table 4-2 ranges from 14.9 to 26.0 g per
70-kg person/d, with a single high consumption estimate of 30.6 g per 70-kg person/day
that might be more appropriate for families of beef producers who home slaughter. Milk
fat consumption from all dairy products ranges from 18.8 to 43 g per 70-kg person/d,
with the lower end of this range appearing best supported at present. Considering fresh
milk only, milk fat consumption is reported to average 8.9 to 10.7 g per 70-kg person/d,
with a single high consumption estimate of 35 g per 70-kg person/d perhaps appropriate
81
-------
TABLE 4-2. RATES OF INGESTION OF BEEF AND DAIRY PRODUCTS
Total consumption
rate ± std. error
(g/70 kg person-d)
Percentage
of fat
Fat consumption
rate ± std. error
(g/70 kg person-d)
Reference
124
110.7 ± 1.7
87.6 ± l.la
96.3
66.8 (average)
137.1 (high)
Dairy products
550
308.6 ± 5.3
431.6 ± 5.6
Fresh milk (only)
15
23
(23)
(23)
22
22
7.8
(7.8)
(4.4 implied)
19
26.0 ± 0.3
(20.1 ± 0.3)b
(22.1)
14.9
30.6
43
(24.1 ±
18.8
0.4)c
U.S. EPA (1981b)
U.S. EPA (1984b)
U.S. EPA (1984c)
Berglund (1984)
Fries (1986)c
Fries (1986)
U.S. EPA (1981b)
U.S. EPA (1984c)
U.S. EPA (1984b)
253.5 ± 4.9
305 (average)
1000 (high)
(3.5)
3.5
3.5
(8.9)
10.7
35.0
U.S. EPA (1984c)
Fries (1986)
Fries (1986)
aThe categories established in U.S. EPA (1984c) exclude beef in meat mixtures
(e.g., meat loaf), meat by-products (e.g., wieners), and organ meats. The
basic data set underlying both U.S. EPA (1984b) and U.S. EPA (I984c) was the
USDA National Food Consumption Survey 1977-1978. The basis for the difference
in total dairy products consumption rates noted for U.S. EPA (1984b) and U.S.
EPA (1984c) has not yet been resolved.
bfleef fat consumption rates in parentheses are calculated using percentages of
fat derived from U.S. EPA (1984b).
cThis and succeeding values from Fries (1986) reportedly derived from Breiden-
stein (1984).
"Dairy fat consumption rates in parentheses are calculated using percentages of
fat derived from U.S. EPA (1981b).
82
-------
for dairy farm families. [Age range-specific information is available in both U.S. EPA
(1984b) and U.S. EPA (1984c).]
Differences in beef and dairy fat consumption rates cited above from those used by
Schaum (1984), and revised bioconcentration factors (Section C of Chapter 3), will result
in a reduction in estimated exposure.
G. INGESTION — FISH CONSUMPTION DATA
1. Available Studies
A variety of definitions have been used for fish consumption. Some studies examine
only commercially-caught fish while others do not distinguish between marine and
freshwater fish. Others do not differentiate between fin and shellfish or fresh versus
processed fish. Some data have been published which provide only nationwide averages
while others provide data for regions or states. Consequently, drawing meaningful
comparisons between figures derived from different sources often is difficult.
Different population bases have been used in the various surveys. For example,
nonconsumers often are included in the population base for nationwide or regional
averages. Reporting averages based only on consumers seems preferable on most
occasions, but so long as the basis for a given average is reported unequivocally
modifications may be made as needed for later analyses.
Several recent studies of fish consumption by the U.S. population are summarized
below. These studies for the most part estimate consumption of certain population
subgroups and thus do not indicate an unequivocal need for changes in the average fish
consumption estimates presented by the Ambient Water Quality Criteria for 2,3,7,8-TCDD
(U.S. EPA, 1984), where an average daily consumption of 6.5 g/d per capita of freshwater
and estuarine fish and shellfish was derived from analysis of a survey of fish and
shellfish consumption in the United States (U.S. EPA, 1980). The variety of results do
83
-------
emphasize the need to base consumption assumed in a particular exposure assessment on
studies involving similar populations.
Data from the Nationwide Food Consumption Survey conducted by the USDA in
1977-1978 were not available when the U.S. EPA derived the 6.5 g/d figure for
consumption of freshwater and estuarine fish (Stephan, 1980). The USDA survey obtained
information on both household and individual intake of food products. Interviews were
conducted to determine food consumption in households during the previous week, and
included a 1-day recall plus a 2-day diary of individual food intake. A national
probability sample of households in the continental United States was obtained by means
of approximately 15,000 interviews covering over 36,000 individuals. Supplemental surveys
of households with elderly and low-income individuals were conducted. Separate data
were gathered for Alaska, Hawaii, and Puerto Rico.
Specific information was gathered, including use of specific fish species, as well as
the state of processing (fresh, frozen, etc.). Analysis of the data indicated an average
individual intake as 12 g/d fish and shellfish (edible weight) on a per capita basis
nationwide, although geographic regional averages ranged from 9 to 14 g/d, with highest
consumption in the Northeast (U.S. Department of Agriculture, 1985). Total population
figures, including non-consumers, were used in computing these averages. This survey
also presents fish consumption by age group and season of the year. Other USDA
publications have provided average figures for fresh commercial fish — in 1983, an
average of 6.4 g/d was estimated to have been consumed per person.
The most recent fish consumption data from the National Marine Fisheries Service
(NMFS) report total per-capita fish and shellfish consumption at 6.2 kg/yr (16.9 g/d)
(U.S. Department of Commerce, 1985). This estimate is based on the commercially-landed
fish and shellfish catch only.
84
-------
An earlier report in the series (U.S. Department of Commerce, 1983) gave
consumption of edible weights of fresh and processed commercial marine fish and
shellfish as 9.9 g/d per capita, based on yearly catches, imports, exports and existing
inventories. The recreational catch has been estimated to contribute an additional 3.7 to
5 g/d, based on information from the National Oceanographic and Atmospheric
Administration (as cited in U.S. EPA, 1986a).
U.S. Department of Commerce (1983) also reports 3.7 - 5.3 g/d (edible weight),
marine fish and shellfish are consumed by recreational fishermen, fairly close to an SRI
analysis showing consumption of 5.3 g/d of fish from recreational sources. Cordle (1981)
reported a 90th percentile consumption of 15.7 g/d for Great Lakes region consumers
only and a 99th percentile figure of 36.8 g/d. No average figures were presented.
A National Seafood Consumption Survey was conducted by the NMFS in 1981 with a
panel of 7,500 households (NMFS, undated). The households kept diaries on the amount
of fish and other seafoods purchased for household use, as well as the amount actually
eaten both at home and away from home. Purchase data are broken out by species,
nature of product (fresh, frozen, etc.), region, and a variety of demographic factors.
The Longwoods Research Group (1984) analyzed some of the 1981 NMFS data based upon
frequency of use rather than quantities consumed. This revealed that 82% of all
projected U.S. households eat seafood or fish.
Still earlier (1969-1970), a Market Facts Survey conducted for NMFS revealed a per
capita total fish consumption figure of 16.8 g/d of which 6.1 g/d consisted of fresh and
frozen fin fish. This survey did not discuss explicitly whether portions were based on
edible weight or if recreational sources were considered.
A National Purchase Diary (NPD) Fish Consumption Survey was performed for the
Tuna Research Institute in 1973-1974 with the results based only on actual consumers of
fish rather than total population (U.S. EPA, 1980). The questionnaire was administered
85
-------
to a total of 7,662 families (around 25,000 people) over one year, with 1/12 of the
families responding in a given month to eliminate seasonal effects. Cordle et al. (1982)
later used data from this survey as the basis for their estimate of consumption totalling
18.7 g/d. According to Conner (1984) the NPD Survey data show 6.5 g/d of estuarine
fish and shellfish and 2 g/d of freshwater fish are consumed. SRI International later
reexamined the data tapes of the NPD Survey and found numerous discrepancies (U.S.
EPA, 1980). A corrected version of the data base resulted in an average consumption
figure of 14.3 g/d total fish, with a 95th percentile value of 41.7 g/d. SRI International
also presents average and 95th percentile figures for each sex and different age groups
(U.S. EPA, 1980).
Race and religion, as well as regional factors and age, may have strong impacts on
fish consumption rates. The Market Facts Survey reported seafood consumption by U.S.
blacks and people of Jewish faith to be approximately twice that of whites as a whole
(U.S. EPA, 1980). However, the NPD Research Survey for the Tuna Research Institute
reported only a 13% higher consumption rate among blacks (U.S. EPA, 1980. The NPD
survey also found that orientals consumed fish at a rate 47% above Caucasians. The 95th
percentiles of fish and shellfish consumption typically were a factor of three above
averages for the different population groups.
According to the USDA publication Foods Commonly Eaten bv Individuals: Amounts
per Dav and Per Eating Occasion, consumers of fin fish other than canned, dried or
raw, consumed an average (mean) of 54 g/d/person. It is not apparent what
percentage of fin fish is recreationally caught. The following percentile distribution
was given: 50th percentile - 38 g/d; 90th percentile - 96 g/d; 95th percentile -
132 g/d; 99th percentile - 221 g/d.
The difference between the mean and median (50th percentile) indicates that the
amount of fish consumed is not distributed in a normal pattern among the consumer
population. It is also worthwhile to note the increase in average daily amount over
those presented by the first series of surveys, which included non-consumers as
well. [C.F. Kleiman, "Fish Consumption by Recreational Fishermen: An Example of
Lake Ontario/Niagara River Region" Environ Corporation, prepared for OECM, U.S.
EPA, Washington, D.C., May 20, 1985, page 6.]
86
-------
Puffer et al. (1983) reported the results of a survey of fishing habits and fish
consumption rates among fishermen in the Los Angeles area. Interviewers obtained
information from fishermen at 12 representative locations identified as frequently fished.
A total of 1,059 interviews were conducted from an estimated sport fishing population of
at least 31,000. Approximately half of the fishermen fished one or more times per week,
with 14% of those interviewed fishing three to seven times per week. The majority of
fishermen interviewed consumed the fish they caught. The median amount of fish
consumed by the fishermen themselves was reported to be 37 g/d, with the 90th
percentile at 225 g/d. (The report estimated that at least 100,000 family members of
fishermen shared fish they caught.) These consumption rates are substantially above
those for the general population. These data do not take into account consumption of
fish purchased from stores.
Individuals in other areas are known to have a high intake of sport fish. A study
by the Michigan Department of Public Health (Humphrey et al., 1976) examined the health
status of individuals who consumed at least 30 g/d (annual average) of Great Lakes fish.
The highest recorded fish consumption over the two-year study period was 224 g/d.
The Puget Sound Estuary Program (U.S. EPA, 1986c) developed a guidance document
that included a useful framework for assessing quantities of fish consumed by local
populations; however, no data specific to Puget Sound were presented.
The New York State Department of Environmental Conservation (NYSDEC) uses a
figure of 32.4 g/d in their health advisories, as the average fish consumption for
recreational fishermen, based upon the 90th percentile of nationwide fish
consumption figures (A. Newell, personal communication)...
A survey of users of the 1983 Guide to Eating Ontario Sport Fish (Ontario Ministry
of the Environment, 1984) revealed that Ontario sports fishermen eat locally-caught
fish approximately once every 3 weeks, with an average meal size of 289 g (10.2
oz), corresponding to an average daily figure of 13.8 g/d. A substantial percentage
of respondents (26%) ate at least a pound of fish per meal. However, as this survey
is based on voluntary responses to a questionnaire, it may be subject to self-
selection biases...
87
-------
Finally, a personal communication from R. Sonstegard, who is currently studying
Lake Ontario sports fishermen, indicated that "intensive" fishermen (1 week trips in
spring and fall, and weekends throughout the year) can consume 62 g of salmon and
trout per day. An "average" figure of 31 g/d, with a range of 0-311 g/d, was also
cited. [C.F. Kleiman, "Fish Consumption by Recreational Fisherman: An Example of
Lake Ontario/Niagara River Region" Environ Corporation, prepared for the Office of
Compliance and Enforcement Monitoring. U.S. Environmental Protection Agency,
Washington, D.C., May 20, 1985, pages 6-9].
2. Evaluation
Substantial recent information on fish consumption rates has become available
through the surveys conducted by USDA and NMFS. While these surveys do not indicate
the need for major revision of previous fish consumption estimates, they can provide
more recent information and will allow examination of the fish consumption habits of
particular population subgroups. The data from these surveys would also allow
recalculation of the U.S. EPA's estimate of human consumption of freshwater and
estuarine fish and shellfish. Because current information indicates that some population
groups consume fish at rates much above the national average, this work could be of
significant value in determining the risks from 2,3,7,8-TCDD contamination that may be
encountered by specific population groups.
Based on this review, the 6.5 g/d average consumption rate for freshwater and
estuarine fish and shellfish that has been used in previous U.S. EPA assessments is still
appropriate. To account for individuals who habitually consume larger quantities of fish,
a value of 30 g/d is suggested based on the Los Angeles and Great Lakes data.
EPA (1987a) references the following values of average consumption rate that may
be assumed when site-specific data are unavailable:
(a) 6.5 g/d represents an estimate of average consumption of fish and shellfish
from estuarine and fresh waters by the U.S. population;
(b) 20 g/d represents an estimate of the average consumption of fish and shellfish
from marine, estuarine, and fresh waters by the U.S. population;
88
-------
(c) 165 g/d represents average consumption of fish and shellfish from marine,
estuarine, and fresh waters by the 99.9th percentile of the U.S. population; and
(d) 180 g/d represents a "reasonable worst case" based on the assumption that
some individuals would consume fish at a rate equal to the combined
consumption of red meat, poultry, fish, and shellfish in the U.S.
89
-------
5. POST-EXPOSURE
Within the last few years, two important areas have come increasingly under study
with regard to the fate of 2,3,7,8-TCDD once exposure has occurred; they are
bioavailability and pharmacokinetics. Unavailability, the first of these areas, refers to an
organism's ability to remove 2,3,7,8-TCDD from an ingested or inhaled particle and then
to absorb it. Strictly speaking, bioavailability is a property of both the material to which
an .organism is exposed and the organism's capabilities and pharmacokinetic responses.
However, it is useful to assume for the present that the organism's extraction ability is
constant, and to look at bioavailability as a property of the material. Recent research
has observed that 2,3,7,8-TCDD adsorbed on various substrates can differ in
bioavailability by approximately an order of magnitude. The current state of knowledge
about the causes of bioavailability differences is incomplete, but early hypotheses (based
on a very small data set) hold that the bioavailability of 2,3,7,8-TCDD from various
materials can be related to chemical availability measured by solvent extraction. Most
contaminated soils tested so far (five) show bioavailability in animal tests of about 25%
to 50% that of 2,3,7,8-TCDD in corn oil given by gavage. Three soil samples spiked with
2,3,7,8-TCDD had bioavailabilities in the 40% to 70% range compared with corn oil. Based
on limited data, 2,3,7,8-TCDD in fly ash proved roughly 25% as bioavailable as 2,3,7,8-
TCDD from the solvent extract of the fly ash. (It should be noted that in this
experiment 2,3,7,8-TCDD from both fly ash and solvent extract were recovered from the
rat liver in low quantities, making interpretation of the experiment difficult.) Studies
with soil from one site, and with activated carbon with dioxin added, however, showed
gut bioavailabilities of < 10%, and < 1% compared with 2,3,7,8-TCDD in solvents,
respectively.
The implications of this early work are important, since estimated risk following
intake of an environmental matrix will be proportion to bioavailability. At this point, it
90
-------
is unclear what the long-term implications of bioavailability differences will be to risk
assessment. Some soils have shown high 2,3,7,8-TCDD bioavailability, while bioavailability
in one tested soil is lower. No data on the distribution of contaminated soils by
bioavailability currently exist to allow this difference to be systematically considered, nor
is there an accepted protocol for measuring bioavailability from soil on a site-by-site
basis.
The second important post-exposure area of study is pharmacokinetics. Theoretically,
if one knows what happens to the 2,3,7,8-TCDD once it is absorbed, one can look at
body burdens and back-calculate an average dose, or average exposure. In practical
terms, this procedure can lead to reduced uncertainty in a risk assessment by allowing
calculation of exposure from two independent methods. Currently, our ability to perform
these pharmacokinetics calculations is in its early stages of development. There are
significant difficulties in current approaches to using body burden data to back-calculate
exposure, and at this point verification of certain not-easily-verified assumptions needs
to be done before such calculations can become standard tools for exposure assessment.
It is safe to say that, in the future, the role of pharmacokinetics in exposure assessment
will increase.
A. ABSORPTION FROM ENVIRONMENTAL MATRICES (BIOAVAILABILITY)
I. General Considerations
Following ingestion of a material containing 2,3,7,8-TCDD or other toxic species,
the toxic effect of the material is modified by the degree of absorption, principally in
the small intestine. In several experimental studies, investigators administered 2,3,7,8-
TCDD-containing environmental matrices to experimental animals, and measured
parameters relating to bioavailability. These studies included quantitation of 2,3,7,8-
TCDD in liver and other tissues following treatment; comparison of toxicities of
contaminated environmental materials with pure 2,3,7,8-TCDD; and examination of enzyme
91
-------
induction. The results of these different approaches, their limitations, and needs for
further research are discussed below.
2. Review of Data on Bioavailabilitv
Umbreit et al. (1985, 1986a,b) conducted experiments in guinea pigs, administering
2,3,7,8-TCDD in corn oil, 2,3,7,8-TCDD added to chemically decontaminated soil, or soil
from two industrial sites in Newark, New Jersey (a manufacturing site and a salvage site)
contaminated with CDDs. 2,3,7,8-TCDD was the principal lower chlorinated isomer
(dioxin or furan) present in the soil from the manufacturing site (for which a chemical
analysis was presented). Soil from the manufacturing site was found to have 1,500 to
2,500 ppb 2,3,7,8-TCDD under soxhlet extraction; release under ambient temperature
manual solvent extraction was much lower, reported as ">2.5 ppb." The soil from the
salvage site was reported as approximately 180 ppb 2,3,7,8-TCDD under soxhlet
extraction.
Table 5-1 summarizes the findings of the study, in which groups of two or four
male and two or four female guinea pigs received single gavage doses of the test
materials and were observed until death or sacrifice at 60 days. 2,3,7,8-TCDD in corn oil
or in recontaminated soil (6 /ig/kg in both) proved highly toxic, without similar toxicity
being observed in animals treated with up to twice this dose of 2,3,7,8-TCDD in the soil
from the manufacturing site. The limited data on 2,3,7,8-TCDD levels in the liver showed
much higher levels following administration of recontaminated soil versus contaminated
soil from the manufacturing site.
Umbreit et al. (1986a) thus demonstrated that gavaged 2,3,7,8-TCDD containing soil
from the manufacturing site was substantially less toxic than equivalent doses of 2,3,7,8-
TCDD in corn oil. However, quantitative comparison of the effective doses in this study
is difficult. Approaches to a quantitative comparison are outlined below.
92
-------
TABLE 5-1. GUINEA PIGS RECEIVING A SINGLE GAVAGE DOSE
OF MATERIALS CONTAINING 2,3,7,8-TCDD
Treatment
2,3,7,8-
TCDD
Early
deatha
2,3,7,8-TCDD
ppb in liver"
NEWARK MANUFACTURING
Decontamin-
ated soil6
Contaminated
soil6
Recontamin-
ated soil
Corn oil
TCDD in
corn oil
Decontamin-
ated soil
Contaminated
TCDD in
corn oil
0
3
6
12
6
0
6
0
0.32
6
0/7
0/8
0/7
0/7
6/7
0/7
5/8
NEWARK
0/4
0/4
3/4
NR
NR
NR
0.09 (0.07%)f
18.0 (14%)f
NR
NR
Effects
noted0
SITE
N
N
N
N
TS
N
TS
%weight
gain
0-4 weeks"
70
83
57
51
25
57
37
SALVAGE YARD SITE
NR
0.23 (5.8%)f
NR
N
N
TS
NR
NR
NR
93
-------
TABLE 5-1 (continued).
aGavage deaths excluded from counting; all treatment deaths occurred < 31 days
after dosing.
^Results of analysis of single pooled liver tissue samples from subsets of the
tested animals.
CTS = "Typical signs" of 2,3,7,8-TCDD toxicity including thymic atropy, absence of
body fat, and loss of approximately 40% of body weight. N = "Typical signs" not
present.
^Data for survivors only.
eSoil type at site reported to be medium dense, black, coarse to finegrained sand
fill with some medium to fine gravel, and with traces of silt, organic material,
and cinders. Fill material at site included asphalt. No chemical characteriza-
tion of the soil sample was reported.
'Approximate percent of gavage dose found in liver, assuming liver is 4.7% of body
weight. For the group receiving soil from the salvage site, no weight gain data
were reported and a 100% weight gain (0 to 8 weeks) was assumed, which is consis-
tent with other surviving groups. Liver weight data were not available to allow
more precise calculation.
SSoil type at site reported similar to that at Newark manufacturing site.
Still bottoms were dumped on the site during salvage operations and were incor-
porated into soil. No other characterization of soil sample was reported.
NR = Not reported.
SOURCE: Umbreit et al., 1985, 1986a,b.
94
-------
(1) Guinea pigs receiving 12 fig/kg 2,3,7,8-TCDD jn contaminated soil
experienced no deaths, while five out of eight guinea pigs receiving
6 Mg/kg 2,3,7,8-TCDD in corn oil died, with no groups tested having
lower doses in corn oil. Other authors have provided data on the
toxic effects of 2,3,7,8-TCDD in corn oil which could aid in the
comparison.
McConnell et al. (1984) observed one out of six animals dying at 1
and six out of six animals dying at 3 Mg/kg. Silkworth et al. (1982) observed
three out of six animals dying at 2.5 /*g/kg and no deaths out of six at 0.5
/ig/kg. Comparing these data directly with the Umbreit et al. results would
suggest that the 2,3,7,8-TCDD in the Newark manufacturing site soil was less
effective, by a factor of 10 or greater, in producing toxicity than 2,3,7,8-
TCDD in corn oil.
(2) Umbreit et al. reported a "slightly reduced" weight gain in guinea pigs
receiving 6 Mg/kg of 2,3,7,8-TCDD in Newark manufacturing site soil, and a
"greater reduction" at the 12 /ig/kg dose. No other signs of toxicity were
noted in these groups. The animals receiving 6 /*g/kg 2,3,7,8-TCDD in corn
oil, in contrast, exhibited a marked loss of body weight and showed toxicity
and mortality. Silkworth et al. (1982) also provided data on weights of guinea
pigs receiving 2,3,7,8-TCDD in corn oil. Those receiving 2.5 ^g/kg exhibited a
marked reduction in weight gain among three out of six survivors, while those
receiving 0.5 /ig/kg showed a weight gain comparable to vehicle controls.
Comparison of this weight data with that of Umbreit et al. suggests that the
2,3,7,8-TCDD in corn oil was more than 5 times but less than 25 times as
potent as 2,3,7,8-TCDD in the Newark soil. This comparison assumes that the
effect of the Newark manufacturing site soil on weight gain was due to
95
-------
2,3,7,8-TCDD as opposed to other compounds in the soil. Numerous other
dioxin and furan compounds and other chemicals have been identified in this
soil (Umbreit et al., 1987a). It has not been established that 2,3,7,8-TCDD is
the sole or prime source of toxicity in the soil.
(3) Umbreit et al. presented liver concentrations of 2,3,7,8-TCDD after death or
sacrifice at 60 days following gavage (see Table 5-1). Much lower
concentrations of 2,3,7,8-TCDD were found in the livers of animals receiving
soil from the manufacturing site compared with those receiving the dose in
cordn oil. There are, however, two factors that limit the conclusions than can
be drawn from this comparison.
First, the corn oil group experienced major toxicity and weight loss, particularly
complete loss of body fat. These changes may have affected the partitioning of 2,3,7,8-
TCDD within the body, leading to a higher concentration in the livers of the animals
experiencing toxicity. Second, the animals gavaged with corn oil died early—half were
dead by 26 days, while all of the guinea pigs treated with soil survived to 60 days (with
the exception of one gavage death). The U.S. EPA (1985d) reported a half-life for
2,3,7,8-TCDD elimination of 30 + 6 or 22 to 43 days from two studies in guinea pigs.
Additionally, the U.S. EPA (1985d) stated that elimination in the guinea pig may follow
zero-order kinetics. Differences in elimination due to differences in periods of survival
are likely to have affected the relative quantities of 2,3,7,8-TCDD found in the livers of
the test groups.
Perhaps a more appropriate comparison can be made with the four animals
receiving 0.32 /ig/kg of 2,3,7,8-TCDD in contaminated soil from the Newark salvage site.
These animals experienced no reported toxic signs (weight data not presented) and
survived the full 60-day experiment. Approximately 6% of the gavage dose was found in
the liver of these animals (Table 5-1), while only about 0.06% of the gavage dose was
96
-------
found in the livers of guinea pigs in the 12 pg/kg group receiving the Newark
manufacturing site soil. This would suggest that the 2,3,7,8-TCDD in the manufacturing
site soil was 100 times less bioavailable. However, given the different doses used and
the fact that only a single pooled sample was analyzed for 2,3,7,8-TCDD in each group,
caution must be used in interpreting this comparison.
The 2,3,7,8-TCDD in soil from the salvage site was substantially bioavailable, based
on the single liver tissue analysis. Approximately 6% of the administered dose was
recovered from the livers of these animals at 60 days. This can be compared with data
on hamsters given 2,3,7,8-TCDD in corn oil by McConnell et al. (1984), where
approximately 8% of the 2,3,7,8-TCDD could be recovered in the 1 Mg/kg dose group
among survivors at 30 days.
McConnell et al. (1984) treated Hartley guinea pigs (2.5 weeks old) with single
gavage doses of either 2,3,7,8-TCDD or dioxin contaminated soil from two sites in
Missouri. The 2,3,7,8-TCDD concentrations from the two sites were reported at 700 and
880 ppb respectively; total tetrachlorodibenzofurans (TCDF) concentrations in the soil
were 40 to 80 ppb, and polychlorinated biphenyls (PCB) concentrations were 3 to 4 ppm.
Taking into account the relative toxicities, the authors concluded that toxicity from the
other compounds was likely to be small compared with that from 2,3,7,8-TCDD. The
results of the study are shown in Table 5-2. Livers were analyzed for 2,3,7,8-TCDD at
death or sacrifice at 30 days following treatment. Treatment deaths occurred between
5 and 21 days post-gavage.
Guinea pigs that died exhibited severe loss of body fat, markedly reduced thymus
and testicle size, and adrenal hemorrhage. No adverse affects were noted in animals
treated with decontaminated soil. For 2,3,7,8-TCDD in corn oil and for both
contaminated soils, there were clear dose-responses in mortality. The calculated LD5Q
97
-------
TABLE 5-2. TOXICITY OF TCDD CONTAMINATED SOIL
2,3,7,8-
TCDD in liver (ppb)
f% Admin. dose)a»b
Estimated LD50
TCDD dose
Group # Treatment /*g/kg
1 Corn oil 0
2,3,7,8-
2 TCDD in 1
corn oil
3 3
Times Beach
4 soil c»d 1.3
5 3.8
6 12.8
Minker Stout
7 soilc'd 1.1
8 3.3
9 11.0
Uncontamin-
10 ated soild 0
Uncontamin-
ated soil with
11 2,3,7,8-TCDD 10
added3
Dead/
treated
0/6
1/6
6/6
0/6
1/5
5/5
0/6
2/6
6/6
0/5
0/6
Alive at Animals
30 d dying
ND
1.6 +.02 4.1
(7.5%) (19%)
13.3 + 2.3
(26%)
(™&b)
1.0 + 0.1 3.2
(1.3%) (4.0%)
34.3+6.0
(13%)
*(<«%)
1.4 + 0.3 2.0 + 0.1
(2.0%) (2.8%)
25.7 + 5.2
(11%)
ND
\
45.3 + 8.4
(21%)
1.75
(95% CL
1.26-2.24)
7.15
(95% CL
4.90-9.40)
5.50
(95% CL
3.45-7.55)
Source: McConnell et al. (1984).
98
-------
TABLE 5-2 (continued).
aThe percentages of the administered doses found in the liver were calcu-
lated assuming a liver weight of 4.7% of body weight. For survivors,
the percentages are underestimated to some degree because data on
weight gain over the study period were not available.
''Animals were observed for up to 30 d when survivors were sacrificed.
cSoil was sifted through a wire mesh to remove gravel particles. No
details on soil type or composition presented. The soil was contaminated
by waste oil containing 2,3,7,8-TCDD and other dioxins and furans.
The presence of residual waste oil in the soil was not specified.
^Soil was administered as a gavage dose in distilled water.
99
-------
values for the two soil types were lower than the LD5Q for 2,3,7,8-TCDD in corn oil by
a factor of three- to four.
There was a dose-response between the liver concentration of 2,3,7,8-TCDD and the
gavage dose; the details of this, relationship are complex. Animals dying during the
experiment had liver concentrations a factor, of 1.4 to 3.2 higher than animals in the
same dose groups who survived. 30 days. This observation makes quantitation of the
dose-response relationships difficult (all or most of the animals in the low-dose groups
survived the experiment, while all of the animals in the high-dose groups died). When
the; liver concentrations of 2,3,7,8-TCDD in animals dying early at the middle and high-
dose groups are compared, there appears to be a greater-than-linear increase in liver
concentration with dose for the Times Beach and Minker Stout soil groups, with a 3.3-
fold increase in dose producing a 10- to 13-fold increase in liver concentration.
Liver concentrations of animals in the different dosing groups can best be compared
among groups that experienced similar mortality.
(1) Animals in dose groups in which all animals died within 30 days: 2,3,7,8-TCDD
in corn oil (group 3), approximately 20% of the administered dose was in the
liver. For the soil-treated groups (groups 6 and 9), 13% and 11% of the doses,
respectively, were in the liver. Comparison of these data suggest that 2,3,7,8-
TCDD was approximately twice as available through corn oil as through soil.
(2) Animals surviving the 30-day experiment (in groups where at least 4 out of 6
survived): For 2,3,7,8-TCDD in corn oil (group 2), 7.5% of the administered
dose was in the liver. For soil-treated animals (groups 4, 5, 7, and 8) < 3.6,
1.3, < 4.2, and 2.0% of the doses, respectively, were in the liver. Comparison
here would suggest that 2,3,7,8-TCDD was approximately four times as available
through corn oil as through soil.
100
-------
The authors note that the differences in liver concentrations observed in the study
may reflect varying partitioning of the 2,3,7,8-TCDD among internal organs, since dying
animals suffered major loss of body weight and fat content. In addition, surviving animals
would have had greater opportunity to metabolize and excrete 2,3,7,8-TCDD due to a
longer lifetime.
Umbreit et al. (1986a) reported additional chemical analyses of the Times Beach soil.
Soxhlet extraction of the Times Beach soil yielded a similar quantity of 2,3,7,8-TCDD to
the solvent extraction reported by McConnell et al. (1984). This is in contrast to the
Newark manufacturing site soil used in the Umbreit et al. (1987a) experiments, where
only a small fraction of soxhlet-extractable 2,3,7,8-TCDD was extractable by the solvent
extraction methodology used by McConnell et al. (1984).
McConnell et al. (1984) also reported an experiment in which groups of six Sprague-
Dawley rats were given single gavage doses of 2,3,7,8-TCDD in corn oil or dioxin-
contaminated soil from the Minker site. Induction of aryl hydrocarbon hydroxylase
(AHH) in the rat livers was measured at sacrifice 6 days after dosing. Experimental
doses ranged from 0.4 to 5.0 Mg/kg 2,3,7,8-TCDD. Measured AHH induction was similar
for groups receiving 2,3,7,8-TCDD in corn oil or receiving contaminated soil containing
nearly equal doses of 2,3,7,8-TCDD. For example (based on the rate of formation of 3-
hydroxybenzo[a]pyrene), AHH activity was measured at 1,269 pmole min~^ mg~' for the
group receiving 5 A*g/kg 2,3,7,8-TCDD in corn oil and at 1,230 pmole min~l mg~^ for the
group receiving 5.5 /Jg/kg 2,3,7,8-TCDD in contaminated soil. For the five dose groups,
the AHH activity for the soil group ranged from 50% to 110% of the activity in the corn
oil group.
The McConnell et al. rat data indicate that the bioavailability of 2,3,7,8-TCDD from
the Minker site soil was at least 50% of that of equivalent doses of 2,3,7,8-TCDD in corn
oil.
101
-------
JLucier et al. (1986) provided additional information on the induction of hepatic
enzymes in rats by the 2,3,7,8-TCDD contaminated soil from the Minker site tested by
McConnell et .al. (1984). AHH induction was similar for the groups of rats receiving
,2,-3,7,8-T.CDD .in .corn oil .and,contaminated soil (within a factor of two) over a broader
range of doses (0.015 /ig/kg to 5 Mg/kg) than reported by McConnell et al. In a second
enzyme assay using the same animals, UDP glucuronyltransferase activity was found to be
slightly higher -in groups receiving 2,3,7,8-TCDD in corn oil than groups receiving equal
.doses in contaminated soil.
.Liver concentrations of 2,3,7,8-TCDD for the rats were also reported. For the corn
oil vehicle the liver concentrations were 40.8 ± 6.5 ppb at the 5 /ig/kg dose and 7.6 ±
2.5 ppb .at the 1 pg/k% dose. Assuming that the liver comprises 4.0% of body weight,
the retention rates for the 5 ;and 1 /ig/kg doses were 33% and 30%, respectively. In rats
receiving 2,3,7,8-TCDD in contaminated soil, the 5.5 jig/kg group had liver concentrations
of 20.3 ± 12.9 ppb, and the 1.1 MgAg group had concentrations of 1.8 ± 0.3. Thus,
retention rates for the 5.5 and 1.1 /ig/kg groups are estimated at 14% and 7%,
respectively. These data indicate that liver retention in the soil group was 20% to 40%
of that in the corn oil vehicle groups.
Umbreit et al (1986b) report additional studies of mortality in guinea pigs treated
with soil containing 2,3,7,8-TCDD from Newark (manufacturing site) and Missouri (Times
Beach) previously tested by Umbreit et al (1985, 1986a) and McConnell (1984),
respectively. Guinea pigs received a single gavage dose of a soil suspension and were
observed for 60 days. After autopsy, deaths were classified as whether or not they
appeared to be due to TCDD toxicity. Substantial mortality (25% overall) from conditions
not attributed to TCDD was observed across all groups. Table 5-3 contains the study
observations.
102
-------
The data for both the Newark and Missouri sites are similar in trend for the
previous data on these sites; and clearly indicate the greater toxicity of the Newark soil
for given equal administered doses of 2,3,7,8-TCDD. With larger groups of guinea pig
studied, a toxicity-related death was observed in both the 5 and 10 Mg/kg dose groups
for Newark soil while no deaths were observed in corresponding dose groups (6 and 12
A*8/kg) with fewer animals in Umbreit et al (1986a).
TABLE 5-3. COMPARISON OF MORTALITY IN GUINEA PIGS
FOLLOWING A SINGLE DOSE OF CONTAMINATED SOILS
Treatment TCDD
(Mg/kg)
Decontaminated Soil
Newark Soil
Times Beach Soil
Recontaminated Soil
0
3
5
10
1
3
10
10
# Deaths
Attributed to
TCDD Toxicity
0
0
1
1
2
1
8
19
# Animals
Treated Minus
Deaths Not Due
to TCDD
13
12
14
10
16
13
11
19
Source: Umbreit et al (1986b)
103
-------
Comparing groups within this study, similar mortality (1 or 2 deaths in 10 to 16
animals) was seen in both the 5 and 10 /Jg/kg Newark groups and the 1 and 3 jig/kg
Missouri groups. These results suggest that the toxicity of these materials, differs by an
order of magnitude or less. As noted above the degree to which toxicity from thesesoils
can be attributed to 2,3,7,8-TCDD in the presence of numerous other related toxic
compounds is not known.. 2,3,7,8-TCDD tissue concentrations were not reported in this
work..
In another comparative study Umbreit et al. (1987b) compared the Newark
manufacturing site and Times Beach soils in the induction of aryl hydrocarbon
hydroxylase (AHH) in rats. While the use of only single dose levels prevents
detailedanalysis, the two soils proved quite similar in their ability to induce AHH. The
explanation for the difference, in this finding from those observed in the toxicity studies
discussed above is not clear, but may relate to the presence of other toxic and/or AHH
inducing compounds.
Umbreit et al. (1987a) report a reproductive toxicity study with soils from the
Newark manufacturing site and salvage yard previously studied by Umbreit (1986a).
Female mice were treated thrice weekly with soil from these sites, with treatment
continuing through fertilization to weaning of pups. The total doses of 2,3,7,8-TCDD
received by the mice were 720 pg/kg in manufacturing site soil, and 86 /*g/kg in salvage
yard soil. A corn oil vehichle group and are contaminated soil group received a total of
225 /jg/kg.
Deaths in animals showing "classic signs" of TCDD toxicity were observed in the
corn oil and recontaminated soil groups, and indicate appreciable bioavailability of
2,3,7,8-TCDD. Deaths were also observed in animals receiving manufacturing site soil but
the authors did not observe "classic signs" of TCDD toxicity. Fewer live pups born and
fewer pups surviving until weaning were observed in the manufacturing site soil group
104
-------
compared with those receiving decontaminated soil. TCDD completely blocked
reproduction in the corn oil and recontaminated soil groups. The results of this study
demonstrate acute and reproductive effects occurred in animals receiving manufacturing
site soil. However, these effects were of a lesser magnitude than those seen in animals
treated with 2,3,7,8-TCDD in corn oil at a dose three fold lower. The authors note the
presence of substantial quantities of other toxic substances in the manufacturing site soil
(chemical analyses presented). No toxic effects were noted in animals treated with
salvage site soil, who received a much smaller 2,3,7,8-TCDD dose. The data does not
allow a quantitative evaluation of the bioavailability of 2,3,7,8-TCDD.
Kaminski et al. (1985) and Silkworth et al. (1982) reported the results of a series of
studies on the toxicity of soot containing dioxin and furan compounds from a fire
involving transformer fluid containing PCBs. Hartley guinea pigs (500 to 600 g) received
single oral doses of soot in an aqueous vehicle, a soxhlet extract of the soot in the same
vehicle, or 2,3,7,8-TCDD in either an aqueous vehicle or corn oil.
The soot was reported to contain 2.8 to 2.9 ppm 2,3,7,8-TCDD and 124 to 273 ppm
2,3,7,8-TCDF. The total polychlorinated dibenzofuran content was estimated at 5,000
ppm. Animal weights and mortality were recorded for 42 days, at which point the
survivors were sacrificed and LD5Q values were calculated. Blood chemistry and a
pathologic examination were performed at sacrifice. Table 5-4 summarizes results from
these experiments.
Silkworth et al. (1982) noted that the LD50's for contaminated soot and soot extract
were similar at 410 and 327 equivalent mg/kg, indicating that the matrix had only a small
effect on toxicity. If expressed in terms of the content of 2,3,7,8-TCDD, the LD5Q from
soot is 2.5 Mg/kg, which is a factor of seven below the LD5Q for 2,3,7,8-TCDD in an
aqueous vehicle, suggesting that other compounds contributed to the toxicity of the soot
and soot extract.
105
-------
TABLE 5-4. GUINEA PIG MORTALITY AND WEIGHT CHANGES FOLLOWING
TREATMENT WITH CONTAMINATED SOOT OR 2,3,7,8-TCDD
Contaminated soot in aqueous vehicle - Males
Dose eq. Dose,2,3,7,8- % Weight gain Estimated
Treatment soot mg/kg TCDD /*g/kg Mortality (0-42d) LD5Q (95% C.I.)
Methyl
cellulose
vehicle
Active
carbon
Untreated
Contami-
nated soot
H
H
ti
Active
carbon
Contami-
nated soot
n
n
n
0
0
0
1
10
100
500
Contaminated
0
1
10
100
500
0
0
0
0.003
0.03
0.3
1.5
soot in aqueous
0
0.003
0.03
0.3
1.5
N2
N
N
N
N
N
N
vehicle - Females -
N
N
N
N
2/6
48
38
37
37
34
18
-1
1st Experiment
20
22
13
7
-5
106
-------
TABLE 5-4 (continued).
Contaminated soot in aqueous vehicle - Females - 2nd Experiment
Dose eq. Dose 2,3,7,8- % Weight gain Estimated
Treatment soot mg/kg TCDD jig/kg Mortality (0-42d) LD5Q (95%C.I.)
Active
carbon
Contami-
nated soot
n
n
tt
ti
Control
(unspec)
Soot
extract
M
soot equiv.
II
it
ti
0
250
500
750
1000
1250
Soxhlet extract
0
4
20
100
500
1000
0
0.7
1.5
2.2
2.9
3.6
of contaminated
0
0.01
0.06
0.3
1.5
0/6
0/6
3/6
6/6
6/6
6/6
soot in aqueous
0/6
0/6
0/6
0/6
4/5
6/6
35
6
-10
_*
_*
_*
vehicle - Females
41
38
28
21
_*
_*
410 mg/kg
(281-604)
327 mg/kg
(183-583)
107
-------
TABLE 5-4 (continued).
TCDD in aqueous vehicle - Females
Dose eq. Dose 2,3,7,8- % Weight gain
Treatment soot mg/kg TCDD /ig/kg Mortality (0-42d)
C.I.)
Untreated
Methyl
cellulose
2,3,7,8-TCDD
tl
tt
2,3,7,8-TCDD
n
n
n
Untreated
Corn oil
2,3,7,8-TCDD
2,3,7,8-TCDD -
n
H
ti
n
0
0
0.1
0.5
2.5
12.5
20.0
TCDD
0
0
0.1
0.5
2.5
12.5
20
0/6
0/6
0/6
0/6
0/6
0/6
4/6
in corn oil - Females
0/6
0/6
0/6
0/6
3/6
5/5
6/6
39
31
28
29
25
33
11
39
22
37
24
7
_*
_*
Estimated
LD50 (95%
19 Mg/kg
(15-23)
(1.2-5.4)
*Major weight loss seen before death.
Source: Silkworth et al. (1982)
^Data on survivors only, data read from graph.
2N: No deaths mentioned by authors, group size 4-6
•'As quantities of chlorinated dibenzofurans were present, 2,3,7,8-TCDD, alone, may not
be responsible for toxicity. The 2,3,7,8-TCDD dose is based on 2.9 ppm in soot.
108
-------
The authors stated that they adopted an aqueous vehicle in these experiments
because it was nontoxic and provided a stable suspension of soot; they regarded this
vehicle as more appropriate for modeling of human exposure conditions than an oil
vehicle. The data from these experiments also demonstrate that use of an oil vehicle
leads to substantially greater 2,3,7,8-TCDD toxicity than does an aqueous vehicle.
Comparison of mortality and weight loss in groups of female guinea pigs receiving
500 A*8/kg of soot or the equivalent amount of soot extract suggests that the extract
may be somewhat more toxic; however, all six animals died in the 1,000 /xg/kg soot
group, while four out of five died in the 500 /ig/kg extract group. Taken together,
these dataindicate that the soxhlet extract of soot in an aqueous vehicle was between
one and two times as toxic as the soot itself. It is likely that a larger difference in
toxicity would have been observed if the soot extract was in an oil vehicle.
Van den Berg et al. (1983) fed small groups of male Wistar rats fly ash from a
municipal incinerator (pretreated with HC1) containing dioxins and furans, a soxhlet
extract of the fly ash, or a purified extract of the ash that was obtained using column
chromatography. Table 5-5 lists the concentrations of dioxin and furan groups in these
materials. 2,3,7,8-TCDD was present as 3.3% of the TCDD isomer group in the fly ash
extract. (The authors did not specify whether this reference was to crude or purified
extract.) 2,3,7,8-TCDF was present as 17.9% of the tetra-CDF isomer group in the
extract. The rats were fed 2 g/d fly ash mixed with diet or the residual from 2 mL/d
extract after the extract was mixed with diet and the solvent was evaporated. The
animals were exposed to the treated diet for 19 days, and then sacrificed, and the liver
tissue was analyzed for the presence of dioxins and furans. Table 5-6 shows the average
concentrations of isomer groups found in the livers. (Results were not presented
separately for 2,3,7,8-TCDD or 2,3,7,8-TCDF.) Table 5-7 gives the percentages of the
109
-------
TABLE 5-5
CONCENTRATIONS OF PCDD AND PCDF ISOMER-GROUPS IN FLY-ASH EXTRACTS
(DILUTED WITH ACETONE) AND FLY-ASH
Extract (I)
(crude)
Extract (II)
(purified)
Fly-ash (III)
Tetra-
CDD
322
475
245
Tetra-
CDF
400
609
314
Penta-
CDD
493
580
422
Penta-
CDD
485
624
500
Hexa-
CDD
500
539
562
Hexa-
CDF
438
587
660
Units
ng/mL
ng/mL
ng/g
Source: van den Berg et al. (1983)
110
-------
TABLE 5-6
CONCENTRATIONS OF PCDD AND PCDF ISOMER GROUPS
IN LIVERS OF RATS FED FLY ASH MATERIALS
(ng/g)1
Tetra-CDD Tetra-CDF Penta-CDD Penta-CDF Hexa-CDD Hexa-CDF
Fly ash 1.5
Fly ash
extract
(crude) 2.4
Fly ash
extract
(purified) 0.8
8.1 17.8 39.2 36.8 64.2
21 20.0 60.5 58.6 102.4
3.5 2.7 10.9 5.1 13.8
Source: van den Berg et at. (1983)
^Each Entry is the average for two rats.
Ill
-------
TABLE 5-7
PERCENTAGE OF CUMULATIVE DOSE OF DIOXINS AND FURANS
PRESENT IN RAT LIVER FOLLOWING 19 DAYS EXPOSURE IN DIET
Terra- 2,3,7,8- Tetra- 2,3,7,8- Penta- Penta- Hexa- Hexa-
CDD TCDD CDF TCDF CDD CDF CDD CDF
% cum. dose
in liver —
fly ash1 0.11 0.9 0.4 0.3 0.2 0.7 0.3 0.7
%cum. dose
in liver —
fly extracts2 0.16 3.7 0.9 1.0 1.2 2.9 3.0 5.3
Ratio of perct.
in liver
ash/extract) 0.7 0.2 0.4 0.3 0.2 0.2 0.1 0.1
Source: van den Berg et al. (1983)
1 Average retention in two treated animals.
2Average of retention for extract and purified extract groups, total
four animals.
112
-------
cumulative administered doses found in the liver at sacrifice, and the ratios between the
percentages in liver for fly ash and fly ash extracts.
Approximately 1% of the 2,3,7,8-TCDD dose from fly ash was retained in the liver,
and approximately 4% of the dose of this isomer from fly ash extract was so retained.
The corresponding percentages for 2,3,7,8-TCDF are 0.3 and 1.0. Data on the retention of
isomer groups in adipose tissue were presented for the extract-treated groups but not for
the fly-ash-treated group. The concentrations of the various isomers in adipose tissue
are comparable to, or less than, the concentrations in liver tissue.
The U.S. EPA (1985d) reported a half-life for elimination of 2,3,7,8-TCDD in the rat
of 20 days at high dose. If a similar half-life is assumed in this experiment, the
quantities of 2,3,7,8-TCDD in the animals at the end of the 19-day feeding experimen
would be significantly less than the absorbed dose, but still of the same order of
magnitude. However, the recovery percentages in this study are low for both the fly ash
and fly ash extract groups in comparison with other studies in which 2,3,7,8-TCDD was
administered to rats. Fries and Marrow (1975) fed rats diets containing 7 or 20 ppb of
2,3,7,8-TCDD for a period of up to 42 days. After 14 days of feeding, the rat livers
contained an average of 32% of the cumulative administered dose; at 28 days, 21% of the
dose; and at 42 days, 18% of the dose. Thus, in the van den Berg et al. study, the liver
retention of 2,3,7,8-TCDD for the fly ash extract group is a factor of five to eight below
what could be anticipated for the Fries and Marrow data, and the liver retention in the
van den Berg group fed soot is a factor of 20 to 30 lower than that seen by Fries and
Marrow. Data from Kociba et al. (1976), Rose et al. (1976), and Kociba et al. (1978) lead
to similar conclusions to those from the Fries data regarding the fraction of cumulative
2,3,7,8-TCDD dose retained in the rat liver.
113
-------
An explanation of the low level of recovery for the animals receiving the soxhlet
extract of soot is not apparent. It is possible that the presence of multiple compounds
affected absorption or metabolism in the rats fed soot and soot extract.
A second approach to the van den Berg et al. data is to compare the ratios of liver
concentrations for dioxins in fly-ash-treated animals to the concentrations in extract-
treated animals. These ratios, based on measurements in small numbers of animals,
indicate a substantial bioavailability of dioxin and furan compounds from the tested fly
ash. This availability varied among the different isomers with the value of 0.3 for
2,3,7,8-TCDD, indicating that this isomer was three times as available from fly ash
extract as from fly ash.
Van den Berg (1985) fed fly ash (pre-treated with HC1) to Wistar rats, guinea pigs,
and Syrian golden hamsters. Fly ash was mixed with standard laboratory diet at 2.5% by
weight, and animals were allowed to eat ad libitum. The amount of fly ash consumed by
each group of five rodents was determined by the authors. For each species there were
three groups of animals each fed fly ash for approximately 32 days (group I), 60 days
(group II), or 94 days (group III). Concentrations of dioxin and furan isomer groups in
the food were presented, and include 1.4 ng/g TCDD compounds and 2.1 ng/g TCDF
compounds.
The authors presented calculated recovery percentages for the cumulative dose of
specific isomers in the rodent liver. For 2,3,7,8-TCDD in guinea pigs, 3.7%, 0.9%, and
1.4% of the administered dose was recovered in the liver in groups I, II, and III,
respectively. The 32-day (group I) recovery percentage is somewhat higher than seen in
the lower dose groups receiving 2,3,7,8-TCDD contaminated soil in McConnell et al.
(1984). The value in hamsters was approximately 2% (only reported for group II), and
analytical problems prevented this determination in rats. No other TCDD compounds
were quantitated. Similarly, for 2,3,7,8-TCDF, guinea pigs showed retention of 4.7%, 2.2%,
114
-------
2.5% of the administered dose in groups I, II, and III, respectively. For both 2,3,7,8-
TCDD and 2,3,7,8-TCDF the recovery percentages in guinea pigs at 32 days were
approximately a factor of 4 to 15 higher than that observed in the van den Berg et al.
(1983) study in rats.
Other TCDD compounds that were present showed comparable or somewhat lower
retention, averaging 1% to 2% over the animals groups. No TCDD or TCDF compounds
were detected in hamster liver or analyzed for in rat liver. Higher chlorinated isomers
most typically showed retention in the range of 2% to 5% in rat liver and 1% to 3% in
guinea pig liver, with the exception of 2,3,4,7,8-PnCDF (9.8%, 8.3%, and 11.3% in the
hamster groups). Few other compounds were found in hamster liver, but 2,3,4,7,8-PnCDF
was found with a recovery of 5% to 8% and 2,3,4,7,8-HxCDD was found at 3% to 7%.
As with other experiments in which the retention of dioxins in the liver has been
determined, these percentages place a lower bound on the bioavailability of the dioxins
but, because not all dioxin is localized in the liver, do not permit bioavailability to be
estimated without knowledge of the elimination of the administered dose over time and
the quantity of dioxins in the remainder of the organism. No positive control group
receiving 2,3,7,8-TCDD was included for comparison.
Poiger and Schlatter (1980) conducted several experiments in Sprague-Dawley rats
(180 to 220 g) in which liver concentrations of tritium label from 2,3,7,8-TCDD were
determined using various doses and vehicles. All experiments consisted of a single
gastric intubation of 2,3,7,8-TCDD-containing material, followed by animal sacrifice at
predetermined times. The doses used were well below the LD5Q in the rat (the maximum
dose applied was 5 /jg/kg), and no deaths or toxic effects were reported.
In a preliminary experiment, rats were treated with 14.7 ng/rat 2,3,7,8-TCDD in
ethanol. Table 5-8 shows the percentage of recovery of the administered dose at various
times.
115
-------
TABLE 5-8. PERCENTAGE OF TRITIUM-LABELED 2,3,7,8-TCDD IN RAT LIVER
FOLLOWING ADMINISTRATION OF 14.7 ug DOSE IN ETHANOL
Number of
animals
2
7
6
2
2
2
Time after treatment
(hours)
8
24
48
72
96
120
Percentage of
dose in liver
33.2
36.7+ 1.2
30.8 ± 2
22.3
17.5
17.5
SOURCE: Poiger and Schlatter, 1980.
116
-------
These data indicate substantial localization of 2,3,7,8-TCDD in the rat liver, with a
decrease of a factor of two in the fraction of the dose in the liver between 1 and 4
days. Poiger and Schlatter (1980) conducted all further studies with sacrifice at 24 hours
to maximize the recovery of 2,3,7,8-TCDD from the liver.
In a second experiment, the authors administered 2,3,7,8-TCDD doses in ethanol
ranging from 15 to 1,070 ng/rat to groups of six rats. They found a graded increase in
percentage retained in the liver from 37% + 1% at the 15 ng dose to 51% ± 4% at 280 ng.
•
At the high-dose point, the percentage may have fallen (42% + 10% at 1,070 ng).
In a further experiment, 2,3,7,8-TCDD was administered at low dose in a series of
vehicles. These results are shown in Table 5-9. These data demonstrate that
TABLE 5-9. PERCENTAGE OF TRITIATED 2,3,7,8-TCDD DOSE IN THE LIVER
24 HOURS AFTER ORAL ADMINISTRATION OF 0.5 ml OF VARIOUS MEDIA
Formulation TCDD dose
(ng)
50% ethanol 14.7
Aqueous suspension
of soil after
TCDD contact for:
8-15 hr (room temp) 12.7, 22.4
8 d (30 °C - 40 °C) 21.2, 22.7
Aqueous suspension
of activated carbon 14.7
# Rats % Dose in
liver
7 36.7 + 1.2
17 24.1+4.8
10 16 ±2.2
6 < or = 0.07
Source: Poiger and Schlater (1980)
administration of 2,3,7,8-TCDD in soil reduced the retention of the dose in the liver to
66%, or 44% of the retention seen with 2,3,7,8-TCDD in ethanol. The lower value, 44%
was obtained for soil that was aged for 8 days at 30-40 °C following addition of 2,3,7,8-
117
-------
TCDD. This observation is consistent with the findings of other studies reported here
that 2,3,7,8-TCDD from environmental soil (naturally aged) was generally less available
than 2,3,7,8-TCDD freshly added to clean samples of these soils. The aqueous suspension
of 2,3,7,8-TCDD in activated carbon showed little evidence of bioavailability; this is
supported by the authors' measurements showing that 2,3,7,8-TCDD was only
slightlyextractable from the activated carbon matrix by various solvents. In contrast,
58% to 70% of 2,3,7,8-TCDD could be recovered from soil samples by washing with
hexahe/acetone (4:1 v/v).
•
Poiger and Schlatter (1980) also presented results from several skin application
experiments with TCDD-containing materials using rats and rabbits (not reviewed here).
Bonaccorsi et al (1984) reported the results of a study of gut absorption of 2,3,7,8-
TCDD from soil taken from the Seveso, Italy accident site. Soil containing 81 ± 8 ppb
2,3,7,8-TCDD from the "highly contaminated" area in Seveso was administered to albino
male rabbits (2.6 ± 0.3 kg) in daily gavage doses for seven days. Additional samples of
clean soil were spiked with 2,3,7,8-TCDD in the laboratory to yield 10 and 40 ppb
contamination levels and were administered to rabbits following the same protocol. For
comparison, rabbits were also treated with 2,3,7,8-TCDD in solution in acetone-vegetable
oil (1:6) or alcohol-water (1:1). Rabbits were sacrificed on the day after treatment
stopped and liver concentrations of 2,3,7,8-TCDD were measured. Table 5-10 contains the
results of these experiments. The authors did not remark on the presence or absence of
toxicity in the treated rabbits. EPA (1985) reports values for the single dose LD$Q of
2,3,7,8-TCDD in rabbits of 115 and 275 /ig/kg. The total doses received by the rabbits
in this study were approximately 54, 107, and 215 us/kg over seven days. Based on this
comparison, there is a likelihood that toxic effects occurred in the Bonaccorsi work, and
noted above, toxicity has the potential to affect the tissue concentrations of 2,3,7,8-
TCDD. For this reason the most appropriate comparisons among these data are between
/
118
-------
groups showing similar liver concentrations of 2,3,7,8-TCDD, which may then be inferred
to have experienced similar toxic effects.
TABLE 5-10. GUT ABSORPTION OF 2,3,7,8-TCDD IN RABBITS AFTER 7-DAY
TREATMENT
Treatment
Vehicle
(/ig/dayb)
Acetone-Oil
Mixture
Alcohol or
Acetone-Oilc
Alcohol
Lab Con-
taminated
Soilb
ll it
n M
Seveso Soilb
It M
2,3,7,8-
TCDD Dose
20
40
80
20
40
80
80
160
Number of
Rabbits
5
16
5
7
13
10
7
7
Mean TCDD in
Liver ppb
0.26 ± 0.07
1.1 +0.3
2.7 ± 0.5
0.26 ± 0.08
0.81 ± .31
1.5 ± 0.2
0.88 + .28
2.2 ± 1.0
% Admin. Dose
in Livera
24.14
51.07
62.77
24.14
37.5
34.91
20.43
25.63
alne percentage or the administered dose found in the liver was calculated
assuming a liver weight of 5.0% of body weight; liver weights were not
reported by authors.
bl-2g soil administered by gavage in 10ml water. Soil sifted through wire
mesh.
cThe two vehicle groups are not broken out.
119
-------
That this method of comparison is desirable can also be seen from the Bonaccorsi
data, where both solvent vehicle groups and the spiked soil groups show an increase of
the fraction of the dose in the liver at the higher administered doses. However, it
should be mentioned that use of two different solvent vehicles complicates interpretation.
Similar liver concentrations of 2,3,7,8-TCDD were seen in the 40 jtg/d solvent vehicle
and 80 ng/d Seveso soil groups. Comparing the percentage of liver retention in these two
groups indicates absorption from Seveso soil was 40% of that from the solvent vehicle.
Using the same approach, comparison of the 80 A«g/d solvent vehicle and 160 A*g/d
Seveso soil groups indicates that absorption from the soil was 41% of that from the
solvent.
The same approach can be used to compare absorption from the solvent vehicle and
from the spiked soil. In this case the 40 /ig/d solvent vehicle group had the liver
concentrations closest to either the 40 or 80 pg/d spiked soil groups. Comparison of the
percentage of dose in the liver indicates absorption from spiked soil is 68-73% of that
from the solvent vehicle. Bonaccorsi et al (1984) work conducted with either aged or
non-aged spiked soil but do not present data to allow a comparison of these groups.
Shu et al. (1987, as cited by Leung and Paustenbach, 1987) study of 2,3,7,8-TCDD
from the Missouri site tested by McConnell et al. (1984). Their paper "reports an oral
bioavailability of approximately 43% in the rat dosed with environmentally contaminated
soil from Times Beach, Missouri. This figure did not change significantly over a 500-fold
dose range of 2 to 1450 ng 2,3,7,8-TCDD per kg of body weight for soil contaminated
with approximately 2, 30 or 60 ppb of 2,3,7,8-TCDD. The data from this study is not now
available to the Exposure Assessment Group for review.
3. Summary
Table 5-11 summarizes data that are pertinent to the bioavailability of 2,3,7,8-TCDD
from environmental matrices. Studies of bioavailability, which examined soil samples,
120
-------
soot, and fly ash, have utilized three methodologies: measuring acute toxicity, retention
of 2,3,7,8-TCDD in the liver, and induction of hepatic enzymes.
Among the five samples of soil from contaminated sites that have been tested, three
have shown substantial bioavailability, e.g., 25% to 50%, when compared with 2,3,7,8-TCDD
in corn oil gavage. A fourth soil sample was compared with 2,3,7,8-TCDD administered
in a solvent vehicle, and fell in this range. The fifth soil, tested by Umbreit et al.
(1986a,b; 1987a,b) showed bioavailability substantially less than the other soils tested.
While difficult to gauge quantitatively, dioxin from this soil may be an order of
magnitude less available than from the other soils.
Additionally, three samples of soil spiked with 2,3,7,8-TCDD have been tested for
bioavailability, including one sample in which the 2,3,7,8-TCDD was incubated with soil at
an elevated temperature. The 2,3,7,8-TCDD added to these soil samples proved to be
highly available (e.g., 40% to 70%).
In one study, soot from a transformer fire containing dioxins and furans proved
similarly toxic to a soxhlet extract of the soot in an aqueous vehicle. However, the soot
extract may have proved more toxic if delivered in corn oil, as was 2,3,7,8-TCDD in the
soil studies. The availability of 2,3,7,8-TCDD and other dioxins and furans from
incinerator fly ash have been addressed by van den Berg et al. in extended feeding
studies. In these studies, liver retention of 2,3,7,8-TCDD from either fly ash or fly ash
extract proved low, with availability from fly ash being approximately 25% of that from
the extract.
121
-------
TABLE 5-11
SUMMARY OF DATA ON THE BIOAVAILABILITY OF 2,3,7,8-TCDD
FOLLOWING INGESTION OF ENVIRONMENTAL MATRICIES
Author
Material
Species
Dosing
Observation
Umbreit et al.
(1986a,b)
Soil
Newark
Manuf.
Site
Guinea
Pig
Single
Gavage
McConnell et al.
(1984)
Soil
Newark
Salvage
Site
Recontami-
nated Soil
Soil Times
Beach, MO
Guinea
Pig
Guinea
Pig
Guinea
Pig
Single
Gavage
Single
Gavage
Single
Gavage
McConnell et al.
(1984)
Soil
Minker
Site, MO
Guinea
Pig
Single
Gavage
2,3,7,8-TCDD in soil <10%
as toxic as in corn oil,
based on lethality and
weight loss.
2,3,7,8-TCDD in the manuf.
site soil had retention in
liver approx. 1% as great as
with salvage site soil.
Liver retention similar
to 2,3,7,8-TCDD in corn
oil from lower dose
McConnell et al. (1984)
data.
Toxicity similar to
equal dose of 2,3,7,8-TCDD
in corn oil.
data indicate
2,3,7,8-TCDD in soil appox.
25% as toxic as in corn oil.
Comparing animals dying
early, liver retention of
2,3,7,8-TCDD in soil group
approx. 50% of that in corn
oil vehicle group.
Comparing animals
surviving experiment,
liver retention of 2,3,7,8-
TCDD in soil group approx.
20% of that in corn oil
vehicle group.
^059 data indicate soil
approx. 30% as toxic
as 2,3,7,8-TCDD in corn
oil.
122
-------
Author
Material
TABLE 5-11 (continued).
Species Dosing
Observation
McConnell et al.
(1984) and Lucier
Soil
Minker
Site, MO
Rat
Poiger and
Schlatter
(1980)
Bonaccorsi
(cite in
McConnell et
al., 1984)
Kaminski et al.
(1985) and
Silkworth et
al. (1982)
van den Berg
et al.(1983)
Soil with
2,3,7,8-TCDD
Soil Seveso
Accident Site
Soot
from Fire
Rabbit
Guinea
Pig
Incinerator
Fly Ash
Rat
Comparing animals dying
early, liver retention
approx. 50% of that in
corn oil vehicle group.
Comparing animals surviving
experiment, liver retention
approx. 25% of that of
corn oil vehicle group.
Single Introduction of AHH and
Gavage UDP glucuronyltrans-
ferase activity > 50%
of that in groups receiving
2,3,7,8-TCDD in corn oil.
Liver retention 20-40%
of that in rats receiving
equal dose of 2,3,7,8-TCDD
in corn oil.
Liver retention approx.
40-70% of that in ethanol
vehicle groups.
2,3,7,8-TCDD 30% as bio-
available from soil as from
solvent vehicle.
Single LD5Q data indicate soot
Gavage containing dioxins and
furans approx. equal in
toxicity to soxhlet extract
of soot in aqueous
vehicle.
19 day Liver retention of 2,3,7,8-
Feeding TCDD from ash and ash
extract 1% and 4%
respect., indicating
2,3,7,8-TCDD approx. 25%
as avail from ash as from
extract. Both ash and
and extract retentions
are low compared with other
feeding and gavage studies.
123
-------
TABLE 5-11 (continued).
Author Material Species Dosing Observation
van den Berg
et al. (1985)
Incinerator
Fly Ash
Guinea
Pig
Feeding 4%, 1% and 1% retention
of total dose in liver
following feeding for
32, 60, and 94 days,
respect.
Van den Berg Incinerator Hamster Feeding 2% of total dose retained
(1984) Fly Ash 60 Day in liver following feeding.
Poiger and 2,3,7,8-TCDD Rat Single <0.1% retention in liver.
Schlater on activated Gavage
(1980) Carbon
124
-------
The individual studies reviewed have a variety of limitations, as discussed in the
preceding text. A notable limitation was that some experiments were conducted at using
highly toxic doses of 2,3,7,8-TCDD, so that determination of bioavailability was
complicated by wasting and early death of the test animals. It should also be noted
that, while the relative retention of 2,3,7,8-TCDD in the liver can serve as an
appropriate indication of differences in bioavailability between samples, the percentage of
dose found in the liver only places a lower bound on absorption. This is particularly
relevant to experiments where animals have been maintained for many weeks after dosing
and an undetermined quantity of 2,3,7,8-TCDD has been excreted.
Finally, toxicity data for mixtures for which both toxicity and bioavailability of
individual compounds may vary are difficult to interpret quantitatively in terms of
bioavailability.
As presented in U.S. EPA (1985c), Rose et al. (1976) determined gut absorption of
2,3,7,8-TCDD in a 1:25 mixture of acetone to corn oil (by volume) in the rat. In both
single dose and multiple dose experiments, measured absorption was approximately 85%.
Assuming that absorption from pure corn oil is similar to that from this mixture, and
assuming that absorption in other species for which data are not available is similar, the
85% factor can be applied to the data presented here to obtain an approximate range for
typical 2,3,7,8-TCDD absorption from soil. Using this factor, the estimated relative
bioavailability of 2,3,7,8-TCDD from soil is 25% to 50% and, when compared with corn oil,
provides an estimate of gut absorption of 20% to 40% of ingested 2,3,7,8-TCDD in soil.
This estimate is comparable with the 20% to 26% absorption from 2,3,7,8-TCDD treated
soil from the work of Poiger and Schlatter (1980).
Recognizing these limitations, the weight of evidence indicates that 2,3,7,8-TCDD is
often highly available from environmental materials. However, in one tested soil sample
the compound was substantially less bioavailable. While the data are too sparse to allow
125
-------
a prediction as to whether a particular environmental sample will prove more or less
bioavailable, one important suggestion has emerged. In the two samples that have proved
least bioavailable (the Umbreit et al. (1986a) manufacturing site soil sample, and 2,3,7,8-
TCDD on activated carbon tested by Poiger and Schlatter (1980)) the 2,3,7,8-TCDD was
largely resistant to solvent extraction. This was not the case for more bioavailable
materials.
Further research, using short-term experiments in which animals are handled under
identical conditions and are fed dioxins in different media, is needed for an improved
comparison of absorption between different environmental samples. Acutely toxic doses
should be avoided to ensure that tissue concentrations are directly interpretable.
Experiments studying both tissue retention and enzyme induction should prove valuable
for this research. Whole-body levels of 2,3,7,8-TCDD need to be related to liver
concentrations, and the effects of metabolism need to be addressed. The vehicle of
administration has been shown to affect acute 2,3,7,8-TCDD toxicity, and vehicle effects
. need to be considered in designing experiments.
B. PHARMACOKINETICS AND BODY BURDEN OF DIOXINS
The pharmacokinetic profiles of CDDs and other related compounds, such as the
polychlorinated dibenzofurans, are quite complex. However, a thorough analysis and
understanding of these pharmacokinetic data could prove very helpful in ensuring that
exposure assessments for CDDs are reliable. In addition, such profiles may be useful in
providing information to help understand and apply the data from animal studies to
human exposure and risk assessments.
Wroblewski and Olson (1985) reported differing levels of responses to CDDs by
various species. It may be that some of these differences can be explained by examining
and quantifying the species differences in disposition and metabolism (Wroblewski and
Olson, 1985; King et al., 1983). The varying responses to various isomers may also be
126
-------
explained by differences in the disposition and metabolism (Burant and Hsia, 1984). It
may be well to note that the definition of "disposition" may have to be extended to
include suborgan or even subcellular sites in order to more fully describe some of the
noted differences. As will be discussed in greater detail later in this section, CDDs have
been implicated to bind with very specific loci within cells. The structure and
concentration of these intracellular receptors appear to be under genetic control, and
thus may exhibit considerable inter- and intra-species variation.
Pharmacokinetic analysis may also allow for predicting the time required for
eliminating the body burden after exposure ceases. With sufficient data and proper
understanding, these analyses can account for various exposure and physiologic
conditions.
The redistribution of a CDD among the various tissues and organs, which may occur
during elimination, can be accounted for and tracked. Effects on disposition, which may
result from altered physiology, such as from sudden weight loss or from lactation, can be
incorporated and thus adequately considered in exposure and risk assessments. Lactation
(Astrila, et al., 1981) and pregnancy (Nau and Bass, 1981) are known to accelerate the
removal of CDDs from the body. This increased elimination may be at the expense of
accumulation by the embryo, fetus, or offspring.
In the future, as more becomes known about the mechanisms of action of these
compounds, it may become necessary to extend the traditional physiologic pharmacokinetic
analysis to subcellular sites. The risk assessor would be greatly aided by predictions and
descriptions of toxin disposition at sites of toxicity. This factor becomes particularly
significant if concentrations at target sites deviate significantly from linear relationships
with ambient concentrations.
One potentially powerful and practical application of pharmacokinetic analysis is to
estimate exposure levels from body burden data. The goal here would be to use as much
127
-------
human data as are available, in order to avoid uncertainties that arise when making
extrapolations from animal data.
Some of the applications mentioned thus far are more easily applied than others for
these compounds. Significant data exist regarding these compounds, and some attempts
have been made to apply pharmacokinetic analyses to exposure assessments. Other
applications will require more research and data gathering. It is well to note that
assumptions are made when making predictive statements based on pharmacokinetic
analyses. The rest of this chapter will discuss some present and future applications and
some of assumptions involved.
1. Body Burden: Estimate of Exposure
"Commoner Approach": Commoner et al. (1985, 1986) discussed ways to calculate the
intake of 2,3,7,8-TCDD per day from human adipose tissue data. A major factor of the
data base for these compounds is that significant documentation exists regarding the
amounts that are found in human adipose tissue. After making certain assumptions one
may then "back-calculate" what the exposure was to give such levels in the adipose
tissue.
A commonly used function to describe repeated dosing, as would occur with chronic
ingestion, would be:
D = (M)(ka)(l - e-kat) (5-1)
where D = dose, M = mass of material taken in, ka = absorption rate constant, and t =
time after first dosing.
C = D/V =(M)(ka)(l - e'kat) (5-2)
where C = concentration, and V = volume of body (or compartment, etc.) If one assumes
steady-state conditions (material coming into the body equals material leaving the body)
128
-------
for relatively long periods of time,
Css = M/V (5-3)
where Css = concentration at steady-state. Commoner et al. defines I as the intake per
day based on the steady-state concentration:
I . Css (ke)a (5-4)
where ke = first-order elimination rate constant, and a = fraction of body which is
adipose tissue.
Note that the Css used here is the actual concentration measured in adipose tissue
samples. The adipose tissue fraction is assumed to be 0.14, and ke is determined from
the half-life of the compound in the body according to:
ke = (In 2)/t1/2 (5-5)
Some of the assumptions that are inherent in this approach are as follows:
(1) After a long period of time, steady-state is reached within the adipose tissue.
This assumption greatly simplifies these calculations, because the absorption
rate constant, ka, need not be independently determined. This assumption only
applies for exposure under relatively stable conditions for long periods. If the
half-life of this compound in the body is 5 years, it would require 16 years for
the concentration to reach 90% of steady-state. This further assumes stable
exposure conditions for those 16 years.
(2) Whole-body concentration and adipose tissue concentration are linearly related,
with the adipose tissue fraction being the constant of proportionality.
Normally, one finds that the constant called "fraction of the dose absorbed,"
Fo, is related to several factors, including bioavailability, gut contents, other
chemicals in the mixture, etc. From studies with similar compounds (King et
al., 1983) it appears that a multicompartment model analysis best describes the
129
-------
data. Thus, assuming that the adipose tissue fraction is the constant of
proportionality may introduce some error.
(3) Determining ke from the biological half-life (tj/2) implies a simple, one
.compartment model. Given the work on 2,3,7,8-TCDF and the tendency for
CDDs to partition in lipid, it seems unlikely that such a one compartment
analysis would be adequate. This could result in an overestimate of ke, which
would in turn result in an overestimation of the daily input, I.
(4) The parent compound is the toxin and thus the species of interest. This
assumption is reasonable and appears to be supported by several studies
(Burant and Hsia, 1984; Wroblewski and Olson, 1985). Also, studies indicate
that only the parent compounds have a tendency to bioaccumulate (Weber et
al., 1982).
This analysis cannot account for the redistribution among various body organs
resulting from physiologic and biochemical changes such as weight loss, nor for the
variation in ke which may exist among individuals. Further, this analysis cannot be used
to determine any individual organ, cellular, or sub-cellular dispositions.
In summary, this method is designed to calculate, from adipose tissue
concentrations, the average daily intake. Because of some simplifying assumptions that
are inherent, such as assuming a single compartment and linear, first-order elimination,
this method may miscalculate the actual input.
2. Physiologically Based Pharmacokinetic Modeling
An alternate method for describing and predicting disposition within the body is to
use physiologically based pharmacokinetic models. These models take into account
physiologic and biochemical processes such as blood flows, metabolism, and renal
clearance, and describe the body according to its normal anatomy. Physiologically based
pharmacokinetic models can, given adequate data, predict disposition from one exposure
1.30
-------
scenario to another and even from species to species. One such model was developed for
2,3,7,8-TCDF (King et al., 1983) and is described here with some modification for the
dioxins. First, the anatomic regions of the body are: blood, liver, fat, skin, and muscle.
The other organs are all lumped together as the "carcass." Input may be by a variety of
routes, but for the purposes of this discussion, is considered to occur through the
gastrointestinal system by continuous or chronic dosing. Only the parent compound is
tracked in each organ because of its presumed toxicity. The pertinent equations are:
Liver:
VL(dCL/dt) = QLCB -(QLCL)/RL-(kmCL)/RL+ D (5~6)
where V = Volume of liver, QL = blood flow to liver, CL = concentration of TCDD in
liver, Cg = concentration of TCDD in blood, RL = partitioning ratio of TCDD between
liver and blood, km = first-order elimination rate constant, D = dosing function.
Fat:
vF(dCF/dt) = QF(CB -CF/RF) (5-7)
Terms are analogous to those for liver. Skin, carcass, and muscle equations are
analogous to that for fat.
Blood:
VB(dCB/dt) = QLCL/RL + QFCF/RF + QmCm/Rm + QSCS/RS + QCCC/RC - QBCB(5-8)
Equations are also written to account for metabolite formed and excreted through
the bile.
Parameters:
(V) Organ volumes: Literature
(Q) Blood flows: Literature
(R) Partition ratios: Estimated from data
(further discussed below)
131
-------
(km) Metabolic clearance rates: Estimated from data
(in vol/time)
(ka) Absorption rate constant: Estimated from data
Obviously, such an approach requires a great deal of data simply to estimate the
parameters. Some parameters, such as the partitioning ratio, can be determined from in-
vivo or in-vitro animal studies or from mathematical analysis of time versus
concentration profiles (King et al., 1983).
The partitioning ratios for humans would have their initial estimates set at values
very close to other species. For this compound it may be reasonable to assume that the
major difference among the species is metabolism and not partitioning ratios.
The clearance rate, km, can be determined in various animal species and then scaled
to humans (King et al., 1983). A more acceptable approach would be to set the
elimination rate constant from human elimination data. The elimination profile after
exposure ceases would be necessary in order to calculate this parameter.
Some assumptions may be made to simplify the model for use in estimating intake
from fat (adipose) tissue concentrations determined in actual human samples:
If one assumes fat tissue to be in steady-state (the limitation of this assumption
was discussed previously), then
CF = CF,SS (5-9)
and
VF = d£F - 0
dt (5-10)
and thus,
Qc = (QF/RF)CF,SS (s-ii)
Solving for
CB =(CF,SS)/RF (5- 12)
132
-------
If one assumes the other various organs to be in steady-state, then the equation for
liver becomes
0 = dCL/dt = -(kmCL)/VLRL -(QLCL)/VLRL + (QLCF>ss)/VLRL + (FoMka)/VL(5- 13)
Then
(F0Mka)/VL =(kmCL)/VLRL (5-15)
and
F0Mka =(kmCL)/RL (5-16)
and
= CB (5-n)
and at steady-state
CB = (CF,ss)/RF (5-18)
Thus,
I = MkaF0 =(CF>sskm)/RF (5-19)
With MkaFo defined as the intake function, the last equation resembles the equation
used by Commoner et al. (1985, 1986). In this equation km is the rate constant for
metabolic clearance. Rather than using the body fraction as in Commoner et al. (1985,
1986), the fat-to-blood partitioning ratio is used. The intake calculated in this manner
could be significantly different from that calculated by the Commoner Method.
Thus, it can readily be seen that three important factors that govern the
bioaccumulation of these compounds are bioavailability, lipid-to-blood partitioning, and
rate of elimination.
To estimate intake, several pieces of information are needed. Obviously, the
concentration of dioxin in the fat tissue is needed. Such information may be available
from various "adipose tissue" banks. However, some assumptions have to be made when
using such data. First, the assumption of steady-state is made. The basis for this is
that it is assumed that regular and repeated daily dosing occurs. Another assumption is
133
-------
that composite samples are representative for all persons at an exposure site. Given the
possible differences that may exist regarding absorption, metabolism, etc., between
individuals, the accuracy of such an assumption may be questioned. Another parameter
needed is the fat-to-blood partition ratio. This may be difficult to obtain for humans,
but may be estimated from animal data. There is evidence that for similar compounds
this parameter is nearly equal in all the species tested (King et al., 1983, Kahn, 1987).
The most difficult parameter to estimate for humans is the elimination rate
constant. One way to obtain this parameter would be to directly monitor humans for the
disappearance of the chemical of concern from the body after exposure was known to
have stopped. An alternate method would be to determine the parameter experimentally
in animals, and extrapolate to humans. This technique must be carefully evaluated before
application, however. Given the rather significant differences that exist among species,
care must be taken to ensure that the animal parameter is determined in species which
are representative of humans. Further, extrapolation methods for this parameter are not
uniformly accepted, and thus some uncertainty may arise.
Animal studies need to be performed to simulate steady-state conditions and validate
such methodology for determining intake. The animal experiments should be performed in
several species, at several doses, and under different conditions of bioavailability (see
Section A of Chapter 5) and non-steady-state conditions.
The major advantage of global physiologically based pharmacokinetic models is that
they may be modified to account for any number of circumstances. Obviously, at non-
steady-state conditions the mass balances for the model are considerably more complex;
thus, more parameters need to be determined. However, there may be a number of
circumstances where the global model is necessary.
There is ample evidence, for example, that during lactation (Astrila et al., 1981)
CDDs are mobilized from the fat stores and eliminated through the milk. Thus, it is of
134
-------
interest to both mother and child to track the course of the chemical, for example
2,3,7,8-TCDD. A reasonable assumption would be that the 2,3,7,8-TCDD mobilizes from
the mother's fat tissue, reaches the mammary glands, and there enters the milk. Suitable
mass-balance equations are written and added to the model. Given the proper
parameters, this model could also be modified and used to describe the 2,3,7,8-TCDD
concentration in the milk of exposed cows.
Similarly, if one needed to follow 2,3,7,8-TCDD passage across the placenta, suitable
mass-balance equations would be added to the model. All of these modifications would
require laboratory experimentation to estimate the necessary parameters and for ultimate
validation. However, when and if it becomes obvious that information is needed in these
circumstances, the necessary parameters would be estimated from appropriate experiments
and the model utilized.
Ultimately, this model could be expanded to include various subcellular sites. There
is a growing body of evidence which indicates that the possible mechanisms of action of
2,3,7,8-TCDD involve interaction and/or binding of the chemical with subcellular sites
(Hannah et al., 1986; DiBartolomeis et al., 1986). Thus, as more becomes known about
mechanisms of action, exposure assessments will need to be extended to these various
subcellular sites. A well formulated and validated physiologically based pharmacokinetic
model could be an excellent method to accomplish such extended exposure assessments.
It also appears that CDDs may alter normal metabolic and physiologic processes such
as cholesterol metabolism (DiBartolomeis et al., 1986) vitamin A bioavailability (Hakansson
and Ahlborg, 1985), lipid metabolism (Swift et al., 1981; Albro et al., 1986), biliary
excretion (Herman et al., 1986), and thyroxine activity (McKinney et al., 1985).
Thus, in the future the model can be modified to account for such alterations
induced by CDDs. Such pharmacodynamic models ultimately relate disposition with effect,
and can be a great aid in risk assessment.
135
-------
With the steady-state assumptions, Equation 5-19 may be used to estimate the fat
concentration at steady-state given an average intake over time, for example per day.
Alternatively, given the concentration in the fat, an individual's daily intake may be
estimated. In doing so, total body clearance is substituted for metabolic clearance. To
use Equation 5-19 to estimate intake, several pieces of information are needed.
Obviously, the concentration of the CDD in the fat is needed. These data are generally
available for humans from various "adipose tissue" banks. For animals, these data may be
determined when conducting experimental studies. Another parameter needed is the fat-
blood partition ratio. This may be difficult to obtain for humans, but may be estimated
from animal data. There is evidence that for similar compounds this parameter is nearly
equal in all tested species (King et al., 1983). Total body clearance is the most difficult
parameter to estimate for humans. One way to obtain this parameter would be to
directly monitor humans for the disappearance of the chemical from the body after
exposure was known to have stopped. From these data, the elimination rate constant can
be determined and thus some estimate of clearance can be made. Elimination rate
constants and clearance may be related according to:
CL = (ke Vd) (5-20)
where CL = clearance, ke = elimination rate constant, and Vj = volume of distribution.
For a one-compartment model,
ke = In 2/t1/2 (5-21)
where tj/2 represents the half-life of the chemical in the body.
The volume of distribution may be estimated as in King et al. (1983) by summing
the product of organ volume times tissue-blood partition ratio for each of the individual
organs.
From these equations and laboratory data, some approximation of clearance in the
human may be made, as follows:
136
-------
(1) From animal experiments, determine elimination rate constant and
partition ratios. If data in humans exist, the elimination rate constant
need not be extrapolated from animals. (In this section, human half-lives
are extrapolated from animal data and compared to published human data.
The elimination constant used to calculate daily intake is taken from the
published data. The extrapolation result is compared merely to assess the
possible utility of such an extrapolation procedure.)
(2) From known organ volumes and partition ratios, calculate the volume of
distribution for 2,3,7,8-TCDD.
(3) From the volume of distribution and the elimination rate constant,
determine the clearance.
(4) Relate clearance in animal species to clearance in the human according to
body weight, to the 0.7 power, or determine clearance using available
half-life in the human and estimating the volume as described above.
(Determining elimination rate constants directly from human data reduces
the uncertainty associated with interspecies scaling.)
Obviously, careful consideration must be made when extrapolating any parameter
from animals to humans. It may be that both partition ratio and clearance should be
extrapolated from several animal species, if possible. As another possibility, the
extrapolations can be made from available data in non-human primates. Preliminary data
from rhesus monkeys are available (Bowman, et al., 1987) and have been used here.
Human half-lives for 2,3,7,8-TCDD were calculated using various values estimated from
the preliminary data in monkeys. The calculations were based on a fat-to-blood partition
ratio of 100 based on the monkey data (Bowman, et al., 1987) and human data (Kahn,
1987) and a monkey volume of distribution was calculated, as described above, to be 51.4
137
-------
liters. The human volume of distribution was calculated to be 720 liters. Table 5-12
summarizes the resulting half-lives:
TABLE 5-12. ANIMAL VS. HUMAN CLEARANCE AND HALF-LIVES OF TCDD
Monkey
clearance
(L/yr)
28.47
21.9
16.43
1-3.14
Human
clearance
(L/yr)
182.0
137.6
104.0
82.5
Half-Life
in monkey
(yr)
1.24
1.65
2.19
2.74
Half-Life
in human
(yr)
2.76
3.63
4.79
6.07
Commoner et al. (1985) reported that a survey of the available literature revealed
an average human half-life for 2,3,7,8-TCDD of 4.95 years; Poiger and Schlatter (1986)
found it to be 5.8 years. Thus, these clearance values are appropriate for estimating the
average intake of 2,3,7,8-TCDD based on human adipose tissue data. A survey of human
adipose tissues from the human adipose tissue bank reveals a wide range of concentration
of 2,3,7,8-TCDD. Data were gathered from a variety of geographical locations across the
United States and categorized according to three age groups. As an example, the data
used here were reported for persons over 45 years of age. The 45-year-old age group
could be approximated to be in steady-state if exposures over that period were fairly
constant regardless of living conditions and geographical moves during the span. The
time required to reach steady-state can be determined from the following equation for a
one-compartment model:
log [(Css - C) /Css] = - kt/2.303 (5-22)
-------
From this equation it may be determined that it would require approximately four
half-lives, or 20 years, to reach 90% of the steady-state concentration, and seven half-
lives, or 35 years, to reach 99% of the steady-state concentration.
3. Calculation of Daily Intake.
As discussed, a growing body of evidence may imply that there exists an overall
body burden of CDDs, including 2,3,7,8-TCDD in the general population of the United
States. If true, a number of important issues arise. First, sources of such a widespread
body burden need to be identified. Second, if such background exposures exist, then a
calculation of the carcinogenic risks based on such background levels need to be
performed.
To review, some basic questions require resolution in order to determine daily
intake and subsequent risk estimates from such background exposure. First, is it
reasonable to assume, after examining the available data that a body burden of 2,3,7,8-
TCDD exists in the general population of the United States? Second, if such a body
burden exists, what are the average daily intake levels that result in such background
levels? The following sections address these questions, attempt to identify and estimate
the uncertainties associated with such assumptions and calculations, and try to estimate
ranges of risk from such putative intake levels.
a. Data
Because of their high lipid solubilities CDDs tend to preferentially partition into
adipose tissues and reside there for long periods of time. Depending upon the exact
elimination rates they can then remain in human adipose tissue for well over 25 years.
Thus, the presence of CDDs in adipose tissue is good evidence for previous exposure.
Further, if steady state conditions are assumed, the adipose tissue levels may be used to
calculate average daily intake as described earlier in the chapter. As a result, most
investigators have taken fat biopsies to determine the body burden of dioxin in humans.
139
-------
iBlood .concentration can
-------
those individuals having an exposure history could have been low, the exposure could
have been sporadic, and could have occurred long before the monitoring. Thus the levels
of 2,3,7,8-TCDD due to phenoxy herbicide exposure, in those persons, could have had
little impact on their total concentration observed at the time of monitoring. In other
words, the concentrations in these persons might have returned back to background
levels.
Ono (1986) reports 9 ppt as an average concentration in the adipose tissues of 13
Japanese people with no known exposure to 2,3,7,8-TCDD containing substances. Persons
in rural areas of Georgia and Utah with no known exposure had 7.1 ppt in their adipose
tissues (Patterson et al., 1986). Grahm (1986) reported similar levels for samples taken
at autopsy.
Results of the National Human Adipose Tissue Survey (NHATS) (Stanley 1986)
showed 6.2 + 3.3 ppt in the composited adipose tissue samples taken at random from
throughout the U.S. at surgery or autopsy.
d. Conclusions Regarding Body Burden Data
Technical Resources, Inc. in support of EPA, reviewed all available data regarding
body burden in the U. S. Population (U.S. EPA, 1987). The open literature was first
reviewed and, where appropriate and necessary, authors were contacted and asked for the
original raw data. At this time, the Exposure Assessment Group staff conclude that it
would be reasonable to assume that background levels of 2,3,7,8-TCDD exist in some
persons in the U.S. It is also possible that background levels exist in the population at
large (i.e., virtually all persons in the U.S.). Review of the limited available data
indicate that an upper bound background level of 2,3,7,8-TCDD in the adipose tissue of
the U.S. population at this time is estimated to be about 6.7 ppt., but it is not possible
at this time to state what an average body burden would be for the U.S. population at
large.
141
-------
e. Calculation, Assumptions, Uncertainties, and Actual Parameter Values
The method chosen to calculate daily intake from adipose tissue concentrations is
developed and described in detail earlier in the chapter. Two major assumptions go into
the simplification of the model for purposes of calculations from the available data.
First steady state conditions are assumed. Given that the elimination of the compound
from the body is probably at least five years, it could take well over 15 years to reach
such steady state conditions. Thus, the assumption of steady state could be considered
reasonable only if background environmental concentrations are relatively similar
throughout the nation. Under those conditions even if people move from one geographic
area to another, exposure concentrations would be relatively constant. Also implicit in
the scenario of steady state is an assumption that the bioavailability of 2,3,7,8-TCDD is
virtually the same through the U.S. As discussed in section 5-A, this does not appear to
be the case. Therefore, even if environmental concentrations are the same for a person
moving from one area of the country to another the amount absorbed into the body
could vary significantly. Hence, steady state conditions may not be constantly
maintained. The result of this is that the adipose tissue concentrations measured at any
one time may not reflect actual steady state concentrations. Alternately, concentrations
measured at one time, although at steady state, may not reflect concentrations of a
steady state reached with a previous exposure to a form of 2,3,7,8-TCDD with different
bioavailabilty. Errors resulting from this assumption could either overpredict or
underpredict daily intake. Without knowing the bioavailability of the 2,3,7,8-TCDD at the
various locations to which an individual has been exposed and the length of time that he
or she may have been exposed at these locations it is not possible to measure the
amount of uncertainty associated with this assumption.
The second major assumption is that 2,3,7,8-TCDD is eliminated from the body by
monophasic kinetics. Data gathered within only a few years of exposure might not
142
-------
reveal any second or even third elimination phases that might exist. (It should be noted
that compounds with large fat-to-blood partition ratios frequently behave according to
two compartment kinetics with the fat acting as a "deep" second compartment). The
result from an error in this assumption is that the half-life would be erroneously
underestimated. As a result, the daily intake levels required to reach the steady state
levels in the adipose tissues would be overestimated. The impact of this can be more
easily minimized. Intake values can be calculated using a range of half-lives. Choosing a
range with sufficiently long half-lives assures that possible slower phase elimination
kinetic constants would also be included.
Also, the elimination kinetics are assumed to be constant over the entire life of the
individual. This may not be accurate in many cases. Variation in renal function,
metabolic capabilities caused by disease and exposure to other compounds would alter the
kinetics of elimination over the lifetime. For the purposes of these calculations, due to
the lack of relevant data, it is assumed that the elimination rate constant is relatively
stable with time. Again, by choosing half lives far greater than those estimated in the
literature one is assured of including those individuals with reduced elimination rates.
f. Parameters Chosen
The total fat volume for a 70 kg adult person was assumed to range between 5
liters and 14 liters. A common assumption is 10 liters. Based on the discussion above
an upper bound concentration of 2,3,7,8-TCDD in adipose tissue at steady state of 6.72
ppt was chosen. Based on several reports (Commoner et al., 1985; Poiger and Schlatter,
1980) and from extrapolation from elimination data in non human primates, a half-life of
approximately 5 to 8 years is assumed. Because of the possible underestimation of this
value as discussed earlier, however, a wider range of 5 to 30 years was chosen. As
discussed previously, the value of the partition coefficient between adipose and lean body
tissues was set at 100.
143
-------
ig. Daily Intakes Calculated
The smallest fat volume (5 liters), an adipose tissue concentration of 6.72 ppt and
±he longest half-life in the range (30 years) were used for calculating the lowest
•reasonable daily intake. The largest fat-.volume (-10 liters), the same concentration in the
adipose tissue (6.72 ppt) and the shortest half-life in the range (5 years) were used for
calculating the highest reasonable daily intake. A fat volume of 7 liters, the same
adipose tissue concentration of 6.72 ppt and a half-life of 10 years were arbitrarily
chosen :as a .reasonable expectation between extremes of the range. Table 5-13 shows ithe
.results.
TABLE 5-13. CALCULATED AVERAGE DAILY INTAKE
tl/2
(yrs)
5
5
10
20
30
30
Fat Vol
(L)
10
7
7
7
7
5
Vol of Dist
(L x 100)
14
10
10
10
10
7
,Conc in fat
(pg/gm)
6.72
6.72
6.72
6.72
6.72
6.72
Daily Intake
(pg/kg)
0.51
0.36
0.18
0.09
0.06
0.04
.In summary these calculations resulted in estimates of daily intake of 2,3,7,8-TCDD
between 0.04 picograms per kilogram body weight per day (pg/kg) and 0.51 pg/kg. The
"reasonable expectation" value is 0.18 pg/kg. This latter value is in reasonable agreement
with those reported by others (Geyer et al., 1986; Graham et al., 1985). Also, the values
calculated above are similar to those calculated using formulas discussed in a recent
paper by Leung and Paustenbach (1987).
h. Impact of Daily Background On Risk
Calculations are then performed as follows to estimate the upper bound risk that
could result from the background intake levels calculated in the preceding section:
144
-------
A potency based on animal studies of 1.5 x 10 is multiplied by the average daily
intake (adjusted for fraction absorbed) to calculate the estimated upper limit of risk.
The value of the potency is discussed in Chapter 6. Because the average daily intake was
determined from steady state conditions as described above the exposure duration is
assumed to be over the entire life time. For illustrative purposes the highest calculated
daily intake was used for estimation.
These upper limit risks are then compared with incidence of specific tumors and all
tumors for the general population. Table 5-14 summarizes the results of these
calculations and comparisons. As may be observed from examining Table 14 if 2,3,7,8-
TCDD were assumed to cause only human soft tissue sarcomas, the background intake
levels presently calculated would account for, at most, about 10% of the soft tissue
sarcomas observed in the general population. Similarly the background levels would
account, at most, for 1% of all non-Hodgkins lymphomas, and less than 0.1% of all
cancers in the general population.
i. Recommendations for Future Activities
TCDDs have interesting and important pharmacokinetic characteristics. There appear
to be significant differences between species in several pharmacokinetic properties and
parameters. For example, rodents and primates show very different elimination kinetics.
The impact that this and even the effect of the larger fat compartment of primates on
risk estimates could be elucidated by the use of physiologically based pharmacokinetics.
The role of the lymphatic system on absorption and transport of dioxin within the body
remains an unexplained process. A well formulated and validated physiologically based
pharmacokinetic model will accurately assess the dose received by infants from lactating
mothers with a body burden of TCDDs.
145
-------
TABLE 5-14
RISKS ASSOCIATED WITH BACKGROUND DAILY INTAKE OF 2,3,7,8-TCDD
COMPARED WITH ANNUAL CANCER INCIDENCE IN U.S. POPULATION
Upper Limit
of animal
Potency Daily Dose3
(pg/kg-day)'1 (pg/kg)
Life Time
Incremental
Cancer Risk
Resulting from
Daily Intake
Upper Limit
Annual Cancers
Resulting from
Daily Intakeb
Background
Probability Annual
Cancer in Background
U.S. Cancerb
1.5 x 1(T4 0.96
-4
1.5 x 10
1.5 x 10~4
0.96
0.96
1..47 x 10'4
1.47 x 10~4
1.47 x 10~4
504
504
504
1.9 x 10 '3 6500
(Soft Tissue
Sarcoma)
9.2 x 10'3 32000
(Non-Hodgkins
Lymphoma)
(All Cancers) 965000
aDaily intake (absorbed) converted to applied dose to be consistent with animal derived potency
which is based on applied dose.
bBased on a U.S. population of 240,000,000.
EPA's Exposure Assessment Group (EAG) is developing a physiologically based
pharmacokinetic model for TCDDs that, with properly gathered data for formulation and
validation, will:
o more accurately account for simultaneous exposure by more than one route and which
does not depend on the implicit assumption that absorption fractions are the same over
all concentrations, times, and for all species;
o not be restricted to steady state conditions;
o account for elimination by other than monophasic kinetics and assess the potential impact
of the change in elimination kinetics which may occur throughout life with changing
exposure and physiologic conditions;
146
-------
o account for elimination by more than on simultaneous process including lactational
shedding;
o realistically represent the role that absorption of TCDD by lacteals and its transport by
lymphatics may have;
o with proper data, be extended to food producing animals and thus be used to more fully
evaluate human exposure by this route; and
o help explain and account for obvious pharmacokinetic differences (metabolism, lipid
sequestering, etc.) between species. This is especially important regarding the differences
between rodents and humans.
With regard to risk assessments, there are several potential ways in which pharmacokinetics
may be applied in exposure assessments. In a conventional risk assessment for carcinogens, some
dose-response function is generated, and from that function human risk is calculated at various
exposure concentrations. Usually such a process involves extrapolation from animal high-dose
experiments to calculated risk for animals at low doses, and then further extrapolation from animals
to humans. First, the pharmacokinetic model enables the risk assessor to utilize some internal body
concentration of the parent compound of the metabolite (depending upon the mechanism of action)
as the dose for the dose-response curve. The dose-response function is then calculated from the
model-generated target concentrations. To do this, something needs to be known about the
mechanism of action. At the very least, it needs to be known whether the parent compound or the
metabolite is the carcinogen. An example of such a process was performed for tetrachloroethylene
(Chen and Blancato, 1987), where the
physiological pharmacokinetic model described total metabolite formation. The amount of
metabolite was then used as the dose in the two-stage cancer model to define the
incremental cancer risk for mice and the compound's potency factor. The next step
involves adjusting the parameters so that the pharmacokinetic model describes the target
concentration of the carcinogenic species in human tissues after exposure. This
147
-------
calculated dose is then used with the potency factor to estimate the incremental risk for
humans (Chen and Blancato, 1987). As more detail becomes known about the mechanism
of action, the pharmacokinetic model is further refined to give detailed concentrations of
the toxin at very specific target sites.
148
-------
PART TWO
APPLICATION OF EXPOSURE ASSESSMENT METHODS IN EVALUATING
2,3,7,8-TCDD EXPOSURES FROM SELECTED SITUATIONS
149
-------
6. USE OF METHODOLOGIES TO ESTIMATE EXPOSURE TO 2,3,7,8-TCDD
In .this chapter, the methods and parameter refinements discussed in the previous
chapters will be used to calculate exposures and risks associated with several
hypothetical situations. The set of starting assumptions for these hypothetical situations
(both in terms of parameters and in terms of how sources are related to exposure) are
referred to as "exposure scenarios." Two sets of scenarios are presented in this chapter.
The first set deals with 2,3,7,8-TCDD-contaminated land representing a range of site
types, i.e., contaminated soil situations, controlled-access landfills, and uncontrolled
access landfills (open dump). The second set deals with 2,3,7,8-TCDD-contaminated
emissions from combustion devices (i.e., incinerators) and the disposal of
2,3,7,8-TCDD-contaminated fly ash collected in control equipment. It should be noted
that these scenarios are not meant to represent every possible event that could lead to
high risk. The scenarios were selected to represent a range of plausible conditions, but
they cannot be considered "representative" of the U.S. or region of the U.S. In many
cases, the parameter values were selected on the basis of best judgment rather than data.
Thus, this chapter applies exposure assessment methods to a set of defined scenarios for
purposes of illustrating how they can be applied in site-specific situations. Application
of these techniques to non-site-specific problems, may require collection of extensive
survey data to ensure that the input parameters are truly representative of the area of
concern. Further discussion of related issues is presented, along with details of the
scenarios, later in this chapter.
Although this document is focused toward 2,3,7,8-TCDD, the other dioxin congeners
may also be of concern. Accordingly, the other congeners are discussed briefly with
regard to exposures associated with incinerators. This was feasible since some data was
available on the distribution of dioxin congeners in fly ash. However, no such data was
available for contaminated soil and thus they were not addressed for these scenarios.
150
-------
Exposure is often expressed as a daily contact rate averaged over an individual's
lifetime and body weight. Estimates of exposure, including assessments of uncertainty,
are used for conducting risk assessments. The general equation used to estimate
2,3,7,8-TCDD exposure is as follows:
Exposure = (2,3,7,8-TCDD media concentration x contact rate
x exposure duration x dilution and degradation factor
x distribution factor) / (body weight x lifetime) (6-1)
The above procedure will be used to calculate exposure levels in this document. The
procedure for converting these exposures to doses by consideration of absorption of the
contaminant into the body and estimating risk is discussed in the Appendix.
A. DESCRIPTION OF THE EXPOSURE SCENARIOS FOR CONTAMINATED SOIL AND
LANDFILLS
The first seven scenarios described below deal with situations where soil
surrounding a home or a farm has been contaminated with varying concentrations of
2,3,7,8-TCDD. In scenario 1, a family lives on a 1-acre site with 50% grass cover. This
is the only scenario where farming and fishing are not conducted. It is intended to
represent the more common residential setting. However, it is still considered
"reasonable worst case" in terms of the behavior patterns affecting duration and
frequency of contact. Scenarios 2-4 represent "reasonable worst-case" scenarios where
the soil has 50% vegetative cover and the family living in the area gets much of their
food from their own farm. Scenarios 5-7 represent a more typical situation where the
contaminated soil is largely covered with grass and the family's habits are somewhat
more typical of the population at large.
Scenarios 8-15 deal with a farm situated near an inactive landfill. In scenarios
8-11, the farm is located 100 feet from an inactive, uncontrolled access landfill
containing contaminated soil with no vegetative cover. These factors, along with the
151
-------
behavior of the farm family are meant to represent a reasonable worst case. In
scenarios 12-14, the inactive uncontrolled access landfill has a grass cover, the farm has
been moved to 500 feet from the landfill, and the farm family's habits are considered
more typical of the population at large. Scenario 15 is a capped controlled access
landfill containing 2,3,7,8-TCDD contaminated material (soil or equivalent; no organic
solvents) in the same "reasonable worst-case" configuration as scenario 9.
For each scenario, human exposure may occur by several pathways. It is a
relatively simple exercise to hypothesize situations for each exposure pathway that would
lead to unacceptably high risks. On the other hand, most such situations are relatively
rare, and would affect a relatively small number of people. In exposure assessment in
general, occurrence of a true "worst case," in which many variables approach their
maximum potential for exposure simultaneously, is exceedingly rare, approaching zero
probability of happening in a real-life case. Moreover, the population affected would
also approach zero. For that reason, the concept of "reasonable worst case" is used.
Describing a "reasonable worst case" involves specifying situations where there is
judged to be a reasonable probability of individual events occurring, rather than looking
at a situation which would maximize all exposure pathway risks simultaneously. While
risks for all scenarios and pathways considered in this chapter are summarized later in a
single table (see Appendix), it is very unlikely that people would experience the highest
risk for all exposure pathways simultaneously. It would be reasonable to assume that an
individual could experience the calculated risk of one to several of the pathways
simultaneously.
Tables 6-1 and 6-2 summarize the assumptions concerning the fifteen scenarios for
which exposure and risk estimates have been made. For each scenario, estimates were
made for the following pathways:
o dermal contact with contaminated soil;
152
-------
o inhalation of 2,3,7,8-TCDD vapors;
o inhalation of dust from contaminated soil, both from wind erosion and
vehicular traffic;
o ingestion of contaminated soil;
o ingestion of drinking water contaminated by runoff (stream only);
o ingestion of fish contaminated by runoff;
o ingestion of beef contaminated via grazing on or near the site;
o ingestion of contaminated dairy products from cattle contaminated via grazing
on or near the site; and
o ingestion of contaminated vegetables.
The following additional pathways were considered, but were not included in the
scenario calculations for the stated reasons:
o Ingestion of ground water contaminated from leachate from the landfill.
Contaminant levels proved to be so low under all of the scenarios, even after
hundreds of years, as to be of negligible risk. However, none of these
scenarios considered leaching under conditions where non-polar organics or
other solvents were present as co-contaminants. These situations could
apparently make 2,3,7,8-TCDD more mobile, but to what extent is uncertain
(see section A of Chapter 2). Physical transport of 2,3,7,8-TCDD-containing
particles through a porous subsurface zone can also have an appreciable effect
on mobility of 2,3,7,8-TCDD in soil. In any case, these situations should be
evaluated on a case-by-case basis.
o Dermal contact with contaminated surface water and sediments. These
pathways would result in very small exposure relative to the other pathways in
any but the most extreme cases. In a case where relatively heavy
153
-------
contamination exists in a body of water where frequent swimming or
wading/fishing occurs, this would need to be evaluated on a
case-specific basis.
The seven scenarios summarized in Table 6-1 represent the following situations:
Scenario 1: A 1-acre area of contaminated soil (at 1 ppb of 2,3,7,8-TCDD)
where a family lives long enough for a child to spend a 70-year lifetime.
About 80% of the lifetime of this individual is spent in the vicinity of home.
This is meant to represent a "reasonable worst-case" scenario for a non-
farming residential situation.
Scenario 2: A 10-acre area of contaminated soil (at 1 ppb of 2,3,7,8-TCDD),
upon which is a small stocked pond, some area where cattle graze, and a
garden to grow vegetables. Fish are regularly caught and eaten by a family
living on this farm. The family lives there long enough for a child to spend
his/her 70-year lifetime there. About 80% of the lifetime of this individual is
spent in the vicinity of home. Of the individuals' lifetime food intake, about
40% of the dairy products, 44% of the beef products, and 10% of the
freshwater fish come from the immediate vicinity (i.e., from the pond and
grazing animals). This is meant to represent a "reasonable worst-case" scenario.
The exposure and risks, however, are not added for each exposure pathway,
since it is unlikely that an individual would suffer worst-case exposures in all
pathways simultaneously. It is reasonable to expect, however, that given this
situation, individuals would have a reasonable chance to experience exposures
such as those estimated in one or more pathways (i.e., several, but not all
simultaneously).
Scenario 3: As above for scenario 2, but a one part-per-trillion starting
concentration in the soil.
154
-------
Scenario 4: As above for scenario 2, but a one part-per-quadrillion
concentration in the soil.
Scenario 5: The exposure area is the same as in scenarios 2 through 4, but
the pond is replaced by a stream located immediately adjacent to the exposure
area. This is meant to represent a more "average" case than those above, in
that the exposed individual's actions are closer to the typical case. The
individual obtains 10% of his freshwater fish from the immediate vicinity. The
individual spends only about 40 years in the area. About 50% of the lifetime
of this individual is spent in the vicinity of home. The soil is contaminated at
1 ppb.
Scenario 6: As above in scenario 5, but contamination level is 1 part per
trillion.
Scenario 7: As above in scenario 5, but contamination level is 1 part per
quadrillion.The eight scenarios summarized in Table 6-2 represent the following
situations:
Scenario 8: A 1-acre uncapped landfill with no vegetation. The open landfill
(see Figure 6-1) is located 100 feet away from the exposure area. The
exposure area is a 10-acre farm including a residence, stocked pond, and
pasture. Fish are caught regularly and eaten by a family living on this farm.
The family lives there long enough for a child to spend his/her 70-year
lifetime near the site. About 80% of the lifetime of this individual is spent in
the vicinity of home. Cattle graze in the pasture, which is contaminated by
runoff from the landfill. Of the individual's lifetime food intake, about 44% of
the beef, 40% of the dairy products and 10% of the fish come from the
immediate vicinity (i.e., from the pond and grazing animals).
Scenario 9: As above for scenario 8, but a 10-acre landfill site.
155
-------
Scenario 10: As above for scenario 9, but a starting concentration of 1 part
per trillion in the landfill.
.Scenario 11: As above for .scenario 9, but a concentration of 1 part per
quadrillion in the landfill.
.Scenario 12: The exposure area is the same as in scenarios 9 through 11, but
the pond is replaced by a stream located immediately adjacent to the exposure
area. This is meant to represent a more average case than in 9-11, though,
since the individual's actions are closer to the typical case. The uncapped
landfill in this scenario has been abandoned and grassed over (90% vegetation),
but the soil is contaminated at the surface. The exposure area (farm) is 500
feet from the inactive landfill. The exposed individual fishes from the
adjacent stream, is an average fish-eater, and gets 10% of his freshwater fish
from the stream. The individual only spends 40 years in the area. About 50%
of the lifetime of this individual is spent in the vicinity of home. The
inactive landfill itself is a 10-acre site with a contamination level of 1 ppb.
Scenario 13. As above in scenario 12, but the contamination level is 1 part
per trillion.
Scenario 14. As above in scenario 12, but the contamination level is 1 part
per quadrillion.
Scenario 15. This scenario is meant to be a reasonable worst case treatment
of a closed landfill. The size of the landfill is 10 acres, with 1 ppb
contamination under a 25 cm clean cap, with grass over the cap (90%
vegetation). In other aspects this scenario is the same as scenario 9.
Scenarios 1-15 are diagrammed in Figure 6-1.
-------
Figure 6-1. Landfill Scenarios
ON-SITE; Reasonable Worst Case
Scenario 1:
1 c
/\
ere
Scenarios 2-4:
ON-SITE; Typical Case
Scenarios 5-7:
Pasture
10 acres
OFF-SITE; Reasonable Worst Case
Scenarios 8-11,15:
Bare
Landfill
1-10 acres
100 feet ^
rponcT^
10 acres
OFF-SITE; Typical Case
Scenarios 12-14:
Grass
Covered
Landfill
10 acres
500 feet
Pasture
a
10 acres
^\
157
-------
TABLE 6-1. ASSUMPTIONS FOR CONTAMINATED SOIL SCENARIOS
Scenario
Area (acres) contaminated
Soil concentration
Water body type
Vegetation on-site?
Family
Yrs. at residence
% time at residence
% freshwater fish
diet from contami-
nated water body
% beef from vicinity*
% dairy from vicinity*
Ages for soil
ingestion
1
1
Ippb
none
50%
70
80
NA
NA
NA
2-6
2
10
Ippb
pond
50%
70
80
10
44
40
2-6
3
10
Ippt
pond
50%
70
80
10
44
40
2-6
4
10
1 ppq
pond
50%
70
80
10
44
40
2-6
5
10
1 ppb
stream
90%
40
50
10
44
40
2-6
6 7
10 10
Ippt 1 ppq
stream stream
90% 90%
40 40
SO 50
10 10
44 40
40 40
2-6 2-6
aAverage percent of annual consumption which is home-grown by 900 rural farm households
U. S. Department of Agriculture, 1966).
NA = not applicable.
-------
TABLE 6-2. ASSUMPTIONS FOR LANDFILL SCENARIOS
Scenario
Scenario
9 10 11
12 13 14 15
Access
Cap?
Size (acres)
Waste concentration
Water body type
Vegetation on-site?
Distance (ft) from
site to exposure area
Family
Yrs. at residence
% time at residence
% freshwater fish
diet from contami-
nated water body
% beef diet from
contaminated sourcea
% dairy diet from
contaminated sourcea
Ages for soil
ingestion
uncontrol
no
1
Ippb
pond
0%
100
70
80
10
44
40
2-6
uncontrol uncontrol uncontrol uncontrol
no
10
Ippb
pond
0%
100
70
80
10
44
40
2-6
no
10
Ippt
pond
0%
100
70
80
10
44
40
2-6
no
10
Ippq
pond
100
70
80
10
44
40
2-6
no
10
Ippb
stream
90%
500
40
50
10
44
uncontrol uncontrol cont
no no yes
10 10 10
Ippt Ippq Ippb
stream stream pond
500
40
50
10
44
500
40
50
10
100
70
80
10
44 40
40 40 40 40
2-6 2-6 2-6 2-6
aAverage percent of annual consumption which is home-grown by 900 rural farm households
(U. S. Department of Agriculture, 1966).
159
-------
B. EXPOSURE PATHWAYS
1. General
The parameters shown in Equation 6-1 must be defined for each of the pathways of
concern in a given scenario. In the case of a landfill, inhalation of vapors and dust,
ingestion of soil, contaminated foodstuffs (fish, cattle) and drinking water (from either
surface water or ground water sources), as well as dermal exposure to soil and sediments
are plausible exposure routes. Chapter 4 of this report presented techniques for
estimating inhalation of vapors, indoor levels of contaminated dust, and quantities of soil
ingested. Updated data on fish, beef, and dairy products consumption also are found
there.
The effects of the physicochemical properties of 2,3,7,8-TCDD on transport,
transformation, and bioavailability should be examined soon after undertaking an exposure
assessment. Table 2-1 in this report presents values for the vapor pressure, water
solubility, and octanol/water partition coefficient. These values were used to estimate
intermedia distribution processes such as volatilization, sorption, and partitioning between
aquatic sediments and water (Section B of Chapter 3).
Although the water solubility of 2,3,7,8-TCDD is extremely low, the presence of oil
or solvents would affect the movement and environmental behavior of the contaminant
(Section A of Chapter 2). Although the presence of oil or solvents would contribute to a
worst-case scenario, the methodology for predicting the effects is not well defined, so oil
or solvents are assumed to be absent in the scenarios presented here.
Exposure from inhaling 2,3,7,8-TCDD vapor released from contaminated soil can be
estimated by calculating potential emissions, dilution in the atmosphere, and effects of
photodegradation using methods explained in Section B of Chapter 4.
An issue common to all exposure routes associated with soil contamination is how
the concentration profile over depth affect human exposure. Work by Freeman and
1.60
-------
Schroy (1986) suggests that contaminated soils develop a profile in which the TCDD
levels at the surface are essentially zero and increase with depth. This profile develops
because photodegradation is assumed to deplete the surface more quickly than vapor
diffusion moves TCDD upward. Freeman and Schroy suggest that since the upper layer is
clean, exposure levels will be much lower than implied by the average soil concentration.
However, the clean layer is very thin (<1 cm) and deeper levels may contribute to
exposure if human activities disturb the surface (i.e., digging, playing, working, etc.).
For this reason, the concentration levels used in this report are averages which are
assumed to represent the levels of concern for purposes of human exposure.
2. Exposure Factors Common to More than One Pathway
a. Degradation and Dilution
Exposure calculations for several pathways may be influenced by mobility of
2,3,7,8-TCDD in soil. Apparent mobility in soil is probably the result of volatilization
and erosion (discussed in later sections) rather than leaching. Other factors possibly
affecting 2,3,7,8-TCDD soil concentration include photolysis at the soil surface and
microbial degradation in soil.
The approach used in calculating risk associated with the scenarios of this chapter
was to obtain the average soil concentration over the depth of concern and duration of
exposure. The combined effect of the processes noted above should be 2,3,7,8-TCDD soil
concentration profiles which change only slowly with time and depth. The available
literature on these topics was reviewed in depth in Section A of Chapter 3.
Young (1983) studied a surface contamination site where losses were attributed
primarily to volatilization and photolysis. An overall loss half-life for 2,3,7,8-TCDD
under those conditions was reported to be 10-12 years. However, this study involved
shallow contamination depths. Since photolysis is a surface phenomenon, the results may
apply only to scenarios which involve similar depths, such as those where transport
161
-------
occurs by. surface erosion. Consequently, the effects of photolysis on estimated exposure
have been incorporated into the procedures used to model erosion in Section 6-B-2b
which follows, specifically, through an overall loss rate constant in Equation 6-4.
In. Section 3-A, data of Bumpus et al. (1985) were analyzed further, based on
apparent first-order kinetics, to derive a biodegradation half-life for 2,3,7,8-TCDD equal
to 29- years. This value was specific, to one fungus type and soil condition. Other
investigators report that microbial degradation of. 2,3,7,8-TCDD is very low. or not
detected.. A 29 year half-life, would translate to a 35% reduction in exposure over a
period of 40 years, relative to a situation, where no degradation occurs. Given the
shortage of data on biodegradation and disagreement among existing data, rather than
consider this phenomenon separately, it is assumed that any losses due to biodegradation
are reflected in the overall loss rate constant of Equation 6-6.
The ratio of the 2,3,7,8-TCDD concentration in the soil at the exposure site to that.
of the soil or sediment at the source referred to in this report as the dilution factor.
This factor is common to all exposure pathways except those for inhalation.
Soil, is transported from a contaminated site primarily via windblown dust and
suspended sediment in overland runoff. As shown below, the contributions from
windblown dust to downgradient areas are estimated to be negligible compared to the
sediment in runoff.
Using the assumptions and procedure described below for the particulate emission
calculation, 0.44 mg/year of 2,3,7,8-TCDD is estimated to be released from the site on
dust (for scenario 9; 1 ppb and bare soil). Using the Universal Soil Loss Equation, it
was estimated that approximately 56 mg/year of 2,3,7,8-TCDD could be released via
runoff. Shifting wind directions will further reduce the amount of dust deposited at any
one location around the site. Lacking site-specific data, a wind direction frequency of
0.15 is commonly assumed. This factor reduces the potential dust deposits in one of the
162
-------
standard 16 wind directions from the site to 0.066 mg/year of 2,3,7,8-TCDD. This
emission rate is approximately 0.1% of the total runoff releases.
In situations where the exposure area is upgradient from the contamination source,
soil transport will occur entirely via windblown dust. Although these scenarios were not
explicitly considered in this report, the risk can be estimated by simply dividing the risk
estimated for the downgradient areas by 1,000. This procedure should be applied to only
the landfill scenarios (8-15) for dairy ingestion, beef ingestion, soil ingestion, and dermal
contact with soil.
The 2,3,7,8-TCDD in a field downgradient of the site was computed by assuming
that the 2,3,7,8-TCDD carried to the field becomes mixed with the soil in the field to a
certain depth. The mixing depth depends on activities which disturb the surface, such as
construction, plowing, vehicle traffic, movement of cattle or other animals, burrowing
action of animals, and other biological activity. Mixing depths for fallout plutonium have
been found to be 20 cm on cultivated land and 5 cm on uncultivated forest and rangeland
(Foster and Hakonson, 1987). An intermediate value of 10 cm was assumed here, for all
scenarios. Using this depth and the area of the field (10 acres), the mass of soil into
which the eroded 2,3,7,8-TCDD is mixed can be calculated as:
M = (10 acres) (4,047 m2/acre) (0.10 m) (1,700 kg/m3) = 6.88 x 106 kg (6-2)
The delivery rate of 2,3,7,8-TCDD to the field must also be estimated. The first
step in deriving this value is to use the Universal Soil Loss Equation (USLE) (U.S. EPA,
1976) to estimate the total amount of soil eroded from the site:
A = (R)(K)(LS)(C)(P) (6-3)
163
-------
where A = average annual soil loss (tons/acre-year), R = rainfall and runoff erosivity
index, K = soil erodibility factor, LS = topographical factor representing slope length and
steepness, C = cover and management practice, and P = supporting practices factor.
Average values of these factors were obtained from a survey of over 70 sanitary
landfills in the U.S. (Science Applications International Corp. (SAIC), 1986):
R = 155 -The rainfall and runoff factor represents the influence of precipitation on
erosion. It is derived from data on the frequency and intensity of
storms. A value of 155 is typical of rainfall patterns seen in much of
the midwestern United States.
K = 0.23 - The soil erodibility factor reflects the influence of soil properties on
erosion. A value of 0.23 is typical of sandy loam with 2% organic matter.
LS = 1.5 - The topographic factor reflects the influence of slope steepness and
length on erosion. A value of 1.5 can correspond to a variety of
combinations, such as a 10% slope over a 100-foot length or a 5% slope
over a 1,000-foot length.
P = 1.0 - The supporting practices factor reflects the use of surface conditioning,
dikes, or other methods to control 'runoff/erosion. A value of 1.0 reflects
a compacted surface without control structures.
The final term in the USLE is the cover and management practice factor (C), which
primarily reflects how vegetative cover influences erosion. The values assumed for this
term were based on the scenario descriptions rather than the survey discussed above.
Since the site is bare in scenarios 8 through 11, C is assumed to be 1.0; since the site is
grass-covered in scenarios 12 through 15, C is assumed to be 0.1 (U.S. EPA, 1976).
The survey data could be used in two ways. The average values for each factor
could be multiplied together, which yields R K LS P = 56 tons/acre-year. Alternatively,
the factors for each landfill in the SAIC survey could be multiplied together and then
164
-------
averaged, which yields (R)(K)(LS)(P) = 62 tons/acre-year. Since there may be some
dependency among the factors, the latter method is probably more statistically valid and
was adopted here. Combining this value with the assumptions for C yields the following
erosion estimates:
Scenarios 8-11: A = 62 tons/acre-year
Scenarios 12-15: A = 6.2 tons/acre-year
More sophisticated models are available for estimating erosion rates (i.e., CREAMS),
and should be considered in actual site-specific assessments.
The fraction of eroded soil which enters the field must now be computed. This
quantity is very site-specific, depending largely on local gradients and runoff channeling
patterns. If the field is located immediately downgradient of the site, a large fraction of
the eroded soil may enter the field. In scenarios 9-15, it is assumed that the field is
downgradient from the site, but 100-500 feet away. Channeling patterns over this
distance could divert much of the runoff away from the field. For the reasonable worst
case (scenarios 8 through 11 and scenario 15), this factor was assumed to be 0.5, and for
the more typical cases (scenarios 12 through 14) it was assumed to be 0.1. This factor is
very similar to the sediment delivery ratio used by soil scientists to describe the fraction
of eroded soil which reaches a water body. Wade and Heady (1978) studied 105 major
U.S. river basins and found sediment delivery ratios which ranged from 0.001 to 0.378
with an average of 0.042. The delivery factor used in this report refers to the fraction
of soil reaching a nearby field, which should be greater than the fraction ultimately
reaching a water body. Thus, it is reasonable that the worst-case factor is slightly
greater than the upper end of reported sediment delivery ratios and the typical factor is
slightly greater than the average reported sediment delivery ratio. The wider range of
reported sediment delivery ratios reveals the site-specific nature of erosion phenomena
165
-------
and the potential uncertainty regarding how the values assumed here would apply to a
particular, situation-.
Mow the soil delivery rate, to the field can be-computed: Scenarios 9 through 11:
D.I; - (62- tons/acre.-year) (10 acres) (0:5) (907 kg/ton)
= (-280,000 kg/year) (6-4),
For. the: other, scenarios, a similar, procedure is used, yielding Dj = 28,000 kg/ year for
scenario, 8 and. 15 and^D,], = 5*600 kg/year for scenarios 12 through 14.
In: addition, to. the contaminated, soil, delivered to the field a certain amount of clean
soil: will- also be delivered. (D-2). Assuming that the mass of soil in the field mixing zone
remains constant a soil mass balance, can be; written as follows:
Dj + D2 =-R (6-5)
where R. = removal rate of soil from the field mixing zone. No assumption is made
regarding: whether the soil is removed from the. mixing, zone by accumulation under the
mixing zone or by runoff. The value of D2 was. estimated using the same procedure^ as
illustrated in Equation 6-4 except that the area term represented the area between the
site and. field. This approach assumes that.clean soil enters the-field only from the area
between, the site, and field and that the. same erosion rates-apply to this area as the site.
The" values of D], D-2, and R are summarized in Table 6-3.
166
-------
TABLE 6-3. EROSION PARAMETERS
Scenario
8
9
10
11
12
13
14
15
D!
(kg/yr)
28,000
280,000
280,000
280,000
5,600
5,600
5,600
28,000
E>2
(kg/yr)
42,000
42,000
42,000
42,000
4,200
4,200
4,200
4,200
R
(kg/yr)
70,000
320,000
320,000
320,000
9,800
9,800
9,800
32,000
Dj = Delivery rate of contaminated soil to field
D2 = Delivery rate of clean soil to field
R = Removal rate of contaminated soil from field
Assuming that the eroded soil is perfectly mixed with the field soil to a depth of 10
cm, the resulting concentration of 2,3,7,8-TCDD in the field can be computed on the
basis of the following mass balance:
dC = DjCo - RC - kC (6-6)
dt M M
where C = concentration of 2,3,7,8-TCDD in field soil (kg/kg), C0 = original
concentration of 2,3,7,8-TCDD in site soil (kg/kg), Dj = delivery rate of contaminated
soil to field (kg/kg), M = mass of soil in field down to mixing depth (kg), k =
167
-------
degradation rate constant (year-1), R = soil removal rate from field (kg/yr), and t = time
(year).
The last term in Equation 6-6 represents the rate at which 2,3,7,8-TCDD degrades
by any combination of vaporization, photolysis, or biological degradation. As discussed
previously, this rate constant has not been well established, but appears to be slow,
particularly when the soil is not exposed to sunlight. For all the scenarios, no
degradation was assumed to occur at the landfill site, largely because much of it is
buried and therefore not subject to photolysis. Eroded soil transport, however, is a
surface phenomenon that increases the opportunity for photolysis and increases
vaporization rates. This situation most closely matches the experimental conditions
studied by Young (1983), who derived a rate constant of 0.069 year"', which is also
assumed to apply to eroded soil.
Assuming that the 2,3,7,8-TCDD concentration in the exposure site field soil is
initially zero, Equation 6-6 can be solved to yield:
C = DjSo. [1 - e-(R/M + k)t] (6-7)
R + kM
Equation 6-7 computes C as a function of time (t). The average value over a
70-year exposure period can be computed by integrating C with respect to t and dividing
by the exposed period. An alternative approach is to assume that contamination at the
site has existed for a long period of time, so that the 2,3,7,8-TCDD concentration in the
field has reached steady state, in which case:
C= DjCo. (6-8)
R + kM
1.6.8
-------
This latter approach was assumed to apply for the exposure site worst-case
scenarios 8 through 11 and scenario 15. Using the parameter values discussed earlier,
the dilution factor (C/CO) can now be computed (for scenarios 9-11):
£_ = _D\_ - 280.000 = 0.35 (6-9)
C0 R+kM 320,000 + (.069) (6.88x106)
The average value approach was assumed to be more appropriate for the off-site
typical case (scenarios 12 through 14). The average value of C over a 40-year exposure
period was calculated as 0.008 C0, which implies a dilution factor of 0.008. Since the
soil concentrations are defined for all contaminated soil scenarios (1-7), the dilution
factor becomes 1.0. All soil dilution factors are summarized in Table 6-4.
It should be noted that field studies of fallout plutonium have observed an
enrichment of plutonium levels in stream sediment over that found in upland areas of
water sheds (Foster and Hakonson, 1987). This phenomena is explained by the fact that
as overland runoff enters a deposition area, dense and large particles (i.e., sand, large
aggregates) deposit first and small/less dense particles (i.e., clay, silt) deposit further
downgradient. This causes stream sediment to be richer in fine particulates than upland
soils. Since plutonium preferentially associates with smaller particulates the levels in
sediment are higher than levels in upland soils. A similar phenomenon may occur with
CDDs due to their affinity for organics which are found in the smaller or lighter
particulates. This particle size enrichment was not accounted for in the dilution model
described above. However, we believe that this enrichment effect will be small compared
to the dilution effects. The plutonium scenario involves fallout over the entire water
shed whereas the TCDD scenarios involve contamination areas which are small compared
to the whole water shed. Thus, the contributions of eroded soil from the contamination
areas would be small compared to the contributions from clean areas. The resulting
169
-------
dilution is further enhanced by mixing with clean soil originally found at the site.
Ultimately, field tests are needed to validate this model.
TABLE 6-4. DILUTION FACTORS
Scenarios
Soil
dilution factor
Sediment
dilution factor
Soil, reasonable
worst case (pond)
scenarios !• - 4
Soil, typical
case (stream)
scenarios 5 - 7
Landfill rea-
sonable worst
case (pond)
scenarios 8-11
and 15
Landfill, typical
case (stream)
scenarios 12-14
Soil concentration is
defined.
DF soil = 1.0
Soil concentration is
defined.
DFsoil = 1.0
Use steady state mixing
model to calculate soil
concentrations.
DF soil = 0.051 (Scenario 8)
DF soil=0.35 (Scenario 9-11)
DF soil=0.055 (Scenario 15)
Use non-steady state
mixing model to calcu-
late soil concentration.
DF soil = 0.008
Assume sediment
levels equal
soil levels.
DF sed = 1.0
Use erosion model to
calculate sediment
concentration.
DF sed = 0.001
Assume sediment
levels equal
soil levels.
DF sed = 0.051 (Scenario 8)
DF sed = 0.35 (Scenarios 9-11)
DF sed = 0.055(Scenario 15)
Use erosion model to
calculate sediment
concentrations.
DF sed = 0.001
Finally, a methodology has been presented above which can be used to estimate soil
concentrations resulting from runoff from an upgradient soil contamination source. Site-
specific values were used to illustrate application of the methodology at a hypothetical
but plausible range of sites. We believe that the range of selected values are reasonable,
170
-------
but insufficient data are available to interpret these as representative of the country or
region of the country. Rather, the methodology is best applied on a site-specific basis.
b. Sediment Dilution Factor
The worst-case scenarios involve a pond surrounded by contaminated soil. Since the
pond sediment is derived from locally eroded soils, the concentrations of 2,3,7,8-TCDD in
the sediment are assumed to equal the levels in the soil; i.e., the dilution factor for
sediment is the same as that for soil.
In the typical scenarios, the stream is located 500 feet from the contaminated site.
In this situation, not all of the soil eroded from the site will reach the stream, and the
portion that does will be diluted by clean sediment derived from other portions of the
watershed. If a uniform erosion rate occurs throughout the watershed, this dilution is
equal to the area of the site divided by the area of the upgradient portion of the
watershed. This assumption should be valid in the typical scenarios, since the site is
grass-covered, as can be expected in the remainder of the watershed. The site in all of
these scenarios (5-7, 12-14) is 10 acres, and the watershed is assumed to be 10,000 acres
(this represents the median value found in a survey of 70 landfills [Science Application
international Corp., 1986]). Thus, the dilution factor can be calculated as 0.001.
Alternatively, the dilution can be calculated as follows. The average runoff rate for the
midwestern U.S. is about 15 inches/year (Linsley et al., 1982). For a 10,000-acre
watershed, this yields a stream flow of 18 cfs. The sediment yield can be estimated from
the stream flow as follows (Linsley et al., 1982):
Qs = aQn (6-10)
where Qs = sediment flow rate (T/year); Q = stream flow rate (cfs); a and n =
171
-------
empirical constants, reflecting the vegetation cover in the watershed. Linsley et al.
(1982) recommend using:
a = 3,500 and n = 0.82 for coniferous forest and tall grass
a = 19,000 and n = 0.65 for scrub and short grass
Substituting into Equation 6-10 yields a sediment flow of 37,000 to 124,000 T/year.
Using the earlier estimate of 6.2 T/acre-year for the 10-acre site, the sediment dilution
is calculated to be 1.7 x 10"-* to 0.5 x 10"^. These estimates agree closely with the
1,000-fold dilution indicated by the ratio of the site area to the watershed area.
The soil and sediment dilution factors are summarized in Table 6-4.
c. Body Weight
Throughout all of the pathways considered in this chapter, except that of soil
ingestion, adult human body weight is taken to be 70 kg. This factor was discussed in
more detail in Section B of Chapter 2. The specific assumptions and parameters used in
the soil ingestion pathway will be discussed in detail in a following section.
d. Lifetime
Following widespread practice in exposure and risk assessment, the average adult
lifetime assumed throughout this chapter is 70 years. Even though actuarial data indicate
that the U.S. average lifetime now exceeds 70 years, this convention was continued in
order to simplify comparisons of risk with those calculated in other analyses.
e. Pharmacokinetics
While not employed in this study, physiologically based pharmacokinetic models can
be of great utility in describing the disposition of compounds within the body after
exposure. As such, they are an extension of the exposure assessment into the body.
These higher-resolution exposure assessments provide target doses or concentrations
172
-------
which risk assessors may eventually use in various types of dose-response functions to
better characterize risk. Little is known about the exact mechanism of toxic action of
2,3,7,8-TCDD. However, the model developed in Section B of Chapter 5 has great utility
in describing previous intake from adipose tissue concentrations, provided one assumes
the achievement of steady-state conditions. As more information is gained about the
mechanisms of action, pharmacokinetic models may be used as the basis for
pharmacodynamic models which will further aid in the quantification of risk. The latter
might also be modified to describe and predict other important facets of the behavior of
this compound, such as placental and lactational transfer.
3. Specific Factors bv Pathway
The specific input parameters used in calculating exposure and risk associated with
various pathways are summarized in Table 6-5. The exposure estimates for each scenario
are presented in Table 6-6.
a. Dust Inhalation from Wind Erosion
Dust emissions may occur as a result of wind erosion only; i.e., no disturbance of
the surface due to vehicle traffic or other activity. Emissions caused by surface
disturbance are discussed later in this section. The surface was assumed to be exposed
to the wind, uncrusted, and to consist of finely divided particles. This creates a
condition defined by U.S. EPA (1985a) as an "unlimited reservoir" and results in maximum
dust emissions due to wind only.
173
-------
TABLE; 6-5. FACTORS USED IN EXPOSURE CALCULATIONS
Scenario
Pathway
Soil ingestion
Contact rate (g/day)
Absorption, fraction
Exposure, duration (days)
Body-weight, (kg)
On-site dilution factora
Off -site dilutions factor*3
Dermal exposure- to soil .
Contact rate, (g/day)
Absorption fraction
Exposure duration (days)
Body weight (kg)
On-site dilution factora
Off-site dilution factorb
Vapor inhalation
n
Contact rate (m /day)
Absorption fraction
Exposure duration (days)
Body-weight (kg)
Fat, ingestion
Beef ingestion (g/day)
Dairy ingestion (g/day)
Absorption
Beef exposure duratation (days)
Dairy 'exposure duration .(days)
Body weight (kg)
On-site dilution-factor*
Off-site dilution factor*1
Beef fat/soil dist:
Dairy fat /soil dist.
1/8
1
0:3
1,600
17-
1.0
0.051
1.
0.006
20,000
70
1.0-
0.051
23
0.76.
20,000
70
26
43
0:68
11,000
10,000
70
1.0
0.051
0.4
0.04
2/9
1
0.3
1,500
17-
1.0
0.35.
1
0.005
20,000
70
1.0
0.35
23
0.76:
20,000
70
26
43
0.68
11,000
10,000
70
1.0
0.35
0.4
0.04
3/10
1
0.3
1,500
17-
1.0
0.35
'l
0.005
20,000
70
1.0
0.35
23
0.76
20,000
70
26
43
0.68
11,000
10,000
70
1.0
0.36
0.4
0.04
4/11
1
0:3
1,500
17
1.0
0.35
1.
0.005
20,000
70
1.0
0.35
23
0.75
20,000
70
26
43
0.68
11,000
10,000
70
1.0
0.35
0.4
0.04
6/12
0.2
0.3
910
17
1.0
0.008
1
0.005
7,300
70
1.0
0.008
23
0.75.
7,300
70
14.9
18.8
0.68
6,400
5,800
70
1.0
0.008
0.3
0.04
fContin
6/13
0.2
0.3
910
17
1.0
0.008
1
0.005
7300
70
1.0
0.008
23
0.75
7,300
70
14.9
18.8
0.68
6,400
6,800
70
1.0
0.008
0.3
0.04
ued
7 714
0.2
0.3*
910
17
1.0
0.008
1
0.005
7,300
70
1.0
0.008
23
0.75
7,300
70.
14.9
18.8
0.68
6,400
5,800
70
1.0
0.008
0.3
0.04
}
7/15
1
0.3
1,500
17
-
0.055
1
0.005
20,000
70
_
0.056
23
0.75
20,000
70
26
43
0.68
11,000
10,000
70
_
55
0.4
0.04
174
-------
TABLE 6-5. (CONTINUED)
Pathway
Dust inhalation
o
Respiration rate (m /day)
Absorption fraction
Exposure duration (days)
Body weight (kg)
Fish Ingeation
Ingestion (g/day)
Absorption
Exposure duration (days)
Body weight (kg)
On-site dilution factor3
Off-site dilution factorb
Distribution factor
Surface Water Ingestion
Ingeetion (L/day)
Absorption
Exposure duration (days)
Body weight (kg)
On-site dilution factora
Off -site dilution factor15
1/8
23
0.27
20,000
70
30
0.68
2,600
70
1.0
0.061
5
2
0.5
20,000
70
1.0
0.051
2/9
23
0.27
20,000
70
30
0.68
2,600
70
1.0
0.35
5
2
0.5
20,000
70
1.0
0.35
3/10
23
0.27
20,000
70
30
0.68
2,600
70
1.0
0.35
5
2
0.5
20,000
70
1.0
0.35
4/11
23
0.27
20,000
70
30
0.068
2,600
70
1.0
0.35
5
2
0.5
20,000
70
1.0
0.35
5/12
23
0.27
7,300
70
6.5
0.68
1,500
70
0.001
0.001
6
2
0.5
7,300
70
1.0
0.001
6/13
23
0.27
7,300
70
6.5
0.68
1,500
70
0.001
0.001
5
2
0.5
7,300
70
1.0
0.001
7/14
23
0.27
7,300
70
6.5
0.68
1,500
70
0.001
0.001
5
2
0.5
7,300
70
1.0
0.001
7/15
23
0.27
20,000
70
SO
0.68
2.600
70
-
0.055
5
2
0.5
20,000
70
_
0.055
aRefers to contaminated soil scenarios 1-7.
Refers to scenarios 8-15.
176
-------
TABLE 6-6. EXPOSURE LEVELS ASSOCIATED WITH VARIOUS
EXPOSURE PATHWAYS/SCENARIOS - CONTAMINATED SOIL
(ng/kg-d)
Scenario.
Dairy Beef Fish Soil
inges- inges- inges- inges -
tion tion tion tion
Vapor Dust
inhala- inhala-
tion tion
Drink-
Soil ing-
dermal water
Vegeta-
ble
inges-
tion
1)1 ppb
1 acre
reasonable
worst-case
NA NA- NA.
3.4xlO."S 9.8X10'6 4.2xlO'6 l.lxlO"2 NA
See
text
2)1 ppb
10 acres
reasonable
worst-case
9.6xlO"3 6.4xlO"2 2.2X10"1 S.4xlO"3 1.4xlO'6 3.4xlO'6 l.lxlO'2 6xlO~5
See
text
3)lppt
10 acres'
reasonable
worst-case
9.6xlO~6 6.4xlO'5 2.2xlOr4 3.4xlO"6 1.4xlO'8 3.4xlO'9 l.lxlO'5 6xlO'8
See
text
4)1 ppq
10 acres
reasonable
worst-case
9.6xlO"9 6.4xlO"8 2.2xlO"7 3.4xlO"
3.4xlO"12 l.lxlO"8 CxlO"11
See
text
5)1 ppb
10 acres
typical
2.4xlO"3 1.6xlO"2 2.7xlO'6 4.2xlO"4 8.7xlO'6 1.2xlO'6 4.1xlO"3 2.2xlO"8
See
text
6)1 ppt
10 acres
typical
2.4xlO"6 1.6xlO"B 2.7xlO"8 4.2xlor7 8.7xlO"9 1.2xlO"9 4.1xlO"6' 2.2X10"11 See
text
7)1 ppq
10 acres
typical
2.4xlO'9 1.6xlO'8 2.7X10"11 4.2x10"10 8.7x10'12 1.2x10'12 4.1xlO"9 2.2xlO'14
See
text
NA = Not applicable.
(Continued )
176
-------
TABLE 6-6. (CONTINUED)
Scenario
8)1 ppb
1 acre
reasonable
worst -case
9)1 ppb
10 acres
reasonable
worst-case
10)1 ppt
10 acres
reasonable
worst-case
11)1 ppq
10 acres
reasonable
worst-case
12)1 ppb
10 acres
inactive
site
typical
13)1 ppt
10 acres
inactive
site
typical
Dairy Beef Fish Soil Vapor Dust Drink- Vegeta-
inges- inges- inges- inges- inhala- inhala- Soil ing ble
tion tion tion tion tion tion dermal water in gee -
tion
4.9xlO'4 S.SxlO'3 l.lxlO"2 l.SxlO'4 l.lxlO"6 4.6xlO"7 5.7xlO'4 S.lxlO"6
See
text
3.4xlO"3 2.2xlO"2 7.6xlO'2 1.2xlO"3 S.lxlO"6 S.SxlO"7 3.9xlO'3 2.1xlO'5
See
text
3.4xlO"6 2.2xlO"5 7.6xlO"6 1.2xlO"6 S.lxlO'9 S.SxlO"10 3.9xlO"6 2.1xlO"8
See
text
3.4xlO'9 2.2xlO"8 7.6xlO"8 1.2xlO'9 S.lxlO'12 S.SxlO'18 3.9xlO'9 2-lxlO"11
See
text
1.9xlO'5 l.SxlO'4 2.7xlO"5 S.SxlO"6 8.9xlO"7 5.3xlO"8 S.SxlO"5 2.2xlO"8
See
text
1.9xlO'8 l.SxlO'7 2.7xlO'8 S.SxlO'9 8.9x10' 10 5-SxlO'11 S.SxlO'8 2.2X10'11
See
text
/
(Continued )
177
-------
TABLE 6-6. (CONTINUED)
Dairy Beef Fish Soil Vapor Dust
Scenario inges- inges- inges- inges- inhala- inhala- Soil
tion tion tion tion tion tion dermal
14)1 ppq
10 acres l.OxlO'11 l.SxlO'10 2.7xlO~U S.SxlO'12 8.9xlO~ 1S 6.3xlO~14 S.SxlO"11
typical
15)1 ppb
10 acres neg. neg. neg. neg. 4x10 neg. neg.
capped land fill
reasonable worst-case
Drink- Vegeta-
ing ble
water inges-
tion
2.2xlO"14
See
text
neg. neg.
scenarios assume that the exposure area is downgradient of the contaminated source. If the exposure -area was
located upgradient and all transport was via windblown dust, these exposures would be reduced by a factor of 1,000.
\\ fi
neg. = negligible exposure (<10"e)
178
-------
The flux of dust particles less than 10 um in diameter from surfaces with an
"unlimited reservoir" of credible particles can be estimated as shown below (U.S. EPA,
1985a):
E = 0.036 (1 - V) (Um/Ut)3 F(x) (6-11)
where E = total dust flux of <10 um particle (g/m2 . hr), V = fraction of vegetation
cover, Um = mean annual wind speed (m/s), Ut = threshold wind speed (m/s), and F(x) =
a function specific to this model.
The value assumptions for each parameter are explained below:
o Fraction of vegetation cover (V)—For contaminated soil scenarios 1-4, the site
is assumed to be 50% vegetated (V = 0.5) and for contaminated soil scenarios
5-7, the site is assumed to be 90% vegetated (V = 0.9). Under scenarios 8
through 11 the site is bare; therefore, V = 0. Under inactive landfill scenarios
12 through 15, the site is assumed to be largely covered with grass; therefore,
V = 0.9.
o Mean annual wind speed (Um)—U.S. EPA (1985c) lists the mean annual wind
speeds at 10 meter height for the 60 major cities in the U. S. These values
range from 2.8 to 6.3 m/s, with an average of approximately 4 m/s. This
average value was assumed to apply in all of the scenarios.
o Threshold wind speed (Ut)—This is the wind velocity at a height of 7 m above
the ground needed to initiate soil erosion. It depends on nature of surface
crust, moisture content, size distribution of particles, and presence of
non-erodible elements. It can be estimated on the basis of the following
procedure (U.S. EPA, 1985a):
179
-------
(1) Determine the threshold friction velocity. This is the wind speed
measured at the surface needed to initiate soil erosion. For "unlimited
reservoir" surfaces, U.S. EPA (1985a) suggests that this velocity is less
than 75 cm/s. Thus, a value of 50 cm/s was assumed to be
representative of these types of surfaces.
(2) Estimate the "roughness height." This is a measure of the roughness of
the surface. The surface of the site in scenarios 1 through 4 and 8
through 11 was assumed to resemble a plowed field, which has a
roughness height of 1 cm, and the site in scenarios 5 through 7 and
12 through 15 is grass-covered, giving a roughness height of 2 cm.
(3) Estimate ratio of threshold wind speed at 7 m to friction velocity. Using
a chart provided by U.S. EPA (1985a), this ratio is seen to be 16.5 for a
roughness height of 1 cm and 15 for a roughness height of 2 cm.
(4) Estimate threshold wind speed. Multiplying the friction velocity by the
ratios described in step (3), the threshold wind speed is estimated as 8.2
m/second for scenarios 1 through 4 and 8 through 11 and 7.5 m/s for
scenarios 5 through 7 and 12 through 15.
o Finally, F(x) is determined by first calculating the dimensionless ratio (x)
where x = 0.886 Ut/Um (where Uf is the erosion threshold wind speed (m/s)
and Um is the mean wind speed (m/s)) and finding F(x) from a chart of F(x)
versus x, as provided in U.S. EPA (1985a). For scenarios 1 through 4 and 8
through 11, x is 1.85 and F(x) is 0.45; for scenarios 5 through 7 and 12
through 15, x is 1.7 and F(x) is 0.65.
The dust flux is converted to an emission rate as follows (U.S. EPA 1985a):
Q = CSEA (1 hr/3,600 s) (6-12)
180
-------
where Q = TCDD emission rate (ng/s), Cs = TCDD concentration in soil (ng/g), and A =
site area (m2).
The value assumptions for each of these parameters are explained below:
o TCDD concentration in soil (Cs)—This value depends on the scenario
assumptions, as specified in Table 6-1.
o Site area (A)—This value also depends on the scenario assumptions. For
scenarios 1 and 8, the site is assumed to be 1 acre (4,000 m2), and for all
other scenarios the site is assumed to be 10 acres (40,000 m2).
For all off-site scenarios (8-15), the dispersion was calculated on the basis of the
virtual point source dispersion model, which was derived from the Industrial Source
Complex Model (Turner 1970). This model is described in Chapter 4 (Equation 4-7).
The value assumptions for each of the parameters used in Equation 4-7 are
explained below:
o Wind direction frequency ((j))--Lacking site-specific data, a default value of 0.15
is recommended for this parameter (Turner 1970), which was assumed for all
scenarios.
o Virtual downwind distance to receptor (Lv)—As explained in Chapter 4,
this parameter is computed from Lv = L + 2.55 S, where L = distance to
the receptor from the facility center and S = facility width perpendicular to
the wind. Assuming that the site is a square, the 10-acre site will have sides
equal to 200 m, and the 1-acre site will have sides equal to 63 m. Using
these distances and the 100-foot (30 m) distance from site edge to receptor,
Lv for scenario 8 is 223 m, for scenarios 9 through 11 is 640 m, and for
scenarios 12 through 14 is 762 m.
181
-------
o The vertical dispersion coefficient (<7Z)—This parameter is a function of the air
stability and distance from the source. Lacking site-specific data, Turner
1970 recommends assuming Air Stability Class D. Using this assumption and a
chart of L versus az from Turner 1970, crz for scenario & is 5 m, for scenarios
9 through 11 and 15 is 6 m, and for scenarios 12 through 14 is 10 m.
o Annual average wind speed (Um)—As discussed earlier, this value was assumed
to equal 4 m/second, based on national data, and was applied in; all scenarios.
For the contaminated soil scenarios, both a simple mixing box model- and a
near-field dispersion modek were evaluated1 in Chapter 4. The approach used' for vapor
emissions (described previously in Chapter 4-B) was also applied to dust emissions.
The final step; is to' calculate' the actual exposure resulting from inhalation' of
2,3,.7,8-TCDD-contaminated dust. This is accomplished by application of Equation* 6"-L
The values assumed for each parameter used in the equation are described below:
o Contact Rate--For this pathway,, this parameter is the respiration rate. The
recommended rate is 21 m?/d for. an. average adult who spends 22.4 hours/day
engaged, in light activity, 1.4 hr/d engaged in moderate activity^ and 0.2
hours/day engaged in; heavy activity (see Section B of Chapter 2). This value
was. assumed to apply to all scenarios..
o Exposure duration - For the reasonable worst case scenarios (1-4, 8-ll,:
IS)'- as shown in Tables 6-1 and. 6-2,. the exposed population is. assumed:
to live 70 years at the exposure area and to actually spend 80% of their
time, at this location. Thus, the exposure- duration is calculated as:
ED- = (70 yr)(365 d/yr)(0.8) = 20,000, d (6> 13')
-------
For typical case scenarios (5-7, 12-14), the exposed population is
assumed to live 40 years at the exposure location and to spend 50%
of their time there. Thus, the exposure duration is calculated as:
ED = (40 yr)(365 d/yr)(0.50) = 7,300 d (6-14)
In converting exposure (equation 6-1) to risk (equation A-l), the following
assumption is made:
o Absorption fraction—Schaum (1984), using animal data and information
on fate of particles in the respiratory system, estimated that the
fraction of 2,3,7,8-TCDD absorbed into the body ranges from 0.25 to 0.29. An
average of 0.27 was assumed to apply to all scenarios.
b. Dust Inhalation from Vehicular Traffic
In addition to dust emissions caused by wind erosion, as discussed above, dust can
also be generated by vehicles entering the contaminated site. This was assumed to take
place for the landfill scenarios [the incinerator scenarios are treated separately, since fly
ash is being landfilled (see Section 4.C.e)], although it is assumed that no other activity
occurs that would generate dust, such as unloading.
Several assumptions are made to estimate emission rates of particulate matter and
hence 2,3,7,8-TCDD bound on the particulates. The methodology for estimating the
particulate emission rate is described in section C of Chapter 4. The assumptions
include: silt content of soil = 20%; vehicle speed on the site = 16 km/hr (10 mph);
weight of a vehicle = 12.3 Mg; number of wheels = 10; days of precipitation =110 days.
Since the emission rate is given in terms of kg dust/vehicle kilometers traveled (VKT), it
is necessary to estimate the approximate distances vehicles traveled on the site. The
dimension of a site would provide the approximate distance that a vehicle would normally
183
-------
travel. This distance multiplied by an estimated number of vehicles that may travel on
the site in a day would provide the VKT per day. The estimated dust emission rate is
converted to an equivalent 2,3,7,8-TCDD emission rate based on contaminant
concentration in soil for each scenario. The emission rate used for dust emissions in the
scenario exposure and risk calculations is the sum of the wind erosion and vehicular
traffic contributions.
Dispersion models are used to estimate the ambient air concentrations from the
emission rates. Exposure estimates were then based on the degradation factor, body
weight, and lifetime assumptions discussed previously. The exposure and risk estimates
are presented in Table 6-6 and the appendix, respectively. The calculations show that
for the scenarios used, the contribution to ambient air concentrations by vehicular traffic
is about 1-2 orders of magnitude higher than that of wind erosion alone.
c. Vapor Inhalation
The exposure assessment considered inhalation of 2,3,7,8-TCDD vapor as a pathway
for human exposure. Despite the chemical's extremely low vapor pressure, volatilization
can occur, with effects on the downwind population possible. Example exposure and risk
evaluations will be shown for the eight assumed landfill scenarios. Calculations will also
be shown wherever appropriate.
In all of the scenarios, the first task in risk assessment is to compute the rate of
volatilization. The emission rate thus calculated will be used in dispersion modeling to
estimate ambient air concentrations at various distances from the source. The Industrial
Source Complex (ISC) model approximating the area source as a virtual point source can
be used for dispersion modeling for receptors located at distances greater than 100 m
from the center of the source. For receptors located at distances less than 100 m from
the source, short-range dispersion models are appropriate to estimate ambient air
concentrations.
184
-------
First, to calculate the 2,3,7,8-TCDD vapor volatilization rate, pertinent data can be
listed as follows:
Water/soil partition coefficient, Kj = 4,680 L/kg (Schroy et al., 1985b)
Henry's Law Constant, Hc = 1.6 x 10~5 atm m3/mol (Podoll et al., 1986)
Diffusivity of dioxin in air, Dj = 0.05 cm2/s (Thibodeaux, 1985)
Soil porosity, E = 0.35
Soil density, Ps = 2.65 g/cm3
Soil/air partition coefficient, Kas = 41Hc/K(j
= 1.4 x 10"7 g soil/cm3 air
In order to estimate the average emission rate, the value for the intermediate parameter,
G is calculated as follows, using Equation 4-3:
G = Di(E)4/3/[E + Ps(l-E)/Kas]
= 1 x 10~9 cm2/s. (6-15)
An example calculation will be provided for a case where the initial level of
contamination, C0, is 1 ppb and the size of the landfill is 1 acre. The exposure factor
values listed in Table 6-5 will be used. In addition, for lifetime exposure evaluation the
emission rate can be averaged over the exposure duration, T, of 70 years (2.2 x 10^ s).
Then the average emission flux, Nj, is
Nd = 2KasC0Di(E)4/3/[(3.14)GT]1/2
= 1.3 x 10~18 g/cm2-s = 1.3 x 10'8 ug/m2-s. (6-16)
The above emission flux assumes that the site is contaminated from the surface of
the soil downward and that no clean cover material is applied initially. When soil is
185
-------
contaminated at the surface, the emission rate is initially high, and gradually decreases.
The instantaneous emission rates at various time intervals can be summed up by an
integration technique averaging over the period during which the emission continues.
For a 1-acre site, the dimensions of the site can be approximated by 63.6 m x 63.6
m. For a receptor 500 feet (152 m) away from the downwind edge of the site, the ISC
model can be applied by noting that the actual distance from the center of the facility
to the receptor is 152 m + 63.6 m/2 = 183.8 m. This is the distance needed in obtaining
the values for the standard deviation. The ambient air concentration, Ca, can be
calculated (Equation 4-8) as 9.7 x 10~^ /jg/m , where the default stability class = D,
wind speed = 4 m/s, when the winds are blowing toward the receptor (100% wind
frequency), and other symbols are identical to earlier definitions. The ambient air
concentrations thus calculated are shown in Figures 6-2 and 6-3 for 1 acre and 10 acre
sites, respectively, as a function of distance.
The exposure associated with breathing ambient air was computed assuming the
concentration was reduced by the frequency that wind blows toward the receptor (15%)
and substituting Ca into Equation 6-1. By multiplying the resulting exposure by
absorption fraction and by the cancer potency factor, risk estimates can be obtained.
These risk estimates are given in the Appendix. The changes in exposures encountered
prior to absorption into the blood stream over distance are shown in Figures 6-2 and
6-3.
For an uncovered landfill with the contaminant initially present at the surface,
ambient air concentration, and therefore exposure, is directly proportional to the
contaminant concentration in soil. This relationship persists up to the point where the
air is saturated with 2,3,7,8-TCDD vapor (186 ppb). The concentrations calculated for
the scenarios in this chapter are well below the saturation level.
186
-------
Figure 6-2. Ambient Air Concentration
and Exposure with distance for
I Acre Site.
Figure fj-3. Ambient Air Concentration
and Exposure with distance for
10 Acre Site.
oo
TO
O>
1—
01
CM
T
Ol
.'O
ex
X
1.2
1.0
0.8
0.6
Q.4
0.2
oo
O
60
9
V
c
C
D
B
.
ro
C\J
i—i
I I
9
r
Ol
i-
^
to
O
O.
X
0
50 100 150 200 250
0
oo
i
o
00
3
u
c
o
u
(-1
•H
C
-------
The change in exposure with distance from a 10-acre site is illustrated in Figure
6-3. These calculations assume 1 ppb 2,3,7,8-TCDD in the soil, wind direction frequency
of 15%, and that exposure occurs over 80% of a 70-year lifetime. For scenarios 10 and
11, concerned with concentrations of 2,3,7,8-TCDD in soil at 1 ppt and 1 ppq, ambient
air concentrations and the corresponding exposure levels at the appropriate locations will
decline linearly with declining concentration in soil.
The ISC model cannot be used for estimating the on-site ambient air concentration
of 2,3,7,8-TCDD at the site with contaminated soil, because the model is not applicable
at distances close to the source. The on-site dispersion model given by Equation 4-6 was
used to estimate the concentration for scenarios 1-7. Wind rose data should be used for
site-specific evaluation of annual ambient concentrations. For the most common stability
class D, the standard deviations (a) can be estimated in units of meters as shown below:
ay = 0.1414 x°-894 (6-17)
and
-------
4-7. The on-site ambient air concentration for a 1-acre area of contamination, obtained
by substituting into the box model equation, equals 1.9 x 10"^ pg/m*. For a 10-acre site
at the initial contaminant concentration of 1 ppb, the on-site ambient air concentration,
calculated in the same fashion, equals 6 x 10"' /ig/m^.
In reality, as volatilization proceeds, a non-uniform concentration will be established
along the depth of the contaminated soil. The non-uniformity of the concentration
profile can be reduced still further if the soil surface is disturbed. Disturbance reduces
the effectiveness of the surface layer in acting as a diffusion barrier to pollutant
transport, increasing the emission rate. Thus, removal of surface soil some time after
the initial contamination will increase emissions, and hence exposure. However, it is
very difficult to quantify the magnitude of this phenomenon and the corresponding
increase in exposure to downwind populations.
Controlled or closed landfills typically are capped. The major purpose of capping is
to minimize the generation of leachate by preventing the infiltration of precipitation into
the waste. However, a cap initially free of contaminant will also reduce volatilization
rate, thus reducing exposure.
Rigorous mathematical formulae are available to predict the emission rate at various
thicknesses of cover material (U.S. EPA, 1986a). Because the calculations involved
require a computer for solution, as an approximation one may use a retardation factor, as
suggested in Chapter 4. The vapor emission rate through a cover will not reach steady
state for hundreds of years. The emission rate, which initially is zero, gradually
increases, and finally reaches a steady state when all adsorption sites are saturated.
Taking into account this time-dependent emission rate, a 25-cm cover initially free of
contaminant will reduce the exposure and risk over a 70-year period by a factor of 5.
189
-------
d. Dermal Exposure
The exposures resulting from this pathway were computed using Equations 6-1. All
parameters used in these equations are discussed below.
Dust from contaminated soil can. be carried indoors, potentially causing exposure
resulting, from cleaning and other indoor activities. For this analysis it is assumed that
the 2,3,7,8-TCDD levels in the; soil surrounding a home are equal to the levels in the
dust inside, a home.
Very few data are available on the amounts of soil that accumulate on human skin.
Obviously, such amounts will vary considerably depending on behavior characteristics,
skin condition, soil type, exposed skin area, contact time, etc. Schaum (1984) concluded
that the limited data suggest a daily contact rate of 0.5 to 1.5 mg/cm . Using this
factor in conjunction with estimates of the exposed skin areas and number of days
exposed, the total soil contacted can be computed:
Soil contact rate (mg/d) =
contact rate (mg/cm^-d) x exposed skin area (cm^) (6-19)
For the purposes of this analysis, the contact rate is assumed to be 1 mg/cm^-day
for all scenarios. The exposed skin area, depends on the type of clothing worn and the
age of the exposed individual. Schaum (1984) used a range of 910 to 2,940 cm^ for an
adult, based on typical clothing, and adjusted this number for children proportional to
their body surface area. This age adjustment is small compared to the overall
uncertainty involved in 40- to 70-year exposure periods. Thus, for this analysis, the
exposed skin area was assumed to be 1,000 cm^ for all scenarios. Applying Equation
4-18, the soil contact rate is computed as 1 g/d, which was assumed for all scenarios.
1;90
-------
The exposure duration was derived from the scenario assumptions given in Table
6-1. For scenarios 1 through 4, scenarios 8 through 11, and scenario 15, the exposure
durations were calculated as follows:
Exposure duration = (70 yr)(365/d/yr)(0.8) = 20,000 d (6-20)
For scenarios 5 through 7, and 12 through 14, the exposure durations were
calculated as follows:
Exposure duration = (40 yr)(365 d/yr)(0.50) = 7,300 d (6-21)
Schaum (1984) was able to find only one relevant animal experiment dealing with
dermal absorption of 2,3,7,8-TCDD. This experiment suggested that the absorption
fraction varies from 0.07% to 3%. Differences in skin properties between humans and
rats and differences between experimental conditions and the human exposure scenario
make it very uncertain how well this absorption estimate applies in this analysis. The
geometric average of this range, or 0.5%, was assumed to apply in all scenarios.
The remaining parameters used to calculate exposures and risks, i.e., body weight,
lifetime, potency factor, degradation factor, and dilution factor were all discussed
previously in Section B.2 of this Chapter.
e. Soil Ingestion
Soil ingestion exposure estimates were based largely on the procedures provided by
Schaum (1984). The exposures were calculated using Equation 6-1. All parameters and
value assumptions used in these equations are discussed below. While estimates of soil
ingestion rates are still uncertain, new studies have allowed some refinement of the
procedures presented in the original report. As discussed in Chapter 4, the current
literature suggests that ingestion rates average from 0.2 to 1.0 g/d for young children.
The low end of this range, 0.2 g/d, was assumed to apply in typical case scenarios 5
191
-------
through 7, and 12 through 14. The upper end, 1 g/d, was assumed to apply in reasonable
worst case scenarios 1 through 4, scenarios 8 through 11, and scenario 15.
Inadvertent soil ingestion occurs among adults as well as children. As indicated in
Chapter 4, actual measurements of adult soil ingestion have not been made. However,
Hawley (1985) estimated soil ingestion could be 61 mg/d, based largely on unsupported
assumptions regarding activity patterns and corresponding ingestion amounts. This
ingestion rate is much less than the 200 to 1,000 mg/d assumed for children.
However,the longer exposure periods for adults suggest that adult soil ingestion could be
of the same magnitude as that for children.
The gut absorption of 2,3,7,8-TCDD adsorbed to soil is discussed in detail in Section
A of Chapter 5. The available data suggest that the absorption fraction is 20% to 40%.
The mid-value of this range (30%) was assumed to apply to all scenarios.
Soil ingestion can occur at any age, but is most prevalent among children. The
best available data apply to children ages 2 through 6. On this basis, Schaum (1984)
concluded that this age represents the time when mouthing tendencies and lack of
understanding of personal hygiene will cause the most significant soil ingestion. This
5-year time interval was adopted in this analysis as the basis for computing the exposure
duration. Combining this interval with the scenario assumptions summarized in Tables
6-1 and 6-2, the following estimates were made: Scenarios 1 through 4, and 8 through
11, and 15:
Exposure duration = (5 yrX365 d/yr)(0.8) = 1,500 d (6-22)
Scenarios 5 through 7 and 12 through 14:
Exposure duration = (5 yr)(365 d/yr)(0.5) - 910 d (6-23)
192
-------
The factors 0.5 and 0.8 in the above calculations reflect estimates of the fraction of time
an individual is likely to spend in the exposure area.
The body weight for a child aged 2 to 6 is an average of 17 kg, which was assumed
in all scenarios.
The remaining factors used to compute exposures and risks (2,3,7,8-TCDD dilution
factor and degradation factor) were discussed previously in Section B.2. of this chapter.
f. Ingestion of Beef and Dairy Products
Grazing animals ingest soil as they forage, in amounts ranging from 2% to 15% of
total dry matter intake, as noted in Section C of Chapter 3. Some fraction of the
contaminated soil deposited on neighboring land by runoff thus is likely to be ingested
by animals pastured there. Dust evolution and redeposition from a contaminated site
onto forage generally is a less important source for bioaccumulation through cattle, as
shown in Section B.2.a. of this chapter, and adds a negligible amount to the intake of
cattle in these scenarios.
The contributions of beef and dairy products to 2,3,7,8-TCDD in the human diet
were calculated by substituting input values from Table 6-5 into Equation 6-1. Rationales
for the dilution factors used are provided in Section B.2.a. and the results are
summarized in Table 6-4.
For the reasonable worst case scenarios the exposure duration was estimated
assuming that the exposed individuals received 44% of their beef diet from the
contaminated area. This value represents the average percent of annual consumption
which is home-grown by 900 rural farm households (U.S. Department of Agriculture,
1966). Accordingly 11,000 days is estimated:
(70 yr)(365 d/yr)(.44) = 11,000 d (6-24)
193
-------
The 44% diet assumption was. also applied to. the 40 year typical case scenario resulting
in an exposure duration of 6400. daySv Similarly, for the dairy products ingestion
scenarios;, a; 40% (U.S.D.A., 1966) diet assumption; was used yielding a 10,000 day exposure
duration? for the reasonable worst-case; scenario: and 5800 day exposure duration for. the
typical! case scenarios. Nate that the percent diet assumption could have been applied to
either; tfie exposure; duration estimate or- the consumption rate estimate. When the
exposure factors; are' multiplied together, either approach yields the same- exposure
estimate.
Bioaccumulation of. 2,3,7,8-TCDD. in cattle is well documented. (Section C of Chapter
3). Equation. 6-1, used to. calculate, exposure, takes this into account in the distribution1.
factor term.. The ratio of 2,3,7,8.-TCDD; in beef fat to that in soil where the animals
were pastured was also discussed in Section C of Chapter 3, as was the ratio- of
2,3y7,8-TCDD in milk fat. to that in soil. Use-of a beef fat/soil bioaccumulation factor of
0.3' to 0.4 and a milk fat/soil bioaccumulation factor of 0.04 were recommended in the
absence, of site-specific data.
Fries: and Marrow (1975) estimated that 50%. to 60% of 2,3,7,8-TCDD was absorbed
by rats from feed. Rose et al. (1976) estimated that 86% of 2,3,7,8-TCDD in a. mixture of
acetone and corn oil fed by gavage to rats was absorbed. The average of this range, or
68%, was assumed to apply to beef and dairy products.
Exposures calculated under the various scenarios are presented in< Table. 6-6, and
risks, are- presented in the Appendix.
g. Ingestion of Fish.
Potential exposure through ingestion of freshwater fish was calculated using
Equation- 6-1. While significant new information on fish consumption is. described, in
Section F of Chapter 4; the national average intake of freshwater fish still appears well
represented by the value 6.5 g/d used by Schaum (1984). The reasonable worst-case
1,94,
-------
scenarios (2-4, 8-11, 15) use a somewhat higher intake value, 30 g/d, which attempts to
represent a small subpopulation fishing from nearby water bodies and consuming fish
caught.
Two fish consumption behavior patterns are assumed in this report. In the typical
case, the exposed individual consumes an average of 6.5 g/d of freshwater fish of which
10% is derived from the contaminated source (a relatively small stream). In the
reasonable worst case, the exposed individual consumes 30 g/d of freshwater fish of
which 10% is derived from the contaminated source (a relatively small pond).
The reasonableness of these assumptions can be analyzed as follows. The
consumption rate for the typical case represents an average U.S. value. A one meal
serving of fish weighs about 100 g. Thus in the typical case a little over two meals of
contaminated fish would be consumed per year:
(6.5 g/d)(365 d/yr)(0.1)/100 g/meal - 2.4 meals/yr (6-25)
Assuming that the weight of the fillet is half that of the whole fish, this would imply
that two 200 g (about a half pound) fish are caught from the stream per year. This
catch rate and fish size appear reasonable for a small stream and an average consumer.
In the reasonable worst case, a consumption rate is used which represents a person
actively involved in sport fishing and consuming relatively large amounts of fish. For
this situation, about 11 meals of contaminated fish would be consumed per year:
(30 g/d)(365 d/yr)(0.1)/100 g/meal - 11 meals/yr (6-26)
This number of meals implies that eleven 200 g fish are caught from the pond per year.
This catch rate and fish size appear reasonable for a small pond and an active fisherman.
195
-------
For the reasonable worst-case scenarios, the exposure duration was estimated as
2600 days:
(365 days/yr)(70 yr)(0.1) - 2600 d. (6-27)
For the typical scenarios, the exposure duration was estimated as 1500 days:
(365 d/yr)(40 yr)(0.1.) = 1500 d (6-28)
Note that the 10% of diet assumption could have been applied to either the exposure
duration estimate or the consumption rate estimate. When the exposure factors are
multiplied together either approach yields the same exposure estimate.
Calculating the concentration of 2,3,7,8-TCDD in water as a result of contaminant
leaching from sediment provides interesting information on bioaccumulation in fish and
other aquatic organisms. The methodology is shown in Section B of Chapter 3. This
calculation represents the "water-to-fish" portion of fish body burden. In addition,
aquatic organisms will pick up contamination from contact with sediments, ingestion of
suspended and bottom sediments, and ingestion of other contaminated organisms. The
latter contribution practically always predominates, but the influence of both are
accounted for by measuring "fish/sediment distribution factors." Based on the discussions
in Section C of Chapter 3, fish/sediment distribution factors are in the range of 1 to 10,
but a single value of 5 was used in calculating exposures associated with this pathway
for scenarios 2 through 15. The results are presented in Table 6-6. Fries and Marrow
(1975) found that 50% to 60% of 2,3,7,8-TCDD was absorbed by rats from feed. Rose et
al. (1976) found that 86% of 2,3,7,8-TCDD in a mixture of acetone and corn oil fed by
gavage to rats was absorbed. The average of this range, or 68%, was assumed to apply
to human gut absorption of 2,3,7,8-TCDD from fish.
196
-------
h. Water Ingestion-- Surf ace Water
Overland runoff could result in contamination of surface waters near the site. The
transport model described in Chapter 3 will be used to estimate the extent of release of
2,3,7,8-TCDD from sediment to water body. It is straight-forward to use the transport
model once all the mass transfer coefficients are evaluated. These transfer coefficients
include the sediment side coefficient, the water-side coefficient, and the overall mass
transfer coefficient for the air-water interface. Basically the sediment-side and
water-side coefficients can be estimated from Equations 3-5 and 3-6. Examples of the
worst case and a typical case will be separately considered in evaluating the effect of
sediment on the water quality.
In the worst case, the water body will be assumed to be a 1-acre lake (64 m x 64
m) surrounded by soil with 1 ppb 2,3,7,8-TCDD. As described previously, the sediments
at the bottom of the lake are assumed to reach an equilibrium level at the same
concentration as the soil (1 ppb). To estimate the sediment- and water-side mass
transfer coefficients, the following assumptions are made: The water and sediment
temperature is ambient (15° - 25°C); the depth of the water body is 5 m; the lake water
has no significant inflow or outflow; the thickness of sediment is 1 cm; and the bottom
of the lake is covered by sediment with a porosity of 50% and with 1% organic matter.
These assumptions provide the water-side coefficient of kw = 0.63 cm/hr and the
sediment-side coefficient of ke = 8 x 10"^ cm/hr.
As presented in Section B of Chapter 3, the overall mass transfer coefficient
between the water and air phases is KLa = 0.725 cm/hr. At the contaminant
concentration in sediment of 1.0 ppb, the concentration of the contaminant in the lake
body can be calculated from Equation 3-4, and is 2.3 x 10"^ ug/L. This concentration
compares with an equilibrium concentration that would exist in water if an equilibrium
between the sediment and water were established; or 1.0 ug/kg/4,680 kg/L = 2.1 x 10"^
197
-------
Mg/L. The transport model therefore indicates a concentration in water of about two
orders of magnitude lower than the equilibrium concentration.
The risk from drinking surface water contaminated at 2.3 x 10"^ /*g/L can be
computed using the procedure presented in the Appendix. Assuming:
o the absorption fraction is 0.5 (Fries and Marrow, 1975);
o the exposure duration is 2Q,000 days;
o the water consumption rate is 2 L/d (U.S. EPA, 1984a);
o a 70-year lifetime; and
o an average body weight of 70 kg;
the risk is estimated to be 9 x 10 . This should be considered as a worst-case
estimate, since it assumes a high level of sediment concentration and a long exposure
duration. Also, it is rare for people to use untreated water from small ponds as a
drinking-water source. It is much more likely that wells would be used, which would
have substantially reduced 2,3,7,8-TCDD levels.
In a typical case, .the analysis deals with the sediment contaminating the bottom of
the river stream. As previously shown, the concentration of 2,3,7,8-TCDD in the river
sediment is diluted by a factor of 1000 from the original concentration to a
concentration of 0.001 ppb. It is assumed that the size of the contaminated area is
similar to .the size of the lake. Also, the depth of the flowing river is assumed to be 5
m with the bottom sediment at a thickness of 1 cm. The average water flow rate is 1
m/s.
Calculations similar to that just shown for the pond water body will yield the
concentration in the river as 3 x 10"^ A*g/L. with an obviously negligible risk (3 x 10"^
at the potency slope of 0.156 (ng/kg d)"1).
198
-------
i. Ground Water Contamination
Improper disposal of toxic contaminants will result in releases of such contaminants
into the environmental media, including ground water. Although the concentration of
2,3,7,8-TCDD in many wastes is very small, it is possible that leachate being generated
from the 2,3,7,8-TCDD-containing landfill site can contain the contaminant and enter the
ground water. The amount of leachate generation and the extent of ground water
contamination will be dependent on many site-specific characteristics, such as the nature
of the waste, the magnitude of precipitation, and the transport characteristics of the
ground water media.
As precipitation falls on the ground, the water will penetrate the surface of the soil
to form leachate. Precipitation includes all forms of water deposited on the earth's
surface, including mist, rain, hail, sleet, and snow. Precipitation will become either
infiltration or surface runoff, or will return to the atmosphere by evapotranspiration.
That portion of the precipitation that infiltrates will affect leachate formation and reach
ground water. The amount of infiltration is affected by many factors. These include
thickness of the saturated layer, moisture content of the soil, magnitude of compaction,
macrostructure of soil, vegetative cover, temperature, freezing of soil moisture, and
entrapped air.
When the infiltrated water reaches ground water, the contaminant will be impacted
by flowing ground water to migrate downgradient by advection, and will be dispersed by
diffusive and dispersive actions, and will be retarded in the ground water medium,
depending on the adsorptive capacity of the medium.
Depending on landfill design, leachate may directly enter ground water or may be
subject to retardation by the unsaturated zone before entering ground water. In order
to assess the extent of migration of 2,3,7,8-TCDD-containing leachate in ground water, a
ground water fate and transport model (Hwang, 1986) is used. There are several
199
-------
parameters that need to be assumed in using the model. These parameters, necessary to
estimate the retardation coefficient, include ground water velocity, the concentration of
the contaminant in leachate, the precipitation infiltration rate, the aquifer depth, the
porosity of the aquifer, and the organic carbon content of the ground water medium.
According to Darcy's law, ground water velocity is a function of hydraulic
conductivity, hydraulic gradient, and effective porosity. Todd (1980) lists representative
hydraulic conductivities for a variety of geologic materials. Gravel material has hydraulic
conductivities ranging from 150 to 450 m/d. Representative values of hydraulic
conductivity for sandy material are shown to be between 3 and 45 m/d. Hydraulic
conductivities for sandstone range from 0.2 to 3 m/d. The ground water velocity used in
the model should be site specific. For a productive alluvial aquifer with hydraulic
conductivity of 100 m/d, a hydraulic gradient of 10"^ (which is within the values
commonly observed in field conditions [Freeze and Cherry, 1979]), and an effective
porosity of 0.2, the ground water velocity is calculated to be 5.5 x 10~^ cm/s, which is
used in the fate and transport model. The hydraulic conductivity used is within the
range for sand and gravel.
The mean annual precipitation in the United States ranges from 5 to 100 inches
(Wisler and Brater, 1959). The highest precipitation occurs along the northern Pacific
coast, where rainfall in excess of 100 inches is not unusual. A belt of very low annual
rainfall lies just east of the coastal mountains, where the rainfall ranges from .5 to 10
inches. Data compiled by Todd (1983) show that a considerable amount of precipitation
is lost to the atmosphere by the process of evapotranspiration, and becomes surface
runoff. The portion of precipitation that becomes ground water recharge is generally
less than 10%, and at times is less than 1%. The infiltration rate is relatively high in
the Lake Tahoe basin region, amounting to nearly 25% of the precipitation. However,
this high percentage is an isolated case. As a typical worst-case infiltration rate, 15% of
200
-------
the worst-case precipitation is assumed to become infiltration in the model simulation.
This corresponds to an infiltration rate of 1.2 x 10~6 cm/s, averaged on an annual basis.
The organic carbon content of the ground water aquifer material is needed in
estimating the pollutant retardation factor. Typical values of the organic carbon content
are reported to be from 0.4% to 10.0% (U.S. EPA, 1985f). The report also notes that low
carbon contents of less than 1% are typical of those found in deep aquifer materials.
Bedient et al. (1984), in their study of ground water quality, attributed high levels of
organics at a well some distance from the studied waste pits to the extremely low levels
of organic carbon in the aquifer (less than 0.02%). Although such an organic carbon
level is considered extremely low, the use of this value will reflect a conservative
estimate for contamination concentration because estimated retardation factors would
predict migration farther away from the pit.
At the soil organic carbon-water partition coefficient of 486,000 cm^/g organic
carbon (Schroy et al., 1985b), and organic carbon content (OC) = 0.0002 for ground water
media, the retardation factor becomes Rj = 973, calculated from R
-------
The concentrations in ground water at several distances from the place of
contamination are calculated at the release period of 100 years using the semi-analytical
three:-dimensional area source model (Hwang, 1985, 1986; Codell, 1982).
The results of these simulations are. tabulated below (Table 6-7).
TABLE 6-1. SIMULATED CONCENTRATIONS AT WELLS
Distance of receptor from. Concentration in
the downgradient edge ground water (mg/L)
of the source (m) after 100 years
15.2 2.8 x 10'17
50 2.3 x 10~17
152 8.8 x 10-18
The results show that the concentrations at the. locations of 15.2 m and 152 m
correspond to the exposure levels of 8 x 10"^ and 3 x 10"'^, ng/kg-d (risks of
2 x 10~13 and 9 x 10"14 at the potency slope of 0.156 (ng/kg d)~J, respectively.
Although the ground water has traveled, about 1.7 x 10^ m of distance over a 100-year
period, the plume concentrations at all locations are relatively small, according to the
simulation. This is principally due to the high retardation factor for the contaminant,
which of course is related to the hydrophobia nature of 2,3,7,8-TCDD.
This simulation indicates that because of very high retardation of 2,3,7,8-TCDD by
the ground water media, the concentration in ground water will be small, and ground
water is not a significant exposure pathway for 2,3,7,8-TCDD migrating from landfill
sites. However, as noted earlier, if cosolvents are present in the landfill or if
channeling occurs, the mobility of 2,3,7,8-TCDD may be significantly increased. This, is
an especially important caution when applying the methods of this report to landfill
situations at actual sites. Although the solubility of 2,3,7,8-TCDD would probably not
increase enough in a one-phase system (i.e., dissolving in benzene-saturated water) to
202
-------
make it extremely mobile, where two separate phases are present (e.g., organic solvents
present at well over saturation levels), the 2,3,7,8-TCDD may behave quite differently.
Evaluating the potential for ground water contamination at any site must include an
analysis of the cosolvent status.
j. Fruit and Vegetable Ingestion
The literature reviewed in Section 3.D. showed plant-to-soil ratios varying from
< 0.1% to > 100%. The most recent data (Sacchi et al., 1986) show an uptake ratio of
about 2% for above-ground plants (beans and maize). Cocucci et al. (1979) found ratios
of about 100% in below-ground vegetables. However, these tests were conducted on
plants exposed to 2,3,7,8-TCDD during the Seveso incident, which may not reflect normal
exposure conditions. Greenhouse tests briefly reported by Wipf et al. (1982) indicate a
ratio of between 1% and 1.5% for underground vegetables (carrots). For purposes of a
preliminary estimate, a 2% ratio was assumed to apply to all vegetables and fruit.
The average daily ingestion rate of fruits and vegetables is 280 g/d (U.S. EPA,
1984c). The U.S. Department of Agriculture (1986) reported that approximately 50% of
vegetables eaten in rural farming areas are home-grown. This value was assumed to
represent the portion of a person's vegetable diet derived from the contaminated source.
Fries and Marrow (1975) found that about 50% of the 2,3,7,8-TCDD was absorbed by rats
from feed. This value was assumed to represent the absorption occurring in humans from
vegetables. Finally, assuming 1 ppb 2,3,7,8-TCDD in the garden soil and 55% absorption,
the 70-year risk can be calculated as:
(plant (soil
(ingestion uptake (fraction concen- (cancer
Risk = rate) ratiol home-grown) tration) potency) (6-29)
(body weight)
« (280 e/d) (0.02) (0.5) (1 ng/g) (0.156 kg-d/ne)
70 kg
6 x 10
-3
203
-------
Such risk estimates are linearly proportional to the soil concentration. If th& soil level
were 1 ppt, the risk estimate would be reduced, to 6 x 10"^.
Several points should be considered in evaluating, the above calculation:
o 2% uptake is much less than, uptake levels measured by several investigators
(as noted above these were rejected because they were based on either
s
non-edible portions of the plant or derived from studies conducted in
connection with the;Seveso incident).
o Only one study (Wipf et al., 1982) was found which actually measured
2,3,7,8-TCDD levels in fruit itself (fruit is used here in the technical
sense to refer to plant parts consisting of the seed and surrounding, pulpy
tissue, which would include corn, tomatoes, grain, etc.) Wipf et al., report that
the levels of TCDD (unspecified isomers) in the edible portions of apples,
pears, peaches, corn and other fruit were less than the detection limit of
about 1 ppt compared to about 10,000 ppt in the soil. The Wipf study would
suggest that the 2% uptake assumption over-estimates the risk, of fruit
ingestion.
o Two studies (Young, 1983 and Sacchi, 1986) did find substantial uptake
levels in the aerial portions of plants. It is unclear whether this contradicts
the findings of Wipf et al., 1982, or whether it is reasonable to expect
different levels in the fruit, than the other aerial, portions of the. plant.
o The literature does generally suggest that higher uptake levels have been
observed in underground plant parts than above-ground, plant parts- The
logical conclusion that this implies, is, if data, were available, risks should be
calculated, in a way which distinguishes between underground and above ground
vegetables, rather than using an across-the.-board uptake value such as. 2%. As
an example of the influence which this consideration could make, an estimate
204
-------
of the possible risk associated with potato ingestion was made. Only one
study (Cocucci et al., 1979) examined uptake in potatoes where it was
estimated as about 40%. As noted earlier, it is unclear if Coccuci results are
applicable to normal gardening scenarios, but for purposes of this example it is
assumed to be applicable. The average ingestion rate of potatoes is about 13
g/d - excluding french fries and potato chips (Pennington, 1983). Using these
parameter values and the same values for the other parameters as were used in
the above calculation, the risk associated with potato ingestion is 6 x 10"^.
Coincidentally, this risk level is equal to the risk calculated earlier (using the
2% value) for all fruits and vegetables. This suggests that using this
approach (i.e., the uptake values from the Coccuci paper applied to
underground portions of plants) would lead to much higher estimates of risk.
Due to the high degree of uncertainty associated with this pathway, the exposures
were not presented in Table 6-6 or in the Appendix. However, if they were computed
using a 2% uptake as shown above and not making any distinction between typical and
worst case behavior patterns, the upper bound incremental risks associated with the
various scenarios would be as follow. For scenarios 1, 2, 5, 8, 9 and 12 where the soil
level is 1 ppb the risk is 6 x lO"-'. For scenarios 3, 6, 10 and 13 where the soils levels
is 1 ppt the risk is 6 x 10"^. For scenarios 4, 7, 11 and 14 where the soil level is 1
ppq the risk is 6 x 10"9. For scenarios 8,9,10, and 11, the risks for the population
living 100 feet away from the site would be further reduced by a factor of 0.35, the
dilution factor. For scenarios 12, 13, and 14, the 2,3,7,8-TCDD concentration at the
population area 500 feet away from the combination site is diluted by 0.008 from the
original concentration. Hence, the risks would also be reduced by 0.008. Finally, for
scenario 15 the risk would be negligible since this involves a capped landfill which would
205
-------
prevent 2,3,7,8-TCDD exposure to the plants (assuming the cap is thicker than the root
depth and remains clean).
C. DESCRIPTION OF EXPOSURE SCENARIOS FOR INCINERATION
A broad spectrum of combustion devices have been tested to determine the presence
of polychlorinated dibenzo-p-dioxins (PCDD) in flue gas and ash. 2,3,7,8-TCDD is
suspected of being formed as a low-level by-product in the process of burning many
types of waste or fuel material. The combustion devices where literature test results
showed presence of 2,3,7,8-TCDD or PCDD in flue gas or ash include municipal waste
incinerators, coal-fired utility boilers, wood stoves and fireplaces, hazardous and chemical
waste incinerators, and the internal combustion engine. Recent reports compiled by U.S.
EPA (1986) also list other combustion devices such as smelters, sewage sludge
incinerators, wire reclamation incinerators, and drum and barrel furnaces. In addition,
Sheffield (1985) mentions forest fires and several types of chemical production facilities
as sources. Des Rosier and Lee (1986) cite PCB transformer fires as a potential source
of 2,3,7,8-TCDD formation. Buser (1979) notes formation of dioxins from the laboratory
pyrolysis of chlorobenzenes. It should be noted that the main focus of this section is to
address the potential of health risk associated with the emissions of 2,3,7,8-TCDD and fly
ash generated from the incineration of municipal waste. Municipal incinerators were used
in the example scenarios since there are emissions data available and these types of
incinerators are currently of interest.
The available literature information on emission tests shows that 2,3,7,8-TCDD
emissions released into the atmosphere consist of vapors and particulate matter, and that
fly ash collected in air pollution control equipment contains some amount of
2,3,7,8-TCDD. Although the exact ratio of vapor to particulate 2,3,7,8-TCDD emissions
has not been systematically studied, the mere presence of the two-phase releases
complicates the multimedia exposure analysis. There is an indication that the long-range
206
-------
transport of 2,3,7,8-TCDD favors the presence of 2,3,7,8-TCDD in the vapor phase (Eitzer
and Hites, 1986).
The vapor-form stack emissions will disperse in the air as will the particulate-form
stack emissions. Exposure to humans results from inhalation of 2,3,7,8-TCDD vapors and
particulates, ingestion of contaminated water, fish, soil, homegrown food, and dermal
contact.
Ash is produced from non-combustible material as bottom ash, which is also referred
to as residue, and from the remains of the refuse or fuel after burning as airborne fly
ash which is collected in paniculate control equipment. The bottom ash produced from a
high temperature oxidation process is commonly referred to as slag. Some incinerators
operate at sufficiently high temperatures to produce slag. Not all incinerators produce
slag. Many newer hazardous waste incinerators operate at slagging temperatures. Most
existing hazardous waste incinerators do not. Very few, if any, municipal waste
incinerators operate at slagging temperatures. One paper reports the bottom residue
produced from a municipal mass burning incinerator as slag (Nottrodt, 1986). Nottrodt
(1986) and Hay et al. (1986) indicate that the total CDD in bottom ash slag is negligible.
However, MRI (1985) reports levels of 2,3,7,8-TCDD in bottom as high as 1.5 ng/g.
Actually, bottom ash as used in the MRI report refer to a combination of both bottom
and fly ash, so it is not clear how much 2,3,7,8-TCDD is actually present in the real
bottom ash. Due to this uncertainty and the fact that fly ash is clearly of greater
concern, the risk from slag is not addressed further in this report. Land disposal of the
collected fly ash in a landfill is covered in the incinerator scenarios discussed later.
There are basically three types of incinerators used to incinerate municipal wastes:
mass burn, modular, and refuse derived fuel (RDF) incinerators. The mass burn
incinerators predominate. Mass burn incinerators are field-erected and can process
municipal waste without major preprocessing steps except removing large items which will
207
-------
not .go through the feed system. These incinerators can range in size from 50 to 3000
tons per day (TPD) of waste input (Radian Corporation, 1986). Modular incinerators are
package units and can range in size from 5 to 100 TPD. The 'refuse-derived fuel
incinerators process wastes to produce refuse derived fuel and combust the RDF in a
waterwell boiler.
1. Summary of Incineration Scenarios
Four scenarios for municipal waste incinerators are described here and apply to the
calculations in the remainder of the chapter. Two of the scenarios (16 and 17) deal with
population habits which involve cultivation of some farm products with a stagnant pond
in the area which may be used for stocking fish or for other purposes. This scenario
•may represent a reasonable-worst case for some population. The other two scenarios (18
and 19) use habits which also involve obtaining some farm products from the impact "area,
but have a flowing stream in the area so that flowing water may have an affect of
diluting soil and water contamination in the stream. This scenario may represent more
typical for some population at large. Of the four scenarios, two (16 and 18) deal with a
large (3000 TPD capacity) incinerator, while the other two (17 and 19) deal with a
smaller 120 TPD facility. The major assumptions for the incinerator facilities are
summarized in Table 6-8, while the major assumptions about population habits are
summarized in Table 6-9.
The scenarios themselves describe a hypothetical family living near a municipal
waste incinerator. The incinerators used in this analysis are modeled after an existing
facility in Hampton Roads, Virginia and planned facility in Tampa, Florida. Buildings at
the Hampton Roads incinerator cause downwash effects which cause the modeled maximum
concentrations from emission to occur at 200 m from the incinerator. Since the
hypothetical family is assumed to live on a small farm it is unreasonable to expect them
to be located in the urban area where the ground level concentrations are maximum.
208
-------
TABLE 6-8. PARAMETER VALUES FOR INCINERATOR FACILITIES
IN FACILITIES IN SCENARIOS 16-19
Capacity (TPD)
Heat Value of Waste (BTU/lb.)
Height of Buildings (m)
Controlled Emission Factor
for 2,3,7,8-TCDD (jig/kg)
Emission Rate for 2,3,7,8-TCDD
(g/s)
Vapor Form Emission Rate(g/s)
Particulate Form Emission Rate
(g/s)
Stack Parameters:
No. of Stacks
Height (m)
Diameter (m)
Gas Temperaturerature (°K)
Gas Velocity (m/s)
16
3000
4815
42
0.001
3xlO~8
1.9xlO~8
l.lxlO"8
4
46a
4.lb
470
11.3
Scenarios
17
120
4815
27.3
0.289
1.8xlQ-7a
l.lxlO-7a
0.7xlQ-7a
2
27.4a
1.2a
543
12
18
3000
4815
42
0.001
3xlO-8
1.9X10'8
l.lxlO'8
4
46a
4.1b
470
11.3
19
120
4815
27.3
0.289
1.8xlO-7a
l.lxlQ-7a
0.7xlO-7a
2
27.4a
1.2a
543
12
aEach stack.
''Note that a heat balance calculation indicates that the amount of combustion gas generated
requires four stacks each with a 4.1 m diameter to maintain the assumed stack gas velocity.
209
-------
TABLE 6-9. PARAMETER VALUES FOR
POPULATIONS IN SCENARIOS 16-19
Parameter Scenarios 16 & 17 Scenarios 18 & 19
Distance to Exposure .8 .8
Area (km)
Years at residence 70 40
% time at residence 80 50
% freshwater fish
from contaminated
source 10 10
% beef diet from
contaminated source a 44 44
% dairy diet from
contaminated source a 40 40
Ages for soil ingestion 2-6 2-6
a Average percent of annual consumption which is home grown by
900 rural farm households (U. S. Department of Agriculture, 1966).
210
-------
Figure 6-4. Incinerator Scenarios
Scenarios: 16,17 (Reasonable Worst Case)
O
Incinerator
0.8 Km
Exposure
Area
Scenarios: 18,19 (Typical Case)
Incinerator
0.8 Km
/\ / f Stream
Exposure
Area
211
-------
The nearest distance where a small farm could be located was assumed to be about 0.5
miles from the incinerator; for this reason, the hypothetical family is located 0.5 miles
(0.8 km) from the incinerator. Figure 6-4 is a schematic of the scenarios.
2. General Calculations and Factors Used
Since stack emissions are not covered in the methods discussed in Chapters 2-5,
which deal essentially with land-based material, some background information on
municipal incinerators and emission transport are briefly described below. This section
discusses the selected values (see Table 6-8) for size and emission factor for the
incinerator facilities in scenarios 16-19 from the perspective of what is known about
existing facilities in general. In addition, the methods used to calculate concentrations
a. Summary of Emissions Data and Vapor/Particulate Distribution
In 1977 researchers in the Netherlands reported that CDDs are found in fly ash and
flue gas of municipal incinerators (Olie et al., 1977 ). These findings led Dow Chemical
Company to issue a report suggesting the widespread presence of CDDs in emissions from
various combustion devices (Bumb et al., 1980). The sources listed by Dow include
hazardous waste incinerators, municipal waste incinerators, industrial boilers, and
fireplaces. Since the issuance of this report, many researchers have begun to
characterize the nature and amount of CDD compounds which can be present in flue gas,
fly ash, and slag. The results of these studies are briefly summarized in Tables
6-10 through 6-12. Since the major emphasis of this report is to assess the exposure to
2,3,7,8-TCDD, the summary presented is not inclusive of all literature information dealing
with the analysis of other types of CDDs. Many references which do not contain
specific analyses for 2,3,7,8-TCDD are omitted from the summary table.
212
-------
In attempting to quantify the presence of CDD emissions in flue gas, researchers
primarily analyzed fly ash emissions. Although some papers measured the vapor phase
CDDs in flue gas as well as in the particulate phase, most literature does not distinguish
between the two.
In several emission tests, the levels of 2,3,7,8-TCDD in the gas phase were higher
than those measured on captured particulate matter (Hagenmaier et al., 1986; Nielsen, et
al., 1986; Scheidle et al., 1985; EPA 1987b). In one study testing a municipal incinerator,
virtually all CDD emissions in the stack were in the gas phase (EPA 1987b).
More recently, Midwest Research Institute under contract to EPA summarized an
emission data base for stack test results conducted in U.S.A. and Canada (EPA 1987c).
The results show analyses of various congeners of dioxins in stack emissions, but there is
no indication as to what proportion is particulate vs. vapor form. Knowledge of the
relative amounts of particulates and vapors in stack emissions is important because
vapors will behave differently from particulates in the environment.
Table 6-10 lists a collection of data showing emissions of 2,3,7,8-TCDD in the stack
gas of municipal waste incinerators. For emissions other than those for 2,3,7,8-TCDD,
there is a more complete data base compiled for U.S. EPA (Radian Corp., 1983; U.S.
EPA, 1986; U.S. EPA, 1987). Many data in the literature are given in terms of
concentration in stack gas, commonly in nanograms per normal cubic meter (ng/Nm^). To
provide the data in consistent units, these data are converted to emission factors given
in ug of contaminant emission per kg of refuse burned.
This conversion is difficult because many papers do not provide process data, and %
CC«2 in the stack emissions necessary for such a conversion. For example, the stack
flow rate data corrected to 12% CC>2 is presented in units of dry standard cubic feet per
minute, or dscfm, and the concentration of 2,3,7,8-TCDD in stack emissions is given in
213
-------
TABLE 6-10. STACK EMISSIONS OF PCDDS
FROM MUNICIPAL INCINERATORS
Emissions Factor( jig/kg)
2.3,7,8-TCDD Total TCDD PCDD H6CDD H7CDD OCDD Total
Dioxins
0.001-0.0034 0.055-0.24 0.005-0.024 0.016-0.085 0.03-0.14 NR
Av = 0.002 Av = 0.11 Av = 0.013 Av = 0.04 0.07
0.015-0.038* 0.27-0.36* 0.3-5.2* 0.46-2.5* 0.59-2.7* 1.3-2.7*
Av=0.0265
NR 72.8* 107.1* 132.8* 98.6* 17.1* 428.
0.0007-0.01 0.01-0.07 0.25-0.29 0.21-0.37 0.14-0.8 0.48
Av = 0.005
0.0003-0.004 0.022-0.14 NRS NRS NRS 0.022-0.27
Av = 0.002
0.0006-0.003 0.012-0.028 0.002-0.005 0.008-0.02 NR 0.022-0.08
Av = 0.002
NR 0.031-0.3 NR NR NR NR
NR 0.22 0.33 0.72 0.91 1.76
0.015 0.043 NR NR NR NR
0.021 0.045 NR NR NR NR
0.38 1.6 2.9 2.3 3
NR 0.14-0.94 0.28-0.72 0.77-2.1 1.05-2.5 0.9-2.04
0.002 0.032 ND 0.082 0.038 0.013 0.167
Remark
Participate 37%
Vapor 63%
ESP Control
Starved Air
No control
Before Control
20-30% parti-
culate 70-80%
vapor
Literature
Survey
ESP & Wet
Scrubber
ESP Control
Particulate 37%
Vapor 63%
Av of 4 samples
Av of 3 samples
Sample Before
Control
Found in gas
None in ash
fContinued
Reference
Hagenmaier, et
al., 1986, EQV
DeFre'. 1986,
EQV
Hay et al.,
1986
Nielsen et
al., 1986.EQV
Nottrodt,
1986
Nottrodt, 1986,
EQV
Rappe, 1986
Scheidl et
al., 1985
EPA 1987b
EPA 1987b
EPA 1987b
Kilgroe et al
1986
EPA 1987b
EQV
)
214
-------
TABLE 6-10. (CONTINUED)
2,3,7,8-TCDD Total TCDD PCDD H6CDD H7CDD OCDD Total Dioxins Remark Reference
0.0021* 0.0316*
0.0824* 0.0383* 0.0128*
ESP Control MRI, 1987
Chicago
3.02
3.84* 6.05* 7.32* 1.93* 22.1*
ESP Control MRI, 1987
Hampton (1981)
0.289* 1.13*
ESP Control MRI, 1987
Hampton (1982)
0.145
1.04
5.44* 2.32* 0.726* 0.186* 9.68*
ESP Control MRI, 1987
Hampton (1983)EQV
0.089* 2.93* 6.86* 8.09* 7.31* 1.87* 27*
ESP Control MRI, 1987
Hampton (1984)EQV
0.00117* 0.0118* 0.0117* 0.016* 0.023* 0.037* 0.966*
ESP Control MRI, 1987
Peekskill EQV
0.000397* 0.00634* 0.0117* 0.02* 0.0174* 0.0189* 0.0745*
ESP Control MRI, 1987
Tulsa (Units EQV
1&2)
0.000371* 0.000893* 0.000243* 0.000504* 0.000842* 0.0027* 0.00517*
Fabric Filter MRI, 1987
Marion County EQV
0.000056* 0.000621* 0.00827* 0.0104* 0.0141* 0.0339* 0.0727*
Fabric Filter MRI, 1987
Wurzburg EQV
0.069
1.9
5.3 13.8 4.53
1.8 26.85
ESP MRI, 1987
Philadelphia EQV
(NWI)
0.062
1.8
5.2 4.6 2
0.83 14.05
ESP MRI, 1987
Philadelphia EQV
(NW2)
0.0206
Cyclone
Mayport
NR AV=0.013* 0.038* 0.0663* 0.1085* 0.167* 0.393*
Before Control MRI, 1987
Prince Edward EQV
Island
NR = Not Reported, NRS = Not Reported Separately, ND = Not Detected
EQV = Data used in calculating weight percent adjusted to equivalent 2.3.7.8-TCDD.
9 O +
All Data Converted from ng/m to ug/kg by assuming 6000 Nm /T of refuse burned unless otherwise noted by .
215
-------
TABLE 6-11.
PCDDS IN FLY ASH OF COMBUSTION
Source 2,3,7,8-
Categpry TCDD
MWI 0.07-0.1
MWI
MWI
MWI
MWI
MWI
Ontario
Oslo
Paris
Kyoto
Hiroshima.
Manchida
WBS
WBS
MWI 2.3
MWI 0.29
MWI
MWI
Concentration (ng/g)
Total
TCDD PCDD H6CDD H7CDD OCDD
0.9-1.6 0.3-0.5 3.2-9 42-215
<20ppt
0.15 for all dioxin homelogue
No dioxin in bottom ash
47 145 417 781 1861
436 504 668
27 77 149
18 50 142
8 17 38
29 95 149
0.2 0.8 4
ND ND-0.5 ND-1.7 0.1-0.5 0.2-0.4
ND-0.8 ND-4.2 ND-10 0.3-11 0.1-10
67.5 NR 350 1040 650
6.1 NR NR NR 160
+ + + + +
ND-1.5 ND-2.7 ND-26 2-87 6.5-330
Remark Reference
ESP control Nottrodt, 1986,
wet Scrubber EQV
Slag Nottrodt, 1986
No control Hay, 1986
Starved Air
Hay, 1986
Scheidl et al., 1985
Tong & Karasek, 1986
Chimney Ash Clement et al., 1985
Bottom Ash Clement et al., 1985
Lamparski, 1980 EQV
Cavallaro, 1980
Olie et al., 1977
* Tosine et al., 1985
fContinued }
216
-------
TABLE 6-11. (CONTINUED)
DIOXINS IN FLY ASH OF COMBUSTION DEVICES
Concentration (ng/g)
Source
Category
HWI
MWI
MWI
MWI
MWI
2,3,7,8-
TCDD
ND-110
0.4
0.065
0.14
100
Total
TCDD PCDD
3300-12000 NR
7.7 NR
1.12 NR
2.74 NR
NR NR
H6CDD
1300-5600
14
NR
NR
NR
H7CDD OCDD Remark Reference
2000-37000 3000-59000 Rotary Kiln ADL 1981, Bumb,
1980, Crummett, 1982
28 30 "
NR 35.5 Cavallaro, 198
NR 0.6
NR NR Single Test Ahling, 1982
Pyrolysis
HWI
MWI
MWI
MWI
MWI
1.2-2.5 2-2
NR
NR
NR
NR
Dioxin Work
Group, 1981
5.2 273.8 555
608.8 169.4
1.4 91.5 346.7 334
192.6
0.74 23.6
0.1
2.9
46.3
8.3
27.8 11.7
9.3 5.9
37.3 NW Unite MRI, 1985
Fly Ash
174 EC Units MRI, 1985
Fly Ash
10.7 NW Units MRI, 1985
Bottom Ash
8 EC Units MRI, 1985
Bottom Ash
MWI: Municipal Waste Incinerator
WBS: Wood Burning Stove
HWI: Hazardous Waste Incinerator
+ : Present
ND : Not Detected
FA : Fly Ash
* : This is an analysis of municipal waste, not fly ash
NR: Not Reported
EQV: Data used in calculating the weight percent adjusted to equivalent 2,3,7,8-TCDD
Bottom Ash: Operationally combined fly ash and bottom ash from MWI.
217
-------
TABLE 6-12. EMISSIONS OF CDDS FROM COMBUSTION
DEVICES OTHER THAN MWI
Emission Factor (/ig/kg)
Source 2,3,7,8- Total
Category TCDD TCDD
Boiler NR 0.1
Boiler NR 0.13
245T NR <100-3000
IWI <0.0006
IB (coal) 0.2 0.69
IB(RDF) 0.11 0.36
IB(coal) 2 5.8
IB(RDF) 1.3 3.7
HWI 1.
UB(coal) NR ND
PCDD HgCDD H7CDD OCDD Remark Reference
0.69 0.37 0.5 0.07 RDF -only Hahn, et al., 1986
1.28 1.28 1.06 0.13 RDF-NG Hahn, et al., 1986
<300-1200 <200-2000 <200-300 <200-1400 * Ahling, 1977
<0.0006 <0.0008 <0.0006 <0.0006 Brenner et al., 1986
NR NR NR NR U.S. EPA, 1987b
NR NR NR NR U.S. EPA, 1987b
NR NR NR NR U.S. EPA, 1987b
NR NR NR NR U.S. EPA, 1987b
ND ND ND ND U.S.EPA, 1986
RDF: Refuse-Derived Fuel, 245T: Combustion of 2,4,5-T Formulation
NG: Natural Gas
NR: Not Reported
* : emission factor in /ig/kg of 2,4,5-T formulation combusted
IWI: Industrial Waste Incineration,
HWI: Hazardous Waste Incinerator
IB: Industrial Boiler
UB: Utility Boiler
ND: Not detected
218
-------
grains/dscf, then the conversion of the concentration data to an emission factor, EF, in
/ig/kg can be done as follows:
(% of CO2VStack gas flow rate in dscfm)
EF(/ig/kg) = (gr/dscf) 1.43xlO~9 Refuse process rate ton/hr (6-30)
The concentration and emission factors can be expressed in many different units.
In this case, appropriate conversions were necessary to present the data on a consistent
basis.
In the absence of information on the relation between the flue gas generation rate
and process weight rate, an average value of 5000 Nm^ flue gas per ton of refuse burned
was used, since existing data from European and American test results show that the gas
generation rate ranges from 4000 Nm^ to 6000 Nm^ per ton of refuse burned. The
emission factor for 2,3,7,8-TCDD ranges from 0.002 - 0.289 /*g/kg. Emission factors for
other CDD compounds, as well as available information on use of particulate control
devices are shown in Table 6-10.
Data concerning the analysis of 2,3,7,8-TCDD and other CDD compounds in fly ash
collected by air pollution control devices are tabulated in Table 6-11. Fly ash from
municipal incinerators and other combustion devices is also included in the table.
For municipal waste incinerators, the levels of 2,3,7,8-TCDD in fly ash emissions
collected in control equipment range from 0.065 - 5.2 ng/g with an average value of 1.3
ng/g. The Swedish experiment (Ahling and Lindskog, 1982) showing the 2,3,7,8-TCDD
level of 100 ng/g was excluded in the averaging process, since it is an unexplained
outlier.
Table 6-12 summarizes emission factors for combustion devices other than municipal
waste incinerators. Very few data are available for 2,3,7,8-TCDD emissions from other
combustion devices. Since test results are presented in the literature on a weight or
volume basis, it is difficult to convert the literature information to an emission factor.
219
-------
Additional data are available for CDD emissions from modular starved-air and RDF-fired
municipal waste facilities (MRI, 1987) and from incineration devices other than municipal
waste combustors (U.S. EPA, 1986).
There are virtually no data concerning the division of 2,3,7,8-TCDD emissions
between the vapor and particulate phases for stack emissions in the U.S.A. There are
three European studies analyzing the vapor and particulate phase emissions from a
municipal waste incinerator (Hagenmaier, et al., 1985; Nielsen, et al., 1986; Scheidl, et al.,
1985). Nielsen et al., (1986) states that the reported distribution of CDDs and CDFs
between particulate and vapor phases in flue gases varies widely, but it is generally
recognized that an average of 20-30% is in the particulate phase, while 70-80% is in the
gas phase. For our scenarios we assumed 63% to be in the vapor phase and 37%
particulate, since these values are reported in two studies (Hagenmaier et al., 1985 and
Scheidl et al., 1985).
In scenarios 16 and 18 the refuse rate of 3000 TPD corresponds to 2.7 x 106 kg/d.
Since downtime will be required for maintenance and repair, the incineration rate
represents an average value over the time period of consideration. At this average rate
of refuse combustion, the emission rate of 2,3,7,8-TCDD is 0.001 ug/kg (2.7 x 10^ kg/d) =
2700 A*g/d, or 3 x 10~° g/s [see Table 6-8]. This emission is assumed to consist of about
37% particulate form and 63% vapor form (Hagenmaier et al., 1985; Scheidl et al., 1985;
Nielsen et al., 1986). Vapor emissions of 2,3,7,8-TCDD are then 1.9 x 10~8 g/s and the
particulate emissions are 1.1 x 10~° g/s. Similar calculations can be performed for the
120 TPD plant in scenarios 17 and 19. Since emissions are assumed to be discharged into
the atmosphere from two stacks, each stack emits a total amount of 2,3,7,8-TCDD at a
rate of 1.8 x 10~7 g/s as shown previously [see Table 6-8]. The emission from each
stack will consist of 1.1 x 10~7 g/s of 2,3,7,8-TCDD as vapor and 0.7 x 10~7 g/s of
2,3,7,8-TCDD as particulate.
220
-------
b. Municipal Waste Incinerator Capacities in the U. S.
There are currently a total of 45 mass burn facilities, 56 modular incinerators, and
10 RDF incinerators in the U.S., with a combined incineration capacity of about 49,000
tons per day (TPD). The mass burn facilities make up about 68 percent of the total
incineration capacity at present, with modular incinerators making up about nine percent
and RDF facilities 23 percent. There are a few existing mass burn incinerators with a
capacity as high as 3000 TPD. Typical air emissions control is electrostatic precipitation
which removes primarily particulate matter (Radian Corporation, 1986).
There are 210 proposed, planned, or partly constructed municipal waste incinerators
with a total combined capacity of about 193,000 TPD (Radian Corporation, 1986). Of
these, 118 are mass burn facilities, 24 are modular incinerators, and 31 are RDF
facilities. The planned mass burn facilities range in capacity from 100 TPD to several
large incinerators with capacities in the neighborhood of 3000 TPD. Most plan either
ESPs or baghouses to control particulate emissions (Radian Corporation, 1986), with many
of the planned facilities including dry scrubbers to reduce emissions of acid gases
according to state and local air pollution regulations.
Mass burn incinerators are being designed and built with and without energy
recovery. Energy recovery involves producing steam which in turn is used to produce
energy. Steam production at a controled rate is an important consideration for system
operation in planned facilities. Modular units are starved-air, two-stage systems where
waste is first in contact with lean air followed by completion of combustion. Although
the capacity of a single modular unit may be limited to less than 100 TPD, a facility can
process waste in excess 100 TPD by operating two or more of these units simultaneously.
Energy recovery steam plants require cooling towers, where steam plumes formed by large
amounts of water vapor evaporated from these towers may interfere with the dispersion
221
-------
characteristics of the incinerator plumes. It would be worthwhile to consider this effect
on dispersion characteristics of incineration emission.
The scenarios in this chapter use two mass burn facilities, one large one (3000 TPD)
and a smaller one (120 TPD).
c. Air Dispersion Modeling
The estimation of ambient air concentrations resulting from the 2,3,7,8-TCDD
emissions requires dispersion modeling. Dispersion modeling involves the elevation of the
virtual origin obtained by adding the plume rise (Ah) to the actual stack height (hs).
The effective stack height (H) is then H = Ah + hs. There are numerous methods for
calculating Ah. It is not unusual to get the spread in the answers for Ah using
different equations. The Holland and Briggs equations take the forms of (Wark and
Warner, 1981):
h = (Vsd/u)[1.5 + 0.0096 (Qh/Vsd)J (Holland) (6-31)
JJUKQQ1/3
h = u (Briggs) (6-32)
where Vs = stack gas speed, d = stack diameter, Qj, = heat emission rate associated with
stack gas, u = wind speed,
K = 1.58 - 41.4 (dtf/dz) (6-33)
where dd/dz is the potential temperature gradient,
F = g Vs d2(Ts - Ta)/4 Ta (6-34)
222
-------
where Ts, Ta = stack and ambient air temperatures, respectively. To apply the Briggs
equation for Ah, the stability class must be known. For neutral stability or d0/dz = 0,
the Briggs equation provides the plume rise Ah = 242 m for the 3000 TPD facility
scenarios. Hence the effective stack height can be H = 46 + 242 = 288 m. Such an
effective height is applicable to a flat terrain with no precipitation assumed. For all the
scenarios under consideration, however, the heights of the buildings are as tall as the
stack heights, which triggers the building downwash effects in the application of the
Industrial Source Complex (ISC) Model. An option available in an advanced dispersion
model is the capability to simulate precipitation along with building downwash caused by
a short stack height (EPA, 1986e). This feature is not available in the current ISC. The
results of sample dispersion modeling for the scenarios considered can be found in U.S.
EPA (1986e). The deposition rate of particulate matter and the ambient air concentration
of various pollutants emitted at a given rate are given as a function of distance and the
sector of the area.
The particle distribution is needed to estimate the deposition rate of particulate
matter, because the settling velocity is a function of particle size and density. The
particle size distribution assumed for air dispersion modeling which incorporates the dry
and wet deposition is 6.7% for particles greater than 10 urn in diameter, 26.7% for
particles between 2 and 10 /*m in diameter, and 66.6% for particles less than 2 nm in
diameter.
Table 6-13 is an output of air dispersion modeling for particulate emissions of
2,3,7,8-TCDD from the 120 TPD municipal waste incinerator. The total emission rate of
2,3,7,8-TCDD (from the two stacks combined) used in the model estimation [see Table
6-8] is 1.3 x 10~7 g/s. Several case studies were made to compare the results of the ISC
model having the option of precipitation and building downwash, with the short-term ISC
223
-------
TABLE 6-13. AIR DISPERSION MODELING OF
PARTICULATE-FORM 2,3,7,8-TCDD EMISSIONS
Case No.
1
2
3
4
5
Maximum
deposition rate
Oig/m2.yr)
0.048
0.136
0.168
0.154
0.13
Distance of
maximum
deposition
(m)
200
200
200
200
200
Angle of
maximum case no.
Deposition
225°
360°
280°
180°
180°
(ISCST) model without the options (U.S. EPA, 1986e). The model output results are
shown for the deposition rate averaged over the days when precipitation occurs (Case 1);
for the dry settling case in which the yearly deposition calculated with the wet
deposition available in the program removed (Case 2); for the deposition rate with
precipitation and dry settling occurring according to the local meteorological conditions
[assumed to be southeastern Virginia] (Case 3); for the deposition rate obtained from the
short-term ISC model averaged on an annual basis in which no precipitation is accounted
for (Case 4); and for the dry deposition rate with wet deposition removed using the stack
tipdown wash option, but at the same location where the ISC model calculates in Case 4
(Case 5). This analysis shows that the maximum deposition occurs when both wet and
dry deposition are considered according to local meteorological conditions (Case 3).
Thus, Case 3 was used to estimate dispersion for both the 120 TPD and 3000 TPD
incinerators. The actual dispersion modeling utilized site-specific meteorological data to
estimate the deposition rate and the ambient air concentrations (U.S. EPA, 1986e). The
224
-------
details of the methodology used in treating the wet and dry deposition phenomena can be
found in the literature.
The deposition rates for particulate-form 2,3,7,8-TCDD and the ambient air
concentrations of vapor-form and particulate form 2,3,7,8-TCDD in the surrounding area
were estimated from the air dispersion modeling for the distance where the maximum
deposition and concentration are likely to occur, for 0.8 km where the populations in the
scenarios are assumed to be, and for 100 km which is considered relatively removed from
the facility. The results of computation corresponding to conditions with precipitation
and dry deposition affected by the local meteorological conditions are shown in Table
6-14. No photodegradation or desorption from fly ash are assumed for the emissions in
traveling from the stack exit to the receptor location. This assumption is consistent
with recent experimental observations by SRI (Mill et al. 1987) under contract to EPA.
The proportion of vapor-form and particulate-form 2,3,7,8-TCDD available for dispersion
in the ambient air is assumed to be the same as reported for the stack emissions.
2,3,7,8-TCDD vapor may condense or adsorb onto particulates as it cools during transport
through the air, increasing the fraction in the particulate phase. This is not expected to
happen, however, since experimental data by Eitzer and Kites (1986) showed that all of
2,3,7,8-TCDD in the ambient air as being present in the vapor-phase fraction collected in
polyurethane foam installed in the back end of the collection train.
The collection train used by Eitzer and Hites consisted of high volume filter paper
with a pore size of 0.1 /«n, followed by polyurethane foam. Some speculate that
2,3,7,8-TCDD collected on the filter paper may have been desorbed from particulates by
the high volume of air passing through and subsequently absorbed in the polyurethane
adsorbent as vapor. Others argue that, if this is possible, desorption would have
occurred in the ambient air where a large volume of air surrounds the particulates. Still
others argue that, since equilibrium is already established between particulate and vapor
225
-------
form 2,3,7,8-TCDD in the ambient air, further .desorption by the same flowing air in
.equilibrium would be minimal, and hence most of .2,3,7,8-TCDD is in the vapor phase.
Tjhe .apparently high fraction of vapor-phase 2,3,7,8-TCDD in the ambient air could
•be due to .the large volume of air compared to available paniculate surface area. As an
example, an ambient air concentration of particulate matter at 100 A*8/m^ represents a
paniculate to air volume ratio of .approximately 1/10 . An assumption of monolayer
.adsorption of 2,3,7,8-TCDD vapors on the particulate will further increase this volume
ratio. At a partition coefficient of 10^ for distribution between particulates and air, the
amount of 2,3,7,8-TCDD present in the vapor phase is larger by a factor of 104. Hence,
this calculation shows that the percentage of 2,3,7,8-TCDD on particulates at equilibrium
with a volume of ambient air will be negligible.
The ratio of particulate phase to vapor phase 2,3,7,8-TCDD may also be affected toy
losses during transport from the incinerator stack. For example, particulate will be lost
from the plume due to deposition, which would cause changes in the particulate/vapor
ratio.
Due to the uncertainty surrounding this issue, it was decided to ;assume that the
vapor/particulate ratio remains the same during transport from the stack and highlight
.this issue for future analysis.
The effect of photolysis on the amount of 2,3,7,8-TCDD vapor photodegraded in
traveling from contaminated site to receptor population would be minimal: For Scenarios
12 through 15 where it is assumed that the receptor population is located 500 feet away
from the .site, the travel time of volatilized vapor is about 0.011 hours at an average
wind speed of 10 mph. The amount of 2,3,7,8-TCDD vapor degraded in the traveling this
distance is about 0.1% of the original amount at a half-life of 6 hours. For scenarios 8
through 11 with the population assumed to be 100 feet away, the fraction of 2,3,7,8-
226
-------
TCDD photolyzed would be smaller. For these scenarios, the effect of vapor-phase
photolysis on the result of exposure estimation may be neglected in a practical sense.
d. Surface Water Contamination
Two cases were evaluated for the extent of contamination of surface water by the
emissions from the incinerator. In the first case, used in scenarios 16 and 17, a
one-acre pond seven meters deep was considered with a negligible amount of inflow and
outflow water compared to the amount of water in the pond. In the second case, used
in scenarios 18 and 19, the impact on a 300 meter wide, five meter deep river flowing at
the rate of 0.5 m/s, was considered. Both water bodies are located at 0.8 km from the
incinerator. This river is much larger than the one assumed in the land-related scenarios
described in Sections A and B of this chapter. This is because incinerators are more
likely to be located in urban areas where large rivers are common.
The 2,3,7,8-TCDD emissions that reach the water bodies upon dispersion in the
atmosphere will be partly particulates and partly vapors. The characteristics of stack
emissions are assumed to remain unchanged when the emissions reach the water body, so
the ambient air concentrations above the surface of the water bodies are as shown in
Table 6-14. These concentrations assume no photodegradation in the atmosphere in the
process of pollutant transport from the stack(s) to the surface of water bodies.
A photolysis half-life of 6 hours (Mill et al, 1987) (k = 0.12 hr'1), has been
estimated for the vapor phase 2,3,7,8-TCDD and experimentally supported under simulated
sunlight conditions (Mill et al, 1987). Assuming an average wind speed of 10 miles per
hour, the travel time to the water body is 0.05 hr, very short compared to the 6 hour
half-life. The half-life for 2,3,7,8-TCDD in particulate form is much longer, (Mill et al,
1987). Hence, no correction for degradation will be applied to either the vapor-form or
particulate-form 2,3,7,8-TCDD affecting the water body.
227
-------
TABLE 6-14. SUMMARY OF AMBIENT AIR CONCENTRATION
AND DEPOSITION RATE IN THE VICINITY OF THE
FACILITIES IN SCENARIOS 16-19
Max. Cone.
at 200 m
(*/m3
Max. Deposition
Rate
(ug/m2.yr)
Cone, at
0.8 Km
(*/m3)
Deposition Cone, at
Rate at 0.8 Km 100 Km
(ug/m2.yr) (g/m3)
Deposition
Rate at
100 Km
(ug/m2.yr)
Vapor Form 2.1xlO'16
2,3,7,8-TCDD
(3000 TPD)
Participate- 1.2x10"16 3.7xlO"3
Form 2,3,7,8-
(3000 TPD)
Vapor Form l.SxlO"13
2,3,7,8-TCDD
(120 TPD)
Particulate- l.lxlO'13 0.168
Form 2,3,7,8-
TCDD
(120 TPD)
1x10
-15
7.1x10
-18
6.5xlO'16 6.3xlO~4 4.1xlO"18 2.2xlO'7
8.3x10
,-14
9.1x10
-16
4.9x10'14 0.028
6.2xlO'16 2.7xlO"B
The processes responsible for transferring 2,3,7,8-TCDD in the atmosphere into
the water body are dry deposition, wet deposition and vapor absorption. The first two
processes are responsible for the removal of particulates, and the last process
removes the vapor-form 2,3,7,8-TCDD from the atmosphere. U.S. EPA, 198la) provides
the ranges of dry deposition rates for different particle sizes which will allow
estimation of the concentration of 2,3,7,8-TCDD in the water body caused by the dry
deposition. The two-resistance mass transfer theory was used to estimate the
concentration of 2,3,7,8-TCDD in the water body caused by the absorption of vapors
into water.
The deposition rate of particulate-form 2,3,7,8-TCDD will depend upon the
characteristics of the particles, the particle size distribution, and the rate of
precipitation. For particles with a true density of 1.5 g/cm^, the settling velocities in
228
-------
the air for the size fractions comparable to the particulate emissions being considered is
given in Table 6-15.
TABLE 6-15. COMPARISON OF PARTICLE SIZE AND SETTLING VELOCITY
Particle Size (um) Settling Velocity (cm/s)
0.8 0.025
6 0.25
16 0.7
Source: Wark and Warner, 1981.
The dry deposition velocities rates are given by Turner 1970. Deposition velocities
are minimal for particles with radii in the range 0.1 - 1.0 /zm and increase for both
larger and smaller particles. For particles having radii in the range of 2 - 10 /«n, the
deposition velocities are 1-5 cm/s; for particles having radii less than 0.1 pm, the
velocities are 0.02 - 0.8 cm/s; for particles in the range of 0.1 - 1.0 /mi, the velocities
are 0.1-1 cm/s. The wet deposition velocity cited by the authors is 70 - 100 cm/yr.
Particle size distribution is needed to apply these deposition velocities for various size
ranges. Holton (1985) used a deposition velocity of 0.005 m/s in his risk calculations for
PCB emissions from hazardous waste incinerators. The deposition velocity of
particulate-form 2,3,7,8-TCDD is also estimated as 0.025 cm/s from the results of the ISC
modeling incorporating deposition due to dry settling and wet deposition.
(1) Vapor Absorption
The absorption of vapor phase 2,3,7,8-TCDD will be caused by the concentration
driving force between air and water phases. For the pond scenarios (16 and 17) an
approach to equilibrium is possible between the air and water phases after a long contact
time. From Henry's law constant for 2,3,7,8-TCDD, which is 1.6 x 10~5 atm-m^/mol, the
equivalent water-phase concentration in equilibrium with the air phase, Cw*, is 1.0 x
229
-------
10~15 g/m3)/(1.6 x 10'5 x 41) = 1.5 x 10'12 g/m3,. or 1.5 x 10~18 g/cm3 in the 3000 TPD
case,, and. 1.3 x 10"^ g/cm3 in the 120 TPD case. Shifting winds and unsteady
incinerator feed conditions will cause the 2,3,7,8-TCDD concentration in air over the
pond to be variable which, may prevent the achievement of equilibrium. The site-specific
wind rose frequency used in the dispersion modeling available in ISC will reflect this
phenomenon- to some extent. Hence, the maximum water-phase concentrations achievable
in. the pond, resulting from absorption from the vapor-phase 2,3,7,8-TCDD would be 1.5 x
1.0"I8 g/cm3 and 1.3 x 10~16 g/cm3 in the 3000 and 120 TPD cases, respectively. The
rigorous evaluation of the concentration in the pond water will require consideration of
simultaneous absorption of vapors into water and disappearance from the water phase by
adsorption on sediments occurring under changing meteorological conditions.
For the river scenario, the vapor emissions coming in contact with the river water
(scenarios 18 and 19) will be absorbed into the water to the extent that mass transfer
rates allow. This absorption, rate, Q (in g/s) into the water body can be calculated
from the two-resistance mass transfer theory which takes the form:
(Cw* - Gw) A (6-35)
where Cw = concentration in water, (g/cm3), Cw = water equivalent concentration of air
phase concentration, (g/cm3), KOL = overall mass transfer coefficient expressed in the
liquid phase unit, (cm/s), and A = surface area where absorption occurs, (cm2). The
mass transfer coefficient is assumed to be KOL = 2 x 10~4 cm/s (Thibodeaux, 1979). The
conservative assumption was made that Cw is negligible relative to Cw* due to the
dilution provided by the river flow. The Cw* values were identical to those calculated
above for the pond scenarios.
At the plume location where the maximum concentration occurs, the diameter of the
230
-------
plume is about 120 m at the surface of the river. Hence, the water surface area
available for absorption is 120 m x 300 m = 3.6 x 10** m . Hence
Q = (2 x 10'4 cm/s)(1.5 x 10"18 g/cm3)(3.6 x 108 cm2) (6-36)
= 1.1 x 10~13 g/s for the 3000 TPD case
Q = 9.4 x 10~12 g/s for the 120 TPD case (6-37)
The water flow rate in the river is 5 m x 300 m x 0.5 m/s = 750 m3/s. The vapor
absorption rate and the water flow rate can be used to calculate the contaminant
concentration as 1.5 x 10~13 /ig/L for the 3000 TPD facility and as 1.3 x 10"11 A»g/L for
the 120 TPD facility.
(2) Particulate Deposition
Particulates that entered the water body will slowly settle to its bottom, and will
remain as sediment. The change in the particle settling velocity as the particles enter
the water from the air depends upon the buoyancy effect and the viscosity of the
medium. For the pond scenarios (16 and 17), while particles are in suspension in water,
the amount of 2,3,7,8-TCDD on the particles should be accounted for as part of the
contaminant concentration in water. For the deposition velocity of 0.00025 m/s in the
air, the settling velocity in the water will be changed to
0.00025 m/s[(2.5 g/cm3 - 1 g/cm3)/0.95 cp/(2.5 g/cm3 - 0 g/cm3)/0.018 cp]
= 2.84 x 10"6 m/s (6-38)
where the densities of particles, water and air are assumed to be 2.5 g/cm3, 1 g/cm3,
and negligible (0) relative to particle density, and the viscosities of water and air media
231
-------
are 0.95 and 0.018 cp, respectively. Particles will remain suspended in the water for an-
average of 28 days:
7m/(2.8.4 x 10'6 m/s x 3600 s/hr x 24 hr/d) = 28.5 d (6-39)
In the case of the 3000 TPD incinerator the concentration of 2,3,7,8-TCDD in the
pond due to dry and wet deposition of particulate matter can be estimated as
(6.5 x 10T16 g/m3) (0.00025 m/s)(28.5 d x 24 hr/d x 3600 s/hr)/7m (6-40)
= 5.7 x. 10-14 g/m3
At 0.8 km distance, the lateral standard deviation is approximately ay = 60 m. Hence.
the diameter of the plume where deposition occurs is approximately two standard
deviations and is 120 m with an area of 1.1 x 104 m2. This area is larger than the
surface area of the pond (4047 m2). Hence no correction for further dilution due to a
smaller mass transfer area than the pond surface is required. A similar calculation can
be made for particulate form of 2,3,7,8-TCDD emissions from the 120 TPD MWI which
results in a calculated concentration in the pond due to the particulate-form emissions of
4.3 x 10'12 g/m3.
For the river scenarios (18 and 19), the settling of particulate-form 2,3,7,8-TCDD
will again occur at the rate of 0.025 cm/s at the concentrations of 6.5 x lO"1** g/m3
[3000 TPD] and 4.9 x 10~14 g/m3 [120 TPD]. The settling velocity is a function of
particle size and density which should be determined experimentally. Since this
information is site-specific, a representative settling velocity back-calculated from the
output of the ISC model incorporating the dry and wet deposition is used. Settling on
an area of 3.6 x 104 m2 is equivalent to a particle deposition rate of
0.00025 m/s (6.5 x 10'16 g/m3) (3.6 x 104 m2) = 5.9 x 10'15 g/s (6-41)
232
-------
in the 3000 TPD case, and 4.4 x 10~13 g/s in the 120 TPD case. This results in the
contaminant concentration in water as 7.9 x 10"^ A»g/L for the 3000 TPD case:
5.9 x 10-15 g/s x 103 Oig/L/g/m3)/750 m3/s <• 7.9 x 10'15 pg/L (6-42)
Similarly, the concentration in water for the 120 TPD case can be estimated as 5.9 x
10"13 /ig/L.
The total concentrations of 2,3,7,8-TCDD in the river water due to absorption of
vapor-phase emissions and deposition of particulate emissions into the water body become
(for the 3000 TPD case):
1.5 x ID'13/ig/L + 7.9x 10'15/ig/L = 1.6 x ID'13 Mg/L (6-43)
Similarly, the river water concentration for the 120 TPD case becomes 1.4 x 10"" Mg/L.
Similarly, the total concentration of 2,3,7,8-TCDD in pond water due to absorption of
vapor phase emissions and deposition of particulates is 1.6 x 10"' /ig/L and 1.3 x 10"'
fjtg/L for the 120 TPD and 3000 TPD facilities, respectively.
The above discussion indicates that the vapor emissions contribute considerably more
to the contamination levels in water than do the particulate emissions. The result is
surprising because intuitively (and as supported by our calculation) the rate at which the
2,3,7,8-TCDD enters the water is greater via the particulate route than the vapor route.
However, the particulates settle out relatively quickly and the vapor is assumed to remain
dissolved in the water. Under these assumptions, the vapors contribute more to the total
amount either dissolved or suspended in the water than do the particulates. Calculations
show that the percentage of 2,3,7,8-TCDD concentrations in water attributable to
particulates is about 3 to 5%. In reality much of the initially dissolved 2,3,7,8-TCDD
derived from vapors in the air could be absorbed into the sediment. These losses were
233
-------
not accounted for and could substantially reduce estimates of water concentration and
resulting risks from either fish ingestion or water consumption. However, a preliminary
calculation for transport of the contaminant from water to sediment shows that this
transport rate is rather slow due to the diffusional resistance in the sediment layer,
thereby favoring an equilibrium situation between vapor and water phases. Future
analysis should attempt to address this issue more quantitatively.
e. Land-Disposed Ash
The incinerator generates both bottom ash and fly ash which can be ultimately
land-disposed. The bottom ash, as noted earlier, will not be considered further, since no
data are available indicating PCDD content. One affect of the mixing is to reduce the
average concentration of 2,3,7,8-TCDD initially present in fly ash. In actual incinerator
facilities, the fly ash and bottom ash are often mixed before disposal. If bottom ash has
negligible amounts of 2,3,7,8-TCDD in it, this would result in a dilution of the
concentration of 2,3,7,8-TCDD, with fairly straightforward effects on the predicted risks.
This effect can be noted in the incinerator fly ash and bottom ash reported for the
incinerator in Philadelphia (MRI, 1985). The bottom ash reported in Table 6-11 for this
incinerator is actually a mixture of fly ash and bottom ash being mixed operationally at
the facility. Although the analysis of fly ash is given as reported in Table 6-11, the
analysis of 2,3,7,8-TCDD in bottom ash (not the mixture) is not separately given. It is
also assumed that the fly ash produced from the incinerator contains the same amount of
2,3,7,8-TCDD at the time it is disposed of in a landfill as that being collected in control
equipment
The uncontrolled emission factor for emissions of particulate matter from a
v
municipal waste incinerator is 30 Ib/T as given in U.S. EPA, 1985. At a particulate
emission control efficiency of 99%, the emission factor after control is 0.3 Ib/T. For a
3000 TPD capacity incinerator, the daily amount of fly ash collected for disposal is 3000
234
-------
TPD x (30 - 0.3) Ib/T = 89,100 Ib per day. For the 120 TPD capacity the amount of fly
ash collected is 3,564 pounds per day.
The concentration of 2,3,7,8-TCDD in fly ash reported in the literature varies from
0.07 - 100 ng/g as reported previously [see Table 6-11]. These data were mostly taken
from municipal incinerators in European countries. The cause of the variation can not
be identified based on the data in the literature. Except for a Swedish study (Ahling
and Lindskog, 1982), the literature suggests that the concentration range of 2,3,7,8-TCDD
in fly ash is 0.07 to 5.2 ng/g. An average value of 1.3 ng/g excluding the Swedish study
was used for this. Only fly ash is used in the averaging process.
As can be noted in Table 6-11, data on the amount of CDD isomers in ash are
limited. Also, it is difficult to compare the effects of analytical efficiencies on the
results regarding the amount of 2,3,7,8-TCDD in fly ash. Since the CDD components
including 2,3,7,8-TCDD are not easily extractable, MRI used a 16 hour benzene extraction
to remove the CDDs from the fly ash (MRI, 1985). One plausible explanation for lower
values of 2,3,7,8-TCDD in fly ash reported in European Studies may be lower extraction
efficiencies resulting from different extraction procedures including a shorter extraction
time. Since all literature do not report the exact procedure investigators used in
v
extracting CDDs, it is difficult to make a generalized conclusion.
The exposure scenarios associated with the land disposal of 2,3,7,8-TCDD were
discussed in Sections A and B of this chapter. Where reasonable the same procedures
were applied to incinerator ash; however, some modifications were required. Since ash is
produced on a continuous basis, the quantity disposed in a landfill will increase over
time. As a result, the exposure levels associated with ash disposal will also increase
over time. In actual incinerator facilities, the fly ash and bottom ash are often mixed
before disposal. If bottom ash has negligible amounts of 2,3,7,8-TCDD in it, this would
result in a dilution of the concentration of 2,3,7,8-TCDD, with fairly straightforward
235
-------
effects on the predicted risks. Ideally, the releases and associated exposures would be
computed, on a daily basis and summed over the exposure period to account for unsteady
exposure, rates. In order to simplify the calculations, it was assumed that the ash was
generated at a constant rate over 70 years, so as an average ash amount, it would be
half of that generated in 70 years, or the amount generated, in 35 years4 Assuming that
the ash is landfilled at. a depth of 10 feet with a bulk density of 0.7 g/cm^, landfills of
2.6 acres and 66 acres would be created from the 120 TPD and 3000 TPD incinerators
respectively. Accordingly, the landfill scenarios 9, 12, and 15 were adopted for ash
disposal with sizes changed to 2.6 and 66 acres (for the 35 year volume for the 120 TPD
and 3000 TPD incinerators respectively) and contamination level changed to 1.3 ppb. In
all other respects the ash disposal scenarios are identical to scenarios 9, 12, and 15
described in Sections A and B of this chapter. These three scenarios were selected
because they represent a range of conditions from an uncontrolled access landfill with
reasonable worst-case population characteristics (scenario 9) to an uncontrolled access
landfill with more typical population characteristics (scenario 12), to a controlled access
landfill with cap (scenarios 15). This approach resulted in five fly ash disposal scenarios
defined as follows:
(1) Scenario 16 is an uncontrolled access landfill receiving ash from the 3000 TPD
incinerator. The site conditions and population habits are considered
reasonable worst case as defined earlier for scenario 9.
(2) Scenario 17 is an uncontrolled access landfill receiving ash from the 120 TPD
incinerator. The site conditions and population habits are considered
reasonable worst case as defined earlier for scenario 9.
(3) Scenario 18 is an uncontrolled access landfill receiving ash from the 3000 TPD
incinerator. The site conditions and population habits are considered typical
as defined earlier for scenario 12.
236
-------
(4) Scenario 19 is an uncontrolled access landfill receiving ash from the 120 TPD
incinerator. The site conditions and population habits are considered typical
as defined earlier for scenario 12.
(5) Scenario 20 is a controlled access, capped landfill receiving ash from the 3000
TPD incinerator. The site conditions and population habits are considered
reasonable worst case as defined earlier for scenario 15.
Using the procedures described in Sections A and B of this chapter the ash disposal
exposures were calculated (see Table 6-16). Exposures in Table 6-16 are based on
contact rather than absorbed doses.
The particulate emissions included those generated during unloading of fly ash at
the disposal site, the spreading of the material unloaded, and transportation of fly ash in
trucks, and particulates airborne due to winds at the disposal area. The materials
handling of ash (unloading, transporting by trucks, spreading, etc.) will cause greater
emissions than would be caused by wind alone. Air dispersion modeling along with the
methods described in Section C of Chapter 6 is used to estimate the ambient air
concentrations from each activity and wind dust. The total particulate phase
concentrations are used to estimate the exposures shown in Table 6-16.
Incinerator ash could also lead to human exposure when it is blown from trucks
transporting ash from the incinerator to the landfill. Such fugitive emissions would be
diluted over a long distance which would reduce potential exposures considerably
(compared to a stationary point source). Future analysis should attempt to estimate such
emissions and resulting exposure more definitely.
f. Soil Contamination
Soil can be contaminated by the fallout of particulate phase emissions and
adsorption of the vapor phase 2,3,7,8-TCDD on soil. This adsorption will occur when the
vapor phase plume reaches the ground level and comes in contact with soil. For a
237
-------
conservative estimate, the soil phase concentration associated with the maximum air
phase concentration was evaluated. The deposition rates are previously shown in Table
6-14. In all scenarios involving the 3000 and 120 TPD MWIs, the maximum deposition
rate occurs at a distance (200 m) very close to the incinerator because of the building
washdown effects coupled with precipitation washdown. The maximum deposition location
can be possibly closer than 200 m in reality. The distance of 200 m is a calculated
distance constrained by the limitation of the computational algorithm built in the ISC.
Since soil ingestion by children, and farming with a pond or stream in an area very close
to the incinerator or at proximity of buildings within 200 m is considered unlikely,
exposure from soil ingestion is evaluated at the site of the hypothetical family 0.8 km
from the stack. The difference in deposition rate between the maximum deposition rate
and the rate at 0.8 km from the stack is approximately a factor of 6.
It should be recalled that the particle settling velocity is dependent upon particle
size and density. The particle size distribution is needed for the spatial distribution of
the fallout. Over the period of paniculate fallout, the amount of 2,3,7,8-TCDD will build
up from zero to the amount accumulated up to the fallout time. For 70 year fallout
time, the total amount of the contaminant fallen out per unit area is 4.4 x 10~^2 g/cm2
in the 3000 TPD case, and 2 x 10~10 g/cm2 in the 120 TPD case. Hence, for situations
starting with zero and ending with these values, the average amounts of fallout over the
70-year period would be half, or 2.2 x 10~12 g/cm2 and 1 x lO"1^, respectively. For a
soil bulk density of 1.7 g/cm3, a soil surface density of at 1.7 g/cm2 will be used. In
essence, this procedure assumes that the particulate is mixed uniformly to a depth of
1 cm. The 2,3,7,8-TCDD concentration on the soil surface for the average exposure
period is
2.2 x 10-12/1.7 = 1.3 x 10~12 g/g, or 1.3 x 10'9 mg/g (6-44)
in the 3000 TPD case, and 5.9 x 10"8 mg/g in the 120 TPD case.
238
-------
This concentration will be decreased under the influence of desorption by
precipitation water, biodegradation, volatilization, and photolysis. It is possible that the
action of microorganisms is minimal on the soil surface, while the exact extent of
photodegradation on the soil surface is difficult to predict at this point. In these
scenarios, the disappearance rate constant provided by Young (1983) was used. Young's
rate constant, 0.069 yr~l, is derived from actual field conditions which include
photodegradation, biodegradation and sediment loss, and runoff. When the disappearance
follows the first order kinetics, the total amount of deposition per unit area, M,
remaining on soil at any time t can be expressed by:
M - F [l-e"kt]/k (6-45)
where F = constant deposition rate, and k = first order disappearance constant.
Hence, the concentrations in soil resulting from the deposition of particulate-form
2,3,7,8-TCDD given above should be corrected to account for gradual disappearance
using Equation 6-45. For the case of the 3000 TPD:
M = (6.3 x 10'14 g/cm2 . yr/0.069 yr'1 [1-e -0.069(70)]
= 9.1 x I-'13 g/cm2 (6-46)
Similarly, the mean deposition remaining in soil over the exposure period is 4 x
10" H g/cm2 in the 120 TPD case. As discussed at the beginning of this subsection,
incorporation of the deposited particles into soil to a depth of 1 cm will yield an average
concentration of 5.4 x 10" 1^ mg/g for the 3000 TPD site, and 2.4 x 10~& mg/g for the
120 TPD site.
239
-------
The vapor-phase 2,3,7,8-TCDD in contact with the soil will affect the contaminant
concentration on the soil surface. Under the assumption of an equilibrium between the
soil-air interface, the soil-air partition coefficient can be used to estimate the
contaminant concentration on soil in contact with the vapors. The soil-air partition
coefficient is:
Kas » 41H/Kd - 1.4 x 10~7 mg/cm3 air/mg/g soil (6-47)
Hence, for the 3000 TPD MWI case, the concentration of 2,3,7,8-TCDD on soil due
to the vapor-phase TCDD is:
Csm - 1 x 10~15.g/m3 x (103 mg/g)/ (1.4 x 10'7 g/cm3 x 106 cm3/m3)
= 7.1 x 10~12 mg/g (6-48)
It is likely that disappearance of the adsorbed contaminant will occur by the
process of photodegradation, biodegradation and erosion. The extent of disappearance
will be dependent upon the elapsed time after the vapor impacts the soil, and upon the
kinetic rate of the process. Hence, this concentration represents the maximum soil
concentration attainable under equilibrium with vapors. Similarly, the maximum soil
concentration in equilibrium with vapors for the 120 TPD site is 5.9 x lO"1^ mg/g.
In these scenarios, particulate deposition provides a concentration in soil larger
than or comparable to the concentration caused by vapor adsorption. As particulates fall
down and are mixed into soil, additional adsorption is possible from vapors approaching
the concentrations attainable by vapor adsorption on a short-term period. This would be
particularly true at the initial period of the incinerator operation, because the amount of
particulate fallout is not significant compared to contribution from vapors. If, on the
240
-------
other hand, the particulate fallout contributes to a higher contaminant concentration than
vapor adsorption, no further adsorption of vapors is possible. Hence, the resulting soil
concentration will correspond initially to the values determined by particulate fallout.
Particulate deposition contributes to soil contamination more than vapor adsorption over
a longer period of particulate deposition or the lifetime of an incinerator.
D. INCINERATOR EXPOSURE PATHWAY CALCULATIONS
Table 6-16 is a summary of the calculated exposures for the various incinerator
scenarios, broken out by exposure pathway. The assumptions used to obtain these
exposure estimates are discussed below and the parameter values are listed in Table 6-17.
1. Inhalation of Ambient Air
The downwind concentrations provided in Table 6-14 can be used to estimate
lifetime exposures associated with inhaling ambient air containing 2,3,7,8-TCDD in vapor
and particulate forms. The total exposures are based on the combined effects of vapors
and particulates.
Using the air concentration listed in Table 6-18 and parameter values listed in Table
6-17, the inhalation exposures at 0.8 km for scenarios 9-12 were calculated. These are
also included in Table 6-18.
For example, the exposure to the particulate-phase 2,3,7,8-TCDD for Scenario 16 can
be calculated from the given ambient air concentration and the exposure parameter values
in Table 6-17:
6.5 x KT10 (103)(23)(20,000/25,550)(1/70) - 1.7 x 1(T7 ng/kg-d (6-49)
For purposes of comparison with the inhalation exposures at 0.8 km ( as shown in
Table 6-18) the inhalation exposures at point of maximum ground level concentration
(200m) and at 100 km are shown in Table 6-19.
241
-------
TABLE 6-16. EXPOSURES ASSOCIATED
WITH INCINERATOR EXPOSURE PATHWAYS/SCENARIOS
A. EXPOSURES (ng/kg-day) ASSOCIATED WITH STACK EMISSIONS
Dairy Beef fish Soil Vapor Participate Soil Drinking Vegetable
Scenario Ingestion Ingeation Ingeation Ingeation Inhalation' Inhalation Dermal Water Ingeation
16)3000TPD JxlO'6 4xlO~B 7 x 10'7 2 x 10"6 6 x 1(T7 3 x 10~7 6 x 10*8 4 x 10"8
reasonable See text
worst case
17)120TPD 6xlO"4 2xlO~S 6xlO"5 8 x NT6 5 x 1
-------
TABLE 6-16. (CONTINUED)
B. EXPOSURES (ng/kg-day) ASSOCIATED WITH FLY ASH DISPOSAL
Dairy Beef Fish Soil Vapor Participate Soil Drinking Vegetable
Scenario* Ingestion Ingestion Ingestion Ingestion Inhalation Inhalation Dermal Water Ingestion
16)3000TPD IxlO"2 6xlO"2 2 x 10~5 3 x 10~3 5 x 10"6 3 x 10~7 1 x 10"2 2 x 10~6
reasonable See text
worst case
17) 120 TPD 2 x NT3 1 x 10"2 4 x 10"6 6 x 10~4 3 x 10"6 2 x 10"7 2 x NT3 3 x 10~6
reasonable See text
worst case
18) 3000 TPD 1 x 10~3 1 x 10"2 4 x 10'9 2 x 10'4 2 x 10'6 9 x 10'8 5 x 10'6 neg.b
abandoned See text
typical
19) 120 TPD 7 x 10"5 6 x 10"4 4 x 10"9 1 x 10"6 6 x 10"7 2 x 10-8 6 x 10"6 neg.
abandoned . See text
typical
20) 3000 TPD neg. neg. neg. neg. neg. neg. neg. neg.
reasonable neg.
worst case
monofill
w/capc
a All scenarios assmume O.Sppb 2,3,7,8-TCDD in flyash.
b Negligible exposure (<10~8).
c Stack emissions part of this scenario is the same as scenario 16.
243
-------
TABLE 6-17
PARAMETER VALUES FOR CALCULATING EXPOSURES
ASSOCIATED WITH INCINERATION
Inhalation
Respiration Rate (m-'/d)
Vapor Absorption Fraction
Participate Absorption Fraction
Exposure Duration (d)
Body Weight (kg)
Soil Ingestion
Ingestion Rate (g/d)
Absorption Fraction
Exposure Duration (d)
Body Weight (kg)
Dermal Contact
Contact Rate (g/d)
Absorption Fraction
Exposure Duration (d)
Body Weight (kg)
Drinking Water
Ingestion Rate (L/d)
Absorption Fraction
Exposure Duration (d)
Body Weight (kg)
Fish Ineestion
Ingestion Rate (g/d)
Absorption Fraction
Exposure Duration (d)
Body Weight (kg)
RWC
Scenarios
16,17,20
23
0.75
0.27
20,000
70
1
0.3
1,500
17
1
0.005
20,000
70
2
0.68
20,000
70
30
0.68
2,600
70
TYPICAL
Scenarios
18,19
23
0.75
0.27
7,300
70
0.2
0.3
910
17
1
0.005
7,300
70
2
0.68
7,300
70
6.5
0.68
1,500
70
(Continued )
244
-------
TABLE 6-17. (CONTINUED)
Beef/Dairy Products Ingestion
Beef fat ingestion (g/d) 26 14.9
Dairy products fat ingestion (g/d) 43 18.8
Absorption fraction 0.68 0.68
Beef exposure duration (d) 11,000 6,400
Dairy exposure duration (d) 10,000 5,800
Beef fat/soil distribution factor 0.4 0.3
Dairy fat/soil distribution factor 0.04 0.04
Effective Grazing Area (m2/d) 22 22
Weathering Rate Constant (1/d) 0.05 0.05
Transfer Coefficient (d/L) 0.009 0.009
245
-------
TABLE 6-18. AMBIENT AIR CONCENTRATIONS
AND EXPOSURES AT 0.8 KM
Paniculate Concen- Vapor Concentration
tration (ftg/m^) (/ig/m^)
Scenario 16 6.5 x 1CT10 1 x 1(T9
Scenario 17 4.9 x 1(T8 8.3 x 1(T8
Scenario 18 6.5 x KT10 1 x IO'9
Scenario 19 4.9 x 10'8 8.3 x 10~8
Exposure (ng/kg-d)
Particulate Vapor
1.7 x 1(T7 2.5 x 10'7
1.3 x 10'5 2.1 x 10'5
6.1 x 10'8 9.4 x 1(T8
4.6 x ID'6 7.8 x 10'6
TABLE 6-19. INHALATION EXPOSURES AT 200 m AND 100 km
EXPOSURES (ng/kg.d)
Scenarios 16 17 18 19
Maximum Particulates 3.1 x 10"7 2.8 x 10~5 1.1 x 10~7 1 x 10~5
(200m) Vapors 5.5 x 10'7 4.7 x 10"5 2 x 10~7 1.7 x 10~5
100 km Particulates 1.1 x 10"9 1.3 x 10'7 3.7 x 10~10 4.8 x 10'8
Vapors 1.9 x 10'9 2.4 x 10~7 6.7 x lO'10 8.5 x 10'8
246
-------
2. Ineestion of Contaminated Soil
The procedures for calculating exposure and risk by soil ingestion were discussed in
Section B.S.d. of this chapter. Using these same procedures, soil concentration listed
below and parameter values listed in Table 6-17, the exposures (see Table 6-20) were
calculated. The corresponding risk values are provided in the appendix.
TABLE 6-20. EXPOSURE FROM INGESTION OF CONTAMINATED SOIL
Soil Concentration (mg/g) Exposure (ng/kg.d)
Scenario 16 5.4 x 10"10 1.9 x 10'6
Scenario 17 2.4 x 10'8 8.2 x 10'5
Scenario 18 5.4 x 10'10 2.3 x 10'7
Scenario 19 2.4 x 10"8 1 x 10"5
3. Dermal Contact .with Contaminated Soil
The procedures for calculating exposure and risk by dermal contact were discussed
in detail in Section B.3. Using the same procedures, the soil concentrations listed below
and the parameter values listed in Table 6-17 the exposures (see Table 6-21) were
calculated. The corresponding risk values are shown in the Appendix.
TABLE 6-21. EXPOSURE FROM DERMAL CONTACT OF CONTAMINATED SOIL
Scenario 16
Scenario 17
Scenario 18
Scenario 19
Soil Concentration (mg/g)
5.4 x 10-10
2.4 x 10'8
5.4 x 10~10
2.4 x 10"8
Exposures (ng/kg.d)
6 x 10-6
2.7 x 10"4
2.2 x 10~6
9.8 x 10~5
247
-------
4. Ingestion of Contaminated Drinking Water
The total concentration in water was derived from vapor absorption and particulate
deposition. Thus it is composed of both a dissolved and suspended -component. The
contamination level was assmumed to remain constant from the source to point where
•water consmumption occurs. Municipal or home treatment may substantially reduce
'exposure levels. The procedures for discussing exposure from drinking water
consmumption were discussed in detail in Section B.3. Using the same procedures, the
water concentrations listed below and the parameter values smummarized in Table 6-17,
•the exposures (see Table 6-22) were calculated. The risk values for each scenario are
shown in the Appendix.
TABLE 6-22. EXPOSURE FROM INGESTION OF CONTAMINATED DRINKING WATER
Water Concentration (0g/L) Exposures (ng/kg.d)
Scenario 16 1.6 x 10~9 3.5 x 10~8
Scenario 17 1.3 x 10'7 2.9 x 10'6
Scenario 18 1.6 x 10'13 1.3 x 10~12
Scenario 19 1.4 x KT11 1.1 x 10'10
5. Ingestion of Contaminated Fish
The procedures for calculating exposure and risk by fish ingestion were discussed in
Section B.3. of this chapter. The same procedures were adopted here except that the
concentration of 2,3,7,8-TCDD in fish was calculated on the basis of a bioconcentration
factor between the fish and water of 10,000 L/kg (Schaffer, 1985) rather than fish and
sediment. Although the fish-sediment approach is generally considered better supported
in the case of 2,3,7,8-TCDD (see discussion Chapter 5) it was more convenient to use the
248
-------
fish-water approach here since estimates of the 2,3,7,8-TCDD concentration in water
were available and the concentrations in sediment were not. Additionally, the bottom
sediments do not become immediately contaminated after the incinerator emissions begin.
Thus, it may take some time before the sediment-fish approach would be valid. Using
the water concentration listed below (dissolved phase only) and the parameter values
listed in Table 6-17, the exposures (see Table 6-23) were calculated. The risk estimates
for each scenario are presented in the Appendix.
TABLE 6-23. EXPOSURE FROM INGESTION OF CONTAMINATED FISH
Water Concentration (mg/g) Exposures (ng/kg.d)
Scenario 16 1.6 x 10'9 6.9 x 10"7
Scenario 17 1.3 x 1
-------
TABLE 6-24. EXPOSURE FROM INGESTION OF CONTAMINATED
BEEF AND DAIRY PRODUCTS
Soil Concen- Beef Exposure
tration (mg/g) (ng/kg/d)
Scenario 16 5.4 x 10~10 3.5 x 1(T5
Scenario 17 2.4 x 10'8 1.6 x 1(T3
Scenario 18 5.4 x 1(T10 8.6 x 1(T6
Scenario 19 2.4 x 10~8 3.8 x 1(T4
Dairy Exposure
(ng/kg.d)
5.2 x 10'6
2.4 x 1(T6
1.3 x 10'6
5.9 x 1(T5
7. Ingestion of Dairy Products due to Particulate Deposition on Fodder
The particulate matter carrying the dioxin compounds emitted from the incinerator
stack can be deposited on the farm fodder, which will become feed material for the cows
grazing in the field. The fodder contamination can also occur through uptake by plant
roots or absorption from ambient air. The 2,3,7,8-TCDD ingested through the
particulate-contaminated fodder will be partially distributed into the cow's system, will
accmumulate in body fat, and is excreted in the fat portion of milk. This situation is
different from the situation considered in Section 6 (where the 2,3,7,8-TCDD ingested by
cows is from the soil rather than the fodder).
Connett and Webster (1987) presented a comparison of exposures to 2,3,7,8-TCDD
emissions from an incinerator via inhalation and ingestion of cow's milk. Their analysis
showed that exposure from ingesting milk of cows grazing on fodder contaminated from
deposited incinerator emissions is about 200 times greater than exposure from ambient air
inhalation of particulate matter. Since many variables affect the result of the analysis,
they presented the variations of the parameters used by different investigators, and the
resulting comparison between the inhalation and milk ingestion.
250
-------
The human intake rate of 2,3,7,8-TCDD from cow's milk can be presented as
Im = Cm Xm Am (6-50)
where Im = intake rate from cow's milk (mg/d); Cm = concentration of 2,3,7,8-TCDD in
cow's milk (mg/L); Xm = milk ingestion rate (L/d); and Am = absorption factor. The
transfer factor (Fm) is defined as the ratio of the concentration in milk divided by the
cow's contaminant intake rate (I). From this definition, the concentration in milk (Cm)
can be expressed as
Cm = Fm I (6-51)
where the transfer factor Fm has the units of d/L, and I is the intake rate of 2,3,7,8-
TCDD, by the cow from ingestion of particulate deposited fodder (mg/d).
Connett and Webster (1987) provided an analysis regarding the deposition rate of
particulates on fodder and a series of first order processes including photolysis,
volatilization and weathering of the deposited contaminant. After neglecting the effect
of photolysis and volatilization for particulate-phase 2,3,7,8-TCDD, they presented an
equation relating the ingestion rate by a cow with other pertinent parameters:
I = G d/KL (6-52)
where G = effective grazing area, (m^/d), d = particulate dioxin deposition rate,
(mg/m^-d), and KL = weathering rate constant, (1/d).
Combining Equations 6-50, 6-51, and 6-52, one can get an expression for estimating
the human intake of 2,3,7,8-TCDD from drinking cow's milk contaminated by the
particulate deposited fodder:
251
-------
Im = Fm G d Xm ATO/KL (6-53)
The parameter values used to estimate the exposure from cow's milk using Equation 6-53
are shown in Table 6-25.
TABLE 6-25. PARAMETER VALUES NEEDED IN EQUATION 6-53
Average Ranee
Transfer Coefficient (d/L) : 0.009 0.0015-0.013
Effective Grazing Area (m2/d) : 22 5.8-60
Weathering Rate Constant (1/d) : 0.05 0.046 - 0.05
Milk Ingestion Rate : 0.36 0.36 - 1.5
Source: Connett and Webster (1987)
The values in Table 6-25 shown in the colmumn under "Average" represents values
recommended by Connett and Webster (1987) and an average value for the milk ingestion
rate as given by them, and those under "Range" represent the ranges of different values
used by several investigators as reported by Connett and Webster (1987).
As an example, the dose from ingestion of cow's milk contaminated by particulate-
deposited fodder will be calculated for scenario 16. The deposition rate of 6.3 x 10"4
ng/m^-yr at 0.8 km exposure distance is equal to 1.7 x 10~" mg/m2 • d. Equation 6-53 in
conjunction with the values in Tables 6-17 and 6-25 can be used to calculate the
exposure (ng/kg • d).
Exposure = (0.009)(22)(1.7 x 10-9)(0.36)/(0.05)(1/70)
(10,000/25,550))x 106 = 1.4 x 10'5 ng/kg • d. (6-54)
The exposures have beem calculated and are shown in Table 6-26.
252
-------
TABLE 6-26. DAILY EXPOSURE FROM INGESTION OF MILK RESULTING
FROM PARTICULATE DEPOSITION ON FODDER
Daily Exposure
ng/kg.d
Scenario 16 1.4 x 10~5
Scenario 17 6.2 x 1(T4
Scenario 18 7.9 x 10"6
Scenario 19 3.5 x 1(T4
The exposures calculated in Section 6, and Section 7 above are compared, and the higher
exposure values for each scenario are reported in Table 6-16. The doses can be
computed by applying the absorption factor. At an absorption factor of 68%, the dose
(absorbed dose) becomes 9.5 x 10~^ ng/kg • d. The corresponding risk values are
presented in the Appendix.
E. RISKS ASSOCIATED WITH TOTAL CDDs VERSUS 2,3,7,8-TCDD
Stack emissions and fly ash contain many types of CDD congeners. Each congener
can exhibit different toxic effects. In an effort to deal with the problem of different
levels of toxicity for different congeners, the equivalency factors for other dioxins in
relation to the most potent 2,3,7,8-TCDD have been developed (U.S. EPA, 1986).
The term "TCDD-Equivalents", as applied to a CDD mixture, is used to refer to the
amount of 2,3,7,8-TCDD which would have the same toxicity as the mixture. If the mass
fractions of each congener in stack emissions or fly ash containing a mixture of CDDs
are known, these emissions or ash can be converted to the mass of 2,3,7,8-TCDD with
equivalent potency. Hence, the weight fraction adjusted according to the equivalency
factor (EF) will be different from that based on mass weight only. The equivalency
factors used in this report were based on U. S. EPA (1986).
253
-------
In this effort, all available literature data with information on the concentration of
2,3,7,8-TCDD in the samples taken were analyzed to obtain the adjusted weight percent
for the total dioxin components. Appropriate data on emissions were found from 13
different incinerators and appropriate data on fly ash were found from six different
incinerators.
The data denoted by "EQV" in Tables 6-10 and 6-11 were smummed to obtain the
averaged mass fraction of each component in stack emissions and fly ash. Table 6-27
shows the results of the emissions analysis. The weight fraction of each congener was
averaged from the available data points. The weight fractions thus calculated were
multiplied by the EF as shown on the table to obtain the adjusted weight percent.
This table shows that, for stack emissions the total CDD constituents are more
toxic than 2,3,7,8-TCDD by a factor of 100/2.7 = 37. This indicates that if the risk
associated with the releases of 2,3,7,8-TCDD only is 10~*> for example, the risk from
TABLE 6-27. WEIGHT PERCENT DISTRIBUTION OF CDD
CONGENER EQUIVALENTS IN MWI STACK EMISSIONS
Sum of Emission
Factors (/ig/kg) % EFa
2,3,7,8
Other tetras
Pentas
Hexas
Heptas
Octas
TOTAL
0.404
7.79
25.87
30.76
16.85
7.32
89
0.45 1
8.7 0.01
29.07 0.5
34.56 0.04
18.93 0.001
8.22 0
99.98
% x EF
0.45
0.087
14.5
1.38
0.019
0
16.44
Adjusted
%
2.7
0.53
88.2
8.40
0.12
99.95
EF = toxic equivalency factor, based on U.S. EPA, 1986.
254
-------
exposure to the total dioxin constituents in the release are higher by a factor of about
37. A value of about 50 has also been suggested assmuming that 2,3,7,8-TCDD comprises
about 2% on the TCDD-equivalents basis (EPA 1987d)
These ratios represent averages derived from nmumerous incinerators. However, the
ratio applicable to any one incinerator appears to vary widely. For example, a MWI in
Marion County, Oklahoma was found to have an adjusted weight percent of 2,3,7,8-TCDD
of 85-92% (Ogden Projects, 1986). This percent would suggest that the total CDD
constituents are more toxic than 2,3,7,8-TCDD only by a factor of about 1.13.
A similar analysis was conducted on the fly ash data (Table 6-28). This analysis
indicates that the adjusted weight percent of 2,3,7,8-TCDD in fly ash is 6.5% which
suggested that the total CDD constituents are more toxic than 2,3,7,8-TCDD only by a
factor of about 15.
The above procedure, using TCDD equivalents to estimate total CDD risks involves
considerable uncertainty. The procedure assmumes that the fate and transport properties
of the CDD congeners are identical with those of 2,3,7,8-TCDD. Since the
chemical/physical properties are not identical, the environmental partitioning, degradation
rates, and human absorption are likely to be different.
255
-------
TABLE 6-28. WEIGHT PERCENT DISTRIBUTION OF
CDD CONGENER EQUIVALENT IN MWI FLY ASH
2,3,7,8
Other tetras
Pentas
Hexas
Heptas
Octas
TOTAL
Sum of Emission
Factors (/igAg)
3.23
92.02
55
393.2
1057.6
797.2
2398.25
%
0.13
3.84
2.29
16.4
44.1
33.24
100
EFa
1
0.01
0.5
0.04
0.001
0
% xEF
0.13
0.038
1.145
0.656
0.044
0
2.013
Adjusted
%
6.46
1.89
56.88
32.59
2.19
0
100.01
a EF = toxic equivalency factor, based on U.S. EPA, 1986.
256
-------
7. UNCERTAINTY EVALUATION
The estimates of 2,3,7,8-TCDD exposure presented in Chapter 6 are organized around
scenarios developed to reflect some of the varied sources of environmental exposure to
this compound. In developing these scenarios, the Exposure Assessment Group tried to
construct examples that are relevant to exposure assessment needs faced by the Agency.
Accordingly, the major focus is on contaminated soil and on landfills containing dioxins
and on incinerators emitting dioxins. The calculations for land related scenarios (1-15)
are of relevance to other land contamination issues and the incinerator scenarios (16-20)
are relevant to other sources of air emissions.
The land-related scenarios address several simplified situations. These scenarios
vary by the degree to which the contaminated material is maintained and controlled, the
distance to human residences, and the presence of water bodies or agriculture. These
physical scenarios are intended to represent either reasonable worst-case situations or
situations believed to be more typical, i.e., to more closely resemble occurrences that will
be commonly encountered in the field. In these scenarios a constant concentration of
2,3,7,8-TCDD is assumed across the contaminated site. In the absence of detailed site-
specific information, use of an average site concentration to calculate exposures is
deemed a useful procedure.
The incinerator scenarios address two model incinerators: an incinerator of
moderate size assumed to have a high dioxin emission factor and a large incinerator
assumed to have a low emission factor. The second incinerator model is intended to
represent technology that may be utilized in the future. The incineration scenarios assess
multimedia exposure from incinerator stack emissions and landfilled (monofilled) fly ash.
These scenarios are intended to illustrate a range of circumstances that may be
encountered, rather than predict exposures that will occur at specific sites. It may be
expected that some factors affecting emissions will vary markedly between sites. As
257
-------
such, it is not meaningful to discuss the uncertainty present in the simplified physical
scenarios; rather, the test of their construction will be whether they prove useful to the
Agency as examples of how to evaluate sites encountered in practice. Most importantly,
the methods used in assessing these scenarios can be applied to a variety of other
physical situations that may be encountered.
Human behavior patterns can strongly affect exposure to toxic materials. For
example, the number of years spent at a site and the quantity of locally caught fish
consumed will proportionally affect exposure through a contaminated fish pathway.
Typically such factors vary markedly among individuals and may also exhibit variation
between different population groups. Accordingly, variations among behavioral parameters
are factored into the exposure scenarios presented. Because the scenarios do not
represent any specific individuals or population, these behavioral parameters are regarded
as part of the scenario formulation. That is, they are presented as illustrative values
rather than estimates. However, survey data often provide a baseline with which the
reasonableness of behavior parameters may be compared. For example, fish consumption
estimates must make sense when compared with the data available on fish consumption in
various U.S. population groups. In the specific pathway discussions in this chapter, the
basis for selection of behavior parameters is addressed.
After the physical site arrangement and human activities are specified, the degree
to which physical and biological processes transport the contaminant to the human
receptors must be evaluated. Determining exposure requires use of measurement data and
mathematical models. Uncertainty can be present in measured values that may not be
accurate or representative, in mathematical models that do not correctly reflect the
processes actually occurring, and in parameters used in models which are also subject to
measurement error.
258
-------
For the parameters used for transport of 2,3,7,8-TCDD via the various pathways, the
estimates made for the exposure scenarios are intended to be best estimates, given the
limitations of our knowledge of the mechanisms in exposure pathways. In the small
number of instances where upper bound exposure assumptions are made, this fact is
explicitly noted. For each pathway assessed, this chapter presents a discussion of the
strengths and weaknesses of the methodologies utilized to generate the exposure
estimates. This analysis attempts to provide a qualitative weight-of-evidence evaluation
as well as an evaluation of how far numerical results may be wrong.
A. CONTAMINATED SOIL AND LANDFILLS
1. Summary of Uncertainties
The assessed exposure pathways can be conveniently divided into those that depend
on direct human or animal contact with contaminated soil (including sediment containing
contaminated soil) and those pathways that do not involve such contact.
The first group includes beef, vegetable and dairy products ingestion, fish ingestion,
soil ingestion, and soil dermal contact. The second group, not involving direct soil
contact, includes vapor inhalation, dust inhalation, drinking water (surface) and drinking
water (ground). A review of the exposure and risk estimates
presented in Tables 6-6, 6-16, and in the appendix shows that the pathways involving
direct soil contact generally have higher estimated risks than those that do not. (The
ground water pathway is not included in these tables, as discussed in Chapter 6 estimated
ground water concentrations of 2,3,7,8-TCDD are very low).
The weight of available data indicates that, however, major uncertainty exists in the
quantity of uptake. Using a moderate amount of uptake as an assumption, the plant
pathway can show exposures of the same order of those of the other food-chain related
pathways. However, due to the quantitative uncertainties plant uptake exposures have
not been estimated in the scenarios.
259
-------
Exposures estimated under the "soil contact" pathways all depend on the estimated
soil or sediment concentrations of 2,3,7,8-TCDD. These estimated concentrations are all
dependent on assumptions about the degradation rate, and in some cases also the erosion
and deposition rates of contaminated soils. The assessment, which is premised upon a
thick layer of contaminated soil being present, assumes that no degradation occurs.
Because 2,3,7,8-TCDD is known to disappear slowly from surface soil it is likely that
concentrations in thick layers of soil will decrease even more slowly. Therefore, in the
absence of directly relevant data, making the assumption that no degradation occurs
below surface is a reasonable procedure which errs towards higher estimates of
concentration.
The off-site soil concentration depends on the landfill erosion rate and the fraction
of eroded soil placed on an adjacent field. The assessment presented is appropriate for
situations where a landfill with a fairly high erosion rate places a significant quantity of
soil on an adjacent field. The erosion estimates used are judged appropriate as they are
taken from a survey of erosion conditions at landfills. Erosion rate and placement of
soil will depend highly on site specific conditions. In a low erosion situation, the
predicted concentration of 2,3,7,8-TCDD in the off-site field could be a factor of 100
lower. On the other hand, soil concentration in the reasonable worst case situation
could not plausibly exceed three times the estimated levels. Off-site pond sediment
concentrations are assumed equal to the estimated off-site soil concentrations and will
exhibit corresponding site- specific variability.
Predicted risks from beef, fish or dairy product consumption depend principally on
three factors: estimated soil concentration, the estimated fish/sediment, beef/soil or milk
fat/soil distribution ratios for 2,3,7,8-TCDD, and the rate of individual consumption of
these foods from the local source. The substantial expected variation in off-site soil or
sediment concentrations of 2,3,7,8-TCDD is discussed above. The fish tissue/sediment
260
-------
distribution ratio has exhibited substantial variation in laboratory and field measurements.
The value of five used in these analyses is towards the high end of measured values, and
a value one order of magnitude lower cannot be ruled out, with substantial fish species-
specific effects being likely. Distribution ratios for beef and dairy products were
estimated using results obtained for polybrominated biphenyls for cattle in contaminated
feed lots. Comparison with laboratory data on cattle fed 2,3,7,8-TCDD are supportive of
the beef/fat distribution ratio used but suggest the milk/fat distribution ratio may be
somewhat low (which would lead to a low human exposure estimate). Dietary
consumption of beef and dairy products was estimated using survey data; survey data
were also available for the proportion of these products that farm families raise for
themselves. Thus the consumption estimates used have a reliable basis. It should be
observed that these estimates apply only to individuals who regularly consume locally
raised beef and dairy products. In summary it is unlikely that on-site exposure estimates
have large, but off-site exposures can be expected to show substantial site specific
variation depending on soil concentration.
Two pathways involve direct human contact with contaminated soil: soil ingestion
by children and dermal contact with soil. The exposures estimated for these pathways
are generally below those estimated for consumption of the animal products discussed
above, but are higher than the "indirect" pathways. Two circumstances could lead to
higher soil ingestion estimates. First, the values used do not specifically account for
children with pica; data are not available for soil ingestion in children identified as
having pica. Second, adult inadvertent ingestion of soil is not assessed, although it
might be argued that exposure at much reduced levels from ages 7-70 would lead to
comparable exposures to those in ages 2-6 for which ingestion is assessed. Gut
absorption of 2,3,7,8-TCDD was estimated using measured absorption in animals fed
contaminated soil. The majority of the small number of tested soil samples showed
261
-------
substantial absorption comparable to the 25% value used the pathway assessment; however
results with one tested soil sample indicate that absorption can be one to two orders of
magnitude lower for some soils. This bioavailability uncertainty may be important for
site-specific risk assessments.
2. Uncertainties in Specific Methods Applied
a. Soil Dilution Factor
The risks estimated for several off-site exposure pathways depend on the
concentrations of 2,3,7,8-TCDD in surrounding soil due to erosion from the soil or
landfill. It can be anticipated that the quantity of material eroded from the soil or
landfill will depend primarily on site-specific climatic, topographic, and soil
characteristics. These factors are reflected in the parameters of the Universal Soil Loss
Equation (USLE) [Equation 6-3]. The USLE is a widely used model in agricultural
situations; the accuracy and precision of USLE estimates have not been separately
reviewed for this report.
The erosion estimates were obtained by using the USLE to calculate soil erosion in
a survey of 70 landfills conducted for EPA (Science Applications International Corp.,
1986). The survey, which was based on interviews with site managers and local officials
and review of USGS topographic maps, ascertained estimates of the parameters related to
rainfall, soil properties, and topography and site management practices. The landfill
survey data can be used to estimate the quantity of material which would have been
eroded from each of these landfills under the scenario assumptions. In scenarios 8-11,
no vegetation was present to reduce erosion (that is, the factor C is equal to 1 in the
USLE). For the 70 landfills, estimated erosion rates ranged from 0.6 to 306 tons/acre-
year, with a mean calculated value of 62 tons/acre-year. The 10th to 90th percentile
range was 2 to 163 tons/acre-year. It is noteworthy that landfills with low erosion
potential may have two orders of magnitude less erosion than 62 tons/acre -year. Changes
262
-------
in the erosion rate directly affect calculations of risk through all food-related pathways.
For the scenarios where grass cover was assumed (12-14), the parameter C = 0.1
was used. In this situation, the Chapter 6 estimates of erosion were reduced by a factor
of 10 to 6.2 tons/acre-year, and all estimates taken from the landfill study are similarly
reduced. Thus, the conclusions about certainty remain the same. It should be noted that
the depth of the contaminated soil will place a limit on the period at erosion, at a
specified rate, may occur. This limit should be considered before applying the long-term
erosion model used here to fields where the contamination by 2,3,7,8-TCDD is not deep.
The next parameter estimated in calculating the dilution factor is F, the fraction of
the soil eroded from the landfill that is assumed to be deposited on the adjacent 10-acre
field present in the scenarios. Two values of F were selected for the reasonable worst
case and typical case: 0.5 and 0.1, respectively. These parameters were selected on the
basis of judgment concerning the fraction of eroded soil that could plausibly be deposited
on an adjacent field. No data for estimating F were available. Much variation in F
would be expected between sites, depending on the gradient of the landfill with respect
to the field and the degree to which runoff is channelized. The estimates of F = 0.5 and
0.1 both reflect direct placement of eroded material on the off-site field; however, this
need not be the case—if the field were not directly down-gradient from the landfill, a
much smaller quantity of material might be deposited.
A simplified mathematical model was developed to address the fate of 2,3,7,8-TCDD
on eroded soil after deposition in the field. Two factors were considered: 1) mixing of
eroded soil with the upper zone of soil present in the field, and 2) degradation of
2,3,7,8-TCDD in field soil due to combined biological, chemical, or physical processes.
The mixing depth was selected as 10 cm, which was judged to be intermediate to
what might occur under different agricultural practices. A half-life of approximately 10
years was selected on the basis of experimental data from one study of 2,3,7,8-TCDD in
263
-------
surface soils. The mathematical model applied assumed even mixing of 2,3,7,8-TCDD into
the soil to a depth .of 10 cm, it is recognized that this is a simplification of actual
processes in a field. Such processes are likely to differ substantially between fields.
The model results, however, show only modest sensitivity to these parameters as the
following observations for the reasonable worst case scenario (62 tons/acre-year, C = 1,
F - .5) show. Changing the assumed mixing depth to 1 cm increased the 70-year average
field soil concentration by a factor of 2.3, while increasing mixing depth to 20 cm, a
value which might be expected in a plowed field, reduces the average soil concentration
by a factor of 1.6. Similarly, increasing the assumed half-life of 2,3,7,8-TCDD from 10
years to 35 years increases average levels by a factor of 1.8, while decreasing the
assumed half-life to 4 years decreased the estimated field concentration by a factor of
2.0. Thus, rather broad variation in these model parameters produces only limited
variation in model estimates.
BAG regards the soil concentration model as a plausible simplified approximation to
field conditions, however no field data are available to assess model predictions.
Based on the preceding discussion, the largest uncertainties in site soil
concentration estimates are probably .due to the erosion rate used for the landfill and the
fraction of eroded material assumed to reach the adjacent field. The values used reflect
relatively high erosion and relatively direct placement of eroded material on the field. If
these conditions are not met, the calculations here are not appropriate. For example, in
a low-.erosion situation (2 kg/acre-year), the soil mixing model would lead to an
estimated soil concentration (dilution factor) a factor of 20 below the value calculated in
Chapter 6. If, additionally, only 5% of the total eroded material was deposited on the
field, the predicted 2,3,7,8-TCDD concentration would be a factor of 200 lower than the
reasonable worst case value and a factor of 40 below the typical case value. On the
other hand, since the field concentration could not reasonably reach a concentration in
264
-------
excess of the landfill concentration, the soil concentration cannot exceed 2.7 times the
value (dilution factor 0.37) calculated in the reasonable worst-case bare site scenarios
(8-11) in Chapter 6. A summary of the uncertainties associated with soil dilution factor
is in Table 7-1.
b. Sediment Dilution Factor
The sediment concentration for a pond near the waste site (scenarios 1-4, 8-11) is
assumed to have the same concentration (dilution factor of 1) as the soil in the field in
which the pond is located. In the absence of site-specific measurements, this is a
plausible assumption. However if substantial quantities of sediment from uncontaminated
sites also reached the pond sediment concentrations would be diluted. Additionally,
biological/photolytic/chemical degradation of 2,3,7,8-TCDD in pond sediments would
reduce this value to an extent depending on the rate of transport of contaminated
sediment to the pond, although the rate of degradation is likely to be slow.
For a stream off-site (scenarios 5-7, 12-14), both the land (assumed to be grass-
covered in these scenarios) and the surrounding watershed are assumed to suffer the
same degree of erosion. Then, the sediment in the stream is assumed to be a
proportional mix of sediment from the 10-acre site and a 10,000-acre area of watershed.
These assumptions lead to a dilution factor of 0.001. This estimate utilized judgment of
the amount of mixing that might reasonably be taking place, while recognizing that
variation between sites and between watersheds would lead to much variation in actual
stream sediment concentrations. Chapter 6 also presents a calculation, leading to a very
similar dilution factor, based on expected run-off from a 10,000 acre watershed and the
estimated sediment burden of a stream carrying this amount of water, and the estimated
sediment eroded annually from a 10 acre landfill. This calculation reinforces the
plausibility of the 0.001 dilution factor used. It is also noted that the stream flow
265
-------
TABLE 7.1. LANDFILL ASSESSMENT
SOIL DILUTION FACTOR
Assumption/
Method.
Approach-
Rationale
Uncertainty
Comments
Quantity of
erosion.from
contaminated
area.
Quantity, of
eroded soil
deposited on
adjacent field.
Mathematical
model.to assess.
rate of 2,3,7,8-
TCDD in soil
placed on field.
Used mean of ero-
sion) estimates
obtained from
applying, universal-soil
loss equation in a
survey of 70 land-
fills.
Mean: 65 Tons/acre--
year.
Values of 50% and
10% used for reason-
able wont case,
more typical-case,
respectively.
Model assumed a
mixing: depth of 10
cm 2,3,7,8-TCDD
well mixed in field
soil, above this,
not below. Half-
life of 10 years for
2,3,7,8-TCDD in
fielf soil assumed.
Universal soil Ion
equation widely used
model for erosion.
Data.for landfills
appropriate for this
assessment.
Judgment on how
much, eroded soil.
could be placed on
adjoining field-
Model judged to be a
physically appro-
priate idealised
description. Mixing,
depth consistent with
agricultural prac-
tices, data on 2,3,7,
8-TCDD profiles! on
soils. Degradation
rate taken from field
data.
Landfill erosion
estimates- ranged from
0.6 to 306 tons/acre-
year.
10th-90th percentiles
from erosion survey 2 to
163 tons acre-year.
Erosion rate will be
highly site-specific.
Could be close to
zero but not much
higher than 50%.
Model not tested in
practice; parameter
estimates are not
precise; However,
model estimates are
not very sensitive to
assumptions on mixing
depth or degradation
rate.
Highly site-specific, based
on-gradient of landfill,
adjoining fields.
Evaluation: Analysis is reasonable for a situation where a contaminated area with a fairly high erosion rate deposits
soil on an~ ad joining field. Largest sources of uncertainty in comparing the scenarios described here with an actual.
field site will be in the assumptions about quantity of soil eroded and placed on the adjacent field. In a low-erosion
situation, the.predicted concentration of.2,3,7,8-TCDD in the field could be a factor of 100 or more lower. The
concentrations on the field could not plausibly exceed about 3 times the estimated levels even in very high erosion
situations: Risk numbers for food-related pathways are sensitive to these values.
266-
-------
calculated for the 10,000 acre watershed (17 ft3/sec) is of a size that could plausibly
support local fishing.
Science Applications (1986) reports watershed areas for most of the 70 landfills for
which erosion rates were calculated. These watershed sizes vary considerably with 10 of
51 watersheds being smaller than 1,000 acres and 16 of the 51 being 100,000 acres or
above (median 11,000 acres; mean 29,000 acres). Thus the assumed value of 10,000 acres
falls in the middle of the observed values. If the landfill were in a smaller watershed,
less dilution of contaminated sediment by stream sediment would be expected. However, if
a much smaller watershed were assumed with a correspondingly lower stream flow, the
stream size may be inadequate to support substantial local fishing. Additionally, however,
if local hydrologic factors allowed significant accumulation of landfill sediment near its
input to the stream, larger values of the dilution factor would be expected.
It should be noted that only the "more typical" and not "reasonable worst case"
scenarios included a stream. If a stream sediment dilution calculation were made using
the reasonable worst case assumption of a bare landfill with a 62 ton/acre-year erosion
rate, the increased erosion would lead to a dilution factor of 0.01 rather than 0.001. A
summary of the uncertainties associated with sediment dilution factors is shown in Tables
7-2 and 7-3.
c. Degradation
As discussed under the soil dilution factor, measured values of near-surface
degradation or removal rates of 2,3,7,8-TCDD were used in calculating off-site soil levels.
It was noted that in the soil mixing model applied, degradation rate had only a modest
effect on predicted soil concentration. No degradation was assumed to occur in the soil
buried in the landfill. No data were available to support estimates of degradation
inside the landfill; however, any such degradation should be less than the estimated half-
life for surface degradation/disappearance of 2,3,7,8-TCDD (estimated at approximately 10
267
-------
TABLE 7.2. LANDFILL ASSESSMENT
SEDIMENT DILUTION FACTOR - PONDS
Assumption/
Method
Contamination
levels in
sediment.
Approach
Level in soil
surrounding pond
assumed to equal level
in sediment.
Rationale
Pond sediment is de-
rived from local soils
and has less turnover
than streams - so
should reflect local
soils.
Uncertainty
Sediment transport to
some ponds may .come
from outside con-
taminated area
causing greater
dilution.
Comments
Worst-case assumption.
Degradation of
2,3,7,8-TCDD in
sediment.
Assumed no
degradation.
Degradation in pond
expected to be less
than for land surface
as some paths of
degradation will be
reduced under water
(e.g., photolysis).
Degradation slow on
land surface. Addi-
tional contaminated
soil will be mixed
with pond sediments
on continuing basis.
Young (1983) observ-
ed degradation rates
on land with roughly
a 10 year half-life.
If these apply to both
sediments in the
pond and to new
sediments reaching
the pond, the average
concentration over 70
years would be re-
duced by 80%.
Worst-case assump-
tion, since it is unknown
how the water
environment will change
the land-based half-life
observed by Young
(1983).
Evaluation: This approach uses worst-ease assumptions in the absence of data. The scenarios in which the pond are
used are the "reasonable-worst-ease" assumptions, so these assumptions do not appear in the other scenarios. Given
the pond itself in the scenarios where it is used in either on the contaminated site or directly downhill adjacent to the
landfill, these are not unreasonable assumptions. Fish consumption risks are directly affected by sediment dilution
factor estimates.
268
-------
TABLE 7.3. LANDFILL ASSESSMENT
SEDIMENT DILUTION FACTOR - STREAMS
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Contamination
levels in
sediment.
Dilution assumed to
be proportional to
landfill area divided
by water shed area.
Assumed a landfill
area of 10 acres and
watershed area of
10,000 acres - yield-
ing a dilution of 1000
fold.
Taking proportions of
source areas is an
accepted method for
calculating sediment
dilution. A 10,000
acre watershed is
near the median size
seen in a survey of
70 landfills. The
runoff from 10,000
acres would support a
stream of fishable
size in many cli-
mates.
Watershed could be
much larger with
greater dilution.
Watershed could be
smaller with less
dilution, but might
not support fishable
stream. Mixing of
sediment in stream
may not be uniform
and higher concen-
trations of 2,3,7,8-
TCDD may be found
near landfill.
A second approach to
calculating dilution
yielded similar results:
Runoff from a 10,000
acre watershed was
calculated and standard
equations were applied to
estimate the quantity of
sediment carried by a
stream of this size.
Sediment dilution was
then calculated by
comparing estimated
landfill erosion rate with
calculated stream
sediment load.
Relative erosion
at landfill site
vs. rest of
stream.
This method assumes
a uniform erosion rate
through the
watershed.
Site-specific infor-
mation would be
needed to evaluate
variable erosion rates.
Landfill survey found
erosion rates to vary
from 2 tons/acre-
year (10th percen-
tile) to 163 tons/
acre-year (90th per-
centile).
See Above.
269
-------
TABLE 7.3. (CONTINUED)
Assumption/
Method
Degradation of
2,3,7,8-TCDD
in sediment.
Approach
Assumed no degrada-
tion.
Rationale
Degradation expected
to be lower than slow
rate observed on land
surface (less photo-
lysis). Additional
contaminated soil will
be added to sediment
continually.
Uncertainty
No sediment degrada-
tion data are avail-
able. Young (1983)
observed degradation
rates on land at 10
year half-life.
Average concentra-
tion over 70 years
would be reduced by
80% if both landfill
and stream sediment
2,3,7,8-TCDD were
degraded at this rate.
Comments
Worst-ease assumption
used. It is unclear what
effect the stream
environment would have
on degradation relative
to Young's observations,
but it may be slow
degradation.
Evaluation: The estimate of stream sediment dilution factor is very site-specific, and therefore applying the stream
scenarios to an actual site should be done with caution. The key 'Uncertainty is .the evaluation of erosion rate from
the landfill vs. sediment load of the stream. The sue of the watershed assumed [i.e., sediment load] is moderate for
landfills, but large enough to have a stream which can support a reasonable fish population. In this sense, the size of
the watershed is small relative to those in actual sites which will support fishing, but there are likely to be many
•streams like this in actual situations. The mixing of the stream sediment load with the erosion loading is not well
understood hi terms of concentrations of resulting downstream sediment, so therefore the uncertainty of the 1000
dilution factor is high (at least one to two orders of magnitude uncertainty). Obviously, uncertainties in the dilution
factor directly translate to uncertainties in the risk estimates for fish consumption.
2.70
-------
years). If a 29-year half-life were assumed inside the landfill (a value which was
calculated in this report for persistence in the presence of a particular kind of mold)
only a modest reduction (less than a factor of 2) in estimated risks averaged over a 40-
or 70-year time period would be predicted.
Because of degradation processes at the surface, it is possible that surface
concentrations of 2,3,7,8-TCDD may be lower than the internal landfill concentrations.
Surface concentrations would depend on a balance between degradation rate and erosion
rate. Some of the data presented in Young (1983) suggest the presence of such an
effect. Recent data also indicate that photolysis occurs on the immediate surface of soil.
It is unclear what lower surface concentrations would mean in terms of risk, since there
are no data indicating how, especially if only on the immediate surface, much of an
effect this might have on 2,3,7,8-TCDD concentrations in materials such as windblown
dust. If surface degradation has a substantial impact on the concentration of 2,3,7,8-
TCDD in surface soil and windblown dust, many of the risk numbers for pathways related
to surface soil (e.g., soil ingestion, dermal contact, dust inhalation, etc.) could also
change substantially downward. At present the assumption of no degradation is
essentially the only alternative, having a moderate, but unquantifiable uncertainty. A
summary of the uncertainties related to degradation is in Table 7-4.
d. Dust Inhalation
Dust emissions from the soil and/or landfill were calculated using an empirically
derived relationship which assumed that the surface was uncrusted and composed of
finely divided particulates, a situation where dust emissions will be maximized for an
undisturbed site. The emission rate was based on the proportion of surface assumed to
be unvegetated. One parameter in the model is the threshold wind speed for erosion,
271
-------
TABLE 7.4. LANDFILL ASSESSMENT
DEGRADATION
Assumption/
Method
Degradation in
landfill.
Approach
Assumed none.
Degradation on
contaminated
soil.
Assumed none.
Rationale
Hydrolysis and oxida-
tion are very slow in
water, and hence ex-
pected to be very
slow in soil (U.S.
EPA, 1985b). While
most soil organisms
appear not to degrade
2,3,7,8-TCDD, one
species of fungus may,
with an apparent
half-life of 29 years
(Bumpus et al., 1985).
Two mechanisms may
serve to renew sur-
face 2,3,7,8-TCDD
levels. Erosion of
surface soil will
expose buried soil.
Movement through
volatilization and
reabsorption may
occur. Need to
balance several fac-
tors makes surface
concentration diffi-
cult to estimate.
Uncertainty
No data exist on
degradation of buried
2,3,7,8-TCDD, how-
ever, degradation
should be less than
slow rate observed for
surface contami-
nation. If buried
2,3,7,8-TCDD degrad-
ed with 29 year half-
life, less than a
factor of 2 change in
70 year average con-
centration would be
predicted.
Some loss of soil
surface 2,3,7,8-TCDD
would be expected
based on observa-
tions of Young (1983),
who noted overall
losses from known
effects like photolysis,
volatili-cation, soil
movement and possibly
biode-
gradation. Young
provided an estimate
of a half-life of
approximately 10
years for 2,3,7,8-
TCDD in the upper
layers of soil, based
on field data.
Comments
Photolysis of 2,3,7,8-
TCDD on soil in presence'
of hydrogen donors (and
light) reported by Crosby
and Wongs (1977). By
extrapolation and
multiple corrections from
2,3,7,8-TCDD photolysis
on plant leaves,
estimated half-life on the
soil surface of 7.2 days
(Thibodeaux and Lipsky,
1985).
Evaluation: 2,3,7,8-TCDD is known to degrade slowly in contaminated surface soil and it is likely that concentrations
in thick layers^ of landfilled soil will decrease even more slowly. In the absence of data, assuming no degradation is a
reasonable procedure that errs towards higher estimates of concentration. However, 2,3,7,8-TCDD at the soil surface
itself may degrade at a rapid rate (see comment above). What effect this surface rate would have on overall
concentration is questionable, and hence the source of significant uncertainty. Uncertainties in degradation rate
directly affect the soil dermal, dust inhalation, and soil ingestion pathways.
172-
-------
estimated in a manner recommended by the model's developers. Since the procedure for
estimating threshold wind speed is complicated, it has not been evaluated here for
empirical agreement, and since the estimated erosion rate depends on the inverse of the
cube of this parameter, uncertainties in the predicted dust emission rates may be of
consequence. Dust emission also depends on the cube of mean annual wind speed. The
mean annual wind speed for 60 U.S. cities varies from 2.8 to 6.3 m/s, a mean of 4 m/s
was used to calculate emissions in Chapter 6. Use of a 2.8 m/s wind speed would have
led to a dust emission estimate a factor of three lower, while use of a 6.3 m/s wind
speed would have led to an estimate a factor of four higher.
Since the empirical validation of the dust emission model has not been reviewed,
uncertainties cannot fully be assessed; however, the assumptions on the character of site
soil tend to maximize emission estimates. In some of the scenarios (1-4, 8-11) it was
assumed that most of the surface remained bare and unvegetated. Such an assumption is
not plausible in many climates and alternate assumptions may be warranted in site
specific assessments.
Dust emissions due to vehicular traffic on an unpaved road was calculated using the
methodology described in the section on land disposal ash from incinerators in this
chapter. Note that the assumption that heavy trucks are present may not be appropriate
for many inactive sites. On the other hand, the calculations assume that only the site
itself contributes contaminated airborne dust. Over time, as erosion and previous
windblown dust contaminate areas adjoining the site, these areas will also release
contaminated, airborne dust, increasing exposures over those calculated here.
Following the generation of emissions estimates, transport models were used to
estimate concentrations of dust reaching individuals in the vicinity. Different models were
used to assess on-site and off-site dust concentrations. For off-site exposures with the
receptor more than 100 m from the site, the ISC model was used, with a virtual
273
-------
downwind distance parameter incorporated to reflect the fact that the site is not a point
source. The ISC model is widely accepted and used throughout EPA as a standard model.
The default model parameters utilized for this analysis are intended to represent typical
meteorologic conditions. Thus, while the quantitative uncertainties of model predictions
are not assessed in this report, there is a strong basis for relying on the ISC model.
On-site exposures, that is, exposures due to living in a contaminated area, are more
difficult to assess, since there is no standard model accepted for modeling on-site air
concentrations. In this analysis, a near-field dispersion model was applied (this model
was also used for off-site exposures within 100 m of the site boundary). The near-field
model is based on theoretical principles, and no empirical validation data are available.
There are however, two supporting analyses that buttress the general predictions of this
dispersion model. First, as shown in Figures 6-2 and 6-3 in Chapter 6, the ISC model
and the near-field dispersion model are in satisfactory agreement (the near-field model
predicts concentrations 40% lower than the ISC model) at the boundary of the site where
both models can be applied. Second, the near-field model is in general agreement with,
but somewhat lower than the simple box mixing model which assumes sideways transport,
without vertical mixing, of contaminants from the site.
Three additional factors, the inhalation rate, exposure duration, and absorption
fraction, enter into the inhalation risk calculation. The inhalation rate utilized, 23
m-Yday, is within the range of standard values commonly utilized for this parameter, and
reasonable variations should not have a marked effect on risks. Exposure duration set at
40 or 70 years is considered as a defined part of the exposure scenario, with other
durations being easily evaluated if desired. Little data exist on absorption of 2,3,7,8-
TCDD following inhalation. However, since much inhaled particulate will eventually enter
the gut, the gut absorption data, discussed below, provide some support for the 27% value
utilized in this assessment.
274
-------
In summary, dust exposure calculations in some "reasonable worst-case" scenarios
assumed site conditions that would tend to maximize dust release from winds, and include
dust emissions due to solid disturbance by vehicles. Table 7-5 includes a discussion of
the uncertainties involved with dust inhalation.
e. Vapor Inhalation
The rate of air emission of 2,3,7,8-TCDD through volatilization was calculated using
a release model previously developed and peer reviewed in connection with an assessment
of PCBs [Equation 4-3]. The model is based on theoretical mass-balance calculations,
utilizing equations for fundamental physical/chemical transport processes. No empirical
data are available to validate the model. Variables upon which the model release
estimates depend are Dj, the diffusivity of dioxin in air; E, soil porosity; Ps, soil density;
and Kas, the soil/air partition coefficient. The parameter Dj is calculable based on
physical laws, while E and Ps depend on site soil characteristics; however, these
parameters would not be expected to vary over a great enough range to strongly affect
emission estimates.
The major source of uncertainty in the predicted emissions pertains to K^, which is
equated to the ratio of Hc (Henry's Constant) to Kj (soil/water partition coefficient).
Chapter 3 reports a range of values for Hc of 1.6 x 1"^ to 4.6 x 10"^ from studies by
two authors. The lower of these values is used in the emissions calculations; if the
higher value were used, the release estimate would be a factor of three higher.
The parameter K^ is calculated using the following empirically supported
relationship:
"
-------
TABLE 7.5. LANDFILL ASSESSMENT
DUST INHALATION
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
2.3,7.8-TCDD
was assumed to
be transported
from the site via
wind blown dust.
Dust generation.
Dust emissions were
modeled via the "un-
limited reservoir
method." Soil sur-
face was assumed to
be finely divided,
uncrusted particu-
late.
Dust generated by
wind and vehicle
traffic.
This is a published
method developed
from field data.
Thought to be the
two principal sources
of dust emissions.
Calculation provides
high estimates of dust
generation rates
because many surfaces
may be crusted, re-
ducing potential for
dust emissions. Model
parameters, e.g.,
widespread, will
exhibit substan-tial
variation.
Vehicle generated dust
calculated assuming
presence of heavy
trucks; if only light
vehicles are present.
Vehicle emissions
reduced by factor of
five.
Vegetation cover.
Vegetation cover was
assumed to vary from
none to complete in
different scenario
calculations.
Inhalation rate. Used 23 mS/d.
Vegetation cover re-
duces dust emission; a
range of values was
used to show impact
on dust emissions.
In range of standard
values used for this
parameter.
Full range of this
variable was utilized
among the several
scenarios, resulting in
a difference of about
two orders of
magnitude.
Not a major source of
variation.
Evaluation: Reasonable approach that tends toward making high estimates of dust generation for an untrafficked site,
but site disturbance could lead to much higher estimates. The ground cover assumptions here are highly site-specific
with a resulting range of about two orders of magnitude in risk.
276
-------
and 4% among different soils. Koc has not been measured for 2,3,7,8-TCDD, and must
itself be calculated using an empirical relationship relating Koc to Kow, the
octanol/water partition coefficient. A careful review of this last relationship was
conducted by Lyman and Loreti (1986), who empirically compared Koc and Kow for 57
organic compounds selected as having the most reliable experimental data. Using the
three regression equations derived by Lyman and Loreti, and the experimental value of
4.24 x 10*> for Kow (presented in Chapter 2) leads to estimates of Koc in the range 4 x
10^ to 7 x 10^. However, in the data analyzed by Lyman and Loreti, the only compound
having a Kow above 10^ showed a Koc more than an order of magnitude below the
regression line estimate, reducing confidence in the relationship for compounds such as
2,3,7,8-TCDD with very high Kow values.
U.S. EPA (1985g) reported laboratory measurements of the soil/water partition
coefficients for 10 soil samples from sites in Missouri and New Jersey (the eight Missouri
samples were from soils contaminated by dioxin containing waste oil, the two New Jersey
samples were from industrial sites). The measured partition coefficients (mean of "SWLP-
R" data) ranged from 4 x 104 to 4 x 106 with a geometric mean of 5 x 105. The total
organic content of these soil samples ranged from 1.5 8%. Since these data show
substantially higher partition coefficients than the soil/water partition coefficient used in
the exposure calculations (Kj = 4,680 for a soil with 1.0% organic carbon content).
In light of the points raised above, the use of the selected value of K.
-------
As with dust emissions, respiration rate and residence time at the site would
influence risk; however, both parameters are within standard ranges used for these
values, and reasonable variation (pertaining to long-term exposure) would have limited
impact on predicted risk.
The absorption fraction for vapor inhalation is estimated at 0.75. Little data exist
to support a specific value for this absorption fraction; however, in consideration of the
high gut absorption of pure 2,3,7,8-TCDD, substantial inhalation absorption is plausible.
Table 7-6 is a summary of uncertainties for vapor inhalation.
f. Dermal Exposure
Estimates of dermal exposure relied on the analysis by Schaum (1984) of the limited
data available on this pathway. The quantity of soil on exposed body surface areas was
estimated. Using two studies of the quantity of soil on children's hands using two
different measurement methods (Lepow et al., 1975 and Roels et al., 1980). Data from
these studies were analyzed to obtain estimates of 0.5 and 1.5 mg/cm2 for the soil levels
on children's hands. A value of 1 mg/cm2 was used in scenario calculations. It is then
assumed that the soil levels on hands are reflective of soil levels on all exposed skin
areas of the body, and that adults have levels of soil on skin surfaces similar to those of
children. No data were available to assess either assumption. It may be argued that
these assumptions are likely to overestimate exposure if other unclothed body areas have
less dirt exposure than hands and if adults have lower dermal soil levels than children.
Depending on the type of clothing worn, exposed adult skin area ranges from 900 to
2,500 cm2. A value of 1,000 cm2 was used in scenario calculations. Exposure durations
were taken as specified in the separate scenarios.
Experimental data on skin absorption for soil containing added 2,3,7,8-TCDD are
available for the rat. Poiger and Schlatter (1980) reported the percentage of applied
radiolabel dose in the liver following placement of a water/soil paste on the skin. Three
278
-------
TABLE 7.6. LANDFILL ASSESSMENT
VAPOR INHALATION
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Air emissions
release rate.
On-site air
concentration.
Used air emission
release rate model
based on physical
calculations of
2,3,7,8-TCDD
partitioning and
transfer, Used factor
previously desired in
a PCB assessment to
account for effect of
a clean cover, if
present.
Short-range disper-
ion modeling used.
Measurement data on
air release are not
available. Models are
commonly used for
estimation of emis-
sion rates. The model
applied has been peer
reviewed and
published by BAG.
Gaussian dispersion
modeling not applic-
able for short-range
dispersion.
Models for air re-
lease rates are not
validated on a field
scale. Models de-
scribe physical setting
fairly accurately.
Substan-tial
uncertainties are
associated with input
parameters in-
cluding Henry's law
constant, partition
coefficients.
Model not validated
on a field scale.
Wind speed and dis-
persion coefficients
are the most uncer-
tain input parame-
ters.
The value for water-soil
partition coefficient most
uncertain. Emis- sion
rate can change by
2 orders of magnitude
depending upon the value
of the partition coeffi-
cient. Wicking effects
have the potential to
increase air emissions
under some circum-
stances; not quantita-
tively evaluated.
Site-specific meteoro-
logic conditions will
influence estimate.
Off-site air
concentration.
Industrial Source
Complex Model used
with modifications
applicable to area
source emissions.
This air dispersion
model is widely
accepted by the
Agency.
Meteorological input
data are site-specific.
Default values used
for risk estimation.
Input data can be site-
specific.
Inhalation rate: See Dust Inhalation, Table 7-5.
Evaluation: Substantial uncertainty is present in air emissions estimates; transport modeling is a smaller source of
uncertainty.
279
-------
doses were utilized: 26 ng, 350 ng, and 1,300 ng; the percentages of the applied dose
found in the liver were "circa 0.05," 1.7 + 0.5, and 2.2 + 0.5. However, only the two
higher dose measurements were in the range for which the authors reported that
"reproducible quantities of radioactivity in the liver were measured." The absorption
estimate in this report utilizes the geometric mean of the extreme values (0.07 and 3.0%)
for an estimate of 0.5% (an absorption fraction of .005). Since this value is a factor of
four below the more reliably established higher dose values, absorption (and thus
exposure) may be underestimated by this factor. However, as the contaminated dirt was
in contact with rat skin for 24 hours, the experiment may over-represent absorption
compared with the case where skin would be washed periodically.
The levels of 2,3,7,8-TCDD in dirt on skin are assumed to be equal to the 2,3,7,8-
TCDD in soil; thus, for the off-site scenarios, these estimates rely on the soil
concentration estimates for which the substantial uncertainties have been discussed above.
In summary, dermal exposure estimations rely on a number of parameters whose values
are not well established; principally: skin levels of dirt, dermal absorption, and off-site
soil concentration estimates are subject to substantial uncertainty or site-to-site
variation. Therefore, the estimates for this pathway can be improved substantially if the
underlying data base is strengthened. Table 7-7 is a summary of the uncertainties
related to dermal contact.
g. Soil Ingestion
Estimated soil ingestion is based on field measurements, using trace elements, of
quantities of soil ingested by relatively small groups of children over brief
periods.Methodological issues in these studies remain to be addressed, particularly
ingestion estimates may have been lower if dietary intake of the trace elements was
taken into account and BAG is conducting research to refine soil ingestion estimates
obtained through trace element measurements. Given the available data, 0.2 g/day is
280
-------
TABLE 7.7. LANDFILL ASSESSMENT
DERMAL CONTACT
Assumption/
Method
Absorption
fraction.
Approach
0.5%.
Rationale
Based on Poiger &
Schlatter (1980).
Uncertainty
May be low since data
supporting upper end
of range (396) more
reliable.
Comments
Percutaneous absorption
is influenced by various
factors, including in-
trinsic skin properties
and environmental con-
ditions.
Exposure
duration.
7,300 - 20,000 days.
Contact rate. 1 mg/cm -day.
Based on typical and
reasonable worst
case scenarios
assumptions.
Studies by Lepow et
at. (197S); Reels
et al. (1980) and
Hawley (1985).
Can vary for popula-
tions in rural
settings (especially
for adults).
Range reported in
literature is .5 to
1.5 mg/cm day.
Actual range is
probably much wider.
Time spent outdoors will
be function of geogra-
phic and climatic condi-
tions.
This approach does not
distinguish between
adults and children or
outdoor and indoor
exposures.
Evaluation: The lack of supporting data and large influence of personal habits make this pathway very uncertain, by
at least two orders of magnitude.
281
-------
used as a typical value for soil ingestion in young children. Due to the behavior known
as pica, some children are known to be high ingesters of various non-food materials;
although no. quantitative data on soil ingestion are available for children known to
exhibit pica, the use of the high-end estimate of 1.0 g/day may better reflect such
behavior.
Because of remaining methodological research needs, no quantitative estimate of the
uncertainties in these estimates is made here.
The gut absorption fraction for 2,3,7,8-TCDD from soil is a second major
determinant of exposure. Data from experiments with rodents, reviewed in Chapter 5,
are consistent with, the 30% absorption fraction utilized for the pathway assessments.
Four of the five tested soils are in agreement with an absorption fraction of this
magnitude. One soil sample showed absorption one or two orders of magnitude lower,
based on limited data. Therefore, sites may be encountered where 2,3,7,8-TCDD in soil is
substantially less available than assumed in the scenario calculations.
Soil ingestion exposure estimates also depend on the duration of the period over
which children are assumed to ingest soil. Data on soil ingestion by age are not
available, and the estimate that significant ingestion occurs between ages 2 and 6, is
broadly supportable on behavioral grounds.
No measurement data are available on soil ingestion in infants (0-2 yrs. old) or in
older children or adults, and no ingestion is assumed for these groups. While some soil
ingestion will occur in these groups, e.g., through contact of soiled hands with food, it is
plausible that such ingestion is of a lesser degree than occurs in early childhood. If
Hawley's (1985) estimate that an adult ingests an average 0.060 g/d of soil is used, after
accounting for differences in exposure duration (70 yr vs. 5, yr) and body weight (70 kg
vs. 17 kg), the adult soil ingestion risk is close to the estimated risk for children (at 0.2
g/d).
282
-------
Considering these uncertainties, the soil ingestion exposure estimates presented for
contaminated soil are plausible. In the landfill scenarios, the estimated soil
concentrations are influenced by uncertainties, previously discussed, in the soil dilution
factor. Table 7-8 is a summary of the uncertainties related to soil ingestion.
h. Beef and Milk Fat Ingestion
Major determinants of 2,3,7,8-TCDD exposure through the intake of bovine products
are the relative concentrations of dioxin present in milk or meat fat and the soil
concentrations in the cattle's pasture or pen. No field data for these 2,3,7,8-TCDD
distribution ratios are available. Jensen et al. (1981) determined 2,3,7,8-TCDD
concentrations in the fat of cattle fed a diet containing 2,3,7,8-TCDD present as a
contaminant of the herbicide 2,4,5-T. The 2,4,5-T was placed in a silica gel before
incorporation into a feed mixture. After 28 days of feeding, the ratio of beef fat to
dietary 2,3,7,8-TCDD concentrations was about 4. However, there was no indication that
a steady state had been reached, and a kinetic model developed by the authors suggests
that 2,3,7,8-TCDD concentrations in fat at 28 days would still be increasing rapidly at
this time. Additionally, there are no data on the absorption of 2,3,7,8-TCDD from the
herbicide-gel diet mix as compared with soil. Based on comparisons in rodents,
absorption from diet may be substantially higher than from soil.
Jensen and Hummel (1982) similarly fed cattle 2,3,7,8-TCDD in 2,4,5-T added to
silica gel mixed into the diet for up to 21 days. These limited data suggest a milk fat
to diet distribution ratio of 2,3,7,8-TCDD of as high as 6 (assuming that 2,3,7,8-TCDD in
milk cream is in the fat component, where the cream is assumed to contain 30% fat).
The authors'- data are suggestive of a relatively rapid approach to a steady-state
concentration of 2,3,7,8-TCDD in milk. Thus, while this ratio is unlikely to be at steady
state, it may be approaching it. As noted above, the relative absorption of 2,3,7,8-TCDD
from the prepared diet and from soil has not been established.
283
-------
TABLE 7.8. LANDFILL ASSESSMENT
SOIL INGESTION
Assumption/
Method.
Approach
Rationale
Uncertainty
Comments
Soil contamination level: see'Soil Dilution Factor (Table 7-1)
Child's inges-
tion rate- (2-6'
yeamold)..
Ingestion rate assum-
ed to vary from 0.2-
1.0 g/d.
The range selected
was primarily baaed
on the results of two
field studies of soil
ingestion in children.
Field study method-
ology not fully
validated. Data from
several sources
indicate this range of
values for small
children.
Pica.children have been
estimated to. ingest
higher quantities (5 g/d).
Ingestion rate
for other, ages.
Absorption
fraction..
Ingestion assumed
to occur only, during
ages 2-6.
Absorption estimated
at 30%.
Mouthing tendencies
strongest and under-
standing of personal
hygiene low during'
ages 2-6.
Based on results of
2,3,7,8-TCDD absorp-
tion for several soil
samples in experimen-
tal animals.
Hawley estimates in-
advertent ingestion
may be 60 ug/d for
adults, which would
lead to an estimated
soil ingestion risk
over 64 years com-
parable to childhood
risk before age 7.
Four of five tested
soil samples showed
absorption of this
magnitude. The fifth
sample indicated much
less absorption. Thus,
estimate may not be
accurate for all soils
encountered.
Differences may exist
in human and rodent
gut absorption.
Adults may inadvertently
ingest soil during
gardening and yard work.
Evaluation: The assumption of inadvertent soil ingestion for 2-6 year1 old children as. 0.2. g/d is a reasonable one based
on current data. This value does not account for children.with pica tendencies. Future studies may revise this
estimate, since the methods used to derive the 0.2 g/d value are still being developed and evaluated. Adult: inadvertent
soil ingestion was not specifically included, although it might be argued that exposure at much reduced levels from
ages 7-10 would lead-to comparable'exposures to those during ages 2-6.
284
-------
These studies are reviewed for comparison with the results of the Fries (1986) study
of beef fat/soil and milk fat/soil ratios for polybrominated biphenyls (PBBs) for cattle on
largely bare dirt lots in Michigan. Fries reported beef fat/soil ratios of 0.3 to 0.4 for
beef cattle and milk fat/soil ratios of 0.02 to 0.06 for dairy cows.
In Chapter 3, the data for PBBs from Fries (1986) were used as surrogates to
estimate the distribution ratio for 2,3,7,8-TCDD. Given the results of Jensen et al.
(1981) and Jensen and Hummel (1982), the difference between the milk fat and dairy fat
ratios in Fries (1986) may not agree well with the situation for 2,3,7,8-TCDD. (The
absolute difference between the 2,3,7,8-TCDD ratios and the PBB ratios probably also
reflects differences in absorption.)
A variety of other studies with chlorinated hydrocarbon compounds (reviewed in
Fries, 1982), while not allowing comparisons between beef fat and milk concentrations in
the same animals, do not suggest that the milk fat distribution ratios should be lower
than the beef fat distribution ratios.
A second approach to distribution factors can be made using the data of Jensen et
al. (1981) and Jensen and Hummel (1982). As noted above, absorption of 2,3,7,8-TCDD
from soil may be less than that from diet. When rodent data are used 'for comparison,
results from Fries and Marrow (1975) indicate gut absorption of 2,3,7,8-TCDD in rats as
50% to 60% of the administered dose. On the other hand, typical 2,3,7,8-TCDD
absorption from soil in rodents, as discussed in Chapter 5, is on the order of 20% to 40%.
Very roughly, absorption from soil appears to be half of that from diet. If this same
difference applies to cattle the results of the Jensen studies can be used to estimate
fat/soil distribution ratios. If absorption from soil is a factor of two lower than from
diet, and if 8% of dry diet is soil, fat/soil ratios will also be a factor of 2/.08 = 25
lower than fat/diet ratios. Accordingly, the data of Jensen et al. (1981) suggest a beef
fat/soil ratio of 0.16; actually, a higher value would be anticipated because the 2,3,7,8-
285
-------
TCDD concentrations in Jensen et al. (1981) were probably not close to equilibrium.
Using the same approach, the Jensen and Hummel (1982) study suggests milk fat/diet
ratios on the order of 0.25, or somewhat higher if the milk fat TCDD concentrations
were not in equilibrium.
Thus, using this second approach to the distribution ratios suggests a beef fat/soil
ratio similar to the 0.3 to 0.4 used in pathway calculations, but indicates that the milk
fat/soil ratio may be an order of magnitude higher than estimated.
Methodologies exist to gather more reliable data in this area, and would aid
substantially in reducing uncertainties.
Several other factors enter into the calculation of exposures through milk and beef
contamination. Data on rates of milk and beef consumption were taken from surveys,
and form an adequate basis for evaluating typical and reasonable worst-case product
intakes.
The fraction of meat or milk intake coming from a local contaminated source was
selected on the basis of a survey of 900 rural farm households (U.S. Dept. of Agriculture,
1966). The values selected are intended to apply to the situation of a farm family that
slaughters its own beef and maintains its own dairy cows. In such a situation, the
fraction of intake from the farm can be expected to be substantial. Additionally, if
other families in the vicinity obtain beef and milk directly from a nearby contaminated
farm, similar percentages may reasonably apply. The above analysis is not applicable to
individuals living near a contaminated site who obtain their beef and dairy products from
regular commercial sources. Such a situation can be expected to occur frequently. If
beef and dairy products raised on a contaminated site are sold commercially, population
risk from these activities should be addressed; such an analysis was not part of the
scenario assessments here.
Beef and dairy 2,3,7,8-TCDD concentrations will vary with the soil contamination
286
-------
level. In the off-site scenarios the uncertainty in estimating soil concentrations
influences estimated risks.
Finally, data on human absorption of 2,3,7,8-TCDD from dietary sources are not
available, data on rodents' absorption from diet are used to estimate human absorption.
While this provides a basis for making a plausible estimate, no quantitative evaluation of
uncertainty is made. Table 7-9 summarizes the uncertainties associated with beef and
milk ingestion.
i. Fish Ingestion
The 2,3,7,8-TCDD concentration in fish tissue is estimated using a distribution ratio
between fish tissue and sediment on the basis of information that sediment levels are a
driving influence on fish tissue levels. As discussed in Chapter 3, many factors influence
2,3,7,8-TCDD levels in fish, these include whether the fish is a bottom or surface feeder
and the position of the fish on the food chain. A variety of site specific factors may
influence 2,3,7,8-TCDD concentrations in fish tissue that occur in different settings.
Therefore the use of a single fish/sediment distribution ratio, as done in the fish
pathway assessment, must be recognized as a broad approximation.
Chapter 3 suggests that typical fish/sediment distribution ratios are in the range
1-10; some reports cited indicate that the range may be somewhat broader. Thus, given
the distribution factor of 5 issued in the pathway assessment, plausible concentrations of
2,3,7,8-TCDD in fish must include values from 1/5 to 2 times the calculated levels. If
information on fish contamination is available for a specific site, then adequate
measurement data should supercede values calculated using distribution factors.
Two estimates of human fish consumption were utilized, 6.5 g/d and 30 g/d. The 6.5
g/d is a national average figure for fresh water and estuarine fish and shellfish. Thus,
this figure represents a typical and relatively low (approximately 15-25 fish meals in a
year) consumption rate. The 6.5 g/d figure is based on data now over a decade old, and
287
-------
TABLE 7.9. LANDFILL ASSESSMENT BEEF AND MILK INGESTION
Assumption/
Method
Approach
Rationale
Soil Contamination Levels: .See Soil Dilution Factor (Table 7-1)
Distribution Factors
Uncertaintuv
Comments
Beef
Used empirical data
for PBBs in feedlot
soils, giving beef
fat/soil ratio of
0.3 - 0.4 (Fries,
1985) as surrogate
for 2,3,7,8-TCDD
beef fat/soil data.
No 2,3,7,8-TCDD fat/
soil .field data
available. PBBs
thought to be a
reasonable surrogate
compound (high lipid
partitioning; slow
removal from body).
Analogy with PBBs
may be in error.
However, alternate
calculation of factor
using laboratory data
on 2,3,7,8-TCDD in
feed and assumptions
on absorption from
soil support this
result.
As discussed in Section
C-2 of Chapter 3, details
of farm management are
very important
(e.g., duration of feed -
lot stay).
Dairy
Empirical data for
PBBs in feed lot soils
gives milk fat/soil
ratios of 0.06 for
primiparous to 0.02
for multi-
parous cows. Used an
average of these, 0.04.
No 2,3,7,8-TCDD fat/
soil field data
available. Beef fat
comparisons suggested
PBBs were reasonable
surrogate.
Analogy with PBBs
may be in error. An
alternate calculation
with laboratory data
on absorption on
2,3,7,8-TCDD from
feed, led to a higher
estimate of distri-
bution ratio.
Same as above (e.g.,
lactating dairy cows
pastured?)
Ingestion Rates
Beef
Dairy
Recommended 14.9 •
26.0 g/d based on
literature.
Recommended 18.8
43 g/d based on
literature. Alter-
natively, 8.9 - 10.7
g/d for fresh milk
only.
Range provided en-
compasses averages
from five studies
based on three
surveys.
Range provided en-
compasses averages
from three studies
based on two sur-
veys.
Shape of distribution
of consumption not
well defined,
particularly the
extremes.
Study and survey
yielding 43 g/d (U.S.
EPA, 1981b) less well
documented than
remaining two studies
and survey. Ignoring
U.S. EPA (1981b),
range would be 18.8 -
24.1 g/d. Again,
distribution not well
defined.
288
-------
TABLE 7.9. (CONTINUED)
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Absorption
fraction.
Adopted single value,
0.68.
Average of 2,3,7,8-
TCDD absorption from
commercial rodent
feed by rate (O.S) and
absorption from
gavage mixture with
corn oil and acetone
by rate (0.86).
No data on absorption
after ingestion of
animal products
available. Human and
rodent absorption may
differ.
Fraction of food
from local source
Beef
Dairy
44% of beef diet from
contaminated source.
40% of dairy products
from contaminated
source.
Data available for %
of annual consumption
homegrown by rural
farm households (44%)
(U.S. Dept.
Agriculture, 1966).
Same data source as
for "beef above.
Likely to be sub-
stantial difference
between individuals.
Some from families
will get no beef from
local source.
Same concern as
noted for "beef"
above.
Evaluation: Analysis based on survey and experimental data on distribution ratio, absorption, and consumption. The
beef estimate is judged more reliable than milk estimate which may be low due to uncertainty in the distribution ratio.
Substantial uncertainty in soil concentration of 2,3,7,8-TCDD in off-site scenarios affects estimates. These estimates
apply only to individuals who raise their own beef or dairy cattle.
289
-------
fish consumption may have risen somewhat .in the .intervening period, however .this value
still appears Jo :be .a reasonably .typical value. This average value includes non-consumers
.of fish .in -the denominators; rates for consumers .of fresh water wish may be substantially
higher. The .value of .3.0 g/d is based on more limited data on fish consumption by
'fishermen :or high consumers of locally caught fish (see Chapter 4). These data
.demonstrate that the .intake for high consumers can exceed 30 g/d, with consumption
-rates/reported .by rsome .individuals being .on the order of .20.0 g/d. Thus 30 g/d is a
^reasonable value for .regular .consumers of locally caught fish.
•Kor.the reasonable worst, case analysis 1.0% of the 3.0 g/d fish consumption was
assumed to be from the local, contaminated .source. For the more typical case a value of
10%.of'.6.5..g/d was adopted. These figures .were chosen using the joint Judgment of "EAG
staff • as .survey .data were .not .available. The figures .selected were regarded as
.appropriate for individuals who made .use of .locally caught fish. Neither figure is
intended as an average .for the rural population. These estimates also assume that the
.contaminated water bodies have adequate .biological capacity to support regular fishing.
Clearly, both fishing habits and biological .capacity .of particular ponds and streams will
-vary-between .localities. Thus, the.appropriateness of exposure estimates .from the fish
consumption pathwa.y is best evaluated on .a .site specific basis where local data may aid
in 'the development of appropriate estimates. The reasonable worst case assumptions,
particularly that only 10% of .fish .eaten comes from the contaminated source may be
substantially low for an individual who engages in fishing to obtain .an important dietary
component.
Rodent data on absorption .of 2,3,7,8-TCDD from the diet were used to estimate
human ;gut absorption for food products. This .approach .provides the basis for generating
plausible estimates of absorption but does not allow a quantitative Table 7-10 evaluation
of uncertainty. For off-site scenarios there is .considerable uncertainty in the estimated
290
-------
stream sediment concentrations as is discussed separately above. Table 7-10 is a
summary of uncertainties related to fish ingestion.
j. Water Ingestion - Surface Water
The surface water concentration of dioxin in the pond scenarios is estimated using
a mass transfer calculation based in the level of 2,3,7,8 TCDD level in pond sediments.
Two resistance mass transfer calculations were used to estimate the movement from
sediment into the water and from water into the air. The assumption is then made that
the loss of 2,3,7,8 TCDD from sediment to water equals the loss from water to air. That
is, the water concentration is assumed to be relatively constant, and the only significant
loss mechanism is assumed to be volatilization. Other potential losses, including
chemical, photolytic or biological degradation in the water column, are assumed to be
negligible. If other loss mechanisms are substantial the estimated water concentration of
2,3,7,8-TCDD would decrease.
The two resistance calculations lead to Equation 3-4 in Chapter 3. This relationship
depends on several parameters subject to uncertainty. The factor Kw, the water-side
mass transfer coefficient above the sediment, was estimated from using a relationship
(3-5) empirically tested in a laboratory tank. Among other factors, Kw is estimated to
be proportional to the 5/4th power of water depth to and to be inversely proportional to
fetch, the length of the water body crossed by the moving air. A limitation in the
application of this relationship is that the experiments (Thibodeaux and Becker, 1982)
utilized a tank much smaller than the pond assessed here. The fetch of the tank was no
more than 2.4 m vs. 64 m for the pond. Furthermore, the test data presented by
Thibodeaux and Becker do not clearly indicate that Kw increases with fetch beyond 2 m,
therefore the possibility exists that Kw is under-estimated by an order of magnitude due
to the extrapolation to the larger pond. Similarly the lab data measured depth effects
291
-------
TABLE 7.10. LANDFILL ASSESSMENT
FISH INGESTION
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Sediment Concentration: See Sediment Dilution Factor (Table 7-1)
Distribution
ratio.
Fish/sediment ratio
chosen. (Single value,
5).
Sediment concentra-
tion judged principal
determinant of fish
tissue concentration.
Ratio chosen also
encompasses fish/
water partitioning as
sediment concen-
tration tends to
determine water
concentrations.
Substantial range in
field and laboratory
measurements (<1-10);
equilibrium not
reached in most lab
experiments; species
specific; factors in
migration, feeding
habits, lipid content
etc., will affect
distribution. Dis-
tribution factor may
depend on sediment
level.
Judgment imposed in
choice of approximate
mid-point of range 1-10.
A less conservative
single value of 2 might
be justifiable as repre-
resenting preponderance
of higher values from
several studies.
Consumption
rates.
Used "customary* 6.6
g/d for typical
scenarios and 30 g/d
for reasonable worst -
case scenarios.
6.5 g/d is most
commonly cited single
figure in EPA studies,
while 30 g/d
derived from sizable
study of fish
consumption by recre-
ational fishermen.
Differences in fish
consumption between
population groups has
been shown. Much
individual variation in
consumption has been
shown. Many rural
families will make no
use of locally caught
fish.
In actuality, the popu-
lation is likely to be
divided into those who
eat no freshwater/
estuarine fish and those
who eat more than the
6.6 g/d.
Fraction of fish
consumption
from local
sources.
chosen.
Values chosen by
judgment in the
absence of data, based
on how often fish
were thought to be
caught from the
scenario water bodies.
Productivity in fish
consumption between
population groups has
been shown. Much
individual variation in
consumption has been
shown. Many rural
families will make no
use of locally caught
fish.
Large uncertainty; in
site-specific situations
local data must be used.
Evaluation: Overall, the analysis represents a reasonable compromise between values for various inputs. However,
substantial overall uncertainty exists in the estimates resulting from uncertainties in sediment levels, distribution ratios
and fish consumption habits of local populations.
292
-------
only up to 0.8 m and the assessed pond is assumed to have a depth of 5 m, and the
extrapolation of the depth effects is uncertain.
The factor K.La, the overall mass transfer coefficient for the air-water interface
was calculated using two resistance theory, requiring the use of the Henry's law
coefficient and assumptions about wind speed and other parameters. These methods for
estimating volatilization from the water surface have been tested using comparisons of
predictions with measured emissions from liquid in waste disposal facilities (summarized in
Hwang, 1985). The predicted volatilization based on two resistance theory calculations
were generally within a factor of two of the measured values, with deviations reaching
an order of magnitude in some comparisons. It should be noted that these empirical
tests were conducted with compounds much more volatile than 2,3,7,8-TCDD.
The parameter Ke, the sediment-side mass transfer coefficient, was estimated using
a relationship depending on the diffusivity of 2,3,7,8-TCDD and the sediment depth and
porosity. The empirical support for this relationship is not reviewed in this report.
Finally, the estimated water concentration depends on Kj, the soil/water partition
coefficient for 2,3,7,8-TCDD. There is considerable uncertainty in this parameter as
discussed in the section of this chapter discussing air emissions of 2,3,7,8-TCDD from
soil.
If, as calculated in Chapter 6, Kw is substantially higher than both Ke and
then Equation 3-4 for the water concentration, Cw, of 2,3,7,8 TCDD is closely
approximated by the simplified relationship:
Fe_Ce
Cw = KLaLd (7-2)
Thus, when Kw is large it does not affect estimated water concentration which is
then directly or inversely proportional to the other factors discussed here. The water
293
-------
concentration, Cw, is also proportional to the sediment concentration, Ce, of 2,3,7,8-
TCDD, which will .also involve substantial uncertainty for off-site calculations.
In summary, the simplified model used to calculate the levels of 2,3,7,8-TCDD in
water was developed theoretically and has not been validated. Several of the parameters
entering into the model are themselves subject to substantial uncertainty. Table 7-11
-summarizes these uncertainties.
The analysis for the scenarios with a stream are analogous to the pond calculations.
It is assumed that a one acre area of stream bed is contaminated with 2,3,7,8-TCDD at
the concentration of .01 times the concentration in the landfill (see discussion of
sediment dilution factor). For comparison this area would be equivalent to a length of
0.8 km for a stream 5 m wide.
If 2,3,7,8-TCDD is present in sediment to an average depth of 1 cm, 1 acre of
contaminated sediment has a mass (at 1.7 g/cm^) of 6.9 x 104 kg. Noting the .001
dilution factor, 69 kg of contaminated soil from the landfill must have reached the
stream to produce this level of contamination. For comparison the estimated eroded
material from the landfill (10 acre/grass covered) from Chapter 6 is 5600 kg/yr. Thus
the estimate that only one acre of stream bottom is contaminated is modest. A much
larger area of contamination would be consistent with the landfill erosion estimates
leading to higher estimates of 2,3,7,8-TCDD release. On the other hand, the assumption
used to calculate the sediment dilution factor are uncertain in regard to a stream of the
substantial size being assessed here. Additionally, the assumed stream depth (5 m) maybe
higher for the average depth of the stream envisioned in the scenario, and stream depth
has an inverse relationship to the estimated water concentration of 2,3,7,8-TCDD.
The use of Equation 3-4 to calculate the water concentration of 2,3,7,8-TCDD relies
on steady state assumptions that can be less strongly defended in the case of .a stream
flowing over sediment than in a pond situation. In consideration of the issues mentioned
294
-------
TABLE 7.11. LANDFILL ASSESSMENT
SURFACE WATER CONTAMINATION
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Pond sediment
concentration.
Water concen-
tration - pond
See Sediment Dilution
Factor - Pond. (Table
7.2)
Used mass transfer
calculation to esti-
mate water concen-
tration as a function
of sediment concen-
tration. Model
depends on assumption
that water concentra-
tion is steady state
reflecting balance
between sediment to
water and water to
air transfers.
Model calculations
address physical
processes involved in
determining transfer
to dioxin to the water
column. Moni-
toring data to esti-
mate water concen-
tration vs. sediment
concentration were
not present.
Model has not been
field tested, al-
though field data
exist to support the
water to air transfer
estimates at least for
more volatile
compounds. Some
model parameters e.g.,
soil water partition
coefficient have
substantial
uncertainty.
In these scenarios es-
timates of risks from
drinking pond water are
not presented because
this practice is con-
sidered unlikely to occur
commonly.
River sediment
concentration:
Water concentra-
tion - river.
See Sediment Dilution
Factor - stream.
(Table 7.2)
Used mass transfer
calculation similar to
stream approach but
adapted to a river
flowing at 1 m/s.
Contaminated sediment
assumed to cover 1
acre area of stream
bed.
See above.
See above. The
assumption that
sediment covers only
1 acre area of stream
bed may be low given
the assumed landfill
erosion rates and
sediment dilution
factor sediment could
be contaminated over
much larger area.
Evaluation: Applied a simplified theoretical model that has not been field validated, several model parameters are
subject to substantial uncertainty. For river calculations the model assumptions assumed area of contamination is also
uncertain, however in the river scenarios calculated water concentrations are so low that uncertainties may not have
practical importance.
295
-------
above, the stream 2,3,7,8-TCDD concentrations have substantial added uncertainty beyond
the uncertainty present in the pond water calculations [see Table 7-11].
k. Ground Water Contamination
Contamination of ground water through 2,3,7,8-TCDD leaching from a landfill into a
shallow underlying aquifer was assessed. This assessment made several assumptions that
may lead to high estimates of ground water contamination. Contaminated leachate was
assumed to directly enter ground water under the entire contaminated area. Ground
water was assumed to move with a high speed of approximately 1 mile/year. The
leachate reaching ground water was assumed to be saturated with 2,3,7,8-TCDD at 8
ng/L. Because of the high soil/water partition coefficient for this compound, in the
absence of cosolvent effects, a leachate concentration below the solubility level may be
anticipated for a landfill containing 1 ppb 2,3,7,8-TCDD or less.
To calculate the retardation factor for 2,3,7,8-TCDD movement in ground water, the
estimated organic carbon/water partition coefficient of 486,000 was proportioned down by
the assumed fraction, 0.0002, of organic carbon in the aquifer matrix. Given the low
organic carbon content assumed, the procedure of estimating the media/water partition
coefficient by the fraction of organic carbon present is speculative. Other constituents
of the aquifer matrix may also retain 2,3,7,8-TCDD, leading to a higher matrix/water
partition coefficient. This would lead to a lower estimate of the concentration in ground
water. The value for the organic carbon/water partition coefficient is itself subject to
uncertainty of at least an order of magnitude.
The standard ground water transport model applied under these assumptions
predicted very low levels of 2,3,7,8-TCDD in the water after a period of 100 years. It
should be remembered that the analysis predicting even these low values used higher
range assumptions, however, co-solvent effects were assumed to be absent. A
quantitative analysis of co-solvent effects which could increase transport of 2,3,7,8-
296
-------
TCDD is not included here. For a discussion of these effects see Chapter 3. Table 7-12
summarizes uncertainties related to ground water contamination.
1. Plant Uptake
Estimates of exposure through human or animal consumption of plants growing in
contaminated soil are very uncertain. As discussed in Chapter 3 and Chapter 4 (under
Description of Exposure Scenarios for Landfills) the limited available data indicate that
plants are capable of uptake of 2,3,7,8-TCDD from soil. However, because of the sparsity
of data only a speculative quantitative estimate of uptake was made. The estimate
assumed plant concentrations on average, are on the order of 2% of soil levels.
However, the substantial uncertainty in such a choice should be recognized as some data
indicate that plant roots may have levels exceeding soil levels. See the discussion in
Chapters 3 and 4 for more background on the variability of existing data on plant
uptake.
B. INCINERATION SCENARIOS
All exposures and risks resulting from incinerator stack emissions are closely
proportional to the quantity of 2,3,7,8-TCDD in these emissions. This quantity is the
product of the emission rate per unit waste combusted and the plant size. Wide
variation of the emission rate of 2,3,7,8-TCDD per unit waste combusted have been
reported with this assessment using a high and low emission rate differing by a factor of
300. Some data suggest that emission rates an order of magnitude either above or below
the range used in this assessment occur. Incinerator capacities also exhibit much
variation with the smaller plant used in this assessment having a capacity of 120 tons
per day and the large plant having a capacity of 3000 tons per day; the latter being
selected to represent planned facilities.
Inhalation exposures were estimated using a modified Industrial Source Complex
model. Both building wake effects and precipitation may increase ground level pollutant
297
-------
TABLE 7.12. LANDFILL ASSESSMENT - GROUNDWATER CONTAMINATION
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Concentration of.
2,3,7,8-TCDD in
leachate.
Groundweter.
flow model.
Assumed.leachate
concentration equal to
the solubility of.
2,3,7,8-TCDD in
water. Solubility/of 8
ng/L assumed.
Standard, groundwater
transport model
applied, using a re-
tention factor for
2,3,7,8-TCDD based
on calculated
groundwater media/
water partition
coefficient, an
assumed groundwater
flow rate, etc.
Presented, worst case
for 2,3,7,8-TCDD in
clean, water, noting
that the concentra-
tion, could be higher
for leachate with
substantial organic
content.
Standard modeling
approach used by EPA
to calculate ground -
water concentrations.
In the absence of
cosolvent effects, the
concentration in
leachate is expected
to be lower than 8
ng/L for the soil
concentrations in
range shown in this
document (< 1 ppb).
Some of the
assumptions in the
model were selected
to maximize estimated
transport (e.g.,
groundwater flow
rate). Thus, model
predictions represent
worst case situation
in absence of co-
solvent effects. Some
model parameters e.g.,
partition co-
efficients subject to
substantial uncer-
tainty.
See Chapter III for
discussion of importance
of cosolvent effects; no
quantitative analysis of
these effects has been
developed.
Evaluation:: This-analysis, using groundwater modeling-approach generally used-by EPA, indicated that minimal
transport of 2,3,7,8-TCDD in groundwater occurred. This was true even though the analysis incorporated some
assumptions that would .tend to maximize transport. Co-solvent effects would modify this analysis and increase
groundwater transport, however no quantitative analysis of this effect is yet available. Consequently, one should be
very cautious'of applying the conclusions reached here to an actual site, unless a great deal is known about the nature
of the' landfill and leachate.
298
-------
concentrations and their effects on air concentrations and particulate deposition rates
were evaluated. The uncertainties in air transport modeling are likely to be smaller than
uncertainties in the stack emission rates of 2,3,7,8-TCDD. Inhalation exposures are
assessed for an individual living 1/2 mile from the incinerator stack. Maximum exposure
occurred very close to the facility and were estimated to be five times higher. Exposure
levels were reduced by two to three orders of magnitude at a distance of 60 miles.
Exposure to 2,3,7,8-TCDD through beef ingestion, dairy products ingestion, soil
ingestion by children, and soil dermal contact (listed in decreasing order of estimated
exposure) were evaluated using similar assumptions as in the land-related scenarios.
1. Emissions Data
Data on 2,3,7,8-TCDD stack emissions from municipal waste incinerators (MWI) exhibit
wide variation, with the extremes of the range of data reported in Chapter 6 being 5.6 x
10~5 to .289 Mg/kg (ug 2,3,7,8-TCDD/kg waste combusted). Data for one incinerator
operated without emission controls under starved air conditions showed a total TCDD
release of 72.8 A»g/kg, with no value reported for 2,3,7,8-TCDD for this incinerator. In
comparison 1.1 /ig/ng total TCDD emissions were reported for the incinerator that
showed 0.289 A
-------
incinerator showed 436 ng/g total TCDD;. this suggests a 2,3,7,8-TCDD content of fly ash
of approximately 20 ng/g (based on a rough ratio 20 ng total TCDD/ng 2,3,7,8-TCDD
obtained using the results in Table 6-11).
2.. Selection of Model Incinerator and Exposure Scenario
Incinerators of sizes 120 TPD and 3000 TPD were selected for the scenarios. These
represent a common smaller size and one- of the largest size incinerators. These sizes are
appropriate considering both the existing stock of incinerators and incinerators now being
constructed or planned.
The analysis assumed that the large incinerator had a 2,3,7,8-TCDD emission rate
(0.001 Mg/kg) at the lower end of the range of measured values. If an incinerator of
this size exhibited a high end value of emission, 2,3,7,8-TCDD release would be a factor
of 300 higher. The analysis for the small incinerator scenario used the highest measured
stack emissions of 0.289 /*g/kg. Thus, while even higher release rates might be seen
from uncontrolled incinerators operating under certain conditions, this 2,3,7,8-TCDD
release is a reasonable worst case estimate for an incinerator of this size.
The two incinerator models selected should be regarded as illustrative examples with
the variability of measured emission rates of 2,3,7,8-TCDD kept in mind. No review was
conducted to determine the degree the engineering parameters (e.g., stack height) vary
between incinerators. Table 7-13 summarizes the uncertainties related to plant sizes and
emission rates.
3. Inhalation and Surface Deposition
Using the model incinerators and emissions estimates discussed above, pollutant
dispersion models were used to predict air concentrations and surface deposition rates for
2,3,7,8-TCDD.
The transport model utilized was a modified industrial source complex (ISC) model
which took into account two features that may increase ground level pollutant
300
-------
concentrations.
a) Buildings near the incinerator produce a wake effect leading to higher ground
level concentrations.
b) The modified ISC model incorporated a "tipdown wash" effect through which
precipitation increased surface pollutant concentrations.
The ISC models are standard tools utilized by EPA to estimate pollutant
concentrations from air emission source. A proprietary version of the model was used to
incorporate tipdown wash and EAG has not reviewed the procedure through which this
effect was calculated, although the effects of tipdown wash on the concentration in the
scenario was small.
An assumed particulate size distribution is required to calculate surface deposition
of the particulate emissions. The size distribution assumed is one used by the Office of
Air Quality Planning and Standards in the modeling of incinerator emissions; EAG has not
reviewed the variability of particulate size distributions for incinerators.
Comparison between predicted surface deposition rates using different model
assumptions (with or without effect of precipitation, over different averaging periods)
produced similar results; the estimated quantities of particulate from wet deposition are
smaller than those from dry deposition.
Predicted air concentrations and deposition rates at 1/2 mile (0.8 km) from the site
are used for risk calculation. These concentrations reflect annually averaged
concentrations utilizing local meteorologic data for the two model incinerators (in Florida
and Virginia). It should be noted that the highest 2,3,7,8-TCDD concentrations were
predicted to occur at 200m from the incinerator stack and that these levels were
approximately a factor of five higher than those predicted at 0.8 km. For comparison
estimated concentrations at 100 Km from the site are also calculated and were generally
two to three orders of magnitude lower than levels at 0.8 km. Table 7-14 summarizes
301
-------
TABLE 7.13. INCINERATOR ASSESSMENT
AIR EMISSIONS ESTIMATE
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Incinerator
capacity.
Two plant sizes 120
tons/d, and 3000
tons/d were assessed;
these plants represent
an existing facility in
Virginia and a planned
facility in Florida,
respec- tively.
Smaller plant size is
typical of current
facilities. Larger
plant represents a
large, planned facility.
Existing and planned
plants vary widely in
sice. Many existing
plants are less than
120 tons/d. 3000
tons/d is near upper
limit of planned plant
sizes.
Methodology can be
applied to plants of all
sizes.
Emission rate.
2,3,7,8-TCDD
concentration in
fly ash.
2,3,7,8-TCDD emission
rate for smaller plant
taken as 0.289 ug/kg
waste, based on
measurements at the
Virginia facility. For
larger plant, the
emission rate used is
0.001 ug/kg.
The emission rate for
VA facility repre-
sents the high end of
measured incinerator
emissions. The
emission rate of 0.001
ug/kg repre- sents
the lower end of
measured incinerator
emissions and is
thought to reflect
what can be expected
with new technology.
A single value of 0.55 0.55 ng/g is the
ng/g was used for
this concen- tration
in calcu- lations with
both incinerators.
average of measured
values presented in
Chapter 6, exclud-
ing one high end
point.
Measured incinerator
emission rates cover a
very broad range from
less than 0.001 ug/kg
up to 0.289 ug/kg,
with some data
suggesting that even
higher emissions may
be seen under some
incinerator operating
conditions.
A range of 0.07-2.3
ng/g has been ob-
served with the
exception of the one
high point, 100 ng/g
seen in a pyrolysis
test. Some addi-
tional data on total
TCDD measurements
suggest values well
above 2.3 ng/g may
occur.
Exposure estimates may
be scaled with estimated
emission rate. It should
be noted that if an
incinerator a high end of
the size range were to
have a high end emission
rate, exposures would
substantially exceed
those in either of the
scenarios pre- sented
here.
Evaluation: The scenarios chosen are appropriate for both incinerator size and 2,3,7,8-TDD release. However, the
large variability in both plant size and release rate should be noted. If a large facility had a high emission rate
pathway exposures much higher than those calculated here would be anticipated. Conversely, for a small plant with
low emissions, much lower concentrations would be expected.
302
-------
TABLE 7.14. INCINERATION ASSESSMENT
INHALATION AND SURFACE DEPOSITION
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Estimate
transport of
stack emissions.
Applied industrial
source complex model,
the program-version
used evaluated "tip
down wash" effect of
precipitation and
assumed that a
building of height
equal to the incin-
erator stack was near
by.
This air dispersion
model is widely
accepted by Agency.
Both "tip down wash"
and presence of a
building as tall as the
stack may in- crease
ground level pollutant
concentra- tions.
Input data are site-
specific for meteor-
ologic conditions.
The presence of a tall
building near the
incinerator stack may
not be typical.
Predicted surface
deposition rates with
or without the effects
of precipi- tation
were similar.
P articulate
distribution.
Point where
deposition and
inhalation are
evaluated.
Adopted sice distri-
bution used in OAQPS
modeling for incin-
erators.
Inhalation exposures
and surface deposi-
tion evaluated at
distance 1/2 mile (0.8
km) from in-
cinerator stack.
Data was specific to
incinerators.
Exposures at this
distance represent
rather direct ex-
posure to stack
emissions, but not the
maximum estimat- ed
pollutant concen-
trations.
Variability of sioe
distribution has not
been assessed. Size
distribution will have
strong influence on
deposition rates.
Exposures at 200m
from stack were cal-
culated to be five
times greater than at
0.8 km. For compari-
son, concentrations at
100 km from stack
were also calculated
and are 2 to S orders
of magnitude lower
than at 0.8 km.
Inhalation rate.
See landfill dust
emissions chart.
Evaluation: The analysis is appropriate for determining air concentrations and surface deposition rates on a family
relatively near incinerator stack emissions. The assumption that a building of height equal to the incinerator stack is
present leads to higher estimated concentrations near facility. Concentrations 1/2 mile from stack are used in exposure
calculations. At longer distances exposures will be substantially lower.
303
-------
the uncertainties related to inhalation and surface deposition.
4. Surface Water Contamination
In different scenarios either a small pond or a large river are assumed to be located
at a distance of 1/2 mile (0.8 km) from the incinerator stack. The sizes and locations of
these water bodies should be seen as examples that can be modified in site-specific
assessments. 2,3,7,8-TCDD transfer to the water bodies through vapor absorption and
through wet and dry particulate deposition were estimated.
Vapor absorption into river water was estimated using two-resistance theory for
mass transfer from the air to the water body which was briefly discussed under the
surface water section of the landfill scenarios. As noted there, calculations of this type
have been shown to provide results correct to an order of magnitude or better for
compounds more volatile than 2,3,7,8-TCDD, for which data is available. The calculation
assumes that dissolved phase losses to sediment or biodegradation do not occur; in reality
some such loss is possible.
The mass transfer calculation made the simplifying assumption that water levels of
2,3,7,8-TCDD are assumed to be low in comparison with equilibrium concentrations. That
this assumption is appropriate can be seen by comparing estimated river water 2,3,7,8-
TCDD concentrations to the concentrations calculated for the pond water.
Average river water concentrations are obtained by dividing the vapor absorption
rate by the quantity of water flow. This same methodology can be applied to rivers or
streams of differing size.
Vapor absorption into lake water was estimated using Henry's Law and assuming an
equilibrium relationship between 2,3,7,8-TCDD in the air and water. The equilibrium
water concentration was calculated using the estimated average vapor concentration of
2,3,7,8-TCDD obtained by applying the wind frequency factor to the modeled vapor
concentration. This approach, while judged reasonable may not fully account for the
304
-------
time varying nature of vapor concentration. This calculations does not take into account
any loss of 2,3,7,8-TCDD to sediments; such loss would lead to lower estimated water
concentrations.
Pond water levels of 2,3,7,8-TCDD due to deposition of contaminated particulates
are calculated based on the deposition rates from the ISC model discussed above.
Concentrations of contaminated particulate in the water depend on the effective period
for which the particulate material remains in the water column. As data are not
available to estimate this lifetime, a simplified theoretical model is used. The deposition
rates from the ISC model allow calculation of an effective settling velocity of the
particles in the air. Then physical principles allow calculation of the settling rate of the
same particles in the water column. The particles are assumed to remain in the water
until sufficient time has passed for them to settle to the bottom.
As noted, this is a simplified model and data do not exist to establish that direct
settling is the principal loss mechanism for particulate in water. Particle agglomeration,
effects of biota or photolysis in water could alter the model predictions. Furthermore,
the model does not take into account mixing of particulate within the water column or
resuspension of particulate from the sediment. Levels of 2,3,7,8-TCDD in pond sediment
may accumulate and may then influence water concentrations of 2,3,7,8-TCDD.
Particulate deposition rates into rivers are the same as used in the pond scenario,
but with the plume from the incinerator (as in the vapor case) covering only a calculated
area of the river surface. Water concentrations are obtained from the deposition rate
and water flow rate.
A water/fish bioconcentration factor (BCF) of 10,000 was used to estimate
concentration of 2,3,7,8-TCDD in fish living in the pond and river. The uncertainties in
water concentrations translate directly to uncertainties in fish concentrations. The BCF
of 10,000 represents a median value from a wide range of reported values (Schaffer,
305
-------
1985). In 1984, EPA proposed a BCF of 5,000 for use in the Water Quality Criteria (U.S.
EPA, 1984a). More recently, laboratory studies have measured BCFs of 66,000 for carp
and 97,000-159,000 for fathead minnows (Cook, 1987). The more recent data suggest that
the BCF could be much higher than assumed, and the overall range suggests that this
factor is a major source of uncertainty. Table 7-15 is a summary of the uncertainties
related to surface water concentration.
5. Soil Contamination from Emissions
Both particulate deposition and vapor absorption of 2,3,7,8-TCDD to soil are assumed
to occur at sites 0.8 km from the model incinerators, the same locations evaluated for
the inhalation pathway. Particulate deposition rates are evaluated using results of the
ISC type model discussed above. Disappearance of 2,3,7,8-TCDD from the surface soil
through all loss mechanisms is accounted for by the 0.06 yr"1 rate constant that was
discussed above. Given the values for the deposition rate and the loss rate, the quantity
of 2,3,7,8-TCDD in the soil as a function of time is calculated. It is assumed that the
2,3,7,8-TCDD will remain in the top 1 cm depth of soil. Data to evaluate this assumption
for nontilled soils have not been evaluated.
A partitioning calculation is used to estimate 2,3,7,8-TCDD levels in soil due to
vapor absorption. Here the assumption is made that the soil and air 2,3,7,8-TCDD
calculations will be in equilibrium. This assumption leads to maximum soil concentrations
attainable through vapor absorption for a fixed value of the air/soil partition coefficient.
The time required to approach equilibrium was not estimated, and if this time proves
substantial compared to the estimated half-life of about 10 years for loss of 2,3,7,8-TCDD
from soil, soil levels will be overestimated. Estimated soil concentrations are
proportional to the soil/air partitions coefficient which was estimated from the soil/water
partition coefficient for 2,3,7,8-TCDD which is subject to substantial uncertainty as
discussed earlier. The uncertainty in the assumed rate of soil ingestion, dermal contact,
306
-------
TABLE 7.15. INCINERATOR ASSESSMENT SURFACE WATER CONTAMINATION
BY INCINERATOR STACK EMISSIONS
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Location of
water body.
Either a small pond
or a large river are
assumed to be present
at 0.8 km from the
incinerator stack.
Addresses potential
exposure pathway.
Site-specific situation.
Distance of 0.8 km
reflect* direct
exposure to stack
emissions but not
maximum ground level
concentrations.
2,3,7,8-TCDD
vapor concen-
tration and
particulate de-
position rate.
See: Inhalation and
Surface Deposition.
(Table 7.14)
Pond water con-
centration due to
vapor absorp-
tion.
Estimated by Henry's
Law calculation which
assumes equilibrium
has been reached.
Model reflects
physical processes
affecting water
concentration.
Absorption to
sediment not
accounted for.
Equilibrium may not
be achieved in
practice. These would
lead to lower
estimates of water
concentration.
Pond water
concentration
due to par-
ticulate
deposition.
Assumed that partic-
ulate deposited onto
water surface will
remain suspended for
a period equal to
calculated direct
settling time for
particulate through
the water column.
Idealized physical
model which accounts
for the phenomenon
that small particu-
lates will remain in
water column for
longer period.
Movement of water
and loss mechanisms
for five particulate
sices may alter
estimates. Model is
not field validated.
307
-------
TABLE 7.15. (CONTINUED)
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
River water con-
centration due to
vapor.
Applied mass transfer
calculation account-
ing for water move-
ment. No steady
state assumption is
made; instead calcu-
lation relies on water
concentration being
low (relative to
potential steady state)
to estimate transfer.
Model reflects phy-
sical process that will
affect water
concentration.
Measurement data are
not available.
Partition coeffi-
cients and other
model parameters
subject to uncer-
tainty. Model is not
field validated.
River water
concentration
due to par-
ticulate
deposition.
Used particulate
deposition rate and
water flow rate to
calculate concen-
tration.
Physically appro-
priate.
Appropriate calcula-
tion for average river
water concen-
tration given esti-
mated deposition rate.
Bioconcentra-
tion factor.
10,000.
Median of reported
values.
Range of reported
values very wide.
Recent data suggest it
could be as high as
159,000.
Evaluation: Calculations of water concentration due to vapor absorption are uncertain due to uncertain parameters in
model, and the fact that the models have not been field validated. Particulate concentration in pond also uncertain as
settling rate calculation is not known to be appropriate. River water particulate concentration better established as
this estimate does not rely on unproven models. Bioconcentration factor is a major source of uncertainty, possibly
causing an underestimate of risk by as much as a factor of 16.
308
-------
beef exposure and dairy exposure are discussed under the landfill scenarios. Table 7-16
is a summary of the uncertainties related to soil contamination levels due to incinerator
emissions.
6. Dairy Product Exposure Following Deposition on Plants
The calculation of exposure through milk produced by cows grazing in the vicinity
of an incinerator utilizes the model developed by Connett and Webster (1987). These
authors have adapted a model previously developed for radionuclides to address 2,3,7,8-
TCDD. The major components of the model are:
(a) The deposition rate of 2,3,7,8-TCDD for which this report utilizes the air
transport model estimates discussed above.
(b) The concentration in fodder, which depends on an assumed half-life of 14 days
for 2,3,7,8-TCDD following deposition. This value was based on weathering
considerations by analogy with radionuclide modeling. Experimental data on
2,3,7,8-TCDD were not available for evaluation of this half-life.
(c) The effective grazing area of a cow (surface area of vegetation consumed in a
day, which was estimated using the productivity of a field for hay production
(kg/m^ d) and the food consumption rate of a milk producing, pastured cow.
The authors note that effective grazing area will vary regionally and upper
New York State values were used.
(d) A pharmacokinetic model, based on the experimental data of Jensen et al.
(1981) for lactating cows fed 2,3,7,8-TCDD was used to estimate the transfer
coefficient between fodder and milk assuming 33% bioavailability of the dioxin
from fodder on which particulate was deposited. EAG regards the formulation
of this model as "appropriate" for estimating 2,3,7,8-TCDD exposure via this
pathway. Further, the assumptions made by the authors concerning individual
model parameters are felt to be reasonable; the data for selecting parameter
309
-------
TABLE 7.16. INCINERATOR ASSESSMENT SOIL CONTAMINATION
LEVELS RESULTING FROM INCINERATOR EMISSIONS
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
12,3,7,8-TCDD
deposition rate
to soil (particu-
late)
Soil Concentra-
tion from
participate
deposition.
See Inhalation and Surface Deposition
Constant deposition
over a 70 year period
assessed. 'First order
decay constant 0.06/y
assumed for 2,3,7,8-
TCDD in soil.
2,3,7,8-TCDD assumed
to remain in top 1 cm
of-soil.
70 years selected to
represent long-term
operation. The decay
constant, taken from
Young (1983), was
based on field obser-
vation of disappear-
ance of 2,3,7,8-TCDD
•through all pathways.
A 1 cm mixing depth
represents case where
no agricultural prac-
tices are present.
Field data demon-
strate that environ-
mental half-life of
2,3,7,8-TCDD is a
substantial period of
years but do not
allow precise esti-
mation of decay rate.
Data on mixing depth
of 2,3,7,8-TCDD
applied to soil sur-
face though particu-
late deposition are
not available.
Vapor absorp-
tion by soil.
Soil concentrations
estimated parti-
tioning calculation
using calculated vapor
concentrations (see
Inhalation and Surface
Deposition). The
parti-
tioning calculation
yields maximum soil
levels consistent with
vapor concen-
trations.
A upper bound on
soil concentration
was desired in the
absence of a more
refined model.
Soil concentrations
may be less than cal-
culated by partition-
ing depending on the
time required to
approach equilibrium.
Soil/air partition
coefficient was
derived from soil/
water coefficient
which is subject to
substantial uncer-
tainty.
Soil, beef and dairy
ingestion rate
Duration of ingestion - See Landfill Scenarios.
Gut absorption from soil
Dermal contact rate with soil
Dermal absorption - See Dermal Contact for Landfill Scenarios.
Evaluation: The estimated soil concentration is proportional to 2,3,7,8-TCDD emission rate and depends on air
transport modeling, which are discussed in Tables 7.13 and 14. For the air concentrations assessed, soil concentrations
and associated pathway exposures are likely to be less than those calculated in the case of vapor absorption.
310
-------
estimates, particularly for persistence on vegetation may be weak. Further, as
the model has not been validated with field data, an overall quantitative
evaluation of uncertainty is not possible.
7. Land-Disposed Ash
The quantity of fly ash produced by the modeled incinerators is calculated using a
standard particulate emission factor and factors for control equipment efficiency. The
quantity of landfilled ash generated from 35 years of incineration operation is used as an
average accumulation period to estimate landfill size for this analysis.
The landfill area estimates were derived from the generation quantity and an
assumed disposal depth of 10 ft. Changes in depth or generation time would change the
area estimate proportionally.
The contamination level in the ash (assumed to be 0.5 ng/g) was derived as a mean
from a number of incinerators. The range was 0.07 to 100 ng/g.
Exposures to ash disposal were evaluated for the same pathways as done for
landfills. The same procedures were assumed to apply after changing only the size of
the landfill and contamination level. This assumption could cause uncertainty in several
ways. The physical properties of ash may differ from regular soils causing differences in
transport via surface runoff or windblown dust. The materials handling associated with
ash could cause greater emissions than the soil scenarios where dust generation was
assumed to occur as a result of wind only. The transport of ash in trucks could cause
fugitive emissions leading to exposure. The bioconcentration of fly ash to fish may
differ from that of normal sediments. In general, these differences suggest that the
similar treatment of soil and ash leads to underestimates of risk. The bioconcentration
issue is an exception in that it is unknown if this causes under- or over-estimates of
risk. In the calculation of ingestion exposures, the same bioavailability factor is used for
ash as for soil. There is much less data on availability of 2,3,7,8-TCDD from ash than
311
-------
from soil and as discussed in Chapter 5, the interpretation of the existing ash
bioavailability data presents difficulties. Significant uncertainty exists in this area.
Dust emission from disturbance of ash during the process of landfilling have been
estimated. These estimates take into account the following activities. These activities
are in effect. A scenario for landfilling activities and the extent of these activities was
calculated to reflect the quantity of ash being disposed.
(a) Vehicle traffic: Truck traffic over a unpaved road containing fly ash
particulate generates dust. A standard dust emission factor for traffic on an
unpaved road was obtained from EPA (1985). The authors of EPA (1985)
assigned the road dust model a rating of "A" indicating that it was based on
substantial field data obtained using appropriate methodologies. The
calculations here were based on the passage of 10 heavy trucks per day over
an on-site unpaved road at a 16 kph speed. Ash content of the road was
assumed to be 20%, which would be dependent on the specific character of the
road being utilized. As fly ash may be finer than road dust in general, these
particulates may become airborne to a greater extent than calculated in this
model.
(b) Unloading of ash at the landfill was assessed using particulate emission factor
for handling of silt containing aggregate from EPA (1985). This emission
factor was assigned a rating of "C" by EPA (1985) indicating that extensive,
appropriate field data were not available for the factor. This calculation
requires values for mean wind, drop height and several other parameters.
Values for these parameters were selected to be consistent with the
calculations for wind generated dust and to corresponding to dumping of fly
ash from dump trucks. Parameters used should reflect local conditions in any
site-specific applications.
312
-------
(c) Spreading operations. An emission factor developed for agricultural tilling
(EPA, 1985) was judged the available factor that most closely resembled the
process of spreading fly ash. The emission factor was given a rating of "A" or
"B" by EPA (1985) indicating that substantial appropriate data was available to
estimate the emission factor. However, the use of this factor for spreading of
fly ash has not been validated.
(d) Emissions due to windblown emissions from trucks transporting fly ash on site
(i.e., release from the truck's load) was calculated using an emission factor
approach. However, as it is thought that fly ash is normally wetted before
transportation, an adjustment for this was made. Emissions were calculated by
assuming that a moving truck could be considered as comparable to a pile of
material subjected to wind moving at the speed of the truck (physically
appropriate). However, the adjustment for wetting was based on assumption
that emissions could be calculated, by analogy, with a material pile under the
circumstance where rain fell 364 days/year. Data to evaluate the
appropriateness of this assumption are not available.
Table 7-17 is a summary of the uncertainties related to land disposed ash resulting
from incinerator use.
313
-------
TABLE 7.17. LAND-DISPOSED ASH
Assumption/
Method
Approach
Rationale
Uncertainty
Comments
Quantity of ash
generated .by
incinerator and
landfilled.
Contamination
level.
Treatment of ash
vs. soil.
Used uncontrolled
emission factor to
estimate fly ash
production, assumed
captured for dis-
posal. Quantity of
ash generated in 35
years calculated.
0.5 ng/g.
Assumed ash behaves
same as soil.
Based on standard
emission factor.
Landfill assumed to
receive long-term ash
generation.
Mean of available
data.
Lacked data to
.account for .differ-
ences.
Ash generation/ton
combusted material
may vary between
incinerators.
Assumption that ash
is placed in one
landfill may not be
typical.
Range is 0.07 to 100
ng/g.
Dust generation
during disposal and
transport likely to be
higher.
Evaluation: Range of contamination levels spans three orders of magnitude, so this is a major source of uncertainty.
Treatment of ash like soil may cause slight underestimates of risk.
314
-------
8. REFERENCES
Ackerman, D.G. (1978, April) At-sea incineration of herbicide orange on board the
M/V Vulcanus. EPA-600/2-78-086.
Adams, W.J.; Elaine, K.A. (1986) A water solubility determination of 2,3,7,8-TCDD.
Chemosphere 15(9-12): 1397-1400.
Ahling, B.; Lindskog, A. (1977) Formation of polychlorinated dibenzo-p-dioxins and
dibenzofurans during combustion of a 2,4,5-T formulation. Chemosphere 6 (8):461-468.
Ahling, B.; Lindskog, A. (1982) Emission of chlorinated organic substances from
combustion. In: Hutzinger, O. et. al. eds. Chlorinated dioxins and related compounds;
impact on the environment. Vol. 5, Series on Environmental Science. New York, NY:
Pergamon Press, pp. 215-225.
Albro, P.W.; Corbett, J.T.; Schroeder, J.L. (1986) Effects of 2,3,7,8-tetra-chlorodibenzo-
p-dioxin on lipid peroxidation in microsomal systems in vitro. Chem. Biol. Interact.
57(3):301-313.
Alexander, M. (1977) Introduction to soil microbiology. New York, NY: John Wiley and
Sons, Inc.
Arthur D. Little, Inc. (1981) Study on state-of-the-art of dioxin from combustion
sources. Issued by the American Society of Mechanical Engineers. New York, N.Y.
Astrila, A.U.; Reggiani, G.; Sonvani, T.E.; Raisaneu, S.; Wipf, H.K. (1981) Elimination of
2,3,7,8-tetrachlorodibenzo-p-dioxin in goat milk. Toxicol. Lett. 9:(3)215.
Bartek, M.J.; LaBudde, J.A.; Maibach, H.I. (1972) Skin permeability in vitro: comparison
in rat, rabbit, pig and man. J. Invest. Dermatol. 58:114-123. [as cited in Hawley (1985)]
Bedient, P.B.; Rodgers, A.C.; Bouvette, T.C; Tomson, M.B.; Wang, T.M. (1984)
Groundwater quality at a creosote waste site. Ground Water 22:318-323.
Benfenati, E.; Gizzi, F.; Reginato, R.; Fanelli, M.; Lodi, R.; Tagliaferri,R. (1983)
Polychlorinated dibenzo-p-dioxins (PCDD) and polychlorinated dibenzofurans (PCDF) in
emissions from an urban incinerator. 2. Correlation between concentration of
micropollutants and combustion conditions. Chemosphere 12 (9):1151-1157.
Berglund, R. (1984, September 11) Personal communication to G.W. Dawson National
Cattleman's Association, Denver, CO. [as cited in U.S. EPA (1985)]
Berkow, S.G. (1924) A method of estimating the extensiveness of lesions (burns and
scalds) based on surface area proportions. Arch. Surg. 8:138-148.
Berman, E.F.; Schauss, P.; Fujimoto, J.M. (1986) Comparison of the inhibition of biliary
excretion produced by certain inducing agents including 2,3,7,8-tetrachlorodibenzo-p-
dioxin. J. Toxicol. Environ. Health 17(4): 395-403.
315
-------
Binder, S.; Sokal, D.; Maughn, D. (1986) The use of tracer elements in estimating the
amount of soil ingested by young children. Arch. Environ Health 41:341-345.
Bonaccorsi, A.; diDomenico, A.; Fanelli, R.; Merli, F.; Motta, R.; Vanzati, R.;
Zapponi, G.A. (1984) The influence of soil particles adsorption on 2,3,7,8- tetra-
chlorodibenzo-p-dioxin biological uptake in the rabbit. Arch. Toxicol. Suppl. 7:431-434.
Bowman, R.E.; Schantz, S.L.; Weeasinghe, N.C.A.; Meehan, J.; Pan, J.; Gross, M.L.; Boehm,
K.M.; Van Miller, J.P.; Welling, P.G. (1987) Clearance of 2,3,7,8-tetrachlorodibenzo-p-
dioxin (TCDD) from body fat of rhesus monkeys following chronic exposure. The
Toxicologist 7(l):(Feb) Abstract No.:631b.
Breidenstein, B.C. (1984) Contribution of red meat to the U.S. diet. National Livestock
and Meat Board, Chicago, IL.
Brenner, K.S.; Dorn, I.H.; Hermann, K. (1986) Dioxin analysis in stack emissions, slags
and the wash water circuit during high-temperature incineration of chlorine-containing
industrial wastes. Chemosphere 15 (9-12):l 193-1199.
Brodkin, R.H.; Schwartz, R.A. (1984) Cutaneous signs of dioxin exposure. Am. Fam.
Physician 30(3):189-194.
Bumb, R.R.; Crummett, W.B.; Cutie, S.S.; Gledhill, J.R.; Hummel, R.H.; Kagel, R.O.;
Lamparski, E.; Luoma, E.V.; Miller, D.L.; Nestrick, T.J.; Shadoff, L.A.; Stehl, R.H.;
Woods, J.S. (1980) Trace chemistries of fire: a source of chlorinated dioxins. Science
210:385-390.
Bumpus, J.A.; Tien, M.; Wright, D.; Aust, S.D. (1985) Oxidation of persistent
environmental pollutants by a white rot fungus. Science 228:1434-1436.
Burant, C.F.; Hsia, M.T. (1984) Excretion and distribution of two occupational toxicants,
tetrachloroazobenzene and tetrachloroazoxybenzene in the rat. Toxicology 29:(3)243.
Burkhard, L.P.; Kuehl, D.W. (1986) n-Octanol/water partition coefficient by reverse-
phase liquid chromatography/mass spectrometry for eight tetra-chlorinated planar
molecules. Chemosphere 15(2):163-167.
Buser, H.R.; Bosshardt, H.P. (1978) Identification of polychlorinated dibenzo-p-dioxin
isomers found in fly ash. Chemosphere 7(2): 165-172.
Buser, H.R.; Bosshardt, H.P.; Rappe, C. (1978) Identification of polychlorinated
dibenzofuran isomers in fly ash and PCB pyrolyses. Chemosphere 7:419-429.
Cavallaro, A.; Bartolzzi, G.; Carreri, D.; Bandi, G.; Luciani, L.; Villa, G.; Gorni, A.;
Invernizzi, G. (1980) Sampling, occurrence, and evaluation of PCDDs and PCDFs from
incinerated solid urban waste. Chemosphere 9:611-621.
Chen, C.; Blancato, J.N. (1987) Role of pharmacokinetic modeling in risk assessment:
Perchloroethylene (PCE) as an example. Pharmacokinetics in Risk Assessment - Drinking
and Health, National Acedemy Press, Washington, DC p. 369-390.
316
-------
Clausing, P.; Brunekreff, B.; Van Wijen, J.H. (1987) A method for estimating soil
ingestion by children. Int. Arch. Occup. Environ. Health 59:73-82.
Clement, R.E.; Tosine, H.M.; Ali, B. (1985) Levels of polychlorinated dibenzo-p-dioxins
and dibenzofurans in wood burning stoves and fireplaces. Chemosphere 14(6/7):815-819.
Chlorinated Dioxins Work Group. (1981) Agenda and minutes of the Dioxin Sources
Subgroup of the Chlorinated Dioxins Work Group, Meeting No. 10, March 6.
Cocucci, S.; DiGerolamo, F.; Verderio, A.; Cavallaro, A.; Colli, G.; Gorni, A.; Invernizzi,
G.; Luciani, L. (1979) Absorption and translocation of tetrachlorodibenzo-p-dioxin by
plants from polluted soil. Experientia 35(4):482-484.
Codell, R.B.; Key, K.T.; Whelan, G. (1982) A collection of mathematical models for
dispersion in surface water and groundwater. Prepared for U.S. Nuclear Regulatory
Commission, Washington, B.C., NUREG-0868.
Commoner, B.; Webster, T.; Shapiro, K. (1985) Environmental levels and health effects
of PCDDs and PCDFs. Presented at the 5th international symposium on chlorinated
dioxins and related compounds; September; Bayreuth, FRG.
Commoner, B.; Webster, T.; Shapiro, K. (1986) Environmental levels and health effects
of chlorinated dioxins and furans. Presented at the AAAS Annual Meeting; May;
Philadelphia, PA.
Connett, P.; Webster T. (1987) An estimation of the relative human exposure to 2,3,7,8-
TCDD emissions via inhalation and ingestion of cow's milk. Chemosphere 16(8/9): 2079-
2084.
Connor, M.S. (1984) Comparison of the carcinogenic risks from fish vs. groundwater
contamination by organic compounds. Environ. Technol. (18) 628-631.
Cook, P.M. (1986) Memorandum dated September 2, 1986 to F.W. Kutz of the Office of
Environmental Processes and Effects Research, Washington, DC, from Philip Cook.
U.S. Environmental Protection Agency, Office of Environmental Processes and Effects
Research, Duluth, MN.
Cook, P.M. (1987) Memorandum titled "2,3,7,8-TCDD in aquatic environments," dated
February 4, 1987, to J. Cummings of the Office of Solid Waste and Emergency Response.
U.S. Environmental Protection Agency, from Philip Cook Office of Environmental
Processes and Effects Research, Duluth, MN.
Cook, P.M. (1987a) Memorandum with attachment titled "Bioavailability of polychlorinated
dibenzo-p-dioxins and dibenzofurans from contaminated Wisconsin River sediment to
carp," to Jim Cummings of the Office of Solid Waste and Emergency Response, from
Philip Cook, U.S. Environmental Protection Agency, Office of Environmental Processes
and Effects Research, Duluth, MN.
Cordle, F. (1981) The use of epidemiology in the regulation of dioxins in the food supply.
Regul. Toxicol. Pharmacol. (1) 379-387.
317
-------
Cordle, F.; Licke, R.; Springer, J. (1982) Risk assessment in a federal regulatory agency:
an assessment of risk associated with the human consumption of some species of fish
contaminated with polychlorinated biphenyls (PCBs). Environ. Health Perspect. 45:171.
Crosby, D.G.; Wong, A.A. (1977) Environmental degradation of 2,3,7,8-tetra-
chlorodibenzo-p-dioxin (TCDD). Science 195:1337-1338.
Crummett, W.W. (1982) Environmental chlorinated dioxins from combustion: the trace
chemistries of fire hypothesis. In: Hutzinger et. al. eds> Chlorinated dioxins and related
compounds: impact on the environment. Vol. 5, Series on Environmental Science. New
York, NY: Pergamon Press, pp. 253-263.
Czuczwa, J.M.; Kites, R.A. (1986) Airborne dioxins and dibenzofurans: sources and fates.
Environ. Sci. Technol. 20(2):195-200.
De Fre, R. (1986) Dioxin levels in the emissions of Belgian municipal incinerators.
Chemosphere. 1.4(9-12): 1255-1260,
des Rosiers, P.E.; Lee, A. (1986) PCBs fires: correlation of chlorobenzene isomer and
PCB homolog contents of PCB fluids with PCDD and PCDF contents of soot.
Chemosphere L5 (9-12):1313-1323.
DiBartolomeis, M.J.; Williams, C; Jefcoate, C.R. (1986) Inhibition of ACTH action on
cultured bovine adrenal cortical cells by 2,3,7,8-tetrachloro-dibenzo-p-dioxin through a
redistribution of cholesterol. J. Biol. Chem. 261:( 10)4432.
Diem, K.; Letner, C. (1973) Documenta Geigy (Scientific tables. Ciba-Geigy, Ltd., Basle)
p.528.
Dunagin, W.G. (1984) Cutaneous signs of systemic toxicity due to dioxins and related
chemicals. J. Am. Acad. Dermatology 10(4):688-70.0.
Eiceman, G.A.; Clement, R.E.; Karasek, F.W. (1979) Analysis of fly ash from municipal
incinerators for trace organic compounds. Anal. Chem. 51:2343-2350.
Eiceman, G.A.; Viau, A.C.; Karasek, F.W. (1980) Ultrasonic extraction of polychlorinated
dibenzo-p-dioxins and other organic compounds from fly ash from municipal incinerators.
Anal. Chem. 52:1492-1496.
Eiceman, G.A.; Clement, R.E.; Karasek, F.W. (1981) Variations in concentrations of
organic compounds including polychlorinated dibenzo-p-dioxins and polynuclear aromatic
hydrocarbons in fly ash from a municipal incinerator. Anal. Chem. 53:955-959.
Eitzer, B.D.; Kites, R.A. (1986) Concentrations of dioxins and dibenzofurans in the
atmosphere. Int. J. Environ. Anal. Chem. 27:215-230.
Facchetti, S; Balasso, A; Fichtner, C; Frare, G; Leoni, A; Mauri, C; Vascvo, M. (1986)
Studies on the absorption of TCDD by some plant species. Chemosphere (15) 9-12, pp.
1387-1388.
31S
-------
Farland, W.H. (1987a) Memorandum titled "Issues Associted with RIA Calculations for
Carcinogens - Municipal Waste Combusion Document" dated February 4, 1987, to Randy
Bruins, Environmental Criteria and Assessment Office, Cincinnati, OH. from William
Farland, Director, Carcingen Assessment Group, U.S. Environmental Protection Agency,
Office of Health and Environmental Assessment, Washington, DC.
Farland, W.H. (1987b) Memorandum titled "Absorption fraction when calculating upper-
limit risks due to dioxin exposure," dated September 2, 1987, to Michael Callahan,
Exposure Assessment Group, Washington, DC. from William Farland U.S. Environmental
Protection Agency, Office of Health and Environmental Assessment, Washington, DC.
Feldman, R.J.; Maibach, H.I. (1970) Absorption of some organic compounds through the
skin in man. J. Invest. Dermatol. 54:399-404. [as cited in Hawley (1985)]
Foster, G.R.; Hakonson, T.E. (1987) Erosion losses of fallout plutonium. In: Howard,
W.A.; Fuller.R. Dynamics of transuranics and other radionuclides in natural environments.
NTIS, Springfield, Va, DE 87014456.
Freeman, R.A.; Schroy, J.M. (1985a) Environmental mobility of dioxins. In: Bahner,
R.C.; Hansen, D.J., eds. Aquatic toxicology and hazard assessment: eighth symposium;
Philadelphia, PA: American Society for Testing and Materials.
Freeman, R.A.; Schroy, J.M. (1985b) Environmental mobility of TCDD. Chemosphere
14(6/7):873-876.
Freeman, R.A.; Schroy, J.M. (1986) Modeling the transport of 2,3,7,8-TCDD and other
low volatility chemicals in soils. Environ. Prog. 5(l):28-33.
Freeze, R.A.; Cherry, J.A., Groundwater, Engelwood Cliffs, N.J.: Prentice Hall, Inc.
Fries, G.F. (1982) Potential poly chlorinated biphenyl residues in animal products from
application of contaminated sewage sludge to agricultural land. J. Environ. Qual.
11:14-20.
Fries, G.F. (1985) Bioavailability of soil-borne polybrominated biphenyls ingested by
farm animals. J. Toxicol. Environ. Health 16:565-579.
Fries, G.F. (1986) Assessment of potential residues in foods derived from animals
exposed to TCDD-contaminated soil. Presented at 6th international symposium on
chlorinated dioxins and related compounds; September; Fukuoka, Japan.
Fries, G.F.; Marrow, G.S. (1975) Retention and excretion of 2,3,7,8-tetrachlorodibenzo-
p-dioxin (TCDD) by rats. J. Agric. Food Chem. 23:265-269.
Fries, G.F.; Marrow, G.S. (1981) Chlorobiphenyl movement from soil to soybean plants.
J. Agric. Food Chem. 29:757-759.
Fu, J.K.; Luthy, R.G. (1986) Effect of organic solvent on sorption of aromatic solutes
onto soils. J. Environ. Sci. 112:346-366.
Geyer, H.; Scheunert,; Korte, F. (1986) Bioconcentration potential of organic
environmental chemicals in humans. Regul. Toxicol. Pharmocol. (6) 313-347.
319
-------
Gizzi, F.; Reginato, R.; Benfenati, E.; Fanelli, R. (1982) Polychlorinated dibenzo-p-
dioxins (PCDD) and polychlorinated dibenzofurans (PCDF) in emissions from an urban
incinerator 1. average and peak values. Chemosphere ll(6):577-583.
Graham, M; Hileman, F.; Kirk, D.; Wendling, J.; Wilson, J. (1985) Background human
exposure to 2,3,7,8-TCDD. Chemosphere 14(6/7):925-928.
Greenlee, W.F.; Osborne, R.; Dold, K.M.; Hudson, L.G.; Toxcano, W.A., Jr. (1985) Toxicity
of chlorinated compounds in animals and humans: in vitro approaches to toxic mechanisms
and risk assessment. Environ. Health Perspect. 60:69-76.
Gross, M.L. (1982) Application of mass spectrometric methods to analysis of xenobiotics
in biological systems. IARC Sci. Publ. 39:443-462, Lyon, France
Hagenmaier, H.; Kraft, M.; Jager, W.; Mayer, U.; Lutzke, K.; Siegel, D. (1986) Comparison
of various sampling methods for PCDDs and PCDFs in stack gas. Chemosphere
15(9-12):! 187-1192.
Hahn, J.L.; von Demfange, H.P.; Velzy, C.O. (1986) Effect of boiler operation and RDF
feed stack on emissions of dioxins and furans from an RDF fired spreader-stoker system
in Albany, N.Y. Chemosphere 15(9-12): 1239-1246.
Hakansson, H.; Alborg, U.G. (1985) The effect of 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD) on the uptake, distribution, and excretion of a single oral dose of
[ll,12-34]retinyl acetate on the vitamin A status of the rat. J. Nutr. 115(6):759.
Hannah, R.R.; Lund, J.; Poellinger, L.; Gillmer, M.; Gustafsson, J.A. (1986)
Characterization of DNA-binding properties of the receptor for 2,3,7,8-
tetrachlorodibenzo-p-dioxin. Eur. J. Biochem. 156(2):237.
Hawley, J.K. (1985) Assessment of health risk from exposure to contaminated soil. Risk
Analysis 5(4): 289-302.
Hay, D.J.; Finkelstein, A.; Klicius, R. (1986) The national incinerator testing and
evaluation program two-stage incinerator combustion tests. Chemosphere
15(9-12):1201-1212.
Healy, W.B. (1968) Ingestion of soil by dairy cows. New Zealand J. Agric. Res.
11:487-499.
Holton, G.A. (1985) A comparison of human exposures to PCB emissions from oceanic
and terrestrial incineration. Hazardous Waste and Hazardous Materials 2(4):453-471.
Hryhorczuk, D.O.; Withrow, W.A.; Hesse, C.S.; Beasley, V.A. (1981). A wire reclamation
incinerator as a source of environmental contamination with tetrachlorodibenzo-p-dioxins
and tetrachlorodibenzofurans. Arch. Environ. Health 36(5): 228-234.
Humphrey, H.E.B.; Price, H.A.; Budd, M.L. (1976) Evaluation of changes of the level of
polychlorinated biphenyls (PCBs) in human tissue. Final report of FDA contract No.223-
73-2209, 1976 [As cited in Cordle, et al. (1982)].
320
-------
Hwang, S.T. (1986) Health risk comparison between groundwater transport models and
field data. Environ. Prog. 5:66-70.
Hwang, S.T. (1987) Methods for estimating on-site ambient air concentrations at disposal
sites. Nuclear and Chemical Waste Management, 7:95-98.
Hwang, S.T.; Falco, J. (1986) Estimation of multimedia exposures related to hazardous
waste facilities. In: Cohen, Y., ed. Pollutants in a multimedia environment.
New York, NY: Plenum Publishing Co.
Isensee, A.R.; Jones, G.E. (1971) Absorption and translocation of root and foliage
applied 2,4-dichlorophenol, 2,7-dichorodibenzo-p-dioxin and 2,3,7,8-tetrachlorodibenzo-p-
dioxin. J. Agric. Food Chem. 19(6):1210-1214.
Isensee, A.R.; Jones, G.E. (1975) Distribution of 2,3,7,8-tetrachloro-dibenzo-p-dioxin
(TCDD) in aquatic model ecosystems. Environ. Sci. Technol. 9:668-672.
Jacobs, L.W.; Chou, S.F.; Tiedje, J.M. (1976) Fate of polybrominated bi-phenyls (PBBs)
in soil. Persistence and plant uptake. J. Agric. Food Chem. 24:1198-1201.
Jensen, D.J.; Hummel, R.A.; Mahle, N.H.; Kocher, C.W.; Higgings, H.S. (1981) A residue
study on beef cattle consuming 2,3,7,8-tetrachlorodibenzo-p-dioxin. J. Agric. Food Chem.
29:265-268.
Jensen, D.J.; Hummel, R.A. (1982) Secretion of TCDD in milk and cream following the
feeding of TCDD to lactating dairy cows. Bull. Environ. Contam. Toxicol. 19:440-446.
Kahn, P.D. (1987) Dioxin and dibenzofuran isomer distributions in blood and adipose
tissue of Vietnam veterans who were heavily exposed to Agent Orange and in matched
controls. Submitted to Science for publication.
Kaminski, L.S.; DeCapiro, A.P.; Gierthy, J.F.; Silkworth, J.B.; Tumasonis, C. (1985) The
rule of environmental matrices and experimental vehicles in chlorinated dibenzodioxin and
dibenzofuran toxicity. Chemosphere 14: 685-695.
Kao, J.; Patterson, F.K.; Hall, J. (1985) Skin penetration and metabolism of topically
applied chemicals in six mammalian species, including man: an in vitro study with
benzo(a)pyrine and testosterone. Toxicol. Appl. Pharmacol. 81(3):502-516.
Kapila, S.; Yanders, A.F.; Orazio, C.; Meadows, J.; Malhorta, R.K.; Cerlesi, S. (1987) Field
and laboratory studies on the movement of and fate of tetrachlorodibenzo-p-dioxins in
soil. Chemosphere (in press).
Karasek, F.W.; Clement, R.E.; Viau, A.C. (1982) Distribution of PCDDs and other toxic
compounds generated on fly ash particulates in municipal incinerators. J. Chromatography
239:173-180.
Kenaga, E.E.; Norris, L.A. (1983) Environmental toxicology of TCDD. In: Tucker, R.E.;
Young, A.L.; Gray, A.P., eds. Human and environmental risks of chlorinated dioxins and
related compounds. New York, NY: Plenum Press, pp. 277-299.
321
-------
Kimbrough, R.; Falk, H^ Stehr, S.; Fries, G. (1984) Health implications of 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) contamination of residential soil. J. Toxicol. Environ.
Health 14:47-93.
King, F.G.; Dedrick, R.L.; Collins, J.M.; Matthews, H.B.; Birnbaum, L.S. (1983)
Physiological model for the pharmacokinetics of 2,3,7,8-tetra-chlorodibenzofuran in
several species. Toxicol. Appl. Pharmacol. 67:390.
Kleiman, C.G. (1985) Fish consumption by recreational fisherman: an example of Lake
Ontario/Niagara region. Internal report dated May 20, 1985, prepared by Environ
Corporation for the U.S. Environmental Protection Agency, Office of Enforcement and
Compliance Monitoring, Washington, DC.
Kociba, R.J.; Keeler, P.A.; Park, C.N.; Gehring P.J. (1976) 2,3,7,8-Tetra-chlorodibenzo-p-
dioxin (TCDD): results of a 13-week oral toxicity study in rats. Toxicol. Appl.
Pharmacol. 35:553-573.
Kociba, R.J.; Keyes, D.G.; Beyer, J.E.; Carreon, R.M.; Wade, E.E.; Dittenber, D.A.; Kalning,
R.P.; Frauson, L.F.; Park, D.N.; Barnard, S.D.; Hummel, R.A.; Humiston, C.G. (1978)
Results of a two-year chronic toxicity and oncogenicity study of 2,3,7,8-
tetrachlorodibenzo-p-dioxin in rats. Toxicol. Appl. Pharmacol. 46(2):279-303.
Kuehl, D.W.; Cook, P.M.; Batterman, A.R.; Lothenback, D.B.; Butterworth, B.C.; Johnson,
D.L. (1985) Bioavailability of 2,3,7,8-tetrachlorodibenzo-p-dioxin from municipal
incinerator fly ash to freshwater fish. Chemosphere 14: 427-437.
Kuehl, D.W.; Cook, P.M.; Batterman, A.R.; Lothenbach, D.; Butterworth, B.C. (1987a)
Isomer dependent bioavailability of polychlorinated dibenzo-p-dioxins and dibenzofurans
from municipal incinerator fly ash to carp. Chemosphere 16(4): 657-666.
Kuehl, D.W.; Cook, P.M.; Batterman, A.R.; Lothenback, D.; Butterworth, B.C. (1987b)
Bioavailability of polychlorinated dibenzo-p-dioxins and dibenzofurans from contaminated
Wisconson River sediment to carp. Chemosphere 16(4): 667-676.
Lake, J.L.; Rubinstein, N; Pavignano, S. (1984) Predicting bioaccumulation: development
of a simple partitioning model for use as a screening tool for regulating ocean disposal
of waste. Presented at the Sixth Pillston Workshop, August 12-17, Florissant, CO.
In press: Fate and effects of sediment-bound chemicals in aquatic systems, Dickson, K.L.;
Maki, A.W.; Brungs W. (Eds).
Lamparski, L.L.; Nestrick, T.J. (1980) Determination of tetra-, hexa-, hepta-, and
octachlorodibenzo-p-dioxin isomers in particulate samples at parts per trillion levels.
Anal. Chem. 52:2045-2059.
Lepow, M.L.; Bruckman, L.; Rubino, R.A.; Markowitz, S.; Gillette, M.;
Kapish, J. (1974) Role of airborne lead in increased body burden of lead in Hartford
children. Environ. Health Perspect. 7:99 [As cited in Hawley (1985)].
Lepow, M.L.; Bruckman, L.; Rubino, R.A.; Markowitz, S.; Gillette, M.;
Kapish, J. (1975) Investigations into sources of lead in the environment of urban
children. Environ. Res. 10:415.
322
-------
Leung, H.; Paustenbach, J. (1987) A proposed occupational exposure limit for 2,3,7,8-
tetrachlorodibenzo-p-dioxin. Environmental Health and Safety, Syntex U.S.A. Inc., Palo
Alto, CA., J. Amer. Indust. Hygiene Assoc. (submitted).
Liberti, A.; Brocco, D.; Di Palo, V.; Possanzini, M. (1980) Evaluation of organochlorine
compounds in the emissions of urban incinerators. In: Vol. 3, Series on Environmental
Science, Analytical techniques in environmental chemistry. New York, NY: Pergamon
Press, pp. 157-166.
Liberti, A.; Brocco, D. (1982) Formation of polychlorodibenzodioxins and
polychlorodibenzofurans in urban incinerator emissions. In: Vol. 5, Series on
Environmental Science, Chlorinated dioxins and related compounds: impact on the
environment. New York, NY: Pergamon Press, pp 245-251.
Linsley, R.K.; Kohler, M.A.; Paulhus, J.L.H. (1982) Hydrology for engineers. New York,
NY: McGraw-Hill.
Longwoods Research Group, Limited. (1984) A usage segmentation analysis of the 1981
U.S. seafood consumption study (final report). Prepared for the Fisheries Council of
Canada, [as cited in Kleiman, C.G. (1985)]
Lucier, G.W.; Rumbaugh, R.C.; McCoy, Z.; Hass, R.; Harvan, D.; Albro, P. (1986) Ingestion
of soil contaminated with 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) alters hepatic
enzyme activities in rats. Fundam. Appl. Toxicol. 36:364-371.
Lustenhouwer, J.W.A.; Olie, K.; Hutzinger, O. (1980) Chlorinated dibenzo-p-dioxins and
related compounds in incinerator effluents: a review of measurements and mechanisms of
formation. Chemosphere 9(7/8):501-522.
Lyman, W.J.; Reehl, W.F.; Rosenblatt, D.H. (1982) Handbook of chemical property
estimation methods. New York, NY: McGraw-Hill.
Lyman, W.J.; Loreti, C.P. (1987) Prediction of soil and sediment sorption for organic
compounds. Final Report submitted by Arthur D. Little, Inc. to the U.S. Environmental
Protection Agency, Office of Water Regulations and Standards under EPA contract no.68-
01-6951, June, 1987.
Mackay, D.; Bobra, A.; Chan, D.W.; Shiu, W.Y. (1982) Vapor pressure correlations for
low-volatility environmental chemicals. Environ. Sci. Technol. 16(10):645-649.
Mackay, D.; Paterson, S.; Cheung, B. (1985) Evaluating the environmental fate of
chemicals: the fugacity level III approach as applied to 2,3,7,8-TCDD. Chemosphere
14(6/7):859-863.
Marple, L.; Brunck, R.; Throop, L. (1986a) Water solubility of 2,3,7,8-tetrachlorodibenzo-
p-dioxin. Environ. Sci. Technol. 20(2): 180-182.
Marple, L.; Berridge, B.; Throop, L. (1986b) Measurement of the water-octanol partition
coefficient of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Environ. Sci. Technol. 20(4):397-399.
323
-------
A-.; Wu* P.L.; Adjei, A.; Lindstrom, R.E.; Elworthy, P.H. (1982) Extended
Hildebrand solubility approach and the log-linear solubility equation. J. Pharmacol. Sci.
71(8^849-856'..
Malsumura, F.; Benezet, H.J. (1973) Studies on. the bioaccumulation and microbial
degradation of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Environ. Health Perspect. 5:253-258.
MeConnell, E.E.; Lucier, G.W.; Rumbaugh, R.C.; Albro, P.W.; Harvan, D.J.; Hass, J.R.;
Hacris-,. M.W. (1984) Dioxin.in soil; bioavailability after ingestion by rats and guinea
pigs. Science 223:1077-1079.
McKinney, J.D.; Fawkes, F.; Chae, K.; Oatley, S.; Coleman, R.E.; Briner, W. (1985)
2,3,7^ Tetrachlorodibenzo-p-dioxin (TCDD) as^a potent and persistent thynoxine agonist
a mechanist model for toxicity based on molecular reactivity. Environ. Health Perspect.
61:41-53.
Midwest Research Institute (1985) Emissions test report* city of Philadelphia northwest
and. east central municipal incinerators, Vol. 1 - technical report. Prepared for U.S.
Environmental Protection Agency, Region III..
ilU T.;. Rossi, M,; McMillen, D.; Coville, M.; Leung, D.; Spang.J. (1987) Photolysis of
tetrachlorodioxin and PCBs under atmospheric conditions. Internal report prepared by
SRI, International for the U.S.. Environmental Protection Agency, Office of Health and
Environmental Assessment, Washington, DC.
Miller, G.C; Zepp, R.G. (1987) 2,3,7,8-TCDD environmental chemistry. In: Exner, J.A.;
des Rosiers, P.E., eds. Solving hazardous waste, problems: dioxin. ACS Symposium Series.
In: press.
Nash, R.G'.; Beall, M.L. (1980) Distribution of silvex, 2,4-D, and TCDD applied to turf in
chambers and field plots. J. Agric. Food Chem. 28:614-623.
National Marine Fisheries Service (NMFS). (1971) Market facts survey, Fisheries Statistics
Program, U.S. Department of Commerce, Washington, D.C.
National Marine Fisheries Service (NMFS). (1981) Seafood attitudes and purchase data
available from NMFS. National Seafood Consumption Survey, Fisheries Statistics Program,
U.S. Department of Commerce, Washington, DC.
National Marine Fisheries Service (NMFS). (1983) Fisheries of the U.S. 1982. Current
Fisheries Statistics No. 8300.. U.S. Department of Commerce.
National Marine Fisheries Service (NMFS). (1984) Fisheries of the U.S. 1983. Current
Fisheries Statistics No. 8320. U.S. Dept. of Commerce.
Nau, H.; Bass, R. (1981) Transfer of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) to the
mouse embryo and fetus. Toxicology 20(4):299.
Nielsenv K.K.; Moeller, J.T.; Rasmussen, S. (1986) Reduction of dioxins and furans by
spray dryer absorption from incinerator flue gas. Chemosphere 15(9- 12): 1247- 1254.
324
-------
Nkedi-Kizza, P.; Rao, P.S.C.; Hornsby, A.G. (1985) Influence of organic cosolvents on
sorption of hydrophobic organic chemicals by soils. Environ. Sci. Technol. 19:975-979.
Nottrodt, LA. (1986) Causes for and reduction strategies against emissions of
PCDD/PCDF from waste incineration plants. Chemosphere 15(9-12): 1225-1237.
Olie, K.; Vermeuler, P.; Hutzinger, O. (1977) Chlorodibenzo-p-dioxins and
chlorodibenzofurans are trace components of fly ash and flue gas of some municipal
incinerators in the Netherlands. Chemosphere 6(8):455-459.
Ogden Projects, Inc. (1986) Environmental test report for Walter B. Hall Resource
Recovery Facility Units 1 and 2, Tulsa, OK.
Osborne, R.; Greenlee, W.F. (1985) 2,3,7,8-Tetrachlorodibenzo-p-dioxin enhances terminal
differentiation of cultured human epidermal cells. Toxicol. Appl. Pharmacol. 77(3):434-443.
Paustenbach, D.J.; Shu, H.P.; Murray, F.J. (1986) A critical examination of assumptions
used in risk assessments of dioxin contaminated soil. Regul. Toxicol. Pharmacol.
6:284-307.
Kilgroe, J.A. (1986) Memorandum with attachments titled "The national incinerator testing
and evaluation program: Air pollution control technology", to addresses from James D.
Kilgroe, U.S. EPA Air and Energy Engineering Research Laboratory, RTP, October 15,
1986.
Podoll, R.T.; Jaber, H.M.; Mill, T. (1986) Tetrachlorodibenzodioxin: rates of volatilization
and photolysis in the environment. Environ. Sci. Technol. 20(5):490-492.
Poiger, H.; Schlatter, C. (1980) Influence of solvents and adsorbents on dermal and
intestinal adsorption of TCDD. Food Cosmet. Toxicol. 18:477-481.
Puffer, H.W.; Zen, S.P.A.; Duda, M.J.; Young, D.R. (1983) Consumption rates of
potentially hazardous marine fish caught in the metropolitan Los Angeles area. Project
summary. U.S. Environmental Protection Agency, Environmental Research Laboratory,
Corvallis, OR. EPA-600/53-82-070.
Radian Corporation. (1983) Review and development of chlorinated dioxins and furans
emissions data. Prepared for the Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, NC.
Radian Corporation. (1986) Characterization of the municipal waste combustion industry
(Appendix A). Prepared for the Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, NC, under EPA contract
68-02-4330.
Rappe, C.; Buser, H.R.; Stalling, D.L.; Smith, L.M.; Dougherty, R.C. (1981) Identification of
polychlorinated dibenzofurans in environmental samples. Nature 292:524-526.
Rappe, C.; Marklund, S.; Kjeller, L.O.; Tysklind, M. (1986) PCDDs and PCDFs in
emissions from various incinerators. Chemosphere 15(9-12): 1213-1217.
325
-------
Rizek, R.L. (1978) The 1977-1978 nationwide food consumption survey. Family
Economics Review, p. 3.
Roels, H.; Buchet, J.; Lauwerys, R.R.; Bruaux, P.; Thoreau, F.C.; Lafontaine, A.; Verduyn,
G. (1980) Exposure to lead by the oral and pulmonary routes of children living in the
vicinity of a primary lead smelter. Environ. Res. 22:81-94.
Rose, J.Q.; Ramsey, J.C.; Wentzler, T.H., (1976) The fate of 2,3,7,8-tetrachlorodibenzo-p-
dioxin following single and repeated oral doses to the rat. Toxicol. Appl. Pharmacol.
36:209-226.
Sacchi, G.A.; Vigano, P.; Fortunati, G; Cocucci, S.M. (1986) Accumulation of 2,3,7,8-
TCDD from soil and nutrient solution by bean and maize plants. Experientia 42:586-588.
Sarna, L.P.; Hodge, P.E.; Webster, G.R.B. (1984) Octanol-water partition coefficients of
chlorinated dioxins and dibenzofurans by reversed-phase HPLC using several CIS columns.
Chemosphere 13(9):975-983.
Schaffer, S.A. (1985) Environmental transfer and loss parameters for four selected
organic priority pollutants. Proceedings of the National Conference on Hazardous Wastes
and Environmental Emergencies, Cincinnati, OH., Library of Congress Catalog No. 85-
60904, Hazardous Material Control Research Institute, Silver Springs, Md.
Schaum, J. (1984) Risk analysis of TCDD contaminated soil. U.S. Environmental
Protection Agency, Washington, DC. EPA-600/8-84-031.
Scheidl, K.; Kuna, R.P.; Wurst, F. (1985) Chlorinated dioxins and furans in emissions
from municipal incineration. Chemosphere 14(6/7):913-917.
Schroy, J.M.; Hileman, F.D.; Cheng, S.C. (1985a) Physical/chemical properties of 2,3,7,8-
tetrachlorodibenzo-p-dioxin. In: Bahner, R.C.; Hansen, D.J., eds. Aquatic toxicology and
hazard assessment eighth symposium. Philadelphia, PA: American Society for Testing and
Materials, pp. 409-421.
Schroy, J.M.; Hileman, F.D.; Cheng, S.C. (1985b) Physical/chemical properties of 2,3,7,8-
TCDD. Chemosphere 14(6/7):877-880.
Science Applications International Corp. (1986) Conversion factors for use in estimating
environmental transport of contaminated soil from unregulated disposal sites. Prepared for
the U.S. Environmental Protection Agency, Office of Solid Waste, Washington, DC, under
EPA contract 68-01-7624.
Sendroy, J.; Cecchini, L.P. (1954) Determination of human body surface area from
height and weight. J. Appl. Physiol. 7(1):1-12.
Sheffield, A. (1985) Sources and releases of PCDD's and PCDF's to the Canadian
environment. Chemosphere 14(6/7):811-814.
Silkworth, J.; McMartin, D.; DeCaprio, A.; Rej, R.; O'Keefe.P.; Kaminsky, L. (1982) Acute
toxicity in guinea pigs and rabbits of soot from a polychlorinated biphenyl-containing
transformer fire. Toxicol. Appl. Pharmacol. 65:425-439.
326
-------
Snyder, W.S. (1975) Report of the task group on reference man. International
Commission of Radiological Protection, No. 23, New York, NY: Pergamon Press.
Stephan, C.E. (1980) Memorandum titled "Per capita consumption of non-marine fish and
shellfish" dated July 3, 1980, to J. Stara, U.S. EPA Environmental Criteria and Assessment
Office, Cincinnati, OH., from Charles E. Stephan, Environmental Research Lab, Duluth,
Minn.
Suskind, R.R. (1977) Environment and the skin. Environ. Health Perspect. 20:27-37.
Swift, L.L.; Gasiewicz, T.A.; Dunn, G.D.; Soule, P.O.; Neal, R.A. (1981) Characterization
of hyperlipidemia in guinea pigs induced by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol.
Appl. Pharmacol. 59(3):489.
Thibodeaux, L.J. (1979) Chemodynamics. New York, NY: John Wiley and Sons, Inc.
Thibodeaux, L.J. (1983) Offsite transport of 2,3,7,8-tetrachlorodibenzo-p-dioxin from a
production disposal facility. In: Choudhary, G., eds. Chlorinated dioxins and
dibenzofurans in the total environment. Chapter 5, Boston, MA: Butterworth.
Thibodeaux, L.J.; Becker, B. (1982) Chemical transport rates near the surface of
wastewater surface impoundments and similar water bodies. Environ. Prog.(3) 296-300.
Thibodeaux, L.J.; Lipsky, D. (1985) A fate and transport model for 2,3,7,8-
tetrachlorodibenzo-p-dioxin in fly ash on soil and urban surfaces. Hazardous Waste and
Hazardous Materials 2:225-235.
Thibodeaux, L.J.; Reible, D.D.; Fang, C.S. (1986) Transport of chemical contaminants in
the marine environment originating from offshore drilling bottom deposits—a vignette
model. In: Cohen, Y., ed. Pollutants in a multimedia environment. New York, NY:
Plenum Publishing Corp.
Todd, O.K. (1980) Ground water hydrology, 2nd ed., New York, NY: John Wiley and Sons.
Tong, H.Y.; Karasek, F.W. (1986) Comparison of PCDD and PCDF in fly ash collected
from municipal incinerators of different countries. Chemosphere 15(9-12):1219-1224.
Tosine, H.M.; Clement, R.E.; Ozvacic, V.; Wong, G. (1985) Levels of PCDD/PCDF and
other chlorinated organics in municipal refuse. Chemosphere 14(6/7): 821-827.
Tschirley, F.H. (1986) Dioxin. Scientific American 254(2):29-35.
Tung, L.S.; Freeman, R.A.; Schroy, J.M. (1985) Prediction of soil temperature profiles: a
concern in the assessment of transport of low volatility chemicals in the soil column.
Presented at the 1985 Summer National Meeting of the American Institute of Chemical
Engineers (AlChE); August, Seattle, WA.
Turner, D.B. (1970) Workbook of atmospheric dispersion estimates. PHS Publication No.
999-AP-26 (NTIS PB 191482), U.S. EPA, RTP, NC.
Umbreit, T.H.; Patel, D.; Gallo, M.A. (1985) Acute toxicity of TCDD contaminated soil
from an industrial site. Chemosphere 14:945-947.
327
-------
Umbreit, T.H.; Hesse, E.J.; Gallo, M.S. (1986a) Bioavailability of dioxin in soil from a
2,4,5-T manufacturing site. Science 232:497-499.
Umbreit, T.H.; Hesse, E.J.; Gallo, M.S. (1986b) Comparative toxicity of TCDD
contaminated soil from Times Beach, Missouri, and Newark, New Jersey. Chemosphere,
15(9-12):2121-2124.
Umbreit, T.H.; Hesse, E.J.; Gallo, M.S. (1987a) Reproductive toxicity of female
mice of dioxin-contaminated soils from a 2,4,5-trichlorophenoxyacetic acid manufacturing
site. Arch. Environmental Contamination & Toxicology 16, 461-466.
Umbreit, T.H.; Hesse, E.J.; Gallo, M.S. (1987b) Differential bioavailability of
2,3,7,8-tetrachlorodibenzo-p-dioxin from contaminated soils. American Chemical
Society Symposium Series No. 338, "Solving hazardous waste problems: Learning from
dioxin", American Chemical Society, Washington, D.C.
U.S. Department of Agriculture. (1966) Household food consumption survey 1965-1966.
Report 12. Food consumption of households in the U.S., Seasons and years 1965-1966
Washington, D.C., U.S. Government Printing Office.
U.S. Department of Agriculture. (1985) Food and nutrient intakes: individuals in four
regions, 1977-1978. Report 1-3. Nationwide Food Consumption Survey 1977-1978.
U.S. Department of Commerce. (1985) Fisheries of the United States, April 1985.
Current fisheries statistics No. 8360. [As cited in U.S. EPA (1986b)]
U.S. EPA. (1976) Areawide assessment procedures manual. Vol. 1. Municipal Environmental
Research Laboratory, Cincinnati, OH. EPA-600/9-76-014.
U.S. EPA. (1979) Industrial source complex (ISC) dispersion model user's guide. Office
of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-450/4-79-030.
U.S. EPA. (1980a) Dioxins. Industrial Environmental Research Laboratory, Cincinnati, OH.
EPA-600/2-80-197.
U.S. EPA. (1980b) Seafood consumption data analysis. Prepared for the Office of Water
Regulations and Standards, Washington, DC, by SRI International under EPA contract
No.68-01-3887.
U.S. EPA. (1981a) Aquatic fate process data for priority pollutants. Mabey, W.R.;
Smith, J.H.; Podoll, R.T.; Johnson, H.L.; Mill, T.; Chou, T.W.; Gates, J.; Partridge, I.W.;
Vandenberg, D. Draft report dated July, 1981. Office of Water Regulations and
Standards, Washington, DC. EPA-440/4-81-014.
U.S. EPA. (198Ib) Risk assessment on (2,4,5-trichlorophenoxy)acetic acid (2,4,5-T),
(2,4,5-trichlorophenoxy)propionic acid (silvex), and 2,3,7,8-tetrachIorodibenzo-p-diOxin
(TCDD). Office of Health and Environmental Assessment, Washington, DC.
EPA-600/6-81-003. NTIS PBS 1-234825.
U.S. EPA. (1981c) The potential atmospheric impact of chemicals released to the
environment. Office of Toxic Substances, Washington, DC EPA-560/5-80-001.
328
-------
U.S. EPA. (1983a) Dioxin strategy. Internal report dated November 28, 1983. Office of
Water Regulations and Standards, Office of Solid Waste and Emergency Response in
conjunction with Dioxin Strategy Task Force, Washington, DC.
U.S. EPA. (1983b) Compilation of air pollution factors (AP-42, Supplement No. 14).
Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. EPA. (1984a) Ambient water quality criteria document for 2,3,7,8-tetra-
chlorodibenzo-p-dioxin. Office of Water Regulations and Standards, Washington, DC.
EPA-440/5-84-007.
U.S. EPA. (1984b) Personal Communication, September 4, 1986, Barbara Peterson,
Peterson and Associates, Inc., Bethesda, MD, regarding Office of Pesticides Program
Tolerance Assessment System: Crop to Food Map, Draft Report, August, 1984. (Data
analyzed was compiled in the USDA Nationwide Food Consumption Survey, 1977-78.)
U.S. EPA. (1984c) An estimation of the daily average food intake by age and sex for use
in assessing the radionuclide intake of individuals in the general population. Office of
Radiation Programs, Washington, DC. EPA-520/1-84-021.
U.S. EPA. (1984d) An estimation of the daily food intake based on data from the
1977-1978 USDA Nationwide Food Consumption Survey. Office of Radiation Programs,
Washington, DC. EPA-520/1-84-015.
U.S. EPA. (1984e) Stochastic processes applied to risk analysis of TCDD contaminated
soil: a case study. Internal report dated May 31, 1984; Office of Health and
Environmental Assessment, Exposure Assessment Group, Washington DC.
U.S. EPA (1985) Compilation of air pollutant emission factors, Vol. 1. Office of Air
Quality Planning and Standards, Research Triangle Park, N.C.
U.S. EPA. (1985a) Development of statistical distributions or ranges of standard factors
used in exposure assessments. Office of Health and Environmental Assessment,
Washington, DC. EPA-600/8-85-010. NTIS PB85-242667/AS.
U.S. EPA. (1985b) Dioxin transport from contaminated sites to exposure locations: a
methodology for calculating conversion factors. Office of Health and Environmental
Assessment, Washington, DC. EPA-600/8-85-012. NTIS PB85-214310.
U.S. EPA. (1985c) Rapid assessment of exposure to particulate emission from surface
contamination sites. Office of Health and Environmental Assessment, Washington, DC.
EPA-600/8-85-002. NTIS PB85-192219/AS.
U.S. EPA. (1985d) Health assessment document for polychlorinated dibenzo-p-dioxins.
Office of Health and Environmental Assessment, Environmenatl Criteria and Assessment
Office, Cincinnati, OH. EPA-600/8-84-014F. NTIS PB86-122546.
U.S. EPA. (1985e) Pollutant sorption to soils and sediments in organic/ aqueous solvent
systems. Office of Environmental Processes and Effects Research, Athens, GA.
EPA-600/3-85-050. NTIS PB85-242535.
329
-------
U.S. EPA (1985f) Modeling remedial actions at uncontrolled hazardous waste sites. Office
of Solid Waste and Emergency Response, Washington, D.C. EPA-540/2-85-001.
U.S. EPA (1985g) Proceeding of the eleventh annual research symposium, Leaching
potential of 2,3,7,8-TCDD in contaminated soils. U.S. EPA Hazardous Waste Engineering
Research Laboratory, Cinn, OH. EPA-600/9-85-013.
U.S. EPA. (1986a) Development of advisory levels for polychlorinated biphenyls (PCBs)
cleanup. Office of Health and Environmental Assessment, Washington, DC. EPA-600/6-86-
002. NTIS PB86-232774/AS.
U.S. EPA. (1986b) Development of risk assessment methodology for ocean disposal of
municipal sludge. Office of Health and Environmental Assessment, Environmental Criteria
and Assessment Office, Cincinnati, OH. (ECAO-CIN-492).
U.S. EPA. (1986c) Guidance manual for health risk assessment of chemically
contaminated seafood. Puget Sound Estuary Program, U.S. EPA Region X, Seattle, WA.,
Tetra Tech, TC-3991-07, June 1986.
U.S. EPA. (1986d) National dioxin study, tier 4: combustion sources. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. EPA-450/4-84-014g.
U.S. EPA. (1986e) Air quality modeling analysis of municipal waste combustors. Internal
report dated November, 1986, prepared by PEI Associates, Inc. and H.E. Cramer Company
Inc. for the Office of Air Quality Planning and Standards, Research Triangle Park, NC.
U.S. EPA (1987) Report on 2,3,7,8-TCDD Body Burdens. Submitted by Technical Resources,
Inc. to the Exposure Assessment Group, Office of Health and Environmental Assessment,
Washington, DC., under EPA Contract No. 68-02-4199.
U.S. EPA (1987a) Guidance manual for assessing human health risks from chemically
contaminated fish and shellfish. Draft report C737-01, dated December 1987, submitted to
the Office of Water, Washington, DC, by PTI Environmental Services Inc. under EPA
contract No. 68-03-3319.
U.S. EPA (1987b) National dioxin study, tier 4: combustion sources - engineering analysis
report, Office of Air Quality Planning and Standards, Research Triangle Park, EPA-450/4-
84-014h, September 1987.
U.S. EPA (1987c) Municipal waste combustion study - emission data base for municipal
easte combustors, Office of Solid Waste and Emergency Response, EPA/530-SW-87-021b,
June 1987.
U.S. EPA (1987d) Memorandum and attachments titled "Preliminary risk calculations for
polychlorinated dibenzo-p-dioxins using the ECAO-Cin risk assessment methodology for
municipal waste combustors", from Larry Fradkin, Acting Branch Chief, Systemic
Toxicants Assessment Branch, Environmental Criteria and Assessment Office, to Michael
Callahan, Office of Health and Environmental Assessment, February 9, 1987.
van den Berg, M.; Olie, K.; Hutzinger, O. (1983) Uptake and selective retention in rats
of orally administered chlorinated dioxins and dibenzofurans from fly-ash and fly-ash
extract. Chemosphere 12:537-544.
330
-------
van den Berg, M.; Vroom, A.; van Greevenbroek, M.; Olie, K.; Hutzinger, O. (1985)
Bioavailability of PCDDs and PCDFs adsorbed on fly ash in the rat, guinea pig, and
Syrian golden hamster. Chemosphere 14:865-869.
Vick, R.D.; Junk, G.A.; Avery, M.J.; Richard, J.J.; Svec, H.J. (1978) Organic emissions
from combustion of combination coal/refuse to produce electricity. Chemosphere
7:893-902.
Wade, J.C.; Heady, E.O. (1978) Measurement of sediment control impacts on agriculture.
Water Resources Research: 14:1-8.
Walters, R.W.; Guiseppi-Elie, A.; Yousefi, Z.; Means, J.C. (1987) Sorption of Dioxin to
Soils. Chemosphere (in press).
Ward, C.T.; Matsumura, F. (1978) Fate of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in
a model aquatic environment. Arch. Environ. Contam. Toxicol. 7:349-357.
Wark, K.; Warner, C. (1981) Air pollution--its origin and control. New York, NY:
Harper and Row Publishing Co.
Weber, H.; Poiger, H.; Schlatter, C. (1982) Fate of 2,3,7,8-tetrachloro-dibenzo-p-dioxin
metabolites from dogs in rats. Xenobiotica 12(6):353.
Webster, G.R.B.; Friesen, K.J.; Sarna, L.P.; Muir, D.C.G. (1985) Environmental fate
modelling of chlorodioxins: determination of physical constants. Chemosphere
14(6/7):609-622.
Webster, G.R.B.; Muldrew, D.H.; Graham, J.J.; Sarna, L.P.; Muir, D.C.G. (1986) Dissolved
organic matter mediated aquatic transport of chlorinated dioxins. Chemosphere
15(9-12):1379-1386.
Wipf, H.K.; Schmid, J. (1983) Seveso - an environmental assessment, In: Tucker, R.E.;
Young, A.L.; Gray.A.P. eds., Human and environmental risks of chlorinated dioxins and
related compounds. New York, N.Y.: Plenum Press.
Wipf, H.K.; Homberger, E.; Neuner, N.; Ranalder, U.B.; Vetter, W.; Vuilleumier, J.P. (1982)
TCDD levels in soil and plant samples from Seveso area. In: Hutzinger, O. eds.
Chlorinated dioxins and related compounds: impact on the environment. Vol. 5, Series on
Environmental Science. New York, NY: Pergamon Press, pp. 115-126.
Wisler, C.O. ; Brater, E.F. (1959) Hydrology, 2nd edition, New York, N.Y.: John Wiley and
Sons.
Wolfe, H.R.; Armstrong, J.F.; Durham, W.F. (1974) Exposure of mosquito workers to
fenthion. Mosquito News 34:263. [As cited in Hawley (1985)]
Wroblewski, V.J.; Olson, J.R. (1985) Hepatic metabolism of 2,3,7,8-tetrachlorodibenzo-p-
dioxin in the rat and guinea pig. Toxicol. Appl. Pharmacol. 81:(2)231.
Yalkowsky, S.H.; Flynn, G.L.; Amidon, G.L. (1972) Solubility of nonelectrolytes in polar
solvents. J. Pharm. Sci. 61(6):983-984.
331
-------
Yaitsowsky, S.H.; Valvani, S.C.; Amidon, G.L. (1976) Solubility of nonelectrolytes in
polar, solvents IV: nonpolar drugs in. mixed'solvents. J. Pharm. Sci. 65(10): 1488-1494.
Young;, A.L. (1983) Long term studies on the persistence and movement of TCDD in a
national! ecosystem. In: Tucker, A. eds. Human and. environmental risks of chlorinated
dibxihs and related compounds. New York, NY: Plenum Publishing Corp.
Zeppi R.G.; Miller, G.C.; Herbert, V.R.; Mille, NO.; Mitzel, R. (1988) Photolysis of
octaehlorodibenzo-p-dioxin on soils; production of 2,3,7,8-TCDD. Chemosphere (in press).
332
-------
APPENDIX. RISK ESTIMATES FOR CHAPTER 6 SCENARIOS
333
-------
APPENDIX. RISK ESTIMATES FOR CHAPTER 6 SCENARIOS
The procedure used to calculate an upper-limit incremental cancer risk at low doses
is as follows:
Upper-limit incremental cancer risk = 1 - e"****
= qi*d, when qd < 10~3 (A-l)
where q is the cancer potency factor and d is the dose (discussed further below). The
95% upper confidence limit of the linear slope, factor (qj ) of Dose-Response Function for
2,3,7,8-TCDD is 0.156 kg-d/ng). The derivation of this factor is described in U.S. EPA
(1984a) and further background is provided in U.S. EPA, 1981b. The Agency is currently
reevaluating this potency factor and may change it in the near future.
EPA's Carcinogen Assessment Group (Farland, 1987a) is also currently considering
adjusting the potency factor for exposures that occur to children only (such as soil
ingestion). This is due to the fact that 1) children may have greater sensitivity than
adults and 2) the scaling procedure used in deriving a potency factor uses the surface
area of adults (not children). Currently, no final decision has been made on how to
make this adjustment and no adjustment was made in this document.
Dose as used in Equation A-l is the rate at which a chemical is absorbed into the
body and is typically expressed in units of mg/kg-d. The exposure estimates presented in
Chapter 6 reflect the rate at which chemicals contact the body and do not account for
absorption into the body. Typically, such exposures can be converted to doses by
multiplying the exposure by the absorption fraction (fraction of chemical contacting the
body which enters the body). However, in using Equation A-l, additional consideration
must be given to the differences between the absorption occurring during the human
334
-------
exposure event of interest and the absorption which occurred during the animal
experiment used to derive the potency factor. These differences may be due to
differences between animals and humans and/or differences between the vehicle carrying
the contaminant used in the animal study versus that of the human exposure event. The
absorption which occurred during the animal experiments, which EPA used to derive the
potency factor for 2,3,7,8-TCDD, was estimated to be 55% (Farland, 1987b). Thus, if
55% absorption also occurs during the human exposure event of interest then no
adjustment to the exposure estimate is needed when calculating risk. However, if the
human exposure is different from 55%, the exposure level must be adjusted accordingly.
This is accomplished via the following equation.
Risk = (potency factor) (exposure) (human absorption fraction)/0.55 (A-2)
The risk estimates presented here were calculated using Equation A-2, a potency
factor of 0.156 kg-d/ng, an exposure corresponding to the appropriate level listed in
Table 6-5, and a human absorption fraction corresponding to the appropriate value
listed in Table 6-4.
The risk estimates for all land scenarios are presented in Table A-l and
incinerator/ash scenarios in Table A-2.
335
-------
TABLE A-l. UPPER-BOUND INCREMENTAL CANCER RISKS
ASSOCIATED WITH VARIOUS EXPOSURE PATHWAYS/SCENARIOS:
CONTAMINATED SOILa
Scenario
Dairy Beef
inges- inges-
tion tion
Fish Soil
inges- ingea-
tion tion
Vapor Dust
inhala- inhala-
tion tion
Drink- Veget-
Soil ing able
dermal water inges-
tion
l)l;ppb
1 acre
reasonable
worst-case
NA NA NA SxlO'4 2xlO"6 SxlO"7 2xlO"5 NA
See
text
2)1 ppb
10 acres
reasonable
worst-case
,2xlO"3 IxlO"2 4xlO"2 SxlO"4 SxlO"6 SxlO'7 2xlO"5 9x10"
See
text
3)1 ppt
10 acres
reasonable
worst-case
:2xlO"6 1x10"5 4x10"5 .SxlO"7 SxlO"9 SxlO"10 2x10"8 9xlO"9
See
text
4)1 .ppq
10.acres
reasonable
worst-case
2xlO"9 IxlO"8 4xlO"8 SxlO"10 SxlO"12 SxlO"13 2xlO"n 9xlO"12
See
text
5)1 ppb
10 acres
typical
5xlO~4 SxlO"3 5xlO"6 4xlO"6 2xlO"6 IxlO"7 6xlO"6 SxlO"9
See
text
6)1 ppt
10 acres
typical
BxlO"7 SxlO"6 5xlO"9 4xlO"8 .2xlO"9 IxlO"10 6xlO"9 SxlO"12 See
text
7)1 ppq
10 acres
typical
SxlO"10 .SxlO"9 BxlO"12 4X10"11 .J2xlO"12 IxlO"13 6xlO"12 SxlO"15
See
text
(Continued )
336
-------
TABLE A-l. (CONTINUED)
Scenario
Dairy Beef Fish Soil Vapor Dust Drink- Vegeta-
inges- inges- inges- inges- inhala- inhala- Soil ing ble
tion tion tion tion tion tion dermal water i n g e s t i o n
8)1 ppb
1 acre
reasonable
worst case
9xlO"5 6xlO"4 2xlO"3 IxlO"5 2xlO~7 4xlO"8 SxlO"7 4xlO"7
See
text
9)1 ppb
10 acres
reasonable
worst case
6xlO~4 4xlO~3 IxlO"2 IxlQ"4 7xlO"7 SxlO'8 6xlO"6 SxlO"6
See
text
10)1 ppt
10 acres
reasonable
worst case
6xlO"7 4xlO"6 IxlO"6 IxlO"7 7xlO"10 SxlO"11 6xlO"9 SxlO"9
See
text
11)1 ppq
10 acres
reasonable
worst case
6xlO'10 4xlO'9 IxlO"8 IxlO"10 7xlO"13 SxlO"14 6xlO"12 SxlO"12
See
text
12)1 ppb
10 acres
typical
13)1 ppt
10 acres
typical
4xlO"6 2xlO"5 5xlO"6 SxlO"7 2xlO"7 4xlO"9 5xlO"8 SxlO"9
4xlO"9 2xlO"8 5xlO"9 SxlO"10 2xlO"10 4xlO"12 SxlO"11 SxlO"12
See
text
See
text
(Continued )
337
-------
TABLE A-l. (CONTINUED)
Scenario
14)1 ppq
10 acres
typical
15)1 ppb
10 acres
capped land fill
reasonable worst -case
Dairy Beef FUh Soil Vapor Dust Drink- Vegeta-
inges- inges- inges- inges- inhala- inhala- Soil ing ble
tion tion tion tion tion tion dermal water ingestion
4x10" 12 2x10" U 5x10' 12 3x10' 1S 2x10" 1S 4x10" 15 5x10" 14 3x10" 15
See
text
neg.c neg. neg. neg. 4x10" neg. neg. neg. neg.
alf the 'cancer potency factor recognized by the Agency is revised, the risk associated with any scenario and pathway may be
obtained by multiplying the corresponding entry from Table A-l by the factor (revised potency factor/0.156 kg-d/ng).
These scenarios assume that the exposure area is downgradient of the contaminated source. If the exposure area was
located upgradient and all transport was via windblown dust, these risks would be reduced by a factor of 1,000.
cneg. = negligible risk (<10~8)
338
-------
TABLE A-2. UPPER-BOUND INCREMENTAL CANCER RISKS ASSOCIATED
WITH INCINERATOR EXPOSURE PATHWAYS/SCENARIOS3
A. RISKS ASSOCIATED WITH STACK EMISSIONS
Dairy Beef Fish Soil Vapor Participate Soil Drinking Vegetable
Scenario Ingestion Ingestion Ingestion Ingeation Inhalation Inhalation Dermal Water Ingestion
16) 3000 TPD 3 x 10"6 8 x 10'6 2 x 10~7 2 x 10"7 1 x 10"7 3 x 10~8 1 x 10~8 8 x 10"9 See text
reasonable
worst case
17) 120 TPD 1 x 10~4 4 x 10"4 1 x 10"5 9 x 10"6 1 x 10"6 2 x 10'6 B x 10~7 6 x 10"7 See text
reasonable
worst case
18) 3000 TPD 2 x 10"6 2 x 10"6 2 x 10"12 2 x 10~8 5 x 10'8 1 x 10~8 4 x 10~9 3 x 10"13 See text
typical
case
19) 120 TPD 8xlO'5 9xlO"B 2 x 10"10 1 x 10"8 4 x 10'6 9 x 10"7 2 x 10'7 2 x 10'11 See text
typical
case
a If the cancer potency factor recognized by the Agency is revised, the risk associated with any scenario and
pathway may be obtained by multiplying the corresponding entry from Table A-2 by the ratio (revised potency
factor/0.156 kg-d/ng).
b at 200 m from stack.
339
-------
TABLE A-2. (CONTINUED)
B. RISKS ASSOCIATED WITH FLY ASH DISPOSAL
Dairy Beef FUh Soil Vapor Participate Soil Drinking Vegetable
Scenario Ingestion Ingestion Ingestion Ingestion Inhalation Inhalation Dermal Water Ingestion
16) 3000 TPD 2 x 10'3 1 x 10'2 4 x 10"6 3 x 10"4 1 x 10'6 2 x 10"8 2 x 10~5 4 x 10~6 See text
reasonable ,
worst case
17) 120 TPD 3 x 10'4 2 x 10'3 7 x 10"7 6 x 10'6 6 x 10'7 8 x 10"9 3 x 10'6 6 x 10"7 See text
reasonable
worst case
18) 3000 TPD 2xlO"4 2 x 10"3 7xlO"10 2 x 10"B 6 x 10"7 4 x 10"9 8 x 10"9 2 x 10"9 See text
typical
19) 120 TPD 1 x 10~6 1 x 10'4 7 x 10'10 1 x 10"6 1 x 10'7 1 x 10"9 8 x 10'9 2 x 10'9 See text
typical
20) 3000 TPD neg.c neg. neg. neg. neg. neg. neg. neg. neg.
reasonable
worst case
capped
a If the cancer potency factor recognized by the Agency is revised, the risk associated with any scenario and
pathway may be obtained by multiplying the corresponding entry from Table A-2 by the ratio (revised potency /factor/
0.156 kg-d/ng).
All scenarios assume 1.3 ppb 2,3,7,8-TCDD in fly ash.
c Negligible risk (<10~8).
Stack emissions part of this scenario is the same as scenario 16.
340
r U.SaOVERNUtNTmNTMOOFFICE: IBM. 548-OIO/81TOS
------- |