United States
Environmental
Protection Agency
Off Ice of Solid Waste Off Ice of Air Off Ice of Research EPA/530-SW-87-02ic
and Emergency Response and Radiation and Development September 1987
Washington, DC 20460 Washington, DC 20460 -Washington, OC 20460
xvEPA
Municipal Waste
Combustion Study
Assessment of Health Risks
Associated With Municipal
Waste Combustion Emissions
-------
September 1987
MUNICIPAL WASTE COMBUSTION STUDY:
ASSESSMENT OF HEALTH RISKS ASSOCIATED WITH
MUNICIPAL WASTE COMBUSTION EMISSIONS
For information contact:
Mr. David Cleverly
Standards and Air Strategies Division
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
ACKNOWLEDGMENTS
AUTHORS
D. Cleverly
Pollutant Assessment Branch
Office of Air Quality Planning
and Standards
U.S. Environmental Protection Agency
Office of Air and Radiation
Research Triangle Park, NC 27711
L. Fradldn
Environmental Criteria and
Assessment Office
Office of Health and Environmental
Assessment
Office of Research and Development
Cincinnati, OH 45268
R. J. F. Bruins
Environmental Criteria and
Assessment Office
Office of Health and Environmental
Assessment
Office of Research and Development
Cincinnati, OH 45268
G. E. Wilkins
Radian Corporation
Research Triangle Park, NC
27709
K. K. Fidler
Radian Corporation
Research Triangle Park, NC 27709
P. M. McGinnis
Center for Chemical Hazard
Assessment
Syracuse Research Corporation
Syracuse, NY 13210
G. W. Dawson
ICF Northwest
Richland, WA 99352
R. Bond
ICF Northwest
Richland, WA 99352
CONTRIBUTORS AND REVIEWERS
R. G. Kellam
Risk Assessment
Pollutant- Assessment Branch
Office of Air Quality Planning
and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
R. Morrison
Risk Assessment
Pollutant Assessment Branch
Office of Air Quality Planning
and Standards
U.S. Environmental .Protection Agency
Research Triangle Park, NC 27711
B. Riddle
Human Exposure Model
Pollutant Assessment Branch
Office of Air Quality Planning
and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
N. C. Possiel
Air Dispersion Modeling
Source Receptor Analysis Branch
Office of Air Quality Planning
and Standards
U.S. Environmental Protection Agency
Research. Triangle Park, NC 27.711
H. V. Geary
Air Dispersion Modeling
Salt Lake City, UT 84115
G. J. Schewe
Air Dispersion Modeling
PEI Associates, Inc.
Cincinnati, OH 45246
ii
-------
TABLE OF CONTENTS
Section Page
ACKNOWLEDGMENTS i i
LIST OF TABLES vi
LIST OF FIGURES vi i i
1.0 Introduction 1-1
1.1 References . 1-4
2.0 Summary of Results 2-1
2.1 Introduction 2-1
2.2 Summary of Risk Assessment Methodology and Parameters 2-1
2.2.1 Modeling for Direct Exposure Pathway 2-3
2.2.2 Modeling for Indirect Exposure Pathway 2-4
2.3 Results of Direct Inhalation Exposure Pathway Risk
Assessment .' 2-6
2.3.1 Human Carcinogenic Risks 2-6
2.3.1.1 Ranges and Uncertainties 2-6
2.3.2 Direct Exposure Noncarcinogenic Risks of
Lead and Mercury Emissions 2-12
2.3.3 Human Welfare Effects .., 2-13
2.4 Results of Indirect Exposure Pathway Analysis 2-17
2.4.1 Indirect Exposure to Municipal Waste Combustor
Emissions 2-17
2.4.2 Potential Environmental Effects 2-21
2.5 Potential Reduction of Carcinogenic Risk, Welfare Effects,
and Indirect Exposure with Improved Emission Control ... 2-22
2.6 References 2-24
m
-------
TABLE OF CONTENTS (Continued)
Section Page
3.0 Methodology 3-1
3.1 Introduction and Summary of Methodology 3-1
3.2 Estimating Emissions ; 3-3
3.2.1 Emission Factors 3-6
3.2.1.1 Organic Emission Factors 3-9
3.2.1.2 Metal Emission Factors 3-12
3.2.1.3 Hydrogen Chloride (HC1) Emission Factors . 3-16
3.2.1.4 Uncertainties and Assumptions in
Emission Estimates 3-18
3.2.2 Particle Size Distribution for Deposition
Modeling 3-20
3.3 Estimating Exposure and Evaluating Effects 3-26
3.3.1 Direct Exposure to Municipal Waste Combustor
Emissions .-.,. 3-26
3.3.1.1 Human- Exposure Model 3-26
3.3.1.2 Direct Inhalation Nationwide Cancer
Risk Evaluation 3-29
3.3.1.3 Noncancer Risk 3-33
3.3.1.4 Welfare Effects 3-34
3.3.2 Indirect Exposure from Long-Term Deposition of
Municipal Waste Combustor Emissions 3-36
3.3.2.1 Industrial Source Complex Dispersion
Model ... 3-38
3.3.2.2 Terrestrial Food Chain Model 3-39
3.3.2.3 Surface Runoff Model 3-45.
3.3.2.4 Groundwater Infiltration Model 3-47
3.3.2.5 Dermal Exposure Model 3-47
3.3.2.6 Evaluation of Exposure from
Indirect Pathways 3-48
3.4 References 3-57
iv
-------
TABLE OF CONTENTS (Continued)
Section Page
APPENDICES
APPENDIX A: Human Exposure Model A-l
APPENDIX B: Industrial Source Complex Short-Term Air Dispersion
Model B-l
APPENDIX C: Wet Deposition Model C-l
APPENDIX D: Terrestrial Food Chain Model D-l
APPENDIX E: Surface Runoff Model E-l
APPENDIX F: Groundwater Infiltration Model F-l
APPENDIX G: Dermal Exposure Model G-l
-------
LIST OF TABLES
Table
2-1 Major Parameters Defining the Exposure and Risk Analysis
of Municipal Waste Combustor Emissions 2-2
2-2 Estimated Nationwide Cancer Risk from Inhalation Exposure
to Emissions Under Baseline Control Scenario 2-7
2-3 Impact of the Application of Dry Scrubber/Fabric Filter
Control Devices on the Estimated Nationwide Cancer Risk
from Direct Inhalation Exposure to Emissions 2-8
2-4 Ranges in the Contribution of Pollutants in Municipal
Waste Combustor (MWC) Emissions to the Estimated Total
Annual Cancer Incidence and Maximum Individual Lifetime
Cancer Risk 2-9
*
2-5 EPA-Recommended Toxic Equivalency Factors for Mixtures of
Chlorinated Dibenzo-p-Dioxins (CDDs) and -Dibenzofurans
(CDFs) 2-11
2-6 Comparison of Maximum Modeled Concentrations of Lead to
Ambient Standard for Existing Municipal Waste Combustors ....- 2-14
2-7 Comparison of Maximum Modeled Concentrations of Mercury to
NESHAP Guideline for Existing Municipal Waste Combustors ... 2-15
2-8 Estimated Ambient Air HC1 Concentrations Modeled in the
Vicinity of Municipal Waste Combustors 2-16
2-9 Results of Analysis of the Likelihood of Potential Human
Health Effects from Indirect Exposure to Municipal
Waste Combustor Emissions 2-18
2-10 Results of Analysis of the Likelihood of Potential Human
Health Effects from Indirect Exposure to Municipal Waste
Combustor Emissions 2-19
2-11 Results of Analysis of the Likelihood of Potential Effects
from Indirect Exposure to Municipal Waste Combustor
Emissions 2-20
2-12 Possible Reductions in Ambient Air HC1 Concentrations with
Application of Dry Alkaline Scrubbers 2-23
3-1 Summary Matrix of Emissions Test Data for Municipal Waste
Combustors Used in This Analysis 3-4
3-2 Pollutants Evaluated in the Risk Assessment 3-5
vi
-------
LIST OF TABLES (Continued)
Table Page
3-3 Control Scenarios Modeled for Existing and Projected
Municipal Waste Combustors 3-8
3-4 Organic Emission Factors for Municipal Waste Combustors 3-10
3-5 Municipal Waste Combustion Facilities from Which Test Data
Were Used to Derive Organic Pollutant Emission Factors 3-11
3-6 Metal Emission Factors for Municipal Waste Combustors 3-13
3-7 Municipal Waste Combustion Facilities from Which Test
Data Were Used to Estimate Metals Emission Factors 3-14
3-8 HC1 Emission Factors for Municipal Waste Combustors 3-17
3-9 Particle-Size Distribution Determined in Particulate
Matter Distributions at the Braintree Municipal
Waste Combustor 3-22
3-10 Typical Particle-Size Distribution Determined in
Parti cul ate Emissions at the Wurzburg Mas.sburn
Municipal Waste Combustor 3-23
3-11 Ratio of Metal Emissions As a Function of Particle Diameter .. 3-24
3-12 Pollutants Considered in Analysis of Direct Emissions
from Municipal Waste Combustors 3-27
3-13 Unit Cancer Risk Estimates for Inhalation Exposure to
Specific Chemicals ... 3-31
3-14 Parameters Used to Model Short-Term Maximum HC1
Concentrations and Exposures for Municipal Waste
Combustors 3-35
3-15 Pollutants Considered in the Indirect Exposure Analysis 3-37
3-16 Modeling Parameters for the Model Plant Representing
Projected Municipal Waste Combustors 3-40
3-17 Modeling Parameters for a Municipal Waste Combustor in
Virginia Representing Existing Facilities 3-42
3-18 Results of Analysis of the Likelihood of Potential Human
Health Effects from Indirect Exposure to Municipal
Waste Combustor Emissions 3-49
vii
-------
LIST OF TABLES (Continued)
Table Page.
3-19 Results of Analysis of the Likelihood of Potential Human
Health Effects from Indirect Exposure to Municipal Waste
Combustor Emissions 3-50
3-20 Results of Analysis of the Likelihood of Potential Effects
from Indirect Exposure to Municipal Waste Combustor
Emissions 3-55
vm
-------
LIST OF FIGURES
Figure Pace
3-1 Human Exposure Pathways Evaluated by the Terrestrial Food
Chain Model 3-44
3-2 Ecological Exposure Pathways Evaluated by the Terrestrial
Food Chain Model 3-46
ix
-------
1.0 INTRODUCTION
This volume contains an analysis of potential health risk and
environmental effects due to pollutants emitted from municipal waste
combustion. The information presented herein was developed during a
comprehensive, integrated study of municipal waste combustion. An
overview of the findings of this study may be found in the Report to
Congress on Municipal Waste Combustion (EPA/530-SW-87-021a). Other
technical volumes issued as part of the municipal waste combustion study
include:
• Emission Data Base for Municipal Waste Combustors
(EPA/530-SW-87-021b)
t Combustion Control of Organic Emissions (EPA/530-SW-87-021c)
• Flue Gas Cleaning Technology (EPA/530-SW-87-021d)
• Costs of Flue Gas Cleaning' Technologies (EPA/530-SW-87-021e)
t Sampling and Analysis of Municipal Waste Combustors
•(EPA/530-SW-87-021f)'
t Characterization of the Municipal Waste Combustion Industry
(EPA/530-SW-87-021h)
• Recycling of Solid Waste (EPA/530-SW-87-021i)
In view of the rapid growth predicted in the municipal waste combustion
industry described in "Municipal Waste Combustion: Characterization of the
Municipal Waste Combustion Industry," the U. S. Environmental Protection
Agency (EPA) has a limited opportunity to prospectively evaluate the
potential environmental and health impacts that may result from an expected
proliferation of municipal waste combustion activity in the United States.
In this regard, the agency is developing a methodology for the evaluation of
emissions of pollutants into the atmosphere from the stacks of municipal
waste combustors. The methodology consists of a series of environmental
fate and transport models that utilize the known physical and chemical
1-1
-------
properties of specific pollutants to predict the atmospheric dispersion from
stack emissions, the potential for surface deposition and accumulation; the
movement of the settled pollutants through and into various environmental
media; the potential bioaccumulation of pollutants into trophic systems; the
potential for adverse effects on the vitality of natural ecosystems; and the
potential for adverse effects on human health. The specific models used in
this analysis are: the Human Exposure Model (HEM),' the Industrial Source
Complex Short-Term Air Dispersion Model, the Terrestrial Food Chain Model,
the Surface Runoff Model, the Groundwater Infiltration Model, and the Dermal
Exposure Model.
This report on risk analysis was preceded by a methodology document
presented to the Science Advisory Board (SAB) for review in October 1986.
The document proposed a methodology for a multipollutant, multiple exposure
pathway risk assessment and described mathematical models in detail. Review
and comment by the SAB was favorable for the most part; however, the SAB
recommended continued methodology development in some areas to enhance the
indirect exposure modeling. In response to these comments, the EPA is
continuing.to develop and improve the. models used to analyze indirect
multiple exposure pathways. At the same time, the EPA is beginning to
exercise the methodology to assess municipal waste combustor emissions*
pollutant-by- pollutant.
The utility of the present methodology is limited by a number of gaps
in the available technical data and significant uncertainties in many of the
major analytical parameters, as well as by a need to perform further
research to improve the methodology. For example, the evaluation .of
pollutants is limited by the relatively small number of organic and
inorganic constituents that have been measured in municipal waste combustor
emissions. Another constraint on the methodology is the limited amount of
data regarding the physical and chemical behavior of specific pollutants in
the natural environment, and the adverse impact these pollutants may have on
human health.
The risk assessment described in this report relies on the present
methodology to evaluate the effects of municipal waste combustor emissions
from both existing facilities and from projected facilities that are
1-2
-------
anticipated to be built in the next ten to fifteen years. Direct and
indirect exposure pathways were evaluated. The direct inhalation
quantitative risk assessment considered potential human health risk based on
carcinogenic and non- carcinogenic health effects from direct inhalation of
the pollutants emitted from the stack. In addition, the direct exposure
assessment evaluated other welfare impacts resulting from direct emissions
to the atmosphere. The indirect exposure pathway considered adverse effects
on human health and the environment resulting from exposure to emitted
pollutants that ultimately deposit on soil. A significant limitation to
those analyses is that only the impact of stack emissions was considered.
Potential adverse health effects arising from the land disposal of solid
residues from municipal waste combustors, e.g., fly ash and bottom ash were
not addressed. However, the Agency is currently undertaking a regulatory
analysis of ash disposal.
For the direct inhalation exposure assessment, actual plant data were
used from existing facilities and a limited number of model plants were
developed to represent projected new facilities. In addition, different
combustor design types were taken into consideration in.the analysis. For
the indirect exposure assessment, two facilities were modeled using a
reasonable worst-case analysis. One facility is an existing facility
located in Virginia and the second is a projected model facility located in
western Florida.
Thus far, preliminary indirect exposure analyses have been performed
for only nine pollutants. Because of the preliminary nature of the exposure
results and the continuing methodology development, the EPA believes it
would be premature to produce quantitative risk results for indirect
exposure pathways at this time. However, quantitative results for thirteen
pollutants are discussed in terms of potential health .and welfare effects
due to direct emissions to the air.
The remainder of this report has been organized into two major sections
and several technical appendices. Section 2.0 contains a summary of the
results of the analyses of potential health risk and environmental effects,
and Section 3.0 describes the methodology used in performing the analyses.
Detailed descriptions of the models used in the analyses have been placed in
appendices.
1-3
-------
2.0 SUMMARY OF RESULTS
2.1 INTRODUCTION
This section summarizes the results of the preliminary risk analysis of
pollutant emissions from municipal waste combustors. The analysis involved
complex modeling of many different aspects of the industry and many of the
pollutants emitted from those combustors. To enable the reader to better
understand the results, a brief description of the modeling methodology and
the parameters defining the system to be modeled has been included in
Section 2.2 before the results are discussed. The approaches used to assess
the potential effects resulting from direct and indirect exposure to municipal
waste combustor emissions are described in detail. Section 2.3 discusses the
results of the analysis of human health and welfare effects from direct
exposure to pollutants from municipal waste combustors. Section 2.4 is a
discussion of the general observations regarding the potential health and
environmental effects from indirect exposure to deposited pollutants through
groundwater, surface water, soil, and food. Finally, a discussion of the
potential impact of emission control on reducing risk is presented in
Section 2.5.
2.2 SUMMARY OF RISK ASSESSMENT METHODOLOGY AND PARAMETERS
The major parameters defining the systems modeled to estimate potential
effects from municipal waste combustor emissions are shown in Table 2-1. It
can be seen from the number of major parameters that the modeling results
consist of thousands of numerical estimates.
Two pollutant pathways were modeled: the primary pathway, which is
direct exposure to municipal waste combustor emissions emitted to the
atmosphere; and a secondary pathway, which is indirect exposure to deposited
pollutants. Pollutant concentrations were estimated in air (direct pathway)
and in surface water, groundwater, soil, and food (indirect pathway through
deposition).
2-1
-------
TABLE 2-1. MAJOR PARAMETERS DEFINING THE EXPOSURE AND RISK ANALYSIS OF MUNICIPAL WASTE COMBUSTOR EMISSIONS
ro
i
ro
Exposure Modeling
Direct Emission
to the Atmosphere
Indirect Exposure
from Pollutant
Deposition
surface water
groundwater
- soil
food
Exposure Populations
Pathways of Combustors
Direct Existing
Inhalation
Projected
Ingestion
Dermal
Absorption
Plant Uptake
Design Control
' Types Scenarios
Massburn Baseline
Nonheat Recovery
Controlled
Massburn
Heat Recovery
Modular
Refuse -derived
Fuel (RDF)
Pollutants
Organlcs
chlorinated dibenzo-p-
dioxins/chlorinated
dibenzofurans (CDD/CDF)
chlorophenols
chlorobenzenes
formaldehyde
polychlorinated
biphenyls (PCB)
polycyclic
aromatic
hydrocarbons (PAH)
Inorganics and Metals
arsenic
beryllium
cadmium
chromium (+6)
nickel
mercury
lead
particulate matter
hydrogen chloride
-------
For the direct inhalation pathway, human health and welfare effects were
evaluated using the publicly reviewed and accepted Human Exposure Model (HEM).
For the indirect pathway, potential effects on humans and the environment were
evaluated using methodology currently being reviewed and developed: the
Industrial Source Complex Air Dispersion Model, the Terrestrial Food Chain
Model, the Surface Runoff Model, the Groundwater Infiltration Model and the
Dermal Exposure Model.
2.2.1 Modeling for Direct Exposure Pathway
Modeling for the direct exposure pathway was performed for the existing
population of combustors and for projected facilities, which are facilities
anticipated to be constructed over the next 10 to 15 years. Furthermore,
because the emissions from different.types of combustors varied significantly,
the analyses were performed for the various types of combustor design.
Modeling the estimated pollutant concentrations in the atmosphere and possible
effects from direct exposure to pollutants from the existing population of
combustors was performed by using actual sizes, types, and locations of
existing facilities. Air pollutant concentrations and possible direct
exposure .effects from projected municipal waste combustors were estimated
using model plants located in representative urban, suburban, and rural
locations.
Two control scenarios were modeled: 1) a baseline scenario, constructed
for modeling purposes to represent the status quo (existing controls for the
existing facility and electrostatic precipitators (ESPs) for the projected
facilities) and 2) a controlled scenario representing uniform application of
dry alkaline scrubbers and efficient particulate matter control (fabric
filters) to all facilities.
The following types of human health and welfare effects, were studied for
the direct exposure pathway: cancer, noncancer health effects, and welfare
effects. Noncancer health effects were evaluated by comparing modeled
concentrations of pollutants with pollutant levels associated with adverse
effects. Welfare effects were also estimated by comparison to threshold
effects levels.
2-3
-------
For the direct inhalation pathway, to calculate a probability of cancer,
the exposure estimates obtained from the Human Exposure Model were combined
with estimates of carcinogenic potency ("unit risk estimate"). The unit risk
estimate for an air pollutant is defined by the EPA as a rough but plausible
probability of the upper 95 percent confidence limit of the lifetime cancer
risk that could occur in a population in which all individuals are exposed
continuously from birth throughout their lifetimes (70 years) to a unit
concentration (e.g., 1 ug/m ) of the carcinogen in the air they breathe.
Encompassed in the notion of upper bound is that the computed risk is likely
to be higher than the true risk and that in fact, the true risk is probably
much lower than the computed risk.
By combining the estimates of public exposure with the unit risk
estimates, two measures of excess cancer risk were calculated: nationwide
annual cancer incidence and maximum individual lifetime cancer risk. The
annual cancer incidence is an estimate of cancer cases per year in the total
population residing within 50 kilometers in all directions of the municipal
waste combustor. The selection of 50 kilometers was an arbitrary modeling
consideration. Maximum individual lifetime cancer risk is the probability
that a person residing in the vicinity of the municipal waste combustor, and
exposed continuously to the highest modeled annual average ambient air
concentration of pollutant, will develop cancer over a 70-year lifetime.
2.2.2 Modeling for Indirect Exposure Pathway
The .quantitative analysis of health impacts was limited to those from
direct inhalation of municipal waste combustor emissions. The EPA has begun a
preliminary analysis of the potential for exposure from the deposition of
emitted pollutants and subsequent human contact through indirect exposure
pathways, e.g., ingest ion and dermal absorption.. Because the Agency is
continuing to develop the methodology and incorporate technical comments
received during review by EPA's Science Advisory Board (SAB), the EPA believes
it would be premature at this time to generate quantitative risk estimates
from the indirect exposure modeling. The SAB believes it is necessary to
validate the environmental fate parameters of specific chemicals used in the
analysis before reliable estimates can be made of exposure and risk from
2-4
-------
ingestion of contaminated soil and foods and dermal contact with the soil.
Furthermore, the SAB would like EPA to conduct studies to verify the computer
simulation modeling of the movement of pollutants through the environment and
into the human food chain. In view of these concerns, EPA decided only to
employ the indirect exposure methodology to study whether prolonged stack
emissions of pollutants that deposit over time onto the soil surface could
significantly contribute to the total exposure to individuals living near a
municipal waste combustor, and not to generate a risk assessment of this
exposure.
Instead, of deriving quantitative risk estimates, the methodology was
employed to test the hypothesis of whether indirect exposure pathways could
contribute significantly to the total exposure to stack emissions from
municipal waste combustors. To examine this possibility, two facilities were
used to model possible effects. One was a hypothetical model plant
arbitrarily located in western Florida devised to be representative of planned
massburn energy recovery municipal waste combustors. The second facility, an
existing facility located in Virginia, was employed to represent existing
massburn energy recovery technology. Long-term deposition and exposures over
30 years and 100 years were modeled. Thirty years was chosen to represent the
probable lifetime of a typical facility, and 100 years was selected with the
assumption that once dedicated to that purpose, the same site would be used
for successive replacement of municipal waste combustor units.
As discussed for the direct exposure pathway, two control scenarios were
also modeled for the indirect pathways. The baseline scenario represented
existing air pollution controls for the existing facility and an ESP for the
hypothetical facility. The controlled scenario represented uniform
application of dry alkaline scrubbers and efficient particulate matter control
(fabric filter) to all facilities.
For the indirect pathway, the analysis focused on estimated pollutant
concentrations and possible exposure levels. Because the methodology used to
assess indirect exposure is currently under development, quantitative risk
results for the indirect exposure pathway are not yet available and therefore
have not been presented.
2-5
-------
2.3 RESULTS OF DIRECT INHALATION EXPOSURE PATHWAY RISK ASSESSMENT
2.3.1 Human Carcinogenic Risks
Tables 2-2 and 2-3 summarize the estimated annual cancer incidence and
maximum individual lifetime cancer risk (MIR) resulting from direct inhalation
exposure to ambient air concentrations of pollutants emitted from the stack(s)
of municipal waste combustors. In the tables, the annual incidence and MIR
estimates are disaggregated by control scenario, existing and projected
combustor populations, combustor technology, and pollutant class (organics or
metals). Also shown are the annual cancer incidence and maximum individual
risk (MIR) estimates contributed by pollutant classes and by category of
municipal waste combustor design. For existing combustors, the massburn
nonheat recovery category is associated with the highest risk, and for
projected sources, the RDF units appear to pose the highest risk.
In Table 2-4, the ranges of the estimated cancer risk resulting from
direct inhalation exposure to predicted ambient concentrations of modeled
organic species are summarized. It is apparent from Table 2-4 that 31 to
98 percent of the total risk from direct inhalation exposure is attributable
to stack emissions of CDD/CDF.
2.3.1.1 Ranges and Uncertainties. The risk ranges shown in Tables 2-2,
2-3, and 2-4 reflect several areas of uncertainty in the analysis- One area
of uncertainty is the emissions data which were used to develop emission
factors. Uncertainties associated with these data are discussed in detail in
Section 2.2.1.4. For example, recovery of CDD/CDF from the Modified Method 5
stack sampling trains in some cases has been reported to be as low as
10 percent. Thus, actual emission levels could be higher than reported
emissions by an order of magnitude. However, recovery has also been reported
to be as high as 100 percent. The ranges shown incorporate recovery levels of
10 to 100 percent to account for variation in sampling method.
A basic inconsistency exists between emission factor data and toxicity
data, both of which are needed to estimate risk. Emissions data are available
for mixtures or classes of organic compounds. However, toxicity data have
2-6
-------
TABLE 2-2. ESTIMATED NATIONWIDE CANCER RISK FROM DIRECT INHALATION EXPOSURE TO EMISSONS UNDER BASELINE CONTROL SCENARIO
I'opulacion of Municipal
Waste Combustors
Existing Sources (1985)
Massburn (Non-heat)
Hassburn (Heat Rec)
RDF
Modular
TOTAL ESTIMATED RISK FROM
EXISTING SOURCES (ROUNDED)6
Projected Sources (1993)
Massburn (Heat Rec)
RDF
Modular
TOTAL ESTIMATED RISK FROM
PROJECTED SOURCES (ROUNDED)6
COMBINED TOTAL (ROUNDED)6
Orga
£
Ann. Incid.
1.3 - 30
.2 - it
.1 - 3
.0008 -.01
2 - 40
.3 - 7
.8 - 14.2
.04 - .9
1 - 20
3 - 60
a
tics
Max. Indiv.
io-4 - io-3
10-* - io-3
io-5 - io-3
io-6 - io-4
io-4 - io-3
io-6 - io-5
io-5 - io-4
io-6 - io-5
io-5 - io-4
io-4 - io-3
Met
Ann. Incid.
.2
.04
.2
.01
.5
.3
.1
.01
.4
.9
als1
Max. Indiv.
io-5
io-4
io-5
io-4
io-4
io-6
io-6
io-6
io-6
io-4
Co
Ann. Incid.
1.5 - 30
.2 - 4
.3 - 3.2
.01 - .02
2 - 40
.6 - 7.3
.9 - 14.3
.05 - .9
2 - 20
4 - 60
bined
Max. Indiv.
io-4 - io-3
io-4 - io-3
io-5 - io-3
io-4 - io-4
io-4 - io-3
ID'6 - ID'5
io"5 - io'4
io-6 - io-5
io-5 - io-4
io-4 - io-3
ro
3CDD/CDF, chlorophenols, chlorobenzenes, formaldehyde, PCB, PAH. Organic emissions are based on assumed 20% control
efficiency for both existing and projected source air pollution control equipment.
Arsenic, beryllium, cadmium, chromium +6. Metal emissions are based on estimated efficiency for installed air
pollution control equipment for existing facilities, and a 99% efficient ESP for projected facilities.
°Annual incidence is the aggregate risk of cancer cases per year in populations within 50 km of all municipal waste
combustors in the U.S.
Maximum individual risk is the probability that a person exposed to the highest modeled concentration of pollutants
from a municipal waste combustor to1 which anyone is exposed will develop cancer over a 70-year lifespan.
'Apparent errors in total are due to intentional rounding to one s i c,ni f 'ic.-int figure.
-------
TABLE 2-3. IMPACT OF THE APPLICATION OF DRY SCRUBBER/FABRIC FILTER CONTROL DEVICES ON THE
ESTIMATED NATIONWIDE CANCER RISK FROM DIRECT INHALATION EXPOSURE TO EMISSIONS
i'vi^ilation of Municipal
Waste Combustors
Existing Sources (1985)
Hassburn (Non-heat)
Massburn (Heat Rec)
RDF
Modular
TOTAL ESTIMATED RISK FROM
EXISTING SOURCES (ROUNDED)8
Projected Sources (1993)
Massburn (Heat Rec)
RDF
Modular
TOTAL ESTIMATED RISK FROM
PROJECTED SOURCES (ROUNDED)6
COMBINED TOTAL (ROUNDED)6
£
Orga lies
Ann. Incid.
.08 - 1.9
.01 - .3
.01 - .2
<. 00001
.1 - 2.4
.02-. 4
.05 - 0.9
.002-. 04
0.07-1.3
.2 - 4
Max. Indlv.d
10"5 - 10'4
10"5 - 10"4
lo-6 - io-5
1C'7 - ID'6
io-5 - 10-*
10"7 - 10'6
10'6 - 10'5
10'7 - 10'6
10'6 - 10'5
io-5 - io-4
Metals
Ann. Incid
.05
.01
.03
.001
.1
.2
.04
.001
.2
.3
Max. Indiv.
IP'6
io-6
io-6
io-6
io-6
io-6
1C'6
io-6
io-6
io-6
Co
Ann. Incid
.1 - 2
.02 - .3
.04 - .2
.001
.2 - 2.5
.2 - .6
0.09-0.9
.003-. 03
0.3-1.5
.5 - 4.0
mblned
Max. Indiv.
io"5 - io"4
io"5 - io'4
io-6 - io-5
io-6 - io-6
io-5 - io-4
io-6 - io-6
io-6 - io-5
io-7 - io-6
io-6 - io-5
io-5 - io-4
I
00
aCDDs/CDFs, chlorophenols, chlorobenzenes, formaldehyde, PCB, PAH. Organic emissions are based on 99.5 percent control
efficiency for dry scrubber/fabric filter controls.
Arsenic, beryllium, cadmium, chromium .+6. Metal emissions are based on 95% control efficiency for dry scrubber/
fabric filter controls.
Annual incidence is the aggregate risk of cancer cases per .year in population within 50 km of all municipal
waste combustors in the U.S.
Maximum individual risk is the probability that a person exposed to the highest modeled concentration of pollutants
from a municipal waste combustor to which anyone is exposed' will develop cancer over a 70-year lifespan.
Apparent errors in total are due to intentional rounding to one significant figure.
-------
TABLE 2-4. RANGES IN THE CONTRIBUTION OF POLLUTANTS IN MUNICIPAL WASTE COMBUSTOR (MWC) EMISSIONS TO
THE ESTIMATED TOTAL ANNUAL CANCER INCIDENCE AND MAXIMUM INDIVIDUAL LIFETIME CANCb.il RISK
VO
Pollutant
Chlorinated dioxlns •
and dibenzofurahs
Chlorophenols
Chlorobenzenes
Formaldehyde
Polycycllc aromatic hydrocarbons
Polychlorinated biphenyls
Arsenic
Beryllium
Cadmium
Chromium
Rounded Total6 :
Existing MWC
Annual Cancer
Incidence '
2 to 40
0.0001 to 0.0003
0.009 to 0.02
0.009
0.01 to 0.6
0.02
0.2
0.02
0.2
0.2
2 to 40
Existing MWC
Maximum
Individual .
Risk Range0 '
10"6 to 10"3
1Q~9 to 10"8
10"7 to 10"6
io-fl
10"7 to 10"5
10'8 to 10"5
-7 -4
10 to 10
10"9 to 10"6
10"6 to 10"4
10"7 to 10~4
10"6 to 10~3
Projected MUG
Annual Cancer
Incidence3'
0.8 to 20
0.0001 to 0.0003
0.004 to 0.01
0.02
0.05 to 3.0
0.2
0.1
0.001
0.2
0.1
2 to 20
Projects MWC
Maximum
Individual .
Risk Range0 '
10"6 to 10"4
lO'10 to lO'9
-9 -7
10 to 10
10~8 to 10"7
10'7 to 10"5
10"9 to 10"6
10"8 to 10"7
10'11 to 10'8
10"7 to 10"6
10'8 to 10'6
10"6 to 10"A
The ranges in annual cancer incidence reflect the assumptions made regarding the potential carcinogen-
icity of classes of organic compounds.
Annual cancer incidence is defined as the average number of excess cancer cases experienced annually in
exposed populations.
f*
The ranges in maximum individual lifetime cancer risk reflect differences in emissions and the evaluation
of emissions from municipal waste combustor technologies within the existing and projected categories.
Maximum individual risk is defined as the probability of contracting cancer following lifetime exposure
at the maximum modeled long-term ambient concentration. Probability is expressed as a negative exponent
of 10. A risk of one chancu in 10,000 is expressed as 10
Apparent errors in total are tine to intent iunal rounding io one significant figure.
-------
been developed with respect to human exposure to discrete organic compounds.
Therefore, toxicity data are not available for all the compounds within a
class. Nevertheless, scientists generally agree that structurally related
compounds may exhibit similar toxic effects through similar mechanisms of
toxicity. Therefore, some assumptions are necessary to relate the emissions
data to the risk measures and to account for the potential toxicity of various
compounds in a mixture to which the general population may be exposed. In the
case of CDD/CDF, a method involving toxic equivalency factors (TEF) was used
to convert emissions of a mixture of CDD and CDF congeners to an equivalent
quantity of the most toxic compound in the class, 2,3,7,8-TCDD. The method
for conversion, using weighting factors based on relative toxicities, has been
adopted by the.EPA as an interim procedure to evaluate mixtures of CDD/CDF
(52 FR 11749). Table 2-5 shows TEFs recommended in EPA's interim procedures.
In applying the TEF method to CDD/CDF emissions, another discrepancy
between emissions data and toxicity data had to be resolved. TEFs were
devised with the assumption that certain CDD/CDF isomers having chlorines in
the 2,3,7,8 positions on the benzene rings are the most toxic. However, most
emissions data .are in terms of homologs. Emission tests on a few facilities
have reported data on the emission levels of specific isomers. When TEF
calculations based on isomer-specific data are compared to calculations based
on homolog data at the same facility, the calculations based on homolog data
are usually higher by a factor of 3 to 7. Although it is unclear whether this
same pattern would hold for those facilities for which only homolog-specific
data are available, the ranges shown for CDD/CDF have incorporated these
factors of 3 to 7.
Additional inconsistencies between emissions data and toxicity data exist
for other organics as well. For the classes of polycyclic aromatic
hydrocarbons (PAH), chlorobenzenes, and chlorophenols, cancer potency
estimates have been established only for specific compounds within each class:
benzo(a)pyrene (a specific PAH), hexachlorobenzene, 2,2,4-trichlorophenol and
aroclor 1254 PCB. However, emission data are available only for the class of
compounds as a whole (e.g., chlorobenzenes). Therefore, ranges were formed
based on assumptions of the relative carcinogenicity of each class compared to
the carcinogenicity of the individual compounds. Using chlorobenzenes as an
2-10
-------
TABLE 2-5. EPA-RECOMMENDED TOXIC EQUIVALENCY FACTORS FOR
MIXTURES OF CHLORINATED DIBENZO-P-DIOXINS (CDDs)
AND -DIBENZOFURANS (CDFs)
Toxic Equivalency Factors
CDD
Mono thru di
Tri-CDD
2,3,7,8-TCDD
Other TCDDs
2,3,7,8-PentaCDD
Other PentaCDDs
2,3,7,8-HexaCDD
Other HexaCDDs
2,3,7,8-HeptaCDD
Other HeptaCDDs
OctaCDD
0
0
1
0.01
0.5
0.005
0.04
0.0004
0.001
0.00001
CDF
2,3,7,8-TCDF
Other TCDFs
2,3,7,8-PentaCDF
Other PentaCDFs
2,3,7,8-HexaCDF
Other HexaCDFs
2,3,7,8-HeptaCDF
Other HeptaCDFs
.OctaCDF
0.1
0.001
0,
0,
0,
0,
0,
0.
0
1
001
01
0001
001
00001
2-11
-------
example, the high risk estimate represents the assumption that all
chlorobenzenes have cancer potencies equal to that of hexachlorobenzene,
which, based on test data, Is assumed to be 43 percent of the mixture. The
low estimate 1s based on the assumption that only the single compound,
hexachlorobenzene (of the chlorobenzenes) Is associated with any risk of
cancer.
The range of risk estimates from direct inhalation of formaldehyde
reflects the uncertainty surrounding which tumors Induced in laboratory
animals by exposure to formaldehyde are indicative of formaldehyde's potential
for causing cancer in humans. The low estimate is based on the assumption
that only some of the tumors in animals are indicative of formaldehyde's
potential for causing cancer in humans, and the high estimate is based on the
assumption that all of the tumors are indicative of formaldehyde's potential
2
for causing cancer in humans.
Still another area of uncertainty has not been incorporated in the risk
ranges shown. A significant portion (80 percent or more) of the organic
emissions from the stacks of municipal waste combustors have not been
identified and quantified. Although some portion of the mixture may be
carcinogenic, the carcinogenic fraction, its composition, and its potency
remain unknown. If the unspeciated organics had a carcinogenic potency
equivalent to the average potency of those compounds evaluated, even excluding
COD/CDF, the contribution to the annual incidence estimates could be
appreciable. However, there is no information to quantify this potential
source of risk.
2.3.2 Direct Exposure Noncarcinoqenic Risks of Lead and Mercury Emissions
The potential for noncarcinogenic health risk of lead and mercury emitted
to the atmosphere and subsequently inhaled was evaluated by comparing maximum
modeled estimated annual average concentrations to standards or guidelines.
For lead, the modeled concentrations were compared to the ambient air quality
2
standard (NAAQS) of 1.5 microgram per cubic meter of air (ug/m ) (40 CFR
50.12). Modeled mercury concentrations were compared to the National Emission
Standard for Hazardous Air Pollutants (NESHAP) guideline of 1 ug/m3. Results
of the comparisons are shown in Tables 2-6 and 2-7. The concentrations
2-12
-------
presented in the table are the maximum concentrations modeled for each class
of combustor.
As shown in Table 2-6, the modeling results predicted ambient lead
concentrations resulting from municipal waste combustors emissions to be less
than the ambient lead standard. The modular category had the highest maximum
modeled concentration of 0.9 ug/m , which is 60 percent of the ambient air
quality standard. Individual municipal waste combustors are evaluated during
the permitting process to ensure conformance to the lead standard, including
consideration of background lead concentrations.
Table 2-7 displays the maximum annual average ground level concentrations
of mercury modeled near municipal waste combustors. Estimated ambient air
concentrations of mercury were based on total particulate mass emissions.
Using controlled particulate emission estimates (measured downstream of an
ESP) to predict mercury concentrations yielded maximum modeled ambient air
concentrations for mercury that do not exceed the NESHAP guideline. The
massburn design using a waterwall-type boiler for heat recovery had the
highest maximum modeled concentration of 0.4 ug/m , which is 40 percent of the
NESHAP guideline.
2,3.3 Human Welfare Effects
Among the pollutants found in stack gases from municipal waste combustors
are acid gases. The major acid specie of concern to the EPA is hydrochloric
acid (HC1) because of the magnitude of .emissions. Short-term and long-term
modeled concentrations of HC1 in the vicinity of existing and projected
sources are shown in Table 2-8.
The estimated ambient concentrations were compared to an annual average
concentration of 3.0 ug/m which is associated with materials damage in the
form of corrosion of ferrous metals. The comparisons showed that under the
baseline control scenario, maximum annual average HC1 concentrations predicted
for over 50 percent of the existing municipal waste combustors, exceeded
the materials damage level. The modeling results for the projected facilities
under the baseline scenario indicated that long-term concentrations may exceed
the 3.0 ug/m level for the larger.capacity massburn and RDF units and for all
capacities of modular units, depending on location and local meteorological"
conditions.
2-13
-------
TABLE 2-6. COMPARISON OF MAXIMUM MODELED CONCENTRATIONS OF LEAD TO AMBIENT
STANDARD FOR EXISTING MUNICIPAL WASTE COMBUSTORS
Modeled Maximum Maximum
Annual Average Modeled Cone./
Lead Concentration3 Ambient Standard
Design Type (ug/m )c (%)
Massburn
Nonheat Recovery 0.6 40
Massburn
Heat Recovery 0.8 53
Massburn
Heat Recovery-Refractory
RDF
Modular - Nonheat Recovery
Modular - Heat Recovery
0.3
0.4
0.4
0.9
20
27
27
60
aBased on controlled particulate emissions modeled for existing combustors.
K "3
National ambient air quality standard for lead' is 1.5 ug lead/m (quarterly
average.basis).
cug/m = mierograms of pollutant per cubic meter of air.
2-14
-------
TABLE 2-7. COMPARISON OF MAXIMUM MODELED CONCENTRATIONS OF MERCURY TO
NESHAP GUIDELINE FOR EXISTING MUNICIPAL WASTE COMBUSTORS
Modeled Maximum
Annual Average
Mercury Concentration1
Maximum
Modeled Concentration/
NESHAP Guideline13
Design Type
Massburn
Nonheat Recovery
Massburn
Heat Recovery-Waterwall
Massburn
Heat Recovery -Refractory
RDF
Modular - Nonheat Recovery
Modular - Heat Recovery
(ug/m3)c
0.1
0.4
0.03
0.03
0.01
0.04
w
10
40
3
3
1
4
Based on percentage of mercury in controlled particulate emission estimates
(measured downstream from an ESP) for existing combustors.
NESHAP guideline for mercury is 1 ug mercury/m (30-day average).
c 3
ug/m = micrograms of pollutant per cubic meter of air.
2-15
-------
TABLE 2-8. ESTIMATED AMBIENT AIR HC1 CONCENTRATIONS MODELED IN
THE VICINITY OF MUNICIPAL WASTE COMBUSTORS
Sources
Control
Level
Range of Predicted
Total Annual Annual Average
HCT Emissions Maximum Cone.
(Mg/yr)
(ug/m3)b
Range of Predicted
Maximum 1-Hour
Concentration
(ug/m3)b
EXISTING
SOURCES
Baseline
Controlled
44,900
4,490
1-68
.01 - 7
64 - 2,500
6 - 250
PROJECTED
SOURCES
Baseline 194,400
Controlled
19,400
.7 -
.07 - 9
110 - 170
11 - 17
aRanges of HC1 concentrations reported for all modeled municipal waste
combustion facilities.
ug/m - micrograms of pollutant per cubic meter of air.
2-16
-------
2.4 RESULTS OF INDIRECT EXPOSURE PATHWAY ANALYSIS
2.4.1 Indirect Exposure to Muncioal Waste Combustor Emissions
The indirect exposure analysis was intended to evaluate whether indirect
exposure pathways could contribute significantly to the total exposure to
municipal waste combustion emissions. The indirect exposure analysis
complements the traditional direct exposure analysis by adding the exposure
pathways of ingestion and dermal contact of deposited air emissions, as well
as estimates of possible accumulation of contaminants into the natural
environment. Such a comparison of indirect exposure potential to direct
exposure levels has been done previously in analyses of exposure to
2,3,7,8-TCDD.4
The methodology for modeling indirect exposure to pollutants emitted to
the atmosphere has received favorable review by the EPA's Science Advisory
Board. At this time, however, the Board's technical comments on the
methodology have not been fully incorporated into the various models. Also,
chemical fate parameters selected for use in the indirect exposure models were
from the published literature, but they have not been peer reviewed for this
use. Given the preliminary nature of the methodology and assumptions, the
EPA cannot interpret the results of the indirect exposure analysis
quantitatively at this time. However, results of the indirect exposure
analysis are summarized qualitatively in Tables 2-9, 2-10, and 2-11.
The objective of the analysis was to determine whether indirect exposures
could contribute significantly to the total exposure due to municipal waste
combustor emissions, therefore, the analysis used conservative, potentially
worst-case, assumptions. For example, the analysis defined the most exposed
individuals as a hypothetical farm family that obtained much of their food
supply from the area of maximum deposition of pollutants emitted from the
stack of the two modeled municipal waste combustors, and whose children
ingested 0.5 grams of soil per day. This family was assumed to live just
outside the estimated boundary (200 meters, or about one-tenth of a mile) of
the two modeled combustion facilities. The preliminary conclusions concerning
potential exposures must be interpreted in light of this worst-case, but still
2-17
-------
TABLE 2-9. RESULTS Of ANALYSIS OF THE LIKELIHOOD OF POTENTIAL HUMAN HEALTH EFFECTS FROM INDIRECT EXPOSURE TO MUNICu'AL WASTE COMBUSTOR EMISSIONS
(Exposure by Ingestlon - Inorganic Compounds)
Food IiiRestlon Soil Inaestlon
Control Deposition Low No Low
Facility Scenario Interval Possible Likelihood Data Possible Likelihood
Existing Baseline 30 yc . Pb, Cc
Hg, Nl
(VIRGINIA-120 tpd) 100 yr. Hg Pb, Nl
Cr
Controlled 30 yr. Pb, Cr
Hg, Nl
100 yr. Hg Pb, Nl
Cr
Projected Baseline 30 yr. Hg Pb, Nl
Cr
IM
jl, (WESTERN FLORIDA-3,000 tpd) 100 yr. Hg Pb, Nl
00 Cr
Controlled 30 yr. Hg Pb, Nl
Cr
100 yr. Hg Pb, Hi
Cr
fie Pb
tig
Be Pb
Be
Be Pb
Hg
Be . Pb
Hg
Be Pb
Hg
Be Hg
Be Pb
Hg
Cr, Be
Nl
Cr, Be
Nl
Pb, Cr
Hg, Nl
Cr,. Be
Nl
Cr, Be
Nl
Cr, Be
Nl
Pb, Nl
Cr, Be
Cr, Be
Nl
Surface Uat«r Inaestlon
Fish Ineestlon
No Low No Low No
Data Possible Likelihood Data Possible Likelihood Data
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Pb, Cr
Hg, Nl
Be
Pb, Cr
Hg, Nl
Be
Pb, Cr
Hg, Nl
Be
Pb, Cr
Hg, Nl
Be
Hg Pb,
Cr,
Hg PB,
Cr,
Hg Pb,
Cr,
Hg Pb,
Cr.
Hg Pb,
Cr,
Hg Pb,
Cr.
Hg Pb.
Cr,
Hg Pb,
Cr,
Hi
Be
Nl
Be
Nl
Be
Nl
Be
Nl
Be
Nl
Be
Nl
Be
Nl
Be
Be " Beryllium Cr - Chromium Hg = Mercury Hi •= Nickel Pb = Lead
-------
TABLE 2-10. RESULTS OF ANALYSIS OF THE LIKELIHOOD OF POTENTIAL HUMAN HEALTH EFFECTS FROM INDIRECT EXPOSURE TO MUNICIPAL WASTE COMBUSTOR EMISSIONS
(Human Risk by Ihgestlon - Organic Compounds)
Food
Control Deposition
Facility Scenario Interval Possible
Existing Baseline 30 yr. PCDD
CB
PCB
(VIRGINIA-120 tpd) 100 yr. PCDD
CB
PCB
Controlled 30 yr. PCDD
CB
100 yr. PCDD
CB
f
ro
1 Projected Baseline 30 yr. PCDD
to CB
PCB
(WESTERN FLORIDA-3,000 tpd) 100 yr. PCDD
CB
PCB
Controlled 30 yr. PCDD
CB
100 yr. PCDD
CB
Ingestlon
Low
Likelihood
FA
FA
PCB
FA
PCB
FA
FA
FA
PCB
FA
PCB
FA
Soil InRestlon
No Low
Data Possible Likelihood
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
t
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
Surface Water Ingest Ion
No Low No
Data Possible Likelihood Data
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
Fish InRestlon
Possible
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
PCDD
Low No
Likelihood Data
FA
FA
FA
FA
FA
FA
CB
PCB
FA
CB
PCB
FA
CB » Chlorobenzenes FA - Formaldehyde PCB » Porlychlorthated benzenes PCDD » Polychlorlnated di.benzo-p-dloxi.ns
-------
TABLE 2-11. RESULTS OP ANALYSIS OF TUB LIKELIHOOD OF POTENTIAL ECOLOGICAL EFFECTS FROM INDIRECT EXPOSURE TO MUNICIPAL HASTE COMBUSTOR EMISSIONS
(Inorganic Compound!)
Ecological Rlaks
Terrestrial Impacts Aquatic Impacts
Control Deposition Low.
Facility Scenario Interval Possible Likelihood
Existing Baseline 30 yr. Pb, Cr
Hg. HI
(VIRGINIA-120 tpd} 100 yr. Pb Cr
Hg Nl
Controlled 30 yr. Pb, Nl
Cr
ro 100 yr. Hg Pb, Nl
i
N> Cr
O
Projected Baseline 30 yr. Pb • Pb, Nl
Hg
(WESTERN FLORIDA- 3000 tpd) 100 yr. Pb Cr, 111
Hg
Controlled 30 yr. Hg Pb, HI
Cr
100 yr. Pb Cr
Nl
No Data Possible
Cr» Pb
Be Hg
Cr Pb
Be Hg
Cr Pb
Be Hg
Cr Pb
Be Hg
Cr Hg
Be
Cr Hg
Be
Cr Hg
Be
Cr Hg
Be
Low
Likelihood
Cr, Be
Nl
Cr, Be
Nl
Cr, Be
Nl
Cr, Be
Nl
Pb, Nl
Cr, Be
Pb, Nl
Cr, Be
Pb, Nl
Cr, Be
Pb, Nl
Cr, Be
No Data
Cr«
Cr
Cr
Cr
Cr
Cr
Cr
Cr
Data on chromium were available for certain terrestrial and aquatic pathways showing low likelihood of effects, but were not available for
other pathways (e.g. toxlclty thresholds for soil biota and for predators of aquatic organisms).
Be - Beryllium Cr - Chromium Hg - Mercury Nl - Nickel Pb - Lead
-------
plausible exposure scenario. The general population exposure to deposited
emissions is likely to be less, possibly significantly less, in most
situations.
Subject to the above assumptions and uncertainties, the preliminary
indirect exposure analysis indicates that indirect exposure to long-term
deposited emissions may be comparable to direct inhalation exposure for some
constituents such as CDD/COF, PCB, and chlorobenzenes.
Preliminary analysis indicates that for mercury emissions the principal
indirect exposure pathway is through food ingestion, for example, ingestion of
mercury-contaminated freshwater fish. Also, children in this hypothetical
maximum exposure scenario who are assumed to ingest 0.5 grams of soil per day
may be exposed to levels of deposited lead and mercury emissions that may
indicate a potential health effect.
The preliminary analysis also indicated that indirect exposures to some
pollutants were not of significant concern. The analysis found that indirect
exposure to nickel, chromium (+6), beryllium, and formaldehyde would not
significantly add to the inhalation exposure under the scenarios and time
frames modeled.
2.4.2 Potential Environmental Effects
The indirect exposure modeling results indicated that both lead and
mercury may reach levels in the environment of concern to the vitality of
aquatic and terrestrial plant and animal life. In Sweden, the atmospheric
deposition of mercury emissions has caused mercury levels in lakes in southern
and central Sweden to increase by a factor of five, and mercury levels in
northern lakes to have doubled. As a result, Swedish authorities have
restricted the consumption of some fresh water fish. Swedish authorities have
attributed this increase of mercury tn fresh water systems to emissions of
mercury from stationary combustion sources, including municipal waste
combustors, which essentially corroborates EPA's concerns.
2-21
-------
2.5 POTENTIAL REDUCTION OF CARCINOGENIC RISK, WELFARE EFFECTS, AND INDIRECT
EXPOSURE WITH IMPROVED EMISSION CONTROL
Predicted carcinogenic health risks due to direct emissions to the
atmosphere under two control scenarios are compared in Tables 2-2 and 2-3. As
indicated by the tables, maximum individual lifetime risk and annual cancer
incidence is predicted to be reduced.by about an order of magnitude with the
uniform application of dry alkaline scrubbers combined with fabric filters
when compared to the baseline control estimates. Table 2-12 shows the
potential reductions in HC1 concentration under the same baseline and control
scenarios. Reductions of approximately 90 percent can be seen in predicted
HC1 concentrations with the use of dry alkaline scrubbers. This reduction in
HC1 emissions would bring ambient air concentrations near most municipal waste
combustor facilities below'levels associated with the corrosion of ferrous
metals.
Indirect exposure to deposited pollutants would also be expected to be
reduced by over an order of magnitude. However, as Table 2-9 (previously
presented) shows, even under the controlled/scenario, indirect exposures to
lead and mercury cannot be dismissed for some combinations of soil,
groundwater and surface water characteristics. Because there remains
significant uncertainty concerning the indirect exposure results, exposures
due to these pathways are being investigated further.
It is feasible to reduce mercury emissions from municipal waste
combustors by 80 to 90 percent with the application of dry alkaline scrubbers
in combination with fabric filters or electrostatic precipitators if the
devices are operated at temperatures below 145°C. The controlled scenario
modeled for this analysis, however, assumed only 50 percent control of
mercury.
2-22
-------
TABLE 2-12. POSSIBLE REDUCTIONS IN AMBIENT AIR HC1 CONCENTRATIONS
WITH APPLICATION OF DRY ALKALINE SCRUBBERS
Existing Municipal Projected Municipal
Waste Combustors Waste Combustors
Baseline Estimated 0.1 - 68 0.7 - 88
Maximum Long-Term
HCL Concentration
(ug/m3)
Estimated Maximum 0.01-7 0.1-9
Long-Term HC1
Concentration
Under Controlled
Scenario
(ug/m3)
Percent Reduction 90 . 90
Achievable
(*)
2-23
-------
2.6 REFERENCES FOR SECTION 2.0
1. U. S. Environmental Protection Agency. Interim Procedures for Estimating
Risks Associated with Exposure to Mixtures of
Chlorinated-Dibenzo-p-dioxins and -Dibenzofurans (CDD and CDF). The Risk
Assessment Forum, Washington, D.C. EPA 625/3-87-012, March 1987-
2. U. S. Environmental Protection Agency. Assessment of Health Risks to
Garment Workers and Certain Home Residents from Exposure to Formaldehyde.
Washington, D.C. March 1987.
3. Letter from Henry, J. (Texas Air Control Board), to Kellam, R. (U.S.
Environmental Protection Agency). June 26, 1987. Information on ambient
levels of HC1 associated with corrosion.
4. U. S. Environmental Protection Agency. Estimating Exposures to
2,3,7,8-TCDD. Draft Report. June 1, 1987.
5. Energy from Waste. A study concerning technology, environment and
economics conducted by the National Energy Authority and the National
Environmental Protection Board for the Swedish Government. June 11,
1986.
2-24
-------
3.0 METHODOLOGY
3.1 INTRODUCTION AND SUMMARY OF METHODOLOGY
This section explains the methodology used in the risk assessment of
emissions from municipal waste combustors. In this analysis, emissions test
data from three major types of municipal waste combustors were used to develop
emission factors representative of emissions from typical municipal waste
combustors. The emission factors were used to estimate emissions from both
existing and projected facilities. Exposure assessments were then performed
using a variety of mathematical models.
Exposure to emitted pollutants is "influenced by many factors. Pollutants
emitted from municipal waste combustors into the atmosphere may be distributed
across environmental media (the atmosphere, soil and water) as a result of
complex mechanisms, most of which are just beginning to be understood. Some
pollutants are' emitted in the combustion gas adsorbed onto the surface of
particulate matter, while other pollutants.remain in a gaseous state. Some
pollutants may photodegrade, undergo complex chemical reactions, or combine
with other pollutants when transported through the atmosphere.
Particulate-bound pollutants may ultimately settle on the earth's surface by
the forces of gravity. Precipitation passing through the plume may increase
the rate of surface deposition by washout of adsorbed and gaseous pollutants.
Once deposited on the surface the pollutants may again be physically,
chemically, or photolytically transformed, may persist, or may be transported
by the action of wind and precipitation to other environmental compartments.
Deposited materials may even be resuspended into the atmosphere, and once
again become air pollutants. The net effect is that human exposure to
municipal waste combustor emissions results not only from inhalation of direct
emissions, but also indirectly from ingestion of deposited pollutants in soil,
water and food, and from skin contact with deposited pollutants in soil.
3-1
-------
Detailed experimental evaluation of the environmental fate and
environmental transport of municipal waste combustor emissions have not been
done under actual conditions. Therefore, mathematical models of the fate and
transport of pollutants entrained in the stack exhaust gas are currently the
most feasible alternative to the assessment of human exposure to municipal
waste combustor emissions. These models can also be used to estimate
bioaccumulation in the natural ecosystem, potential accumulation of pollutants
adverse to the promotion of animal and plant life, and accumulation of
pollutants into the human, food chain. The models specifically used in this
analysis include: the Human Exposure Model (HEM), the Industrial Source
Complex (ISC) Dispersion Model, the Terrestrial Food Chain Model, the Surface
Runoff Model, the Groundwater Infiltration Model, and the Dermal Exposure
Model.
Ambient concentrations of pollutants in the atmosphere surrounding
municipal waste combustors and the associated levels of human exposure were
modeled using the Human Exposure Model (HEM). The Industrial Source Complex
(ISC) Model was used to estimate deposition of pollutants on surfaces
surrounding municipal waste combustors. The potential exposures to humans
from ingestion of soil and food contaminated by deposited municipal waste
combustor emissions was estimated using the Terrestrial Food Chain Model. The
Terrestrial Food Chain Model was also used to predict adverse effects on
plants, herbivorous animals, and soil organisms from deposited pollutants.
The potential human exposure to contaminated surface or groundwaters and from
ingestion of fish from these contaminated waters was estimated using the
Surface Runoff and Groundwater Infiltration Models. These models were also
used to predict adverse effects on aquatic organisms living in contaminated
waters. Finally, the potential exposures to humans from dermal contact with
contaminated soils were estimated using the Dermal Exposure Model.
The probability of cancer from direct inhalation was estimated by
applying the carcinogenic potencies for individual pollutants emitted from
municipal waste combustors to the corresponding population exposure levels
predicted by the HEM. Next, noncancer effects resulting from exposure to
direct municipal waste combustor emissions to the air were also evaluated by
comparing ambient concentrations of pollutants to levels associated with
3-2
-------
adverse health effects. Finally, the hypothesis that indirect exposure routes
could contribute significantly to the total exposure to municipal waste
combustion emissions was tested.
In addition to human exposure, the potential for adverse ecological and
welfare effects associated with municipal waste combustor emissions was
examined. These included material corrosion due to acid gas emissions, and
levels of pollutants to which herbivorous animals, plants, soil organisms and
aquatic organisms may be exposed.
The remainder of this section describes the methodology used to
1) estimate municipal waste combustor emissions, and 2) estimate exposure
to, and evaluate the potential effects from, exposure to these emissions.
Section 3.2 presents the methodology used to estimate combustor
emissions based on .emission factors for the various pollutants. Section 3.3
describes the methodology used to estimate direct and indirect exposure to
municipal waste combustors and to evaluate the effects of this exposure on
human health and welfare and the environment. Direct inhalation exposure and
indirect exposure and the associated modeling efforts are discussed in
Sections 3.3.1 and 3.3.2, respectively. The methodology for evaluating
ecological exposures from long-term deposition of pollutants is also presented
in Section 3.3.2.
3.2 ESTIMATING EMISSIONS
While many compliance tests have been conducted for particulate
emissions, a limited number of emission tests for specific pollutants have
been conducted on municipal waste combustors in the United States, Canada,
Western Europe, and Japan. These tests provide the available data on
emissions of pollutants from the three major types of municipal waste
combustors (i.e., massburn, refuse-derived fuel (RDF), and modular).
Table 3-1 summarize* the number of facilities from which emissions data for
various pollutants were available for use in this analysis.* Table 3-2 lists
thirteen of the pollutants in municipal waste combustor emissions that were
evaluated in the risk assessment.
^Because the regulatory analysis required a timely assessment, an early
version of the Emissions Data Base was used for the risk analysis. The
earlier draft was publicly released in January, 1987.
3-3
-------
TABLE 3-1. SUMMARY MATRIX OF EMISSIONS TEST DATA FOR MUNICIPAL
WASTE COMBUSTORS USED IN THIS ANALYSIS
Pollutant
Number of Facilities Tested
Massburn
RDF
Modular
Total
Uncontrolled PM
Controlled PM
Metal s
Arsenic
•Beryllium
Cadmium
Chromium
Mercury
Nickel
Lead
Acid Gases
HCT
HF
Organ ics
Chlorinated Dibenzo-p
di oxi ns/Chl ori nated
Dibenzofurans
Polycyclic Aromatic
Hydrocarbons (PAH)
Pol ychl ori nated
Biphenyls (PCB)
Formaldehyde
Chlorobenzenes
Chlorophenols
10
13
6
5
9
9
5
7
10
10
5
8
3
1
2
1
1
2
2
2
2
5
1
3
1
3
1
2
2
2
2
2
2
2
16
15
11
7
12
13
9
11
14
17
8
19
2
2
2
Z
3
2
2
2
2
1
1
6
7
5
5
Source: Reference 1
3-4
-------
TABLE 3-2. POLLUTANTS EVALUATED IN THE RISK ASSESSMENT
Arsenic
Beryl1i urn
Cadmium
ChTorobenzenes
Chlorophenols
Chromium
Chlorinated Dibenzo-p-dioxins and Chlorinated Dibenzofurans (CDD/CDF)a
Formaldehyde
Hydrogen chloride
Lead
Mercury
Polychlorinated biphenyls (PCB)
Polycycl.ic aromatic hydrocarbons (PAH)
The terms, dioxins and dibenzofurans refer to a group of 75 chloro-
dibenzo-p-dioxin compounds and 135 chlorodibenzofuran compounds,
each having similar chemical and physical properties.
3-5
-------
For the most part, emissions of particulate matter (PM) from municipal
waste combustors have been well characterized. Measurements of PM emissions
have been made at a number of facilities representing a wide range of
combustion technologies. Metals and acid gas emissions from conventional
massburn combustors are also relatively well characterized; however, less data
are available for metals and acid gas emissions from other combustor types.
The most significant data gap in the inventory of municipal waste combustor
emissions is the speciation of halogenated and nonhalogenated organic
compounds. For some of the organic pollutants in particular, data are limited
to only one or two facilities of each type.
A complete description of the tested facilities, testing procedures, and
emissions data used in this report is found in a publicly released draft of
"Municipal Waste Combustion Study: Emissions Data Base for Municipal Waste
Combustors."
3.2.1 Emission Factors
Based on data from emission tests, emission factors were developed to
represent emissions typical of municipal waste combustors. In general,.
emission factors were derived by averaging available emissions data for a
given pollutant and facility type. Emission factors were developed for each
of the pollutants analyzed in this risk assessment including the following
organic pollutants (or classes of organic pollutants): chlorinated
dibenzo-p-dioxins/dibenzofurans (CDDs/CDFs), chlorophenols, chlorobenzenes,
formaldehyde, polychlorinated biphenyls (PCB), and benzo-a-pyrene [B(a)P].
Emission factors also were developed for the following inorganic pollutants:
arsenic, beryllium, mercury, lead, cadmium, hexavalent chromium, and
hydrogen chloride. Different emission factors were developed for each
pollutant to represent emissions from existing and projected municipal waste
combustor populations.
In deriving emission factors a dichotomy was made between existing and
projected municipal waste combustors. This dichotomy was based on the premise
that municipal waste combustors currently marketed represent distinct
improvements in design, combustion efficiency and pollution control when
3-6
-------
compared to municipal waste combustors planned and built just a decade ago.
Subjective judgments were made regarding the selection of certain facilities
to represent potential pollutant emissions from existing and projected
municipal waste combustors.
The difference in emission factors for existing and projected municipal
waste combustors reflect the discriminating use of different emissions test
data. In the case of existing facilities, emission factors were derived using
emissions data from most tested facilities with the exception of specific
state-of-the-art technologies. By contrast, only data from facilities judged
to have combustion technologies representative of newer, more advanced
technologies were used to derive emission factors for projected facilities.
Projected facilities were also assumed to use good combustion controls for
minimizing the formation and emission of organic pollutants.
The following assumptions were made regarding baseline control of
pollutant emissions. For the existing combustor population, the baseline
scenario assumed uniform use of electrostatic precipitators (ESPs) because
available emissions data were collected from ESP-controlled units. A variety
of PM control devices are actually in place on existing units. Baseline
conditions assumed 20 percent control of organic pollutants and current levels
2
of particulate matter control. Control levels for metal emissions (except
mercury) were assumed to be proportional to control levels for PM emi-ssions.
For the projected population, baseline conditions assumed the uniform
applications of ESPs, although some States are requiring more stringent
control. Organic pollutant and PM emissions (with the exception of mercury)
were assumed to be controlled by 20 percent and 99 percent, respectively, and
metal emissions were assumed to be controlled in proportion to particulate
emissions. These control efficiencies reflect optimized control of these
pollutants using these control technologies. No control of inorganic acid
gases such as sulfur dioxide or hydrogen chloride was assumed for baseline
conditions of either existing or projected facilities. The assumptions used
to define the control scenarios considered in this analysis are summarized in
Table 3-3.
The method used to derive the emission factors for the various pollutant
classes are presented in the following subsections.
3-7
-------
TABLE 3-3. CONTROL SCENARIOS MODELED FOR EXISTING AND
PROJECTED MUNICIPAL WASTE COMBUSTORS
Control Efficiency 1%)
Population/
Control Scenario
Control
Technology
Organ ics
Metals
(except Hg) Hg
HC1
Existing
Baseline
Control1ed
ESP
Dry alkaline
scrubber/
fabric filter
20
95
99.5
30
50
0
90
Projected
Baseline
Controlled
ESP
Dry alkaline
scrubber/
fabric filter
20
95
99
99.5
30 0
50 90
For existing municipal waste combustors, baseline control levels for metal
emissions (except Hg) were assumed to be proportional to actual control
levels achieved by existing particulate matter controls, as measured during
emission tes.ts.
3-8
-------
3.2.1.1 Organic Emission Factors. Organic pollutant emission factors,
expressed in grams of pollutant per megagram of waste combusted, are presented
in Table 3-4 for the different municipal waste combustion technologies
modeled. The organic emission factors were developed by calculating for each
tested facility the measured organic pollutant emission rate and then dividing
by the waste combustion rate. For facilities currently equipped with control
devices other than ESPs (i.e., dry alkaline scrubbers in combination with
fabric filters), the measured organic pollutant emission rate was first
adjusted to reflect control levels achieved by ESPs. Emission factors for
each municipal waste combustion technology modeled were determined by
calculating the average emission rate (in g/Mg of waste combusted) for all
facilities, of a given type.
The facilities providing test data used in the calculation of individual
organic emission factors are listed in Table 3-5. As the table shrows, for
some organic pollutants there were very few test data. Also observed from
Table 3-5, data from different facilities were used in determining emission
factors for existing vs. projected municipal waste combustors. Efforts were
made to use the most appropriate data to represent emissions from different
types of combustors and for different scenarios. Some of the existing
facilities, for example Hampton (massburn), may not represent performance
levels expected of new facilities. Test data from this facility were,
therefore, excluded when computing average emission factors for CDD and CDF
for the projected population of massburn heat recovery facilities. However,
for some other organic pollutants, for example, B(a)P, there were not
state-of-the-art data available, so the data from Hampton were used, even
though they may not be considered representative of new facilities.
Most of the tested facilities listed in Table 3-5 are equipped with
particulate controls only. Although little data are available on the
performance of particulate controls in controlling organics, preliminary
results from recent tests on ESPs indicate a range of 0 to 50 percent control
of CDDs and CDFs may be occurring, and data from earlier tests at the
Chicago NW facility indicated 0-45 percent control of other organic compounds
by an ESP. Based on these data, 20 percent control of organics was assumed
to be achieved by ESPs alone without appreciable prior cooling of the flue
3-9
-------
TABLE 3-4. ORGANIC EMISSION FACTORS FOR MUNICIPAL WASTE COMBUSTORS
(g/Mg of waste combusted)3
Organic
CDD/CDFc'd
Benzo-a-pyrene
[B(a)P]
Polychlorinated
Biphenyls
(PCB)
Formaldehyde
Chlorobenzenes
Chlorophenols
Emission
Facility Type Existing
Massburn-Heat Recovery
Massburn-Nonheat Recovery
RDF
Modul ar
Massburn-Heat Recovery
Massburn-Nonheat Recovery
RDF
Modul ar
Massburn-Heat Recovery
Massburn-Nonheat Recovery
RDF
Modular
Massburn-Heat Recovery
Massburn-Nonheat Recovery
RDF
Modular
Massburn-Heat Recovery
Massburn-Nonheat Recovery
RDF
Modular
Massburn-Heat Recovery
Massburn-Nonheat Recovery
RDF
Modular
0.00096
0.0031
0.002
0.000069
0.048
0.048
0.13
0.0038
0.0021
0.0021
T,4
0.0045
5.2
5.2
1.1
1.3
0.32
0.32
0.14
0.018
0.46
0.46
0.18
0.018
Factor .
Projected"
0.000058
N/Ae
0.0012
0.000069
0.048
N/Ae
0.13
0.0038
0.00021
N/AS
1.4
0.0045
2-5*
N/Ae
1.1
-1.3
0.0089
N/Ae
0.14
0.018
0.018
N/Ae
0.18
0.018
Emission factors were derived from an average of available emissions data
for a given pollutant and facility type.
Data used to calculate emission factors for projected facilities exclude
test data for facilities not equipped with state-of-the-art combustion
technology.
CCDD/CDF refers to the class of 75 chlorinated dibenzo-p-dioxins (CDD) and
135 chlorinated dibenzofurans (CDF) compounds.
CDD/CDF emissions are reported as the 2,3,7,8-TCDD equivalent concentration
based on homologue-specific data.
eN/A - Not Applicable. All projected municipal waste combustors are assumed
to be heat recovery facilities.
3-10
-------
TABLE 3-5. MUNICIPAL WASTE COMBUSTION FACILITIES FROM WHICH TEST DATA
WERE USED TO. DERIVE ORGANIC POLLUTANT EMISSION FACTORS
<*>
i
1 MWC Tested
EXISTING WC
1 Massburn
Non Heat Recovery
6 Massburn
Heat Recovery
6 Modular
5 RDF
PROJECTED MrfC
7 Massburn
5 RDF
6 Modular
CDD/COF B(a)P
Philadelphia NW Hampton
Quebec Hampton
Saugus
Chicago
Peeksklll
Hampton
N. Andover
PEI, Dyersburg Cattaraugus
N. Little Rock
May port. One Ida
Cattaraugus
SWARU Albany
Akron
Wright Patterson
Al bany
Niagara Falls
Wurzburg Hampton
N. Andover
Saugus
Peeksklll
Tulsa
Marlon Co.
Chicago NW
SWARU* Albany
Akron
Wright Patterson
Albany
Niagara Falls
PEI, Dyersburg Cattaraugus
N. Little Rock
Mayport, One Ida
Cattaraugus
PCB
Hampton
Chicago NW
Hampton
PEI
Cattaraugus
Albany
SWARU
Chicago NW
Albany
SWARU4
PEI
Cattaraugus
Formaldehyde
Hampton
Peeksklll
Hampton
Peeksklll
Dyersburg
Cattaraugus
Albany
Akron
Nl'agara Falls
Peeksklll
Albany
Akron
Niagara Falls
Dyersburg
Cattaraugus
Chlorobenzenes
Chicago NW
Hampton
Chicago NW
Hampton
PEI
Wright Patterson
SWARU
Chicago NW
Wright Patterson
SWARU3
PEI
Chlorophenols
Chicago NW
Hampton
Chicago NW
Hampton
PEI
Wright Patterson
SWARU
Chicago NW
Wright Patterson
SWARU
PEI
Adjusted to reflect judgment about recent modifteat Ions.
-------
gas. In calculating the baseline emission factors, representing the use of
ESPs only (Table 3-4), organic emissions measurements for the small number of
facilities achieving greater control of organics with dry alkaline scrubbers
in combination with particulate matter control devices were adjusted to
reflect 20 percent control achievable by ESPs.
The organic emission factors in Table 3-4 were applied directly to
estimate organic emissions from municipal waste combustors under the baseline
control scenario. Under the controlled scenario, 95 percent control of
organic emissions was assumed to be achievable through uniform application of
dry alkaline scrubbers combined with fabric filters. This level of control is
supported by recent emission tests conducted by both Environment Canada and
the EPA. Full-scale and pilot-scale systems equipped with dry scrubbers in
combination with a baghouse system (fabric filters) generally achieved better
than 95 percent control of organic emissions when operated at temperatures
below 200°C. To.estimate organic emissions after application of dry alkaline
scrubbers in combination with fabric filters, the emission factors in
Table 3-4 were adjusted as follows:
EFC - EFB
(1 - 0.95)
(1 - 0.20)
where:
EFp = emission factor for controlled scenario
EFg = emission factor for baseline control scenario
0.95 = organic control efficiency for alkaline scrubber
0.20 = organic control efficiency for ESP
3.2.1.2 Metal Emission Factors. Table 3-6 displays metal emission
factors estimated for the various municipal waste combustion technologies.
Metal emission factors are expressed in units of micrograms of pollutant per
gram of controlled particulate emissions. The specific municipal waste
combustion facilities from which test data were used to calculate the
individual metal emission factors are listed in Table 3-7. The test data
shown were used to calculate emission factors representing both the existing
and projected municipal waste combustor populations.
3-12
-------
TABLE 3-6. METAL EMISSION FACTORS FOR MUNICIPAL WASTE COMBUSTORS
(ug metal/g controlled particulate emissions)
Metal
Facility
Type
Emission
Factor
Arsenic
Beryl 1iurn
Cadmium
Chromium (+6)
Lead
Mercury
Nickel
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wai 1
MB Refractory
RDF
Modular
MB Water- Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
335
11.9
90
353
0.091
11.9
65
0.036
686
1,228
106
3,310
9,476
415
10,912
818
13,425
61,105
10,530
72,500
8,163
6,770
868
2,340
6,855
2,812
5,770
1,610
3-13
-------
TABLE 3-7. MUNICIPAL WASTE COMBUSTION FACILITIES FROM WHICH TEST DATA
WERE USED TO ESTIMATE METALS EMISSION FACTORS
Pollutant
Metals
Facility Type
MWC Facilities Tested
u*
i
Arsenic
Beryllium
Cadmium
Chromium
Lead
Mercury
Nickel
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wai 1
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
MB Water Wall
MB Refractory
RDF
Modular
Balt1more> Bralntree, Hampton. Munich, Wurzburg
Tsushima
Akron* Albany
Tuscaloosa, Dyersburg, Prince Edward Island
Bralntree, Hampton, Tulsa, Munich
Tsushima
Albany
Dyersburg
Bralntreei Hampton, Munich, Maimo, Wurzburg
Washington, Alexandria, Nicosia, Tsushima
Albany
Dyersburg, Prince Edward Island
Baltimore, Bralntree, Hampton, Munich, Wurzburg
Washington, Alexandria, Nicosia, Tsushima
Akron, Albany
Dyersburg, Prince Edward Island
Bralntree, Hampton, Tulsa, Munich, Maimo, Wurzburg
Washington, Alexandria, Nicosia, Tsushima
Akron, Albany
Dyersburg, Prince Edward Island
Bralntree, Hampton, Tulsa, Malm6
Tsushima
Akron, Albany
Dyersburg, Prince Edward Island
Hampton, Munich, Wurzburg
Washington, Alexandria, Nicosia, Tsushima
Akron, Albany
Dyersburg, Prince Edward Island
-------
Metal emission factors were derived by first calculating for each tested
facility the ratio of metals to participate emissions in controlled
participate emission measurements. The average metal to particulate ratio was
calculated from the available test data for each combustor technology to
determine baseline emission factors for these metals. The metal emission
factors in Table 3-6 reflect the current control levels achieved by the
particulate control devices in place at each of the tested facilities. In
general, existing particulate controls may be achieving about 99 percent
control of particulate and metals emissions, except mercury emissions which
are assumed in this analysis to be controlled only about 30 percent by
electrostatic precipitators at temperatures above 200°C.
To estimate metal emissions from municipal waste combustors under the
baseline scenario, the emission factors in Table 3-6 were directly applied.
To generate estimates of'emissions under the controlled scenario, fabric
filters were assumed to be capable of achieving 99.5 percent control of metals
emissions except for mercury emissions. Mercury emissions were, assumed to be
only controlled 50 percent. However, tests conducted by Environment Canada on
a pilot-scale dry alkaline scrubber combined with fabric filters demonstrated
that mercury could be controlled by as much as 90 percent if the devices were
operated at temperatures at or-below 145°C. To estimate metal emissions from
municipal waste combustors under the controlled scenario, the emission factors
in Table 3-6 for all metal pollutants except mercury were adjusted as follows:
For existing facilities: 1 - 0.995
EFC=EFB
-CEB
For projected facilities: 1 - 0.995
EFC = EFB
C B 1 - 0.99
where:
EFC » emission factor for controlled scenario
EFg = emission factor for baseline scenario
0.995 - metal control efficiency achieved with particulate controls
under the controlled scenario
3-15
-------
0.99 - metal control efficiency achieved with particulate controls
for projected facilities under baseline scenario
CEg = metal control efficiency achieved with existing particulate
controls for existing facilities under baseline scenario
The mercury emission factors, based on estimated controlled particulate
emissions, were adjusted as follows:
1 - 0.50
EF- = EFB
L ° 1 - 0.30
where:
EFC - emission factor for controlled scenario
EFg - emission factor for baseline scenario
0.50 - mercury control efficiency achieved
with particulate controls under the
controlled scenario
0,30 = mercury control efficiency achieved with
existing particulate controls
3.2.1.3 Hydrogen Chloride (HC1) Emission Factors. Baseline HC1 emission
factors are presented in Table 3-8 for the types of municipal waste combustors
modeled. Emission factors are expressed in units of grams of HC1 per megagram
of waste combusted. To derive this, HC1 emission rates were divided by the
waste combustion throughput for each tested facility. Two sets of emission
factors were estimated from these data: average, and maximum to be used to
model long-term and .short-term averages. Average emission factors were
determined by calculating the average of all HC1 measurements for a given
combustor type. Maximum emission factors represent the highest reported value
of HC1 emissions per quantity of waste combusted for each combustor type. It
should be noted that because the test data are themselves averages of several
runs, generally of 3 to 4 hours duration, this may be a conservative maximum,
and short-term average.
3-16
-------
TABLE 3-8. HC1 EMISSION FACTORS FOR MUNICIPAL WASTE COMBUSTORS
(kg/Mg of waste combusted)3
Facility Type
Massburn-Heat Recovery
Massburn-Nonheat Recovery
RDF
Modul ar
HC1
Average
4.13
4.13
2.57
3.58
Emission Factor
Maximum
7.15
7.15
N/Ab
4.43
akg/Mg = kilogram of HC1 emitted per megagram (metric ton) of waste
combusted.
Not available.
3-17
-------
The baseline emission factors in Table 3-8 reflect zero percent control
of HC1 emissions without the use of dry alkaline scrubbers. These emission
factors were used to estimate annual average and maximum HC1 emissions from
existing and projected municipal waste combustors under the baseline control
scenario.. Ninety percent control of HC1 emissions by dry alkaline scrubbers
is assumed under the controlled scenario (i.e., reflecting use of dry alkaline
scrubbers in combination with fabric filters). HC1 emissions under the
controlled scenario were estimated by adjusting the emission factors in
Table 3-8 as follows:
EF~ = EFB (1 - 0.9)
where:
EF- - emission factor for controlled scenario
EF« - emission factor for baseline scenario
0.9 - HC1 control efficiency achievable under
controlled scenario
3.2.1.4 Uncertainties and Assumptions in Emission Estimates. Emissions
estimates used in the risk assessment are based on emissions test data
collected from a limited number of tests on municipal waste combustors. The
tested facilities vary widely in design and operational performance. In
addition, the tests were conducted with different objectives with the result
that testing protocols and the level of detail of reported data was variable.
Finally, specific sampling and analysis methods were not described in detail
and may not be consistent for all tests. Direct comparison of the test
results for different facilities is therefore difficult. As described above,
average emission factors were developed from the test data to represent
typical emissions from existing and projected municipal waste combustors..
The following uncertainties are noted. First, the emissions data used to
calculate emission factors represent results of tests on emissions from
municipal waste combustor stacks only. Fugitive emissions were not measured
and are not included in the emission factors.
Secondly, the emissions data came from regulatory compliance tests
conducted over a relatively short period of time, and may not be
representative of long-term emissions over the 30-year life of a typical
3-18
-------
municipal waste combustor. Periods of downtime and diminished performance of
pollution control devices over time would contribute to variations in actual
emissions versus tested emissions for a given facility.
Third, emissions data were taken from only a small number of the existing
municipal waste combustor facilities and, therefore, may not be statistically
representative of the industry as a whole in all cases. For example, CDD/CDF
emission estimates from massburn facilities without heat recovery were based
on tests from only one facility, the Philadelphia NW facility. Emissions for
this facility may or may not be representative of those from other nonheat
recovery massburn facilities. However, facilities without heat recovery are
generally older than facilities with heat recovery, so there may be a real
basis for the difference in estimated CDD/CDF emissions.
Similarly, the emission estimates for existing massburn energy recovery
facilities are influenced by emission levels found at the Hampton facility,
which are much higher than those found at other tested massburn facilities
with heat recovery. Without further tests, one cannot determine the extent to
which other facilities may or may not have emissions comparable to Hampton.
The emissions from Hampton were excluded in calculating the average emission
factor for projected facilities with the assumption that the distinct
operational improvements in design and operation at new facilities may make
'Hampton emissions non-representative of projected facilities.
In another case, emission estimates for projected RDF facilities are
influenced by the inclusion of emission data from the SWARU RDF facility
(Hamilton, Ontario). Emissions of CDDs/CDFs are significantly higher at SWARU
than those of other tested RDF facilities, and SWARU had known operational
problems at the time the tests were performed (which recent modifications have
-g
attempted to correct). There are insufficient data for RDF facilities,
however, to statistically determine whether these known problems are unique to
SWARU, and, therefore would make SWARU emission data unsuitable for use in
estimating emissions from new RDF facilities. To account for recent design
and operational improvements at SWARU, the measured emissions of CDDs/CDFs
were reduced 50 percent when included in the emission factor derivation for
projected RDF facilities.
3-19
-------
Some uncertainties are associated with emissions estimates for CDDs/COFs.
For example, recovery of CDD/CDF from the Modified Method 5 stack sampling
traios in some cases has been reported to be as low as 10 percent. Thus,
actual emission levels could be higher than reported by an order of magnitude.
However, recovery has also been reported to be significantly higher. The
ranges of cancer risk results presented in Section 2.3.1 account for this
uncertainty.
The ranges do not account, however, for the fact that a significant
portion (80%) of the organic pollutants emitted from the stacks of municipal
waste combustors have not been fully identified and quantified. Although some
portion of the mixture may be carcinogenic, the carcinogenic potential is
unknown.
In the case of metals, in many tests only the particle phase was analyzed
for metals, while in a few tests metals were measured in both the particle and
vapor phase. Measurements of metals in both phases may result in higher
emission estimates. No attempt was made to account for this uncertainty in
constructing ranges of cancer risk results.
3.2.2 Particle Size Distribution for Deposition Modeling
A limited amount of data are available on particle size
fractionalization of particulate matter entrained in municipal waste
combustor emissions. A series of stack tests conducted by EPA at the
municipal waste combustor in Braintree, Massachusetts provided the data on a
likely distribution of particulate matter by particle size used in the
q
indirect exposure analysis. The Braintree municipal waste combustor was
tested at relatively high particulate emission rates (6.7 kg/hr), in
exceedance of EPA requirements, and the particulate matter control efficiency
of the electrostatic precipitator was calculated to be only 74 percent. Total
particulate concentrations averaged 215 mg/dscm at 12 percent CCL. The
particle size distribution measured at this facility may or may not be
representative of state-of-the-art equipment, but it should provide reasonably
conservative, tending toward the worst-case, data for estimating the surface
deposition of particles in municipal waste combustor emissions.
3-20
-------
Table 3-9 displays the average particle-size distribution measured at the
Braintree municipal waste facility. Recently, tests have been conducted at
the outlet of the fabric filters operating on a massburn, heat recovery
municipal waste combustor in West Germany.10 Table 3-10 summarizes the size
distribution of particulate matter at the Wurzburg, West Germany, massburn
heat recovery municipal waste combustor. In comparison to the Braintree
municipal waste combustor data, the emission rates of particulate matter at
the outlet of the fabric filters averaged 9.0 mg/dscm. (at 12 percent C02)
over 3 test runs. Approximately one-half of the particulate matter emissions
at Wurzburg were less than 0.5 microns in diameter although, over one-third
of the emissions were of a diameter greater than 12 microns.
For purposes of estimating deposition of individual pollutants,
estimations of the relationship between the pollutant concentration and
particle size were assumed. For inorganic pollutants, data from a limited
number of municipal waste combustors have been evaluated to determine the
affinity of individual metals for adsorption onto surfaces of various
particle diameters. In general, the maximum mass of metals emissions was
found to be associated with fine particles,.e.g., two microns or less in
aerodynamic diameter. Table 3-11 is a summary of the particle phase
distribution of metallic elements adsorbed on particulate matter.
Data on the distribution of organic compounds adsorbed onto variously
sized particulates was not available. The lack of data on the relationships
between particle size and organic compound concentration was overcome by
assuming that compounds are distributed in proportion to the percent of total
particle surface area available for adsorption for each particle size category
(assuming particles are perfect spheres). The mass emission rate of the
organic pollutant was distributed by particle size by computing proportion and
12
available surface area for a,given particle size, and holding particle
density constant. The particle weight is proportional to volume if density is
constant. Therefore, the ratio of the surface area to volume is proportional
to the ratio of surface area to weight for a particle with a given radius.
Multiplying this proportion times the weight fraction of particles of a
specific diameter yielded an estimate of the amount of surface area available
for chemical adsorption.
3-21
-------
TABLE 3-9. PARTICLE-SIZE DISTRIBUTION DETERMINED IN PARTICULATE MATTER
DISTRIBUTIONS AT THE BRAINTREE MUNICIPAL WASTE COMBUSTORa'b'c
Geometric Mean Particle Diameter Percent of
(Microns) Total Particle Mass
>15.0
12.5
8.1
5.5
3.6
2.0
1.1
0.7
<0.7 •
12.8
10.5
10.4
7.3
10.3
10.5
8.2
7.6
22.4
aData is an average of 6 stack test runs.
The facility is equipped with an electrostatic precipitator rated at an
average particulate removal efficiency of 74 percent at time of test.
cThe facility recovers steam from the combustion of about 120 metric tons
of municipal waste per day.
Source: Reference 9
3-22
-------
TABLE 3-10. TYPICAL PARTICLE-SIZE DISTRIBUTION DETERMINED
IN PARTICULATE EMISSIONS AT THE WURZBURG
MASSBURN MUNICIPAL WASTE COMBUSTORa'b'C
Particle Diameter Percent of
(Microns) Particle Mass
>12.0 37.8
7..5 - 12 0.0
5.1 - 7.5 1.0
3.5 - 5.1 1.5
2.3 - 3.5 3.6
1.1 - 2.3 2.0
0.7 - 1.1 0.5
0.47 - 0.7 1.0
<0.47 52.6
aData represents one test day, but is typical of performance.
The facility is equipped with a dry venturi scrubber coupled with fabric
filters for air pollution control.
The facility has a capacity of 600 metric tons municipal waste per day.
Source: Reference 10
3-23
-------
TABLE 3-11. RATIO OF METAL EMISSIONS AS A FUNCTION OF
PARTICLE DIAMETER (>2u/<2u)
Pollutant Ratio
Arsenic 25/75
Beryllium 34/66
Cadmium 34/66
Chromium 41/59
Lead 12/88
Nickel 41/59
Source: Reference 11
3-24
-------
Three particle size categories were selected to characterize the data:
greater than 10 microns, 2 to 10 microns, and less than 2 microns. The
fraction of total surface area available for chemical adsorption was
calculated to be 0.03, 0.095, and 0.875, respectively. The specific emission
rate for each organic pollutant considered in the analysis was multiplied by
the fraction of available surface area to estimate the pollutant emission rate
corresponding to the particle size distribution. For example, if the emission
rate of formaldehyde was estimated to be 80 milligrams per second of plant
operation, then the formaldehyde emission rate in consideration of available
surface area would be: (0.030 x 80 mg/s) =2.4 mg/s for particles greater
than 10 microns; (0.095 x 80 mg/s) = 7.6 mg/s for particle sizes ranging from
2 to 10 microns; (0.875 x 80 mg/s) = 70 mg/s for particles less than or equal
to 2 microns in diameter. The organic pollutants emissions were determined on
this basis for purposes of wet and dry surface deposition.
To calculate emission rates of inorganic compounds corresponding to the
particle size distribution, ratios of mass of the compound per particle size
category were assumed. For example, Table 3-1L indicates that approximately
25 percent of arsenic has been found adsorbed to particles greater than
2 microns in diameter, and thus the ratio is 25 percent/75 percent for
>2/<2 microns. If arsenic were emitted at a rate of 2.3 x 10" grams per
second, then the emission rate per particle size cutoff, taking into
-4
consideration available surface area for adsorption, would be 1.4 x 10" g/s
for particles greater than 10 microns; 4.4 x- 10 g/s for particles 2 to
10 microns, and 1.7 x 10 g/s for particles less than 2 microns.
The particle size and surface area distribution were kept constant in the
analysis of inorganic and organic compound emissions resulting from the
application of two distinct sets of air pollution control systems: ESPs and
the combination dry venturi scrubber and fabric filters. Comparisons between
the particle size distributions depicted in Tables 3-7 and 3-8 do show
distinct differences between emissions resulting from ESP control or fabric
filter control. However, it was felt that not enough particle mass
frationalization data exists to generalize a distribution pattern for each
control device, based on these limited measurements. Therefore, EPA has
assumed, for purposes of estimating deposition of specific pollutants, a
unitized particle size distribution of both pollution control systems.
3-25
-------
3.3 ESTIMATING EXPOSURE AND EVALUATING EFFECTS
3.3.1 Direct Exposure to Municipal Waste Combustor Emissions
Direct inhalation exposure to emissions from municipal waste combustors
was evaluated using the Human Exposure Model (HEM). Based on the exposure
levels predicted by the HEM, an estimate of cancer risk, noncancer risk, and
the potential for environmental effects resulting from direct emissions from
municipal waste combustors was made. The pollutants considered in each part
of this analysis are listed in Table 3-12. The key elements of the analysis
of exposure and effects associated with direct emissions from municipal waste
combustors are described below.
3.3.1.1 Human Exposure Model. The HEM is a general model designed to
estimate the population exposed to air pollutants emitted from stationary
sources. The HEM is comprised of an atmospheric dispersion model with
included meteorological data and a population distribution based on Bureau of
Census data. In addition,, the HEM contains a procedure for estimating cancer
risks due to the predicted exposure. Based on source data, the model
estimates the magnitude and distribution of ambient air concentrations of the
pollutant in the vicinity of the source. These concentration estimates are
coupled with the population data to estimate public exposure to the pollutant.
A more detailed description of the HEM is included in Appendix A,
Exposure is the product of the population and the concentration to which
the population is exposed. To form this product, both the concentration and
the population must be known at the same location or.point. The HEM uses the
latitude, and longitude of the facility to determine the population of the
study area. The population data base is comprised of the 1980 Census Data
Base broken down by block group/enumeration district (BG/ED). The population
data base contains the population centroid coordinates (latitude and
longitude) and the 1980 population of each BG/ED in the country (about 300,000
centroids in 50 states plus the District of Columbia). A population centroid
is the population-weighted geographical center of a BG/ED for known geodetic
coordinates. For BG/ED centroids located between 0.2 km and 3.5 km from the
facility, populations are apportioned among neighboring polar grid points. A
3-26
-------
TABLE 3-12. POLLUTANTS CONSIDERED IN ANALYSIS OF DIRECT EMISSIONS
FROM MUNICIPAL WASTE COMBUSTORS
Analysis Pollutants
Cancer Risk Organic
CDD/CDFa
chlorophenols
chlorobenzenes
formaldehyde
polychlorinated biphenyls (PCB) .
polycyclic aromatic hydrocarbons (PAH)
Inorganic
arsenic
beryl1i urn
cadmium
chromium +6
Noncancer Risk Lead
Mercury
HC1
Environmental Effects HC1
CDD/CDF refers to 75 chlorinated dibenzp-p-dioxin compounds and 135
chlorinated dibenzofuran compounds.
PAH compounds are represented by benzo(a)pyrene in this analysis.
3-27
-------
polar grid point is one of the 160 receptors at which concentrations are
estimated by the dispersion model. There are 64 (4 x 16) polar grid points
within this range. Both concentration and population counts are thus
available for each polar grid point. Log-log linear interpolation is used to
estimate the concentration of each ED/BG population centroid located between
3.5 km and 50 km from the source. Concentration estimates for 96 (6 x 16)
grid points (receptors at 5, 10, 20, 30, 40, and 50 km from the source along
each of the 16 wind directions) resulting from dispersion modeling are used as
reference points in the interpolation.
Other information needed by the HEM to predict exposure include the vent
height, vent diameter, gas discharge temperature, gas emission velocity, and
building cross-sectional area for each vent modeled. The emission rate of
each pollutant for whiduexposure estimates are needed must also be specified.
For this analysis, ambient concentrations and exposure were modeled using
the HEM around every murricipal waste combustor currently known by EPA to be
operating in the United "States. Ambient concentrations and exposures were
modeled under both control scenarios. This modeling was performed based on
site-specific meteorology, population, and emissions data. Emission rates of
organics and metals for individual facilities were specified based on the
appropriate baseline and controlled emission factors (see Section 3.2.1)
scaled according to the design waste throughput or controlled particulate
emission rate of each facility. For projected municipal waste combustors,
ambient concentrations and exposure under both control scenarios were modeled
using the HEM for three facilities representing each of the three combustor
types: a 1,000 ton/day massburn, a 1,500 ton/day RDF, and a 250 ton/day
modular combustor. The emission factors developed to represent emissions from
projected municipal waste combustors (see Section 3.2.1) were multiplied times
the model plant waste combustion capacity or controlled particulate emission
rate to generate emission rates. Meteorology and population data for modeling
these three facilities were selected by assuming the following locations:
downtown and suburban sites in Boston, Massachusetts, and Los Angeles,
California, for the massburn and RDF facilities; and downtown and suburban
sites in Erie County, New York, and Ukiah, California, for the modular
combustor. These locations were selected to represent places where these
3-28
-------
types of combustors are expected to be located. Stack parameter data used in
modeling both the existing and projected combustors are given in "Municipal
Waste Combustion Study: Characterization of the Municipal Waste Combustion
Industry."13
The HEM employs a number of simplifying assumptions when computing
exposures to a pollutant, including:
1. It is assumed that most exposure occurs at population weighted
centers (centroids) of block group and enumeration districts (BG/ED)
because the locations of actual residences are not contained in
available databases. The model relies on information provided in a
database developed by the U.S. Census Bureau.
2. It is assumed that people reside at these centroids for their
entire lifetimes (assumed to be 70 years for calculating cancer
risk).
3. It is assumed that indoor concentrations are the same as outdoor
concentrations.
4. It is assumed that plants emit pollutants at the same emission
rate for 70 years. Long-term emission rates are not known.
5. It is assumed that the only source of exposure is the ambient air
and resuspension of pollutants via dust is not considered.
6. It is assumed that there is no population migration or growth.
7. The model does not provide for discriminating exposure situations
that may differ with age, sex, health status, or other situations.
Susceptible population subgroups are not considered.
3.3.1.2 Direct Inhalation Nationwide Cancer Risk Evaluation. The
exposure estimates predicted by the HEM are combined with measures of carcino-
genic potency to estimate the probability of cancer by direct inhalation of
municipal waste combustor emissions. The estimation of'carcinogenic potency
is expressed as the "unit risk." The unit risk estimate for an air pollutant
is defined as the probability of the upper 95 percent confidence limit of the
lifetime cancer risk that could occur in a population in which all individuals
are exposed continuously from birth throughout their lifetimes (70 years) to a
unit concentration (e.g., 1 ug/m ) of the carcinogen in the air they breathe.
The measures of carcinogenic potency (unit risk estimates) of pollutants
emitted from municipal waste combustors were derived by EPA's Carcinogen
3-29
-------
Assessment Group and are listed in Table 3-13. The derivation of unit risk
estimates depends on a a quantitative evaluation of adverse health outcomes
statistically tied to exposure to a chemical as observed in epidemiological,
clinical, toxicological, and environmental research. Evidence for
carcinogenic action is based on the observations of statistically significant
tumor responses in specific organs or tissues. Evidence of possible
carcinogenicity to humans is evaluated according to specific criteria set
forth in EPA's Guidelines for Carcinogenic Risk Assessment. Since risks at
low exposure levels cannot be measured directly either by animal experiments
or by epidemiological studies, mathematical models must be used to extrapolate
from high doses to low doses characteristic of environmental exposures.
There are differences in the amount, type and quality of the data used to
derive the unit risk estimates. Such differences may affect the confidence
that can be assumed in the quantitative estimate of carcinogenic potency and
subsequent quantification of cancer risks. The uncertainties associated with
cancer risk assessment vary with the chemical. Generally, human data provide
the best evidence that a chemical is a carcinogen. In the absence of human
data, animal-to-man extrapolation must be performed. The unit risk estimated
based on animal bioassays is considered an upper bound estimate of the excess
cancer risk over a lifetime in populations exposed to the probable carcinogen.
The concept of equivalent doses for humans compared to animals has little
experimental verification regarding carcinogenic response and this is a major
area of uncertainty. In addition, human populations are more variable than
laboratory animals with respect to genetic constitution, diet, living
environment and activity patterns. The overall uncertainties associated with
unit risk estimates have not yet been statistically quantified. At best, the
linear extrapolation model used to derive unit'risk estimates provides an
approximate but plausible estimate of the upper limit of risk -- it is not
likely that the true risk would be much more than the estimated risk, but the
true risk could very well be considerably lower. The quantitative aspects of
risk assessments should not be construed as an accurate representation of the
true cancer risk, but are best utilized in the regulatory-decision process,
such as setting regulatory priorities based on a relative indication of the
potential magnitude of population risk from chemical exposure.
3-30
-------
TABLE 3-13. UNIT CANCER RISK ESTIMATES FOR INHALATION EXPOSURE TO
SPECIFIC CHEMICALS (risk per ug pol 1 utant/m3 of air)
Pollutant Unit Risk Factor Reference
Arsenic 4.2 x 10"3 EPA 600/8-83-021f
Benzo(a)pyrene 1.7 x 10"3 EPA 540/1-86-022
Beryllium 2.4 x 10"3 EPA 600/8-84-026b
Cadmium 1.8 x 10"3 EPA 600/8-83-025f
Hexachlorobenzene 4.8 x 10"4 EPA 600/8-84-015f
Trichlorophenol 5.7 x 10"6 EPA 440/5-80-032
Chromium (VI) 1.2 x 10"2 EPA 600/8-83-014f
*2378-TCDD 3.3 x 10"5 (pg/m3)"1 EPA 600/8-84-014f
*HexaCDD 1.3 x 10"6 (pg/m3)"1 EPA 600/8-84-014f
Formaldehyde 1.8 x 10"4 Draft Document
PCS 1.2 x 10"3 EPA 540/1-86-004
*The carcinogenic potency of TCDD and HexaCDD is such that the,unit risk
estimates are based on inhalation exposure to 1 picogram (10 g) per
cubic meter of air.
3-31
-------
By combining the estimates of public exposure with the unit risk
estimate, the HEM produces two types of cancer risk estimates. The first,
called maximum individual lifetime cancer risk, relates the risk to the
individual or individuals estimated to live in the area of highest ambient air
concentration as estimated by the dispersion model. These are the
"most-exposed individuals" (MEIs). The second type of risk estimate, called
annual cancer incidence, is a summation of all the risks to people living
within 50 kilometers in all directions of the facility. Fifty kilometers is
the air dispersion modeling radius used in the risk analysis. The cancer
incidence is expressed as cases of cancer among all of the exposed population
after 70 years of exposure; for statistical convenience, it is often divided
by 70 and expressed as cancer incidences per year.
For the existing municipal waste combustor population, the annual cancer
incidence estimates resulting from pollutant emissions at individual
facilities were summed to represent the total nationwide cancer incidence for
each combustor type under the baseline and controlled scenarios. .The highest
maximum individual lifetime cancer risk from direct inhalation of municipal
waste combustor emissions was determined for the baseline and controlled
scenarios by observing the maximum lifetime risk at each modeled facility and
selecting the greatest value. The maximum individual cancer risk reduction
obtainable under the controlled scenario for existing municipal waste
combustors was estimated by subtracting the controlled risk estimates from the
baseline risk estimates.
For the projected municipal waste combustor population, the total
nationwide annual cancer incidence for each combustor technology was estimated
for the baseline and controlled scenarios by multiplying the annual cancer
incidence for each model plant location by the ratio of the projected total
waste throughput for the specific municipal waste combustor technology
category (as estimated in 1993) to the model plant capacity. As in the case
of the existing municipal waste combustor population, the potential cancer
risk reduction that could be achieved under the controlled scenario for
projected municipal waste combustors was estimated by subtracting the
difference between the controlled and baseline risk estimates.
3-32
-------
Ranges of cancer risk resulting from direct inhalation exposure to
predicted ambient concentrations, under baseline and controlled scenarios, of
the carcinogens evaluated in this analysis were previously presented in
Section 2.3.1. The cancer risk estimates were presented as ranges because of
the uncertainties and assumptions in the analysis. As described in
Section 2.3.1.1, these uncertainties include assumptions made concerning
carcinogenic potency of classes of compounds, emissions estimates, and
sampling and analysis methodology.
3.3.1.3 Noncancer Risk; Noncancer risk to humans exposed to direct
emissions from municipal waste combustors was evaluated by comparing the
maximum exposure levels predicted by the HEM to reference levels derived from
consideration of noncancer health effects. This analysis was limited to a
consideration of emissions of lead, mercury, and HC1.
The modeled ambient lead concentrations: were compared to the National
Ambient Air Quality Standard (NAAQS) for lead of 1.5 ug/m . NAAQS are set to
protect the public health and welfare with a margin of safety. Mercury
concentrations were compared to a reference level for ambient air established
in review of the National Emission Standard for mercury, based on health
effects and in consideration of total body burden due to ingestion and
inhalation.
For HC1, modeled concentrations were compared to.the American Conference
of Government Industrial Hygienists (ACGIH) ceiling level of 5 ppm. This is
a level for worker exposure established to be sufficiently low to prevent
toxic injury from exposure to HC1, but on the borderline of severe irritation.
Concentrations of pollutants measured in ambient air may be contributed
by multiple sources. Comparisons with modeled concentrations due to municipal
waste combustors can only be done with reference to air quality in a
particular location. However, in general, having modeled concentrations that
exceed such reference levels indicates a possible adverse effect. Moreover,
approaching such a reference level may also be indicative of adverse effects,
depending on other pollutant sources in the area and general background
levels.
Maximum exposure to annual average emissions of lead and mercury was
estimated for every existing municipal waste combustor in the United States
using the HEM. Site-specific meteorology, population, and emissions data were
3-33
-------
used In the modeling. Emission rates were determined for each facility by
multiplying the emission factors times the controlled particulate emission
rate or design waste throughput of the facility. The results (i.e., maximum
exposure) predicted by the HEM for each individual facility were compared to
the ambient concentration levels associated with noncancer effects presented
above.
In the case of HC1 emissions, maximum short-term (1-hr) exposures were
also estimated for a 500 ton/day massburn and 108 and 200 ton/day modular
combustors. For purposes of specifying meteorological and population inputs,
using a short-term dispersion model, the massburn combustor was located in
Oyster Bay, New York, and the modular combustors were located in Durham, New
Hampshire, and in Portsmouth, New Hampshire, respectively. The source
parameters for this modeling are given in Table 3-14. Emission rates were
specified using the maximum HC1 emission factors presented in Section 3.2.1.3.
The maximum short-term HC1 concentrations predicted for these three combustors
were compared to the worker Threshold Limit. Value (TLV) for HC1 to estimate
whether potential adverse health effects may result from short-term exposure
to acid gases emitted from municipal waste combustors.
3.3.1.4 Welfare Effects. In addition to evaluating effects on humans
from inhalation exposure to. direct emissions from municipal waste combustors,
potential adverse welfare effects on the environment were evaluated.
Specifically, the potential for materials damage resulting from exposure to
HC1 emitted from municipal waste combustors was evaluated.
Modeled ambient concentrations of HC1 surrounding municipal waste
combustors were compared to a level estimated by the Texas Air Control Board
to be associated with observed corrosion of ferrous metals. The estimates
were generated by modeling ambient concentrations expected to occur as a
result of measured emissions from a hydrochloric acid plant. The modeled
concentration at the point of observed corrosion was reported to be 3.0 ug/m .
To estimate the potential for HC1 emissions from municipal waste
combustors to cause materials damage, the exposure levels for HC1 modeled for
the analysis of noncancer risks (see Section 3.3.1.3) were compared to the
concentration level associated with corrosion effects. This comparison was
made for each individual existing combustor based on the modeled estimated
3-34
-------
TABLE 3-14. PARAMETERS USED TO MODEL SHORT-TERM MAXIMUM HC1 CONCENTRATIONS
AND EXPOSURES FOR MUNICIPAL WASTE COMBUSTORS
Facility type
Waste capacity
'
Facility location
Vent height
Vent diameter
Vent discharge velocity
Vent discharge temperature
1
Modular
108 tons/day
Durham, NH
18,8 m
1.2 m
11.5 m/s
455 K
2
Mass-burn
500 tons/day
Oyster Bay, NY
44.2 m
2.1 m
17 m/s
350 K
3
Modular
200 tons/day
Portsmouth, NH
3.7 m
0.9 m
16.25 m/s
437 K
3-35
-------
annual average concentration predicted for that facility. The comparison was
also made for the massburn and two modular combustors for which maximum short-
term (1-hr) HC1 concentrations are available.
3.3.2 Indirect Exposure from Long-Term Deposition of Municipal Waste
Combustor Emissions
Long-term deposition of municipal waste combustor emissions was estimated
using the Industrial Source Complex (ISC) Model. Human exposure to deposited
emissions from municipal waste combustors through ingestion of contaminated
soil, food, or water or through skin contact with contaminated soil were
evaluated for 30- and 100-year scenarios using a series of mathematical
models. These models are the Terrestrial Food Chain Model, the Surface Runoff
Model, the Groundwater Infiltration Model, and the Dermal Exposure Model. The
first three models were also used to evaluate levels of exposure to municipal
waste combustor emissions for herbivorous animals, plants, soil and aquatic
organisms. The pollutants considered in the indirect exposure analysis are
listed in Table 3-15.
For evaluating the hypothesis that indirect exposure to municipal waste
combustor emissions following long-term deposition may contribute appreciably
to total exposure, it is not currently feasible nor practical to apply the
models to every existing or planned municipal waste combustor. Therefore,. EPA
simplified the modeling process by selecting two representative municipal
waste combustors to evaluate the potential cancer risk contribution to the
"most-exposed individual" (MEI). resulting from 30- and 100-year exposure to
deposited emissions from municipal waste combustors. Thirty years was chosen
to represent the operating life of a typical combustor, and 100 years was
selected with the assumption that once a site was dedicated to that purpose,
the facility would be replaced with successive units of the same size and
design. The MEI in this analysis was assumed to be a hypothetical farm family
that obtained most of their food supply from the area of maximum deposition of
pollutants emitted from muncipal waste combustors, and whose children ingested
0.5 grams of dirt per day. This family was assumed to live just outside the
estimated boundary (200 meters, or about one-tenth of a mile) of the modeled
combustion facility.
3-36
-------
TABLE 3-15. POLLUTANTS CONSIDERED IN THE INDIRECT EXPOSURE ANALYSIS
Organics
Metals
CDD/CDF
chlorobenzenes
polychlorinated biphenyls
formaldehyde
lead
mercury
chromium +6
nickel
beryl1i urn
3-37
-------
A model plant was devised to represent technologies typical of massburn
municipal waste combustors currently being planned or constructed. An
existing municipal waste combustion facility operating in Virginia was
selected as the "model" to represent the potential adverse impacts of
emissions from existing municipal waste combustor technology. The model plant
configuration was selected to closely mimic actual planned massburn facilities
in terms of stack parameters, building area, and pollutant emissions, but
reflected the upper end of capacities currently planned. The Virginia
facility was planned in the mid-1970's, and put into operation in 1980. EPA
has accumulated several test reports of emissions of organic and inorganic
pollutants at this facility, more so than any other incinerator site.
However, this facility has known operational problems.
The methodology for modeling indirect exposure to pollutants emitted from
municipal waste combustors is currently under development to reflect recent
review comments received from EPA's Science Advisory Board. Because of the
preliminary nature of the methodology and assumptions, EPA feels that the
results should not be interpreted quantitatively at" this time. Further,
because the objective of the analysis was to test the hypothesis that indirect
exposures could contribute significantly to the total exposure due to
municipal waste combustor emissions, the analysis used conservative,
potentially worst-case, assumptions. The various mathematical models used in
the indirect exposure analysis are described below.
3.3.2.1 Industrial Source Complex Dispersion Model. The first step in
analyzing indirect exposure to municipal.waste combustor emissions was to
model the deposition and ambient air concentrations of pollutants on surfaces
around the Virginia and model plant facilities. This was accomplished using
the Industrial Source Complex Short-Term (ISCST) Model. The ISCST Model, an
extended version of the Single Source (CRSTER) Model, is designed to calculate
ambient concentration or dry deposition values for time periods of 1, 2, 3, 4,
5, 8, 12, and 24 hours. If used with a year of sequential hourly
meteorological data, ISCST can also calculate annual concentration or
deposition values. Because the current ISCST Model has no provision for
calculating wet deposition of the emissions, EPA developed a program to
estimate the effect of precipitation events on the rate of surface
3-38
-------
deposition. A detailed description of the ISCST Model and the accompanying
algorithm developed by EPA to predict wet deposition of pollutants is given in
Appendix B.
Inputs required to run the ISCST Model include hourly meteorological
data, source characteristics and receptor features. Hourly meteorological
data requirements are the mean wind speed measured at height z^, the direction
toward which the wind is blowing, the wind-profile exponent, the ambient air
temperature, the Pasquill stability category, the vertical potential
temperature gradient and the mixing layer height. Source input data
requirements include pollutant emission rates, physical stack measurements
(I.e., height, diameter, base elevation, exit velocity and temperature),
dimensions of nearby buildings, and gravitational settling parameters for
particulate matter. Inputs on the type of receptor grid needed (i.e., polar
or Cartesian) and specific receptor locations.(if different from the standard
receptor array) are also required.
For this analysis, preprocessed hourly meteorological data corresponding
to the modeled sites were used. Emission rates modeled for each combustor are
shown in Tables 3-16 and 3-17 along with other modeling parameters for the two
facilities. Estimates of particle size distribution needed to model dry and
wet deposition were presented in Section 3.2.2. A polar receptor grid was
specified for this analysis.
The outputs for both surface deposition and ambient air impacts from the
ISCST Model consisted of a concentration array for 160 receptors along each of
16 wind directions (computed every 22.5° on a polar grid) and located at
concentric radial distances from each facility of 0.2, 0.5, 1, 2, 5, 10, 20,
30 40, and 50 kilometers. The estimated deposition and ambient air
concentrations at each receptor location was given for each modeled pollutant.
3.3.2.2 Terrestrial Food Chain Model. Humans in the vicinity of
municipal waste combustors have the potential to directly ingest soil
contaminated with deposited emissions, or to consume vegetation and animal
tissues containing the contaminants. The Terrestrial Food Chain Model has
separate components for examining each potential human exposure pathway
illustrated in Figure 3-1. These components include methods for using
empirical data on contaminant uptake by plant or animal tissues to estimate
concentrations, and for integrating these estimates in conjunction with
3-39
-------
TABLE 3-16. MODELING PARAMETERS FOR THE MODEL PLANT REPRESENTING
PROJECTED MUNICIPAL WASTE COMBUSTORS
Massburn Municipal Waste Combustor (heat recovery)
Capacity: 3000 TPD (2727 Mg/d)
Location: Western Florida
Latitude: 27° 57
Longitude: 82° 27
Stack Height: 46 meters (m)
Stack Diameter: 3.1 m
Stack Exit Velocity: 11.3 m/s
Stack Temperature:
(a) 470° K with ESP
(b) 443° K with Dry Scrubc,-s/Fabric Filters (FF)
Building Configuration:
H: 35 m
W: 76 m
L: 42 m
Emission Rates (grams per second):
Organics Baseline Controlled
CDD/CDF I.SOxlO"6' 9.4xlO"8
PCB 6.70xlO"6 4.2xlO"7
B(a)P 1.7 xlO"3 l.lxlO"4
-4 -5
Chlorobenzenes 2.8 xlO 1.8x10
Chlorophenols 5.68xlO"4 3.6xlO"5
Formaldehyde 8.0 xlO"2 S.OxlO"3
3-40
-------
TABLE 3-16. (Continued)
Emission Rates (grams per second) [Continued]:
Baseline
Inorganics >10u >2 - lOu <2u
Arsenic
Cadmi urn
Chromium
Lead
Nickel
Mercury
1.4x10
Beryllium 5.8x10
1.5x10
0.6x10
8.6x10
1.5x10
-4
-6
-3
-3
-3
-3
7.8x10
-3
4.4x10
1.6x10
5.0x10
1.7x10
2.8x10
4.7x10
-4
2.5x10
-5
-3
-3
-2
-3
-2
1.7 xlO
4.4 xlO
1.2 xlO
3.2 xlO
2.67x10
8.9 xlO
2.3 xlO
-3
-5
-2
-3
-1
-3
-1
Controlled
>10u
4.6x10
1.2x10
-5
-6
5.1x10
2.0x10
2.9x10
5.0x10
5.6x10
-4
-4
-3
-4
-3
>2 - lOu
l.'SxlO
5.8x10
1.6x10
6.0x10
9.3x10
1.6x10
-4
-6
-3
-4
-3
-3
1.8x10
-2
<2u
5.9 xlO
1.5 xlO
4.2 xlO
1.1 xlO
8.9 xlO
3.0 xlO
1.66x10
-4
-5
-3
-3
-2
-3
-1
3-41
-------
TABLE 3-17. MODELING PARAMETERS FOR A MUNICIPAL WASTE COMBUSTOR IN
VIRGINIA REPRESENTING EXISTING FACILITIES
Massburn Municipal Waste Combustor (heat recovery)
Capacity: 120 TPD (109 Mg/d)
Latitude: 37° 6 02"
Longitude: 76° 23 28"
Stack Height: 27.44 m (2 stacks)
Stack Diameter: 1.22 m (2 stacks)
Stack Temperature:
(a) 543° K with ESP
(b) 443° K with Dry Scrubbers/Fabric Filters
Stack Exit Velocity: 12 m/s
Building Configuration:
H: 27.28 m
W: 51.88 m
L: 59.5 m
Emission Rates (grams per second):
Oroanics
CDD/CDF
PCB
B(a)P
Chlorobenzenes
Chlorophenols
Formaldehyde
Baseline
1.40x10
1.05x10
1.42x10
5.82x10
2.10x10
6.26x10
-5
-5
-4
-4
-3
-3
Control1ed
8.68xlO"7
6.58x10
8.92x10
3.64x10
1.11x10
-7
-6
-5
-4
4.00x10
-4
3-42
-------
TABLE 3-17. (Continued)
Emission Rates (grams per second) [Continued]:
Baseline Controlled
Inorganics >10u >2 - lOu <2u >10u >2 - lOu <2u
Arsenic 2.4xlO~4 4.8xlO"4 2.0xlO"3 S.OxlO"5 l.SxlO"4 6.4xlO"4
Beryllium 1.6xlO"8 6.4xlO"8 1.6xlO"7 5.4xlO"9 2.2xlO"8 5.2xlO"8
Cadmium 4.8xlO"4 1.4xlO"3 4.0xlO"3 1.7xlO"4 4.4xlO"4 1.2xlO"3
Chromium 2.8xlO"4 l.lxlO"3 1.9xlO"3 9.2xlO"4 3.4xlO"4 6.4xlO"4
Lead 3.0xlO"3 9.0xlO"3 l.OxlO"1 " 9.6xlO"4 3.0xlO"3 3.4xlO"2
Nickel 2.6xlO"4 8.6xlO"8 l.Sxlb"3 8.4xlO"5 2.8xlO"4 5.4xlO"4
Mercury 8.6xlO"4 2.8xlO"3 2.6xlO"2 l.SxlO"4 5.8xlO"4 5.2xlO"3
3-43
-------
Human (1.e., pica)
Animal » Human
MWC Emissions > Deposition » Soil - > Plant > Human
Uptake
Plant > Animal > Human
Uptake
Figure 3-1. Human Exposure Pathways Evaluated by
the Terrestrial Food Chain Model
3-44
-------
typical diet assumptions to give an estimate of potential human dietary
exposure. Other components estimate potential exposure resulting from soil
ingestion in children (termed, "pica"). The model predicts maximum daily
intake values for the MEI based on estimates of human dietary exposure and
pica in children. The MEI was assumed to be an individual residing in the
area of maximum deposition of emissions. In the case of pica exposure, the
MEI was assumed to be a preschool child.
In addition to evaluating human exposure to deposited emissions, the
Terrestrial Food Chain Model also predicts exposure potential to herbivorous
animals, plants, and soil organisms near municipal waste combustors. The
ecological exposure pathways evaluated are shown in Figure 3-2. The model
predicts exposure for these pathways in the same way as for human exposure
except that the MEI is now a sensitive animal, plant or soil organism in the
area of maximal deposition. The Terrestrial Food Chain Model is described in
detail in Appendix C.
3.3.2.3 Surface Runoff Model. Following deposition, pollutants from
municipal waste combustors are subject to dissolution and/or suspension in
runoff after precipitation events. Runoff moves over the surface of the earth
to a surface water body where it mixes with other waters. As a consequence,
humans utilizing water or eating fish from the surface water body may be
exposed to runoff transported contaminants. Adverse effects on aquatic life
in contaminated waters may also be observed. The Surface Runoff Model
includes three successive tiers of analysis beginning with simple but very
conservative estimates and proceeding to more refined analyses if the first
tiers predict unacceptable exposures. Both acute events and chronic exposure
are evaluated using standard approaches to calculate runoff volume and
associated runoff potential. The model calculates maximum receiving water
concentrations of individual contaminants. In the case of human exposure, the
MEI is assumed to be a person residing at the point of maximum deposition of
pollutant emissions from the municipal waste combustor, and consuming water
and eating fish taken from surface waters experiencing maximum contaminant
concentrations from surface runoff of the deposited emissions. In the case of
aquatic organism exposure, the MEI is a sensitive aquatic organism living in
surface waters containing maximum contaminant concentrations. The
3-45
-------
Soil Biota
MWC Emissions —•.—> Deposition » Soil » Plant
Soil Biota -» Predator
Animal
Plant > Animal
Figure 3-2.
Ecological Exposure Pathways Evaluated by
the Terrestrial Food Chain Model
3-46
-------
methodology in this model was originally developed to evaluate impacts from
the application of municipal wastewater sludge to land. The Surface Runoff
Model is described in detail in Appendix D.
3.3.2.4 Groundwater Infiltration Model. Contaminants from municipal
waste combustors dissolved in rain or meltwater from precipitation events may
alternatively infiltrate into the ground and enter the groundwater. As a
consequence, persons using the groundwater may be exposed to groundwater
transported contaminants. Aquatic life inhabiting surface water bodies fed by
the contaminated aquifer could be exposed as well. The Groundwater
Infiltration Model evaluates chronic exposure using standard approaches to
calculate leachate generation and associated groundwater transport in the
unsaturated zone. Similar to the Surface Runoff Model, the Groundwater
Infiltration Model is formulated in three successive tiers which begin with
simple conservative estimates and proceed to more detailed analyses if the
first tiers predict unacceptable exposures. The model estimates maximum
potential concentrations of contaminants at various points of the aquifer,
depending on which analytical tier is being used. The human and nonhuman MEI
are the same as for the Surface Runoff Model. A detailed description of the
Groundwater Infiltration-Model is given in Appendix E.
3.3.2.5 Dermal Exposure Model. Humans may also be exposed dermally to
soils contaminated by municipal waste combustor emissions. The Dermal
Exposure Model estimates a daily dermal intake of contaminants by the MEI
based on the fraction of the contaminant absorbed, contact time (i.e.,
duration of daily exposure), exposed skin surface area, contact amount
(i.e., amount of soil accumulated on skin, and soil contaminant concentration.
This exposure level is then adjusted to account for the relative effectiveness
of an absorbed dermal dose relative to an unabsorbed ingested dose. The MEI
is assumed to be a preschool child playing in soil in the area of maximal
deposition.
It should be noted that there is a fundamental lack of data for
percutaneous absorption of chemicals in human skin from soil. Other factors
important for estimation of human exposure to contaminants by the dermal route
also have many uncertainties. In addition, systemic toxic thresholds or
carcinogenic potencies of chemicals by a dermal route of exposure have not
been delineated by EPA at the present time. ^While this model represents a
3-47
-------
possible approach for the estimation of human exposure associated with dermal
exposure, it is recognized that in most, if not all cases, the available data
will not provide a satisfactory basis for risk exposure calculations. The
Dermal Exposure Model is fully described in Appendix F.
3.3.2.6 Evaluation of Exposure from Indirect Pathways. Because of
the developmental nature of the indirect exposure methodology described
previously, reliable quantitative risk estimates for indirect exposure
pathways have not yet been generated. However, an analysis was performed
using the preliminary methodology to predict the likelihood that indirect
exposure routes could contribute significantly to the total exposure to
individuals living near existing and projected municipal waste combustors,
as represented by the Virginia and Western Florida model facilities. This
limited analysis was performed for indirect exposure to individual
pollutants through four principal routes: food ingestion, soil ingestion,
water ingestion and fish ingestion. The effects of deposition of pollutants
over 30 and 100 years were also considered.
Tables 3-18 and 3-19 summarize the analysis for inorganic and organic
pollutants, respectively. For each pollutant evaluated using the
preliminary methodology, the likelihood that a particular indirect exposure
pathway could contribute significantly to total exposure is subjectively
classified as either "possible" or "low likelihood." The "no data"
classification was used to indicate those pollutants for which insufficient
data were available to apply the preliminary methodology.
Certain aspects of the indirect exposure modeling should be considered
when reviewing the qualitative observations in Tables 3-18 and 3-19. The
selection criteria of both the Virginia and western Florida facilities are
considered as reasonable worst-case in that the facility size, location,
stack parameters, and stack pollutant emissions tended toward maximizing
potential ambient impacts near each municipal waste combustor. Both
facilities are located in humid climates where an excess of 40 inches of
rainfall per year enhanced the effect of wet deposition of pollutants near
each facility. Dryer climates would be expected to produce lower rates of
wet deposition. In the western Florida case the sandy soil characteristics
may have a tendency to maximize the rate of downward movement of pollutants
3-48
-------
TABLE 3-18. RESULTS OF ANALYSIS OF THE LIKELIHOOD OF POTENTIAL HUMAN HEALTH EFFECTS FROM INDIRECT EXPOSURE TO MUNICIPAL WASTE COMBUSTOR EMISSIONS
(Exposure by Ingest ton - Inorganic Compounds)
Food Ingestion
Control Deposition Lou
Facility Scenario Interval Possible Likelihood
Existing Baseline 30 yr. Pb, Cr
Ug, Hi
(VIRGIHIA-120 cpd) 100 yr. Hg Pb, NJ.
Cr
Controlled 30 yr. Pb, Cr
Hg, Hi
100 yr. Bg Pb, NL
Cr
co
i
•** Projected Baseline 30 yr. Bg Pb, Hi
40 Cr
(WESTERN FLORIDA-3,000 tpd) 100 yr. Hg Pb, Nl
Cr
Controlled 30 yr. Bg Pb, Hi
Cr
100 yr. Hg Pb, Ni
Cr
Soil Innestlon
No Low
Data Possible Likelihood
Be Pb
Hg
Be Pb
Hg
Be
Be Pb
Hg
Be Pb
Hg
Be Pb
Bg
Be Hg
Be Pb
Hg
Cr, Be
Ni
Cr, Be
Hi
Pb, Cr
Hg. Nl
Cr, Be
Hi
Cr, Be
Nl
Cr, Be
Ml
Pb, Ni
Cr, Be
Cr, Be
Nl
Surface Water Incest ion
Fish Ingest ion
No Low No Low No
Data Possible Likelihood Data Possible Likelihood Data
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Pb, Cr
Hg, Nl
Be
Pb, Cr
Hg, Nl
Be
Pb, Cr
Hg, Nl
Be
Pb, Cr
Hg, Ni
Be
Bg Pb,
Cr,
Hg PB,
Cr,
Bg Pb,
Cr,
Hg Pb,
Cr,
Hg Pb,
Cr,
Hg Pb,
Cr,
Hg Pb,
Cr,
Hg Pb,
Cr,
Ni
Be
Nl
Be
Nl
Be
Nl
Be
Nl
Be
Nl
Be
Hi
Be
Ni
Be
Be - Beryllium Cr - Chromium Hg » Mercury Nl = Nickel Pb = Lead
-------
TABLE 3-19. RESULTS OF ANALYSIS OF THE LIKELIHOOD OF POTENTIAL HUMAN HEALTH EFFECTS FROM INDIRECT EXPOSURE TO MUNICIPAL WASTE COMBUSTOR EMISSIONS
(Human Risk by Ingest Ion - Organic Compounds)
OJ
tn
o
Food InRestlon
Control Deposit Lon Low
Facility Scenario Interval Possible Likelihood
*
Existing Baseline 30 yr. PCDD FA
CB
PCB
(VIRGINIA-120 tpd) 100 yc. PCDD FA
CB
PCB
Controlled 30 yr. PCDD PCB
CB FA
100 yc. PCDD PCB
CB FA
Projected Baseline 30 yr. PCDD FA
CB
PCB
(WESTERN FLORIDA-3.000 tpd) 100 yc. PCDD FA
CB
PCB
Controlled 30 yc. PCDD PCB
' CB FA
100 yr. PCDD PCB
CB FA
Soil Ingest Ion Surface Water InKestlon
Fish InKestlon
No Lou Ho Low Mo Low Ho
Data Possible Likelihood Data Possible Likelihood Data Possible Likelihood Data
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
Local water* at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
are tidal (not potable)
Local waters at Virginia
' are tidal (not potable)
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD CB
PCB
FA
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
CB
PCB
PCDD
PCDD
FA
FA
FA
FA
FA
FA
CB
PCB
FA
CB
PCB
FA
CB - Chlorobenzenes FA - Formaldehyde PCB - Polychlorlnated benzenes PCDD » Polychlorlnated dlbenzo-p-dloxlns
-------
deposited on the surface to underground aquifers. Soils containing a
greater fraction of humus matter and having a higher carbon content may
impede this downward movement of pollutants.
The analysis described the maximally exposed individuals as a rural
farm family of two adults and.two young children who partially subsisted on
foodstuffs gained by domestic agriculture, freshwater fish caught in local
waters, and potable water from local aquifers and surface water sources.
The farm family lived and undertook all their activities in the area of
maximum dry and wet surface deposition of the pollutants emitted from the
stacks of the municipal waste combustor, which was approximately 200 meters
from the facility. It is extremely likely that this is an atypical
situation, although individuals have been observed to reside as close as
200 meters from municipal waste combustors. It is not expected that
individuals will gain most of their food supply from a location this close
to the facility.
For the food ingestion pathway, the preliminary methodology predicted a
low likelihood of significant contribution to total exposure near the
existing and projected model facilities for lead, chromium, nickel, and
formaldehyde under all deposition and control scenarios. For mercury and
PCB, possible significant contributions to total exposure from food inges-
tion near both model facilities were predicted for some deposition intervals
and control scenarios, while for other deposition intervals and control
scenarios the likelihood of significant contribution to total exposure was
predicted to be low. In the case of mercury emissions, possible significant
contributions to total exposure from food ingestion were predicted for the
projected western Florida model facility for all deposition intervals and
control scenarios. At the existing Virginia model facility, possible
significant contribution to total mercury exposure from food ingestion was
predicted only for the 100-year deposition interval for both control scenarios.
In the case of PCB emissions, possible significant contributions to total
exposure near both model facility were predicted only under the baseline
scenario. Under the controlled scenario at the two modeled facilities, a
low likelihood of significant contribution to total PCB exposure was predicted
for the food ingestion pathway. Of the pollutants modeled, only indirect
exposure to PCDD and chlorobenzene was predicted to result in a significant
3-51
-------
contribution to total exposure from food ingestion near both model facilities
for all deposition intervals and control scenarios modeled. No data were
available to classify potential contribution to total exposure from food
ingestion of beryllium.
For the soil, ingestion pathway, a low likelihood of significant contri-
bution to total exposure was predicted for all deposition intervals and
control scenarios near the existing and projected model facilities for
beryllium, chromium, nickel, chlorobenzene, PCB and formaldehyde. For lead
and mercury, the predicted likelihood of significant contribution to total
exposure from indirect exposure through soil ingestion ranged from possible
to low likelihood for the two model facilities depending on the deposition
interval and control scenario. In the case of lead, the preliminary.
methodology predicted a possible significant contribution to total exposure
from soil ingestion for all deposition intervals and control scenarios with
the following exception. For both model facilities, analysis of the 30-year
deposition interval under the controlled scenario resulted in a prediction
of low likelihood of significant contribution to total exposure. The
analysis predicted similar results for mercury emissions, except that,
unlike for lead, a possible significant contribution was predicted for the
30-year deposition interval under the controlled scenario at the projected
western Florida model facility. For the soil ingestion pathway, only
indirect exposure to PCDD was predicted to result in a significant contribu-
tion to total exposure near the existing and projected model facilities for
all deposition intervals and control scenarios.
The potential for indirect exposure to municipal waste combustor
emissions from surface runoff and groundwater infiltration was not evaluated
for the Virginia facility. This facility is located in a coastal region
where runoff of contaminated soil would affect only nonpotable, tidal water
bodies. For the projected model facility in Western Florida, the
preliminary methodology predicted a low likelihood of significant
contribution to total exposure from surface water ingestion for lead,
chromium, mercury, nickel, beryllium, chlorobenzene, PCB, and formaldehyde.
A possible significant contribution was predicted for PCDD under all
scenarios.
3-52
-------
For the fish ingestion pathway, a low likelihood of significant contribu-
tion to total exposure near the existing and projected model facilities was
predicted for lead, nickel, beryllium, chromium and formaldehyde for all
deposition intervals and control scenarios. A possible significant contribu-
tion to total exposure from fish ingestion near both model facilities was
predicted for mercury, PCDD, chlorobenzenes, and PCB for all deposition
intervals and control scenarios with the following exception. The analysis
predicted that for the projected model facility in western Florida, possible
significant contribution to total exposure from fish ingestion for
chlorobenzenes and PCB would occur only under the baseline scenario.
An example of the effect of the deposition interval on the likelihood
that a particular indirect exposure pathway may contribute significantly to
total exposure may be seen in the case of mercury emissions. Indirect
exposure through the food ingestion pathway at the existing Virginia
facility is predicted to result in possible significant contribution to
total exposure for the 100-year deposition interval, but not for the 30-year
deposition interval, for both control scenarios. However, this was not
observed for the western Florida model facility. In the majority of
pollutants analyzed, the length of deposition interval did not affect the
predicted likelihood that a particular indirect exposure pathway would
contribute significantly to total exposure.
The effect of control scenario on the predicted likelihood that an
indirect exposure pathway would contribute significantly to overall exposure
can also be seen in this limited analysis. For example, indirect exposure
to PCB from food ingestion near both model facilities represents a possible
significant contribution to total exposure under the baseline scenario
(electrostatic precipitators only) but is reduced to a low likelihood under
the controlled scenario (dry alkaline scrubbers combined with fabric
filters). Similarly, for the fish ingestion pathway, the likelihood of
significant contribution to total exposure posed by emissions of
chlorobenzenes and PCB are classified as possible under the baseline
scenario but as low likelihood under the controlled scenario.
3-53
-------
A similar analysis was performed using the preliminary indirect
exposure methodology to predict the likelihood of potential ecological
effects from long-term deposition of municipal waste combustor emissions.
The likelihood of adverse impacts to terrestrial and aquatic life were
analyzed for the Virginia and western Florida model facilities for 30- and
100-year deposition intervals and for baseline and controlled scenarios.
Table 3-2.0 summarizes the analysis of potential ecological effects near
the modeled facilities. Based on the preliminary indirect exposure
methodology, possible terrestrial and aquatic impacts were predicted for
emissions of lead and mercury for some, but not all, of the deposition
intervals and control scenarios modeled for both facilities. Possible
terrestrial impacts from -lead emissions were predicted for the 100-year
deposition interval under the baseline scenario for the existing Virginia
facility and for all scenarios except the 30-year controlled scenario for
the projected western Florida model facility. In the case of mercury
emissions, possible terrestrial impacts were predicted for the 100-year
deposition interval under both the baseline and controlled scenarios for the
existing Virginia facility, and for all scenarios for the projected western
Florida facility. Insufficient data were available to apply the preliminary
methodology to predict the likelihood of terrestrial impacts from beryllium.
Data to analyze likelihood of terrestrial impacts from chromium were available
for certain terrestrial pathways but not for others. Where data were
available, a low likelihood of terrestrial impacts was predicted for chromium.
A low likelihood of terrestrial impacts was also predicted for nickel under
all scenarios.
Possible aquatic effects from lead and mercury emissions were predicted
for all deposition intervals and control scenarios for the existing Virginia
facility. For the projected western Florida model facility, however,
possible aquatic effects were predicted for all scenarios for mercury, while
a low likelihood of aquatic effects were predicted under all scenarios for
lead. A low likelihood of aquatic impacts was predicted under all scenarios
for the two model facilities for beryllium and nickel. As with the analysis
of terrestrial impacts, data on chromium were available for some aquatic
pathways but not for others. For those pathways analyzed for chromium, a
low likelihood of aquatic impacts was predicted.
3-54
-------
TABLE 3-20. RESULTS OF ANALYSIS OF THE LIKELIHOOD OF POTENTIAL ECOLOGICAL EFFECTS FROM INDIRECT EXPOSURE TO MUNICIPAL WASTE COMBUSTOR EMISSIONS
(Inorganic Compounds)
Ecological Risks
Terrestrial Impacts Aquatic Impacts
Control Deposition Low
Facility Scenario Interval Possible Likelihood
Existing Baseline 30 yr. Pb, Cr
Hg, Nl
(VIRGINIA-120 tpd) 100 yr. Pb Cr
Hg Nl
Controlled 30 yr. Pb, Nl
Cr
100 yr. Hg Pb, Nl
Cr
Projected Baseline 30 yr. Pb Pb, Ni
Hg
(WESTERN FLORIDA-3000 tpd) 100 yr. Pb Cr, Nl
Hg
Controlled 30 yr. Hg Pb, Nl
Cr
100 yr. Hg Cr
Pb Nl
No Data
Cr*
Be
Cr
Be
Cr
Be
Cr
Be
Cr
Be
Cr
Be
Cr
Be
Cr
Be
Possible
Pb
Hg
Pb
Hg
Pb
Hg
Pb
Hg
Hg
Hg
Hg
Hg
Low
Likelihood
Cr, Be
Nl
Cr, Be
Nl
Cr, Be
Nl
Cr, Be
Nl
Pb, Nl
Cr, Be
Pb, Nl
Cr, Be
Pb, Nl
Cr, Be
Pb, Nl
Cr, Be
No Data
Cr*
Cr
Cr
Cr
Cr
Cr
Cr
Cr
*
Data on chromium were available for certain terrestrial and aquatic pathways showing low likelihood of effects, but were not available for
other pathways (e.g. toxlclty thresholds for soil biota and for predators of aquatic organisms).
Be = Beryllium Cr - Chromium Hg = Mercury Nl = Nickel Pb = Lead
-------
The limited and qualitative aspects of the indirect exposure analysis
has generally demonstrated that exposure via these pathways can be comparable
in magnitude to exposure to pollutants through the direct inhalation of
ambient air. Thus the analysis, although not quantitative, has served the
purpose of highlighting the importance of considering all possible routes that
humans may be exposed to pollutant emissions from combustion sources. In this
regard, the EPA is continuing to develop and improve the current methodology
so that routine and reliable estimates of exposure and risk can be made.
3-56
-------
3.4 REFERENCES FOR SECTION 3.0
1. Midwest Research Institute. Municipal Waste Combustion Study: Data
Base for Municipal Waste Combustors. Review Draft. EPA Contract
No. 68-02-3817, ESED Project No. 86/19. January 7, 1987.
2. Reference 1.
3. U. S. Environmental Protection Agency. Municipal Waste Combustion
Multipollutant Study. EMB Report No. 86-MIN-02. April 1987.
4. A.-J. Fossa, R. S. Kerr, A. S. Columbus, and R. Waterfall. Air Emissions
Characterization of Municipal Waste Combustors in New York State. New
York State Department of Environmental Conservation. Presented at the
80th Annual Meeting of Air Pollution Control Association. June 21-26,
1987. New York, New York.
5. Reference 1.
6. Klicius, R., D. J. Hay, and A. Finkelstein". The National Incineration
Testing and Evaluation Program. Presented for use by EPA's Science
Advisory Board by Environment Canada. Ontario, Canada. September 1986.
7. Reference 6.
8. Determination of ChloroDibenzo-p-dioxins, Chlorinated Dibenzofurans,
Chlorinated Biphenyls, Chlorobenzenes and Chlorophenols in Air Emissions
at SWARD. Report No. ARB-02-84-ETRD. Ministry of Environment. Ontario,
Canada. 1984.
9. U. S. Environmental Protection Agency. Environmental Assessment of
Waste-to-Energy Process: Braintree Municipal Incinerator. EPA
600/7-80-149. 1980.
10. Hahn, J. L. Air Emissions Testing at the Wurzburg, West Germany
Waste-to-Energy Facility. Presented at the Annual Meeting of the Air
Pollution Control Association, June 1986.
11. Radian Corporation. Municipal Waste Combustion Study Data Gathering
Phase. Preliminary Draft. EPA Contract No. 68-02-3818. 1985.
12. Hart, F. C., and Associates. Assessment of Potential Public Health
Impacts Associated with Predicted Emissions of Polychlorinated
Dibenzodioxins and Polychlorinated Dibenzofurans from the Brooklyn Navy
Yard Resource Recovery Facility. Prepared for the New York City
Department of Sanitation. 1984.
13. U. S. Environmental Protection Agency. Municipal Waste Combustion
Study: Characterization of the Municipal Waste Combustion Industry.
EPA/530-SW-87-021h. March 1987.
3-57
-------
14. 51 FR 33992. Guidelines for Carcinogen Risk Assessment. September 24,
1986.
15. I). S. Environmental Protection Agency. Mercury Health Effects Update.
Final Report. EPA 600/8-84-019F. August 1984.
16. American Conference of Government Industrial Hygienists. Threshold
Limit Values and Biological Exposure Indices for 1986-1987.
ISBN:0-936712-69-4. Cincinnati, Ohio.
17. Letter from Henry, J. (Texas Air Control Board), to Kellam, R, (U.S.
Environmental Protection Agency). June 26, 1987. Information on ambient
levels of HC1 associated with corrosion.
3-58
-------
APPENDIX A
HUMAN EXPOSURE MODEL
-------
APPENDIX A
HUMAN EXPOSURE MODEL
A.I INTRODUCTION
The Human Exposure Model (HEM) is a general model designed to estimate
the population exposed to air pollutants emitted from stationary sources and
the carcinogenic risk associated with this exposure. The HEM is comprised
of (1) an atmospheric dispersion model, with included meteorological data,
(2) a population distribution based on Bureau of Census data, and (3) a
procedure for estimating risks due to the predicted exposure. The only
inputs needed to operate this model are source data, e.g., plant location,
height of the emission release point, the temperature of the off gases.
Based on source data, the model estimates the magnitude and distribution of
ambient air concentrations of the pollutant in the vicinity of the source.
These concentration estimates are coupled with the.population to estimate
public exposure to the pollutant. The HEM then predicts population risks if
a unit risk number determined from health data is input for the pollutant.
Within HEM there are two basic models,-the Systems Applications Human
Exposure and Dosage (SHED) model and the Systems Application Human Exposure
and Risk (SHEAR) model. The SHED model is used primarily for major point
sources where the latitude and longitude are known. It is commonly used, to
evaluate exposure and risk for a source category on a nationwide basis. In
contrast the SHEAR model is used for regional analysis of multiple point
sources and area sources. It is also used to model "prototype" sources
where it is impractical to specify latitude and longitude for individual
sources.
For analysis of nationwide exposure and cancer risks from municipal
waste combustor emissions, the SHED model option in the HEM was used. The
following data are used by the HEM to assess exposures to and -dosages of
atmospheric chemicals:
A-l
-------
• source location
• emissions data
• vent parameters
• atmospheric reactivities of chemicals
• meteorological data
• population distribution data
In addition to these inputs, health data are needed to derive population
risks from population exposure. However, these data are provided in the
form of a unit risk estimate, which is derived external to the exposure
model. These health data.are not presented in this appendix.
The remainder of this appendix describes the input data required to run
the SHED option of the HEM and how the model estimates exposure and dosage.
A discussion of uncertainties and assumptions inherent in the HEM is also
given.
A.2 INPUT DATA
A.2.1 Source Location
The geographic location of the source must be specified in terms of
latitude and longitude.
A.2.2 Emissions Data
The emission rate in Mg/yr must be specified for each modeled
pollutant. (When using SHED, separate modeling runs are required to model
exposure to different pollutants.)
A.2.3 Vent Parameters
Vent parameter data are necessary for dispersion modeling of emissions.
Vent parameter data required include the number of vents, vent height, vent
diameter, gas discharge temperature, gas emission velocity, fugitive
discharge area (if applicable), and building cross-sectional area.
A-2
-------
A.2.4 Atmospheric Reactivities of Chemicals
The HEM. is generally constrained to a study area in a 50 km radius of
the source. In most instances, the reactivity of a chemical is not
sufficiently high to cause significant removal before the material is
dispersed to this distance. For these chemicals, this input to the HEM can
be left blank with the model assuming no atmospheric decay. Some chemicals
(e.g., formaldehyde) do have high reactivity and should be analyzed for this
behavior. When applicable, atmospheric reactivity is specified in terms of
daytime and nighttime decay rate (min~ ).
A.2.5 Meteorological Data
The dispersion computations performed by"the HEM require data.on wind
speed, wind direction, and the intensity of atmospheric turbulence. The
turbulence intensity is represented by the atmospheric stability class.
These data are provided by a permanent data base in the HEM known as the
Stability Array or STAR data set.
Martin and Tikvart developed the STAR program from-routinely collected
meteorological data to generate frequencies and percentage frequencies of
wind direction by speed classes for each stability category. The
specifications of stability categories depending on wind speed and sky cover
2 3
were set up byPasquill and were modified by Turner. The program was
adopted for use at the National Climatic Center (NCC), where archived
records of all national reporting weather stations are kept. The most
up-to-date version of the STAR data from all STAR stations in the country
were used to produce the matrices of STAR frequencies used in the HEM. The
format of the STAR data includes sixteen wind directions, six wind speed
classes, and seven stability categories with categories A, B, C, and D, in
the daytime and categories Dnjqnt> E, an<* F ™ tne nighttime. There are
data sets for 312 stations in the STAR data file. Figure A-l shows the
locations of these STAR sites.
By default, meteorological data recorded at the STAR station nearest to
the source are used in the dispersion modeling for SHED. However, local
meteorological trends and topographic features may be more important factors
A-3
-------
Figure A-l. Location of STAR Sites.
-------
in selecting a STAR station than is the absolute distance between the source
and the station. A STAR station with climatologial conditions most similar
to those of the source of emissions may not be the nearest station, so the
STAR stations may be manually selected to account for such situations.
A.2.6 Population Data
The HEM program uses the latitude and longitude from the input data in
determining the population of the study area. The permanent data base is
comprised of the 1980 Census Data Base broken down by Block Group/
Enumeration District (BG/ED). The population data base contains the
population centroid coordinates (latitude and longitude) and the 1980
population of each BG/ED in the country (about 300,000 centroids in
>k
50 states plus the District of Columbia).
A.3 EXPOSURE AND DOSAGE ESTIMATE SCHEME
The output of the HEM is a concentration array for 160 receptors around
the plant (10 receptors along each of the 16 wind directions). These are
the sum of concentration patterns resulting from all sources within a plant.
The basic approach used in combining the concentration pattern with the
population distribution pattern around a plant is explained in this
subsection. Three terms are defined here that will be used frequently in
the following discussion. A polar grid point is one of the 160 receptors at
which concentrations are estimated by the dispersion modeling. A population
centroid is the population-weighted geographical center of a BG/ED for which
geodetic coordinates are known. A grid cell is defined as the area bounded •
by two radial arcs and two wind directions.
Exposure is the product of the population and the concentration to
which that population is exposed. To form this product, both the
concentration and the population must be known at the same location or
point. The SHED model uses a two-level interpolation scheme to pair the
concentrations with populations prior to the computation of dosages and
exposures. The two-level approach is appropriate because the concentrations
A-5
-------
are defined on a radius-azimuth (polar) grid pattern with non-uniform
spacing. This means that at small radii the grid cells are much smaller
than BG/EDs while at larger radii the grid cells are much larger than
BG/EDs. Interpolation techniques are most appropriately applied by
interpolating values of the factor defined on a coarse network (larger) at
the locations of the finer (smaller) network, thus maximizing the resolution
and minimizing the uncertainties of interpolation. As previously mentioned,
the fine/coarse relationship between polar grid cells and BG/EDs varies with
radius. Hence, the two-level approach allows the BG/EO population to be
interpolated to the grid point when the BG/EDs are larger than the grid
cells and allows the grid point concentration to be interpolated to the
BG/ED centroid when the reverse is true. The details of this approach are
outlined below. The numbers used assume the analysis is to be completed for
a 50 km radius. If the maximum radius changes, so will the radii of the
grid points.
For BG/ED centroids located between 0.2 km and 3.5 km from the source,
populations are apportioned among neighboring polar grid points. There are
64 (4 x 16) polar grid points within this range. Associated, with each of
these grid points, at which the concentration is known, is a smaller polar
sector bounded by two concentric arcs and two radial lines. The boundary
concentric arcs are defined by radii of .10, .35, .75, 1.5, and 3.5 km and
the boundary radial lines are drawn right in the middle of two wind
directions. These boundary lines are represented by the dashed lines in
Figure A-2. Each of the polar grid points are assigned to the nearest BG/ED
centroid identified from the census data set. The population at each
centroid is then apportioned among all polar grid points assigned to that
centroid according to the area of the polar sector associated with the grid
point. For example, all of the polar grid points (labeled b) nearest
population centroid B would be assigned a certain proportion of the people
in B. The ratio between the area of the polar sector and the area assigned
to the population centroid determine the proportionment. Thus, the
population density is assumed to be the same for all polar sectors assigned
to a single centroid. Figure A-2 shows that the grid points closer to the
source are boxed in by much smaller sectors than those further away. Hence
A-6
-------
Figure A-2. Pairing of BG/ED with Concentration Within A 3.5 Km Radius
Population Centroid
Polar Grid
Point
Polar Sector
A, B, C - Locations of BG/ED Centroids
a, b, c - Polar grid points to be allocated a portion of population
reported at centroids A, B, C, respectively.
A-7
-------
the b grid points closer to the center will receive a smaller apportionment
of the people from population centroid B than those further away. Both
concentration and population counts are thus available for each polar grid
point.
Log-log linear interpolation is used to estimate the concentration at
each BG/ED population centroid located between 3.5 km and 50 km from the
source. Concentration estimates for 96 (6 x 16) grid points (receptors at
5.0, 10.0, 20.0, 30.0, 40.0, and 50.0 km from the source along each of the
sixteen directions) resulting from dispersion modeling are used here as
reference points for this interpolation. For each BG/ED centroid, four
reference points are located at the four corners of the polar sector in
which the centroid is located. These four reference points (labeled C,, C-,
C^, and C.) would surround the centroid (C ) as depicted in Figure A-3.
There is a linear relationship between the logarithm of concentrations and
the logarithm of distances for receptors more than 2.0 km away from the
source. This relationship, is used to estimate the concentrations at points
CA, and CA« (see Figure A-3). These estimates are then linearly
interpolated with the polar angle to determine the concentration at the
centroid (C ). Using the two-level approach, concentrations and populations
^
are paired up for the 64 concentration grid points within 3.5 km of the
source and for all BG/ED centroids located between 3.5 km and 50 km from the
source.
The total dosage was then computed as follows:
3 N
Total Annual Dosage (ug/m -person) » £ P^Ci ,
1-1 1
where P. = the population at point i, C. = the annual average concentration
at point i, and N = the total number of grid points and BG/ED centroids with
a specified combination of concentration and population (representing the
entire area within 50 km of the source).
The population exposed to each of a number of concentration levels, L-,
J
was computed by:
A-8
-------
Log-linear Interpolation Scheme
'
r3
i
C4
Given:
A =• The angle in degrees subtended clockwise about the source from
due south to the BG/ED centroid;
A, » The angle from due south to the radial line immediately counter-
clockwise from A, or passing through A if there is an exact match;
A? » The angle from the south to the radial line immediately clockwise
* of Aj (A2 is 0 if it is due south);
R - The distance from the source to the BG/ED centroid;
R, » The distance from the source to the largest circular arc of radius
on grid that is less than R;
R2 » The distance from the source to the smallest circular arc of radius
on grid that is greater than or equal to R;
Cj = The concentration value at (A,, R,);
C2 » The .concentration value at (A,, R2);
C3 = The concentration value at (A2, R,);
C^ » The concentration value at (A2, R2); then
CAj - exp (In Cj •»• (In C2 - In Cj) (In R - In Rj)/("In R2 - In Rj)
CA2 - exp (In C3 + (In C4 - In C3) (In R - In Rj)/(ln R2 - In Rj)
C^A • / /* II f* A \ t A A \ / f A A\
^ I n ^ f I u . lfllfl\«Al/ffl _ A i
V ^"1 T ^V*tt« * l*rt« j \f\ " rt-t if \f\f\ M-t J
Figure A-'3. Log-linear Interpolation Scheme
A-9
-------
N
Exposure to L^ (person) - £
J 1=1
where
0 , if Cj < L.
1 , if C1 > L.
J
The dosage of the fraction of the population that is exposed to
concentrations greater than or equal to each of a number of concentration
levels, L., was computed by using the following summation:
J
3 N
Annual Dosage at L. (ug/m -person) = £ P^C.S.(C.,L.)
J -j_i ill i j
Note that the annual dosage at the minimum concentration within 50 km of the
source will equal the total annual dosage.
The concentration levels at which exposure is to be estimated are
selected by an exponential function coded in the program.
A.4 UNCERTAINTIES AND ASSUMPTIONS
The HEM estimates concentrations of a substance in the ambient air at
specific points around a source. It is known, however, that pollutant
exposure occurs for exposures other than ambient air exposures (i.e., food,
water) and populations move through different microenvironments. Accord-
ingly, the HEM employs a number of simplifying assumptions including:
1. It is assumed that most exposure occurs at population-weighted
centers (centroids) of block group and enumeration districts
(BG/ED) as the locations of actual residences are not contained in
available databases. The model relies on information provided in
a database developed by the U.S. Census Bureau.
2. It is assumed that people reside at these centroids for their
entire lifetimes (assumed to be 70 years for calculating cancer
risk).
A-10
-------
3. It 1s assumed that indoor concentrations are the same as outdoor
concentrations.
4. It is assumed that plants emit pollutants at the same emission
rate for 70 years.
5. It is assumed that the only source of exposure is the ambient air
and resuspension of pollutants via dust is not considered.
6. It is assumed that there is no population migration or growth.
7. The model does not provide for discriminating exposure situations
that may differ with age, sex, health status, or other situations.
8. The model is designed to model for flat terrain.
There is also uncertainty associated with various specifications
required for running the HEM as well as uncertainty associated with the
variables specified. These model specifications and variables include:
1. Given the variations in meteorology each year and with each
location, there is uncertainty as to the extent to which the
meteorological station used to model the site modeled is
representative. This includes the selection of urban or rural for
stability.
2. There is uncertainty associated with the emission estimates and
the plant parameters used to characterize the emission source.
3. If the plant is not correctly located, there will be a.problem in
the matching of population census data with concentrations, which
alters risk estimates.
Simplifying assumptions are made by the HEM in locating population and
computing exposure.
In the case of locating population, the SHED model uses census data to
the BG/ED level. Although this is the smallest defined population unit, it
still places on average approximately 800 people at one point. In reality,
some people will reside closer to the plant and may be more highly exposed.
For others, the modeled exposure will be an overestimate of the actual
exposure.
A-ll
-------
Another problem oc. . when two plants are located within 50 km of one
another. The people who are exposed to both plants will be counted twice in
the summation of the exposed population. Although the number of people will
be an overestimate, the dosage (people times concentrations) is not affected
by the overlap of the study areas. Two people each exposed to a concentra-
tion x from one of two plants is regarded as the same amount of exposure as
a single person exposed to a concentration of x from both. Populations may
be partially doublecounted if sources are close to each other, but aggregate
exposure is not.
The SHED model may also mislocate people if a plant is :cated ne
large body of water. The manner in which SHED spreads the jpulation
concentrations within the'3.5 km radius allows the program to assign p
to areas which may be covered by bodies of water. One potential problt
arise if the program determines a maximum concentration to which people are
exposed that is located over a lake or other body of water. In this case,
it could be argued that no one is actually exposed to this concentration.
For those sources that appear to produce maximum concentrations,
identification of the receptor location and consultation of topographical
maps to verify public exposure is recommended. If the maximum receptor is
located in an area where people would not reside, the next highest receptor
should be used or that plant should be modeled individually with a
different, more specific model.
The concentration patterns used in the exposure computations are
obtained through atmospheric dispersion modeling based on known source
characteristics and weather patterns at nearby stations. Naturally, any
deviations in these estimates from the true pattern directly affect the
exposure results. Thus, if the nearest STAR station is 50 to 100 miles
away, the weather patterns may not be representative of the weather patterns
in the area around the source.
The program also assumes flat terrain. This is especially crucial when
dealing with sources in a valley/mountainous area because the STAR station
will not give adequate information to model the wind patterns accurately.
In addition, the SHED model uses a Gaussian dispersion model to
determine the concentrations around the source. This model is based on how
gases will act when dispersed from a stack. Thus, if the pollutant is
A-12
-------
fibrous or participate, such as asbestos, the assumption that the pollutant
behaves as a gas may misrepresent the actual dispersion pattern.
The assumption is also made that the annual average weather conditions
continually persist and that the source emits at the annual average amount.
At times, these averages may not be representative over shorter terms. It
also assumes that each recorded wind persists in a given direction long
enough to yield dispersion out to 50 km.
There are also time-dependent aspects of the exposure problem. The
exposure program uses a time-averaged concentration pattern for each source,
so that the time dimension is ignored in the computations. If the
population distribution were essentially constant over the averaging time
period, the resulting estimates would be true averages. However, population
distributions are constantly changing as people commute to work, go
shopping, and take longer tp'ps. Particularly in urban industrial centers,
the shifts in populations and concentrations throughout the day may be
highly correlated; thus, the actual exposure may differ considerably from
the value obtained by matching time-averaged concentrations with population
distributions 'based on census addresses. Whether the exposure is over- or
under-estimated depends on whether populations in the vicinity of a source
are drained (e.g., because people leave residences near the source for work
in an urban center) or are augmented (e.g., because of employment near the
source).
A-13
-------
A.5 REFERENCES FOR APPENDIX A
1. Martin, D. 0., and J. A. Tikvart. A General Atmospheric Diffusion
Model for Estimating the Effects on Air Quality of One or More Sources.
Presented at the 61st Annual Air Pollution Control Association Meeting.
St. Paul, Minnesota. June 1968.
2. Pasquill, F. The Estimation of the Dispersion of Windbprne Material.
Meteorology 90:33-49. 1961.
3. Turner, D. B. A Diffusion Model for an Urban Area, Atmospheric
Dispersion Estimates. 6th Printing Rev., Government Printing Office,
U. S. Environmental Protection Agency, Publication No. AP-26. 1964.
A-14
-------
APPENDIX B
INDUSTRIAL SOURCE COMPLEX SHORT-TERM AIR DISPERSION MODEL
-------
APPENDIX B
INDUSTRIAL SOURCE COMPLEX SHORT-TERM AIR DISPERSION MODEL
B.I INTRODUCTION
The Industrial Source Complex (ISC) Dispersion Model combines and
enhances various dispersion model algorithms into a set of two computer
programs that can be used to assess the air quality impact of emissions from
the wide variety of sources associated with an industrial source complex.
For plumes comprised of. particulates with appreciable, gravitational settling
velocities, the ISC Model accounts for the effects on ambient particulate
concentrations of gravitational settling and-dry deposition. Alternatively,
the ISC Model can be used to calculate dry deposition. The ISC Short-Term
Model (ISCST), an extended version of the Single Source (CRSTER) Model, is
designed to calculate concentration or deposition values for time periods of
1., 2, 3, 4, 6, 8, 12, and 24 hours. If used with a year of sequential
hourly meteorological data, ISCST can also.calculate annual concentration or
deposition values. The ISC Long-Term Model (ISCLT) is a sector-averaged
model that extends and combines, basic features of the Air-Quality Display
Model (AQDM) and the Climatological Dispersion Model (COM). The Long-Term
Model uses statistical wind summaries to calculate seasonal (quarterly)
and/or annual ground-level concentration or deposition values. Both ISCST
and ISCLT use either a polar or a Cartesian receptor grid. The major
features of the ISC Model are listed in Table B-l.
The ISC Model programs accept the following source types: stack, area
and volume. The volume source option is also used to simulate line sources.
The steady-state Gaussian plume equation for a continuous source is used to
calculate ground-level concentrations for stack and volume sources. The
area source equation in the ISCST Model programs is based on the equation
for a continuous and finite cross-wind line source. In the ISCLT Model
program, the area source treatment uses a virtual point source
approximation. The generalized Briggs plume-rise formulas are used to
B-l
-------
TABLE B-l. MAJOR FEATURES OF THE ISC MODEL
Polar or Cartesian coordinate systems.
Rural or one of three urban options.
Plume rise due to momentum and buoyancy as a function
of downwind distance for stack emissions.
Procedures for evaluating building wake effects.
Procedures for evaluating stack-tip downwash.0
Separation of multiple point sources.
Consideration of the effects of gravitational settling
and dry deposition on ambient particulate concentrations.
Capability of simulating .point, line, volume and area sources.
Capability to calculate dry deposition.
Variation with height of wind speed (wind-profile exponent law).
Concentration estimates for 1-hour to annual average.
Terrain-adjustment procedures for elevated terrain including
a terrain truncation algorithm.
Consideration of time-dependent exponential decay of pollutants
Method to account for buoyancy-induced dispersion.
A regulatory default option to set various model options and
parameters to EPA recommended values..
Procedure for calm-wind processing.
References 2, 3, 4, 5 and 6.
References 7 and 8.
d
cReference 5.
Reference 9.
B-2
-------
calculate final as well as gradual plume rise.10'11'12'13'14 Procedures
suggested by Huber and Snyder and Huber are used to evaluate the effects
of the aerodynamic wakes and eddies formed by buildings and other structures
on plume dispersion. A wind-profile exponent law is used to adjust the
observed mean wind speed from the measurement height to the emission height
for the plume rise and concentration calculations. Procedures utilized by
the Single Source (CRSTER) Model are used to account for variations in
terrain height over the receptor grid. Except for Urban Mode 3, the
Pasquill-Gifford curves are used to calculate lateral (o ) and vertical
(o ) plume speed. The ISC Model has one rural and three urban options. In
the Rural Mode, rural mixing heights and the o and QZ values for the
indicated stability category are used in the calculations. In Urban Mode 1,
the stable E and F stability categories are redefined as neutral D
stability. In Urban Mode 2, the E and F stability categories are combined
and the o and o values for the stability category one step more unstable
than the indicated stability category (except A) are used. In Urban Mode 3,
the Briggs urban dispersion coefficients derived from McElroy-Pooler
observations are used. Urban mixing heights are used in all three urban
modes.
For purposes of analyzing deposition- of municipal waste combustor
emissions, the ISC Short-Term Model was used. The ISCST dry deposition
18
model is.based on the Oumbauld et al. deposition model. This model, which
19
is an advanced version of the Cramer et al. deposition model, assumes that
a fraction y of the material that comes into contact with the ground
surface by the combined processes of atmospheric turbulence and
gravitational settling is reflected from the surface. The reflection
coefficient yn, which is a function of settling velocity and the ground
surface for particulates and of the ground surface for gaseous pollutants,
is analogous in purpose to the deposition velocity used in other deposition
models. The Cramer deposition model has closely matched ground-level
deposition patterns for droplets with diameters above about 30 micrometers,
while the more generalized Dumbauld et al. deposition model has closely
matched observed deposition patterns for both large and small droplets.
B-3
-------
The input requirements for the ISCST consist of four categories:
• Meteorological data
t Source data
• Receptor data
0 Program control parameters
The remainder of this appendix describes the input data required by the
ISCST Model and the manner in which the model predicts dry deposition
values. The uncertainties and assumptions in the ISCST Model are also
discussed.
B.2 INPUT DATA
B.2.1 Meteorological Data
Table B-2 gives the hourly meteorological inputs required by the ISC
Model short-term computer program (ISCST). These inputs include the mean
wind speed measured at height z,, the direction toward which the wind is
blowing, the wind-profile exponent, the ambient air temperature, the
Pasquill stability category, the vertical potential temperature gradient and
the mixing layer height. In general, these inputs are developed from
concurrent surface and upper-air meteorological data by the RAMMET
20 21
preprocessor program as used by the Single Source (CRSTER) Model. ' If
the preprocessed meteorological data are used, the user may input, for each
combination of wind-speed and Pasquill stability categories, site-specific
values of the wind-profile exponent and the vertical potential temperature
gradient. If the user does not input site-specific wind-profile exponents
and vertical potential temperature gradients, the ISC Model uses the default
values given in Table B-3.
The ISCST program has a rural and three urban options. IN the Rural
Mode, rural mixing heights and the Pasquill-Gifford (P-G) o and o values
for the indicated stability category are used in the calculations. Urban
mixing heights are used in the urban modes. In Urban Mode 1, the stable E
and F categories are redefined as neutral (D) stability, and the P-G o and
B-4
-------
TABLE B-2. HOURLY METEOROLOGICAL INPUTS REQUIRED BY
THE ISC SHORT-TERM MODEL PROGRAM
Parameter
Definition
Mean wind speed in meters per second (m/sec) at height
z, (default value for z, is 10 meters)
AFVR
Average random flow vector (direction toward which the
wind is blowing)
Wind-profile exponent (default values assigned on the
basis of stability; see Table B-3)
Ambient air temperature in degrees Kelvin ( K)
H
m
Depth of surface mixing layer (meters), developed from
twice-daily mixing height estimates by the meteorological
preprocessor program
Stability
Pasquill stability category (1 =» A, 2 =• B, etc.)
ae
az
Vertical potential temperature gradient in degrees
Kelvin per meter (default values assigned on the
basis of stability category; see Table B-3).
B-5
-------
TABLE B-3. DEFAULT VALUES FOR THE WIND-PROFILE EXPONENTS AND
VERTICAL POTENTIAL TEMPERATURE GRADIENTS
Pasquill Stability
Category
Urban
Wind-Profile
Exponent p
Rural
Wind-Profile
Exponent p
Vertical
Potential
Temperature
Gradient (°K/m)
A
B
C
D
E
F
0.15
0.15
0.20
0.25
0.30
0.30
0.07
0.07
0.10
0.15
0.35
0.55
0.000
0.000
0.000
0.000
0.020
0.035
B-6
-------
o values are used. In Urban Mode 2, the E and F stability categories are
combined and the P-G o and oz values for the stability category one step
more unstable than the indicated category are used in the calculations. For
example, the P-G o and QZ values for C stability are used in calculations
for D stability in Urban Mode 2. In Urban Mode 3, stability categories are
not combined, but urban dispersion curves of Briggs are used. These curves,
22
as reported by Gifford, were derived from the St. Louis Dispersion
23
Study. Table B-4 gives the dispersion coefficients used in each mode.
The Rural Mode is usually selected for industrial source complexes
located in rural areas. However, the urban options may also be considered
in modeling an industrial source complex located in a rural area if the
source complex is large and contains numerous tall buildings and/or large
heat sources (for example, coke ovens). An urban model is appropriate for
these cases in order to account for the enhanced turbulence generated during
stable meteorological conditions by the surface roughness elements and/or
heat sources. If an urban mode is appropriate, Urban Mode 3 is recommended
by EPA for regulatory applications. Modes 1 and 2 are generally not used
but are available to the user for historical interest and model evaluation.
B.2.2 Source Input Data
Table B-5 summarizes the source input data requirements of the ISCST
Model for point sources. Source evaluations above mean sea level and source
locations with respect to a user-specified origin are required for all
sources. If the Universal Transverse Mercator (UTM) coordinate system is
used to define receptor locations, UTM coordinates can only be used to
define source locations if a Cartesian receptor array is used. With a polar
receptor array, the origin is at (X»0, Y=0). The X and Y coordinates of the
other sources with respect to this origin are then obtained from a plant
layout drawn to scale. The x axis is positive to the east and the y axis is
positive to the north. Note that the origin of the polar receptor array is
always at X-0, Y=0.
B-7
-------
TABLE B-4. PASQUILL STABILITY CATEGORIES USED BY THE ISC MODEL TO SELECT
DISPERSION COEFFICIENTS FOR THE RURAL AND URBAN MODES
Pasquill Stability Category for the o ,
Actual Pasquill
Stability Category3
A
B
C
D
E
F
Values Used in
Rural Mode Urban Mode
A A
B B
C C
D D
E D
F D
ISC Model Calculations
I Urban Mode 2 Urban
A
A
B
C
D
D
°z
Mode 3b
A
B
C
D
E
F
aThe ISCST program redefines extremely stable G stability as very stable
F stability.
The Briggs urban dispersion curves combine A and B into one "very unstable"
category, and E and F into one "stable" category.
B-8
-------
TABLE B-5. SOURCE INPUTS REQUIRED BY THE ISCST MODEL PROGRAMS
FOR POINT SOURCES
Stacks Definition
Q Pollutant emission rate for concentration calculations
(mass per unit time)
Qt Total pollutant emissions during the period t for which
deposition is calculated (mass)
* Pollutant decay coefficient (seconds )
X, Y X and Y coordinates of the stack (meters)
Z Elevation of base of stack (meters above mean sea level)
h Stack height (meters)
vs Stack exit velocity (meters per second)
d Stack inner diameter (meters)
TS Stack exit temperature (degrees Kelvin)
Mass fraction of particulates in the n settling-
velocity category
V Gravitational settling velocity for particulates in the n
sett!ing-veloci ty category
yn Surface reflection coefficient for particulates in the
n settling-velocity category
hjj Height of building adjacent to the stack (meters)
W Width of building adjacent to the stack (meters)
L Length of building adjacent to the stack (meters)
B-9
-------
The pollutant emission rate is also required for each source. If the
pollutant is depleted by any mechanism that can be described by time-
dependent exponential decay, the user may enter a decay coefficient^. Note
that if SOg is modeled in the urban mode, and the regulatory default option
is chosen, a decay half life of 4 hours is automatically assigned. The
parameters <|»n, Vsn, and yn are only input if concentration or deposition
calculations are being made for particulates with appreciable gravitational
settling velocities (diameters greater than about 20 micrometers).
Particulate emissions from each source can be divided by the user into a
maximum of 20 gravitational sett!ing-velocity categories. Emission rates
may be held constant or may be varied as follows:
• By hour of the day
• By season or month
• By hour of the day and season
• By wind-speed and stability categories
(applies to fugitive sources of wind-blown dust)
Emission rates used by the long-term model program ISCLT may be annual
average rates or may be varied by season or by wind-speed and stability
categories.
Additional source inputs required for stacks include the physical stack
heights, the stack exit velocity, the stack inner diameter, and the stack
exit temperature. For an area source or a volume source, the dimensions of
the source and the effective emission height are entered in place of these
parameters. If a stack is located on or adjacent to a building and the
stack height to building height ratio is less than 2.5, the length (L) and
width (W),of the building are required as source inputs in order to include
aerodynamic wake effects in the model calculations.
B.2.3 Receptor Data
The ISCST Model computer programs allow the user to select either a
Cartesian (X, Y) or a polar (r, 9) receptor grid system. In the Cartesian
system, the x-axis is positive to the east of a user-specified origin and
B-10
-------
the y-axis is positive to the north. In the polar system, r is the radial
distance measured from the origin (X=0, Y=0) and the angle 0 (asimuth
bearing) is measured clockwise from north. If the industrial source complex
is comprised of multiple sources that are not located at the same point, a
Cartesian coordinate system is usually more convenient than the polar
coordinate system. Additionally, if the Universal Transverse Mercator (UTM)
coordinate system is used to define source locations and/or to extract the
elevations of receptor points from USGS topographic maps, the UTM system can
also be used in the ISC Model, calculations. Discrete (arbitrarily placed)
receptor points corresponding to the locations of air quality monitors,
elevated terrain features, the property boundaries of the industrial source
complex or other points of interest can be used with either coordinate
system.
B.3 DRY DEPOSITION CALCULATIONS
Deposition for particulates in the n sett!ing-velocity category or a
gaseous pollutant with, zero settling velocity V and a reflection
coefficient y is given by:
DEP = K Qt VdD (l-yn) ^ (Z^o^x)'1 exp [-0.5(y/py)2] Equation (B-l)
where:
K = a scaling coefficient to convert calculated concentrations to
desired units (default value of 1 x 10 for Q in g/sec and
concentration in ug/m )
Qt = total amount of pollutant emitted during the time period t
for which the deposition calculation is made
V^ =• Vertical Term
D - Decay Term
yn - reflection coefficient for parti culates in the nth settling-
velocity category
mass fra
category
frn - mass fraction of particulates in the n settling-velocity
B-ll
-------
o , o - standard deviation of later, vertical concentration
distribution (m)
X » downwind distance (m)
Y = crosswind distance (m)
A description of the Vertical Term (Vd), Decay Term (D), reflection
coefficient (yn), and dispersion coefficients (o and o ) follows.
B.3.1 Decay Term
The Decay Term, which is a simple method of accounting for pollutant
removal by physical or chemical processes, is of the form:
D = exp (-tyx/u) forty = 0.
or Equation (B-2)
= 0. for * = 0.
(i.e., decay not considered when zero is input forty).
where:
ty = the decay coefficient (sec )
x = downwind distance (meters)
u = wind speed (m/s)
For example, if T,,- is the pollutant half life in seconds,tyis obtained
from the relationship:
ty = 0.693/T1/2 Equation (B-3)
B.3.2 Dispersion Coefficients
24
Equations that approximately fit the Pasquill-Gifford curves are used
to calculate o (meters) and o (meters) for Urban Modes 1 and 2 and the
rural mode. The equations used to calculate o are of the form:
o = 465.11628 (x) tan(TH) Equation (B-4)
where:
TH = 0.017453293 (c - d In x) Equation (B-5)
B-12
-------
In Equations (B-4) and (B-5) the downwind distance x is in kilometers, and
the coefficients c and d are listed in Table B-6. The equation used to
calculate o is of the form:
o - ax Equation (B-6)
where the downwind distance x is in kilometers and o is in meters in
Equation (B-6) and the coefficients a and b are given in Table B-7.
Tables B-8 and B-9 show the equations used to determine o and QZ for
Urban Mode 3. These expressions were determined by Briggs as reported by
25
Gifford and represent a best fit to urban vertical diffusion data reported
26
by McElroy and Pooler. The Briggs functions are assumed to be valid for
downwind distances less than 100 m. However, concentrations at receptors
less than 100 m from a source may be suspect.
B.3.3 Vertical Term
The Vertical Term includes the effects of source elevation, plume rise,
limited mixing in the vertical., and the gravitational settling and dry
deposition of particulates. The Vertical Term is defined as follows:
Vd = [bH + (1 - bf Hv] exp [-0.5((H - Hv)/oz)2] +
•oo
£ (B,B, + B,B.) Equation (B-7)
i-1 L d J *
where:
y = reflection coefficient
b = average value of exponent b for the interval between the source
and the downwind distance x (see Tables B-6 through B-9)
H - h +Ah
h * stack height (m)
Ah » plume rise (m)
H
v s ..
Vsn * Settlin9 velocity of particulates in the n settling-velocity
category (m/s)
B-13
-------
TABLE B-6. PARAMETERS USED TO CALCULATE PASQUILL-GIFFORD 0,
Pasquill
Stability
Category
A
B
C
D
E
F
o (meters) - 465.11628
TH - 0.017453293 (c
c
24.1670
18.3330
12.5000
8.3330
6.2500
4.1667
(x) tan (TH)
- d In x}
d
2.5334
1.8096
1.0857
0.72382
0.54287
0.36191
aWhere o is in meters and x is in kilometers.
B-14
-------
TABLE B-7- PARAMETERS USED TO CALCULATE PASQUILL-GIFFORD o.
Pasquill
Stability
Category x (km)
Aa <.10
0.10 - 0.15
0.16 - 0.20
0.21 - 0.25
0.26 - 0.30
0.31 - 0.40
0.41 - 0.50
0.51 - 3.11
>3.11
Ba <.20
0.21 - 0.40
>0.40
Ca All
D <.30
0.31 - 1.00
1.01 - 3.00
3.01 - 10.00
10.01 - 30.00
>30.00
E <.10
0.10 - 0.30
0.31 - 1.00
1.01 - 2.00
2.01 - 4.00
4.01 - 10.00
10.01 - 20.00
20.01 - 40.00
>40.00
F <.20
0.21 - 0.70
0.71 - 1.00
1.00 - 2.00
2.01 - 3.00
3.01 - 7.00
7.01 - 15.00
15.01 - 30.00
30.01 - 60.00
>60.00
hlf the calculated value of o, exceeds
"«» ± ~. _ _.. _ i i _ ^^^*/*_^ Z
o, (meters)
£
a
122.800
158.080
170.220
179.520
217.410
258.890
346.750
453.850
b
90.673
98.483
109.300
61.141
34.459
32.093
32.093
33.504
36.650
44.053
24.260
23.331
21.628
21.628
22.534
24.703
26.970
35.420
47.618
15.209
14.457
13.953
13.953
14.823
16.187
17:836
22.651
27.074
34.219
5000 m, o is set to 5000
-axb
b
0.94470
1.05420
1.09320
1.12620
1.26440
1.40940
1.72830
2.11660
b
0.93198
0.98332
1.09710
0.91465
0.86974
0.81066
0.64403
0.60486
0.56589
0.51179
0.83660
0.81956
0.75660
0.63077
0.57154
0.50527
0.46713
0.37615
0.29592
0.81558
0.78407
0.68465
0.63227
0.54503
0.46490
0.41507
0.32681
0.27436
0.21716
m.
B-15
-------
TABLE B-8. BRIGGS FORMULAS USED TO CALCULATE McELROY-POOLER oy
Pasquill
Stability
Category o (meters)
A 0.32 x (1.0 + 0.0004 x)"1/2
B 0.32 x (1.0 + 0.0004 x)"1/2
C 0.22 x (1.0 + 0.0004 x)"1/2
D 0.16 x (1.0 + 0.0004 x)'1/2
E 0.11 x (1.0 + 0.0004.x)"1/2
F 0.11 x (1.0 + 0.0004 x)"1/2
aWhere x is in meters.
B-16
-------
TABLE B-9. BRIGGS FORMULAS USED TO CALCULATE McELROY-POOLER PZ
Pasquill
Stability
Category p (meters)
A 0,24 x (1.0 + 0.001 x)"1/2
B 0.24 x (1.0 + 0.001 x)"1/2
C 0.20 x
D 0.14 x (1.0 + 0.0003 x)"1/2
E 0.08 x (1.0 + 0.0015 x)"1/2
F 0.08 x (1.0 +0.0015 x)"1/2
aWhere x is in meters.
B-17
-------
u = wind speed (m/s)
x =• downwind distance (m)
Bj - y1"1 [bHj - (1-b) Hy]
B2 . exp [-O.StHj + Hv)/oz)2]
B3 - y1 [b H2 + (1-b) Hv]
B4.- exp [-0.5 (H2 - Hy)/oz)2]
Hl - 21Hm ' H
H2 . 2iHm + H
Hm - mixing height (m)
B.3.4 Reflection Coefficient
To calculate the Vertical Term, information about the particle size
distribution and the density of the particulates emitted must be known.
Total particulate emissions are subdivided into categories and the
gravitational settling velocity is calculated for the mass-mean diameter of
each category. The gravitational settling velocity for all sizes of
particulates is calculated according to the following equation by
McDonald:27
YS = 2pgr2/9u Equation (B-8)
V = settling velocity (cm • sec )
5 _3
=» particle density (gm • cm )
g » acceleration due to gravity (980 cm • sec )
where:
I
p =» particle density (gm • cm"J)
g =• acceleration due to
r » particle radius (cm)
-4 -1 -1
u - absolute viscosity of air (u - 1.83 x 10 gm • cm • sec )
It should be noted that the settling velocity calculated using '
Equation (B-8) must be converted by the user from centimeters per second to
meters per second for use in the model calculations.
The reflection coefficient y can be estimated for each particle-size
category using Figure B-l and the settling velocity calculated for the
mass-mean diameter. If it is desired to include the effects of
B-18
-------
Q25
2 0.20
>S
U
3
UJ
C/J
O.3U |^^^^™r"^^™^"™T™™T"""1^""™T~™T™"T"""""T™™r"^T>"'™T™""<^"^m
0.15
0.10
0.05
I I
L I I I I » I I I I II
0 0:2 0.4 0.6 0.8
REFLECTION COEFFICIENT rn
Figure B-l. Relationship between the Gravitational Settling Velocity V
and the Reflection Coefficient yn suggested by Dumbauld,
et al.28 n
1.0
sn
B-19
-------
gravitational settling in calculating ambient particulate concentrations
while at the same time excluding the effects of deposition, y should be set
equal to unity for all settling velocities. On the other hand, if it is
desired to calculate maximum possible deposition, yp should be set equal to
zero for all settling velocities. The effects of dry deposition for gaseous
pollutants may be estimated by setting the settling velocity V$n equal to
zero and the reflection coefficient yn equal to the amount of material
assumed to be reflected from the surface. For example, if 20 percent of a
gaseous pollutant that reaches the surface is assumed to be retained at the
surface by vegetation uptake or other mechanisms, yn is equal to 0.8.
B-20
-------
B.4 REFERENCES FOR APPENDIX B
1. U. S. Environmental Protection Agency. User's Manual for Single Source
(CRSTER) Model. EPA-450/2-77-013. 1977-
2. Briggs, G. A. Plume Rise. USAEC Critical Review Series. NTIS
Publication No. TID-25075. 1979.
3. Briggs, G. A. Some Recent Analyses of Plume Rise Observations.
Proceedings of the Second International Clean Air Congress. 1971.
4. Briggs, G. A. Discussion on Chimney Plumes in Neutral and Stable
Surroundings. Atmos. Environ. 6:507-510. 1972.
5. Briggs, G. A. Diffusion Estimates for Small Emissions. USAEC Report
ATDL-106. 1974.
6. Briggs, G. A. Plume Rise Predictions. In: Lectures on Air Pollution
and Environmental Impact Analyses. American Meteorology Society.
1975.
7. Huber, A. H., and W. H. Snyder. Building Wake Effects on Short Stack
Effluents. Preprint Volume for the Third Symposium on Atmospheric
Diffusion and Air Quality. American Meteorological Society. 1976.
8. Huber, A. H. Incorporating Building/Terrain Wake Effects on Stack
Effluents. Preprint Volume for the Joint Conference on Applications of
'Air Pollution Meteorology. American Meteorology Society. 1976.
9. Pasquill, F. Atmospheric Dispersion Parameters in Gaussian Plume
Modeling. Part II. Possible Requirements for Change in the Turner
Workbook Values. EPA-600/4-76-030b. 1976.
10. Reference 2.
11. Reference 3.
12. Reference 4.
j3 Reference 5.
14. Reference 6.
15. Reference 7.
16. Reference 8.
17. Turner, D. B. Workbook of Atmospheric Dispersion Estimates. PHS
Publication No. 999-AP-26. U. S. Department of Health, Education and
Welfare, National Air Pollution Control Administration. 1970.
B-21
-------
18. Dumbauld, R. K., J. E. Rafferty, and H. E. Cramer. Dispersion -
Deposition from Aerial Spray Releases. Preprint Volume for the Third
Symposium on Atmospheric Diffusion and Air Quality. American
Meteorological Society. 1976.
19. Cramer, H. E., et al. Development of Dosage Models and Concepts.
Final Report. U. S. Army, Desert Center Report DTC-TR-609. 1972.
20. Reference 1.
21. Catalano, J. A. Single-Source (CRSTER) Model, Addendum to the User's
Manual. U. S. Environmental Protection Agency. 1986.
22. Gifford, F. A. Turbulent Diffusion - Typing Schemes: A Review.
Nucl. Saf. 17:68-86. 1976.
23. McElroy, J. L., and F. Pooler. The St. Louis Dispersion Study.
U. S. Public Health Service - National Air Pollution Control
Administration Report AP-53. 1968.
24. Reference 17.
25. Reference 22.
26. Reference 23.
27. McDonald, J. E. An Aid to Computation of Terminal Fall Velocities of
Spheres. J. Met. 17:463. 1960.
28. Reference 18.
B-22
-------
APPENDIX C
WET DEPOSITION MODEL
-------
APPENDIX C
WET DEPOSITION MODEL
This appendix presents the approach developed by EPA to model the
effects of wet deposition on municipal waste combustor emissions. The
algorithm developed to predict wet deposition was combined with the existing
ISCST dry deposition model to predict total surface deposition of municipal
waste combustor emissions.
The following equation was used by EPA to model wet deposition:
j) *nKQ
"0eP
-------
TABLE C-l. SUMMARY OF SCAVENGING COEFFICIENTS EXPRESSED PER SECOND OF TIME,
USED IN COMPUTING WET SURFACE DEPOSITION
Precipitation
Intensity
Heavy
Moderate
Light
Particle
<2
1.46 E-3
5.60 E-3
2.20 E-4
Size Cateaorv
>2 and <10
4.64 E-3
8.93 E-4
1.80 E-4
(microns)
>10
9.69 E-3
9.69 E-3
9.69 E-3
Light = less than 0.10 inches per hour.
Moderate = 0.11 to 0.30 inches per hour.
Heavy = equal to .or greater than 0.31 inches per hour.
C-2
-------
squalls; a factor of 0.5 was assigned to represent a 30-minute duration of
precipitation for hours experiencing showers, and a value of 1.0 (60-minute
duration) was assigned for hours with steady precipitation. In computing
the dry deposition which occurs between the period of precipitation, a
factor of (1-f) was used to determine the fraction of the material emitted
during the hour that is subject to dry deposition. Thus, no dry deposition
occurs during hours of steady precipitation.
Few published reports with data on the scavenging coefficients for snow
could be found. Graedel and Franey have computed scavenging coefficients of.
0.5 micron size aerosol particles in the range of 18 to 28 x 10" per
second. Wolf and Dana report an intensity dependent rate of 0.5 x 10" per
second times -the precipitation intensity for 0.5 micron particles, and have
assumed this to define the lower bound coefficients for light and moderate
4
snowfall. However, it is not clear from the literature what the scavenging
coefficients for snowfall would be for the particle size categories used in
this study. Thus, in the wet deposition calculations for this analysis, EPA
has used the same washout coefficients for both rainfall and snowfall
events. Other frozen forms of precipitation, e.g., snow pellets, ice
pellets, and hail, present a relatively smooth.surface to the aerosol
particles and are not likely to be effective scavengers of the pollutants.
Therefore, periods and occurrences of these types of frozen precipitation
are modeled in the dry deposition mode.
The principal assumptions made in computing wet deposition are:
(1) the intensity of precipitation is constant over the entire path between
the source and receptor; (2) the precipitation originates at a level above
the top of the emission plume so that the hydrometers pass vertically through
the entire plume, (3) the time duration of the precipitation over the entire
path between the source and receptor point is such that exactly f, a
faction between zero and one, of the hourly emission Q in Equation (C-l) is
subject to a constant intensity for the entire travel time required to
traverse the distance between the source and receptor, and the remaining
fraction (1-f) is subject only to dry deposition processes.
C-3
-------
C.I REFERENCES FOR APPENDIX C
1. U. S. Environmental Protection Agency. Methodology for Assessment of
Health Risks Associated with Multiple Pathway Exposure with Multiple
Pathway Exposure to Municipal Waste Combustor Emissions. Submitted for
Review to the Science Advisory Board. October 1986.
2. Radke, L. F., P. V. Hobbs, and M. W. Eltgroth. Scavenging of Aerosol
Particles by Precipitation. Journal of Applied Meteorology 19:715-722.
1980.
3. Graedel, T. E., and J. P. Franey. Field Measurements of Sub-Micron
Aerosol Washout by Rain. Proceedings of the Symposium on Precipitation
Scavenging. ERDA Symposia Series 41:503-523. 1977.
4. Wolf, M. A., and M. T. Dana. Experimental Studies on Precipitation
Scavenging. Battelle Northwest annual report. USAEC Report BNWL-1051
' (Part 1), pp. 18-25. 1969.
C-4
-------
APPENDIX D
TERRESTRIAL FOOD CHAIN MODEL
-------
APPENDIX D
TERRESTRIAL FOOD CHAIN MODEL
D.I INTRODUCTION
Contaminants associated with emissions from municipal waste combustors
are subject to deposition on surfaces downwind from the combustor. The
fallout may be deposited on soil and/or vegetation. The Terrestrial Food
Chain Model (TFC) applies to both the human food chain and the ecological
food chain. The pathways in the model assess exposure to humans, animals,
soil biota and vegetation. Terrestrial trophic relationships are numerous
and complex. Humans in the vicin-ity of municipal waste combustors have the
potential to ingest contaminated soil directly or consume vegetation and
animal tissues containing the contaminants. Herbivorous animals, plants,
and soil organisms living where municipal waste combustor emissions are
deposited may also be exposed. Humans or other organisms may be exposed to
the soil-deposited contaminants of municipal waste combustor emissions by
several pathways. For each exposure pathway, it is important to identify
the most-exposed individual organism, or MEI. The MEI is a hypothetical
individual or organism, but care should be taken that the definition is
realistic. The definition of the MEI will vary with each pathway.
In the case of humans, occupational exposures are not considered. It
is assumed that workers involved in the operation of municipal waste
combustors can be required to use special measures or equipment to minimize
their exposure to possibly hazardous materials. This methodology is geared
toward protection of the general public and the environment. While many
individuals of the general public may be exposed to a varying degree, the
MEI is that individual who would be expected to experience the greatest risk
and, therefore, require the greatest protection. For the pathways involving
ecological effects, the MEI is a sensitive plant or animal rather than a
human.
D-l
-------
The remainder of this appendix describes the MEI assumed for each
exposure pathway, the approach for estimating soil concentration and
contaminant uptake In the organisms of Interest, and the assumptions,
calculation steps, and input requirements for estimating exposure for each
pathway. Example calculations of exposure estimates are also given for each
pathway.
D.2 MOST EXPOSED INDIVIDUAL (MEI) ASSUMED FOR EACH EXPOSURE PATHWAY
D.2.1 Crops for Human Consumption
This pathway (deposition-soil-piant-human toxicity) is important
wherever crops for human consumption are grown in a vicinity where emissions
from a municipal waste combustor may be deposited. Uptake of the deposited
contaminants is assumed to occur through the plant roots. Direct adherence
of deposited contaminants or soil to crop surfaces is not considered here.
(Direct ingestion of deposited contaminants is discussed in Section D.2.2.)
The MEI is defined as an individual residing in a region within 50 km of a
municipal waste combustor in the area of maximal deposition of emissions.
The individual who grows a large proportion of his or her own food would
result in highest risk since much of the diet would be potentially affected.
D.2.2 Soil Inqestion bv Children
Human adults may ingest some soil, but the amounts consumed by young
(i.e., preschool, 1-6 years of age) children are much greater. This is
especially true in children exhibiting the behavior known as "pica," the
ingestion of nonfood items; these children constitute the MEI for this
pathway (deposition-soil-human toxicity). Preschool children with pica for
soil are assumed to be exposed in residential areas within 50 km of the
municipal waste combustor in the areas of maximal deposition of emissions.
The exposure is likely to occur in gardens, lawns, landscaped areas, parks
and recreational areas.
D-2
-------
D.2.3 Herbivorous Animals for Human Consumption
Two separate pathways are considered whereby animal products may become
contaminated: 1) deposition-soil-plant-animal-human toxicity and
2) deposition-soil-animal (direct ingestion)-human toxicity. By the first
pathway, row crops (i.e., grains) or other forage crops (i.e., grasses) are
grown on soils contaminated by municipal waste combustor emissions and take
up contaminants through the roots. The crops are then harvested for animal
consumption. By the second pathway, the deposited contaminants adhere to
crop surfaces or remain in the thatch layer on the soil surface. The crop
is then harvested or grazed, resulting in ingestion of deposited particles
or pollutants. In addition to domestic grazers, wild herbivores, such as
deer, may forage grains or grasses within the range of emissions fallout and
be taken by hunters. The MET for this pathway is the human consumer of
these animal products.
D.2.4 Toxicitv to Herbivorous Animals
The pathways for exposure for herbivorous animals are 1) deposition-
soil-plant-animal toxicity, and 2) deposition-soil-animal toxicity (direct
ingestion). For these pathways, it. does not matter whether the animals are
subsequently consumed by humans. The end point of concern is toxicity to
the .animals themselves, which constitute the MEI. It is assumed that
wildlife may forage on lawns, gardens, agricultural areas or forests within
the region of maximal deposition of emissions of the MWC.
D.2.5 Phvtotoxicitv
This pathway is described as deposition-soil-plant toxicity. Toxic
effects in plants are of concern since plants have an integral role in the
terrestrial food chain. The MEI, or vegetation type to be protected, will
ordinarily be the most sensitive plan species for which data are available.
D-3
-------
D.2.6 Toxicitv to Soil Biota or Their Predators
Two pathways are considered here: 1) deposition-soil-soil biota
toxicity, and 2) deposition-soil-soil biota-predator toxicity. The term
"soil biota" is intended to be interpreted broadly. The first pathway
examines effects on a broad range of organisms including microorganisms,
soil invertebrates such as earthworms, and various anthropods living in or
near the soil, as long as potential effects in these organisms can be
related to soil concentrations. The second pathway examines exposures to
predators of these organisms, especially small mammals and birds. These
predators could include insectivores as long as available data permit the
contaminant concentration in the prey to be related to contaminant soil
concentration.
D.3 ESTIMATION OF SOIL CONCENTRATION AND CONTAMINANT UPTAKE BY ORGANISMS
D.3.1 Soil Deposition Rate of Contaminants
The cumulative soil deposition rate of contaminants (in kg/ha) is
determined from the total (dry plus wet) deposition rate of the pollutant
2 1
[g/m (year)] over the total period of deposition from the municipal
combustor by the following equation:
CD•- AD x T x 10 Equation (D-l)
-.
where:
CD = cumulative soil deposition of pollutant (kg/ha)
2 -1
AD = annual deposition rate of pollutant [g/m (year )]
T = total period of deposition (years')
2
10 » conversion factor [m x kg/(ha x g)]
For this methodology, the lifetime of the municipal waste combustor (or
total period of deposition) is considered to be >30 years; however, since
the site is already dedicated for municipal waste combustion, it may be
assumed that the combustor will be replaced. Therefore, the lifetime of the
municipal waste combustor could be as great as 100 years.
D-4
-------
D.3.2 Soil Incorporation of Deposited Contaminants
i
Following deposition, contaminants from emissions of municipal waste
combustors may be incorporated into the upper layer of soil where crops or
other vegetation are grown. If incorporation is accomplished by disking or
plowing of the upper layer of soil, it is assumed that the deposited
pollutants would be mixed into the soil to a depth of -20 cm (8 inches).
If the contaminants persist indefinitely in the upper soil layer, as is
the case for some inorganics, the following relationship exists between
contaminant deposition rates and concentration increment in soil:
LC = (CD x 10) x (B x D)"1 Equation (D-2)
where:
LC = maximal soil concentration increment of pollutant (ug/g DW)
CD = cumulative soil deposition of pollutant (kg/ha)
2 22
10 = conversion factor [(m x ha x ug)/(cm x m x kg)]
D - depth of soil layer (cm).
2
B » bulk density of soil (g/cm )
From Equation (D-2), a soil concentration of 1 ug/g DW corresponds to a
deposition rate.of 2.7 kg/ha. LC represents the concentration increment,
not the total concentration, because it does not take into account
background concentrations of the contaminant that may already be present,
whether natural or from other pollution sources.
Where soil incorporation does not occur, particulate is assumed to be
retained in a shallower, uppermost soil layer. While the actual depth of
this uppermost layer retaining the unincorporated contaminant, is unknown, a
value of 1 cm will be assumed.
D.3.3 Contaminant Loss from Soils
Contaminants may be lost from soils as a result of numerous processes,
including leaching, volatilization and chemical and biological degradation.
These processes may occur simultaneously or at different rates.
D-5
-------
Organic contaminants and ..organic contaminants may be subject to
some or all of these loss processes; thus, it may be extremely difficult to
model overall rate of loss. A simple means to estimate loss is based on
empirical data from soil systems where soil concentrations have been
followed over time. These data may be used to estimate a first-order loss
rate constant for the pollutant. The use of such a rate constant is
recognized to be an over-simplification since the processes involved are
complex and not necessarily or only first-order. Where no basis for an
estimate is available, no loss should be assumed. The maximal soil
concentration of chemicals subject to loss (for all k >0) may be calculated
as a function of the annual deposition rate constant as shown in the equation
below:
where:
LCT = AD x (l-e~kT) x 102 x (B x D x k)"1 Equation (D-3)
LCT = maximal soil concentration of pollutant after time, T (ug/g DW)
2
AD = annual deposition of contaminant [g/(m x year)]
MS » 2.7 x 10 Mg/ha - assumed mass of soil in upper 20 cm
k - loss rate constant (years)"
T = total period of deposition (years)
2 22
10 =• conversion factor [(m x mg)/(cm x g)]
2
B = bulk of density of soil (g/cm )
D = depth of soil layer (cm)
This formula is derived from an environmental application of
toxicokinetic principles.
D.3.4 Contaminant Uptake Relationship in Plants
0.3.4.1 Plant Uptake of Inorganics. Uptake rates of inorganic
chemicals, especially uptake of metals by plants, have been recently
234
reviewed. ' Ryan et al. used linear regression (of plant tissue Cd
concentration against applied Cd) to derive uptake response slopes; i.e.,
the increase in tissue concentration for various crops. [The term "uptake
0-6
-------
response" 1s used to denote an increase in tissue concentration in response
to exposure to a chemical; i.e., the difference in pre- and post-exposure
concentrations.] These authors stated that although the uptake slope could
be altered by several variables, it remained essentially linear.
More recent work has shown that plant response to metals from sludge-
amended soil is curvilinear, approaching a plateau concentration in tissue
as sludge application rate increases; however, metal-adsorptive materials
present in the sludge matrix are thought to be responsible for this effect.
Since no such uptake limiting effect has been demonstrated for deposited
metals, uptake response slopes will be assumed to be linear for this
methodology. The assumption that response slopes are linear means that
dietary intake of a contaminant increases continually with contaminant
application (or deposition) to soil (if some or all of the diet originates
from these soils). Ryan et al. further assumed that response was related to
cumulative Cd application. Using this approach, a limit can be derived for
the cumulative application of a metal based on its dietary threshold level
in humans.
Linear response slopes can be calculated from any data set where tissue
analyses and cumulative metal deposition or application rates have been
recorded, assuming that the metal is not significantly lost over time.
Plant tissue contaminant concentration (in ug/g DW) is regressed against
cumulative contaminant application or deposition (in kg/ha) for the various
treatment levels, including the control, to calculate the uptake response
slope. If contaminant concentrations in soil, LC, were measured rather than
deposition or application rates, the cumulative deposition, CD, may be
calculated based on Equation (D-2). For inorganics that are lost over time,
however, tissue concentration should be regressed against LC.
Wherever possible, uptake response data derived from areas of emission
fallout will be used. For contaminants lacking such data, linear uptake
slopes derived from other types of chemical application (such as sludge or
pesticide additions) will be assumed to apply to deposited contaminants as
well.
If the contaminant is phytotoxic, a maximum tissue concentration for a
given crop will be determined based on available phytotoxicity data, and
assumed as an upper limit to uptake. Phytotoxicity of metals may be altered
D-7
-------
by soil pH. Phytotoxicity data chosen should be appropriate for the soil pH
of the fallout region of the municipal waste combustor if possible. Maximum
concentrations are those associated with severe yield reduction (>75%) or
death of the plant, which would preclude pollutant passage up the foodchain.
D.3.4.2 Plant Uptake of Orqanics. Linear uptake response is also
assumed for organic chemicals and is calculated as described for inorganics,
but with some important differences. Because organic compounds and some
inorganics tend to degrade in soil, plant tissue concentration is usually
expressed as a function of a measured soil concentration, rather than
application or deposition rate, for chemicals subject to loss. Therefore,
the soil concentration (in ug/g DW) is regressed against tissue
concentration to determine the uptake slope.
In addition, because soil concentration rather than deposition rate is
used, and because most of the compounds of concern are xenobiotics, tissue
concentration can be assumed to be zero when soil concentration is zero.
Therefore, the slope reduces to a biocancentration factor that can be
derived from a single data pair.
D.3.5 Contaminant Uptake bv Animal Tissues
Linear response slopes are derived for uptake of inorganics or organics
by animal tissues consumed by humans. Tissue concentration is regressed
against concentration in feed. Tissue concentrations in the literature may
be expressed in dry or wet weight, but dry weight is preferred. For
uniformity in applying this methodology, all slopes should be derived based
on dry weight (moisture free, but including fat) concentrations in tissue
and feed. Conversion from wet to dry weight for various tissues should be
made according to percent moisture values given in USDA (1975) or another
authoritative source.
For lipophilic organics, tissue concentration is often expressed on a
fat basis (ug/g fat). If so, the uptake slope should be expressed on a fat
basis rather than converted to a dry weight basis. Also, the slope for
D-8
-------
organics may be the same as a bioconcentration factor derived for a single
data point3(i.e., animal tissue concentration and feed concentration), as
described previously for plant uptake of organics (see Section D.3.4.2).
D.3.6 Human Diet
Humans may be exposed to crops or animal products that have taken up
the pollutants through the soil or diet, respectively. In order to quantify
potential dietary exposures, it is necessary to estimate the amounts of
various types of foods in the human diet. The most up-to-date and detailed
source of information regarding food consumption habits of the United States
8
population is the FDA Revised Total Diet Study food list. While this food
list provides a very detailed picture of the United States diet, it cannot
be used in its published form for risk assessments of the present type.
Many of the food items listed are complex prepared foods (such as soup or
pizza), rather than the raw commodities (such as crops or meats) for which
contaminant uptake data are available. Therefore, the reanalysis of the
g
Pennington diet used, recently by EPA will also be used in this methodology.
Each item in the Pennington diet (including the infant/junior foods) was
broken down into its component parts, based on information available from
FDA and USDA. ' The percentages of dry matter and fat for each component
were also specified. These components were then aggregated into the
specific commodity groups required for this methodology. A summary of
consumption for each category by each age/sex group is presented in
Table D-l. This analysis should be considered preliminary as it has not
been reviewed by the FDA.
D.4 ASSUMPTIONS AND CALCULATION STEPS FOR ESTIMATING EXPOSURE
D.4.1 Deposition-Soil-PIant-Human Toxicitv Exposure Pathway
D.4.1.1. Assumptions. In addition to the assumptions listed in
Table D-2, assumptions on the percent of diet affected by deposition of
contaminants from municipal waste combustor emissions are made for this
pathway. These assumptions and their potential-.limitations are summarized
in Table D-3.
D-9
-------
TABLE 0-1. AVERAGE DAILY DRY-WEIGHT CONSUMPTION OF FOOD GROUPS, BASED
ON A REANALYSIS OF THE FDA REVISED TOTAL DIET FOOD LIST
I
•—»
o
Food Group
subgroup
Grains and cereals
wheat
corn
rice
oats
other grain
Potatoes
Leafy vegetables
Legume vegetables
Root Vegetables
Garden fruits
Peanuts
.Mushrooms
Vegetable oil
Heats
beef
beef fat
beef liver
beef liver fat
lamb
lamb fat
pork
pork fat
poultry
poultry fat
fish (including fat)
Dairy
dairy fat
Eggs
Other
Cons tiinp lion by
6-11 Months
42.969
6.239
3.003
6.362
0.006
8.391
0.638
2.475
0.893
0.900 .
0.243
0.000
30. 780
3.006
11.885
0.077
0.1012
0.0570
0.127
1.412
4.869
2.253
5.925
0.339
41.021
39.196
3.395
130 992
2 Years
61.820
16.938
4.586
3.758
0.015
13.701
0.482
4.683
0.736
2.001
1.661
0.001
27.119
8.475
8.974
0.109
0.146
0.0314
0.070
4.994
8.264
4.515
1.412
1.200
31.854
16.252
6.788
253.974
14-16
Females
81.980
20. 148
5. 137
1.779
0.018
21.505
1.141
6.459
1.422
3.766
1.374
0.002
41.938
15.322
14.783
0.127
0.168
0.0253
0.056
7.892
10.556
6.997
1.602
2.540
32.088
18.354
5.998
407.624
Age-Sex Group (g dry weight/day)
Years
Males
124.475
26. 358
6.638
4.788
4.197
31.854
1.306
10.651
2.278
4.448
3.162
0.002
63.747
24.640
23. 702
0.200
0.265
0.0180
0.040
10.942
15.698
8.258
1.809
2.676
49.858
29.677
8.103
527.718
25-30
Females
69. 534
14.591
4.776
1.337
9.007
17.481
2.176
7.987
1.491
4.446
1.311
0.003
44.530
15.819
15. 167
0.565
0.748
0.1358
0.302
7.771
10. 590
6.607
1.454
3.501
21.788
14.545
6.286
379.042
Years
Hales
102.207
22.439
6.783
2.044
77.912
28.289
2.164
11.891
2.126
5.943
2.840
0.004
75.746
25.050
27.952
0.423
0.5596
0. 1064
0.237
13.846
20.046
9.681
2.061
4.670
31.772
21.991
9.197
482.937
60-65
Females
65.114
15.502
4.024
2.044
2.419
15.914
2.783
8.454
1.529
4.797
1.115
0.002
37.029
12.995
12. 175
0.443
0.5854
0.0864
0.192
8.036
10,554
6.736
1.464
3.629
19. 108
12. 145
7.000
198. 307
years
Hales
89. 525
20.347
4.396
2.293
25.644
23.837
2.526
11.098
1.866
5.442
2.137
0.002
55.621
20.558
18.619
0.662
0.8762
0.0873
0.194
12.566
16.582
8.169
1.824
4.108
25.501
16.226
10.468
237.888
Source: References 9 and 10.
-------
TABLE D-2. ASSUMPTIONS FOR TERRESTRIAL FOOD CHAIN
Functional Area
Assumptions
Ratifications/Limitations
Soil incorporation of
contaminants
Contaminant loss fro* soils
Plant uptake of inorganics
Animal uptake of contami-
nants
Human diet
If soil incorporation is assumed,
incorporation depth is 20 cm, and the
upper 20 cm soil layer has a dry mass
of 2.7 x 103 Mg/ha.
Soil background concentration is not
considered.
Trace metal contaminants are assumed to
be conserved indefinitely in the upper
layer unless loss constants are avail-
able.
Degradation of organic contaminants is
first-order.
Uptake response slope may vary with pH.
Plant uptake is treated as linear with
application rate until highly toxic
concentrations in tissue are reached.
Plant uptake is treated as linear with
soil concentration until highly toxic
concentrations in tissue are reached.
Animal tissue concentration is treated as
a linear function of feed concentration.
Pure chemicals added to diet can be used
to determine uptake response slope.
The FDA Revised Total Diet Study food
list (Pennington, 1963) is representative
of the United States diet. The age/sex
group with the highest consumption of a
given crop group (typically the 25- to
30-year-old male) is the HE1 for that item.
By this assumption, as soil concentration of 1 ug/g corresponds
to a pollutant application of 2.7 kg/ha. If actual depth (and
mass) is less, impacts on soil biota in the incorporation layer
could be underpredicted (and vice versa). Ramifications for
effects on plants is less clear, since less of the root zone is
contaminated as incorporation depth decreases.
Effects could be underpredicted if background concentration is
not considered.
Although most heavy metals are tightly bound to soil, measure-
ments tend to show that this assumption often overpredicts
concentrations and therefore, probably overpredicts certain
hazards.'
Could over- or underpredict degradation rate, which is complex
and not necessarily first-order.
Uptake response slope determined for one region may over- or
underestimate uptake response slope for a region with different
soil pH.
Nay oyerpredict tissue response to contaminant application if
relationship is truly curvilinear.
if
May overpredict tissue response to contaminant application i
relationship is truly curvilinear.
Could under- or overpredict tissue concentration outside the
observed range of response.
Availability of chemicals in soil may differ; especially, may be
lower, so that uptake is overpredicted.
While very complete and detailed, the Pennington diet provides no
information on variability within age/sex groups.
-------
TABLE D-3. ASSUMPTIONS FOR DEPOSITION-SOIL-PLANT-HUMAN
TOXICITY EXPOSURE PATHWAY
Functional Area
Assumptions
Ramifications/Limitations
Fraction of diet affec-
ted by deposition of
MWC emissions.
All of an individual's
homegrown food could
come from soil with
deposited emissions.
The percentage of
homegrown food in the
diet of the MEI can
be estimated from USDA
survey data on rural
farm households, which
constituted 6 percent
of all households.
May overpredict exposure.
More recent information
if available, might show
significant changes in
both demographics and
gardening habits of
these households.
Source: Reference 12.
D-12
-------
This exposure pathway deals with crops for human consumption. It will
be assumed that home gardeners produce and consume leafy, legume and root
vegetables, potatoes and garden fruits but not grains and cereals, peanuts
or mushrooms. The USDA survey of United States food consumption in
1965-1966 includes data on the percentages of foods consumed that were
12
homegrown for .urban, rural nonfarm and rural farm households. The highest
percentages of homegrown foods were for rural farm households, which
constituted -6 percent of ..all United States households. The rural farm
dweller in the area of maximal deposition within 50 km of a municipal waste
combustor will be taken as the MEI in this pathway. All of the homegrown
food by the M£I could come from soil contaminated with deposited emissions
from the municipal waste combustor. The average percent of annual
consumption that is homegrown for various foods from rural farm households,
is shown in Table D-4.
D.4.1.2 Calculation Method. Uptake response slopes, in units of
ug/g DW (kg/ha) for most inorganics or ug/g DW (ug/g)" for organics or
chemicals subject to loss., are used to determine dietary response to the
deposited contaminant. The main disadvantage of using slopes is that the
slopes for each crop wi-11 likely originate from different experimental
conditions. In order to examine total dietary response to a change in
conditions (such as soil pH), all of the response slopes may need to be
changed. Lack of adequate uptake data for a chemical would preclude this
calculation.
The following steps are taken to calculate daily intakes of
contaminants by this pathway.
Step A. Sort Available Uptake Response Data for All Food Crops
For chemicals showing increased uptake at lower pH (i.e., many metals),
the available response data for the crop should be grouped according to
whether soil pH was <6.0 or >6.0. Studies with pH >6.0 should be used only
if natural soils in the vicinity of the municipal waste combustor have a
neural or alkaline pH.
D-13
-------
TABLE D-4. AVERAGE PERCENT OF ANNUAL CONSUMPTION THAT IS HOMEGROWN
FOR VARIOUS FOODS, RURAL FARMS HOUSEHOLDS
Food Group Percent Homegrown
Milk, cream, cheese 39.9
Fats, oil 15.2
Flour, cereal 1.6
Meat 44.2
Poultry, fish 34.3
Eggs 47.9
Sugar, sweets 9.0
Potatoes, sweet potatoes 44.8
Vegetables (fresh, canned, frozen) 59.6
Fruit (fresh, canned, frozen) 28.6
Juice (vegetable, fruit) 11.0
Dried vegetables, fruits 16.7
Source: Reference 12.
D-14
-------
Step B. Determine Uptake Response Slopes for Each Food Group
Response slopes with units of ug/g DW (kg/ha) or ug/g DW (ug/g) , as
appropriate, are determined for each crop food group shown in Table D-l.
This slope may be determined as a weighted mean of all the available response
slopes where weighting is according to the dry weight consumption of each
crop. If the data do not permit determination of a weighted mean, an
unweighted mean may be taken or, to be conservative, the highest single
value may be chosen to represent that food group.
If the data indicate that a food group consists of some crops that are
relatively high accumulators and some that are relatively low, it may be
worthwhile to subdivide the food group on this basis. If no response slope
can be determined for a particular crop food group, the highest value for
any of the other groups should be assigned to that group.
Step C. Determine Human Daily Intake
The increase in DI (in ug/day) of an inorganic contaminant due to
deposition of municipal waste combustor emissions by this pathway is
calculated from the response data and deposition rate (see Section D.3.1) by
the following equation:
DI = CD x £ (UC. x FC. x DC.) Equation (D-4)
i=l
where:
DI = increment (above background) of daily intake of pollutants
(ug/day)
CD =• cumulative soil deposition of pollutant (kg/ha)
UC. = uptake response slope for i food group [ug/g DW (kg/ha)"1]
FCi - fraction of i food group (crop) assumed to originate from
contaminated soil (unitless)
DC. = daily dietary consumption of i food group (g DW/day)
D-15
-------
The derivation of DI by the deposition-soil-piant-human toxicity
pathway for organic contaminants or chemicals subject to loss is largely the
same as the procedure for inorganics based on the linear response model.
One difference is that uptake data are not segregated on the basis of soil
pH since no reason for doing this has been demonstrated. In addition, the
procedure for calculating response slopes for organics differs somewhat from
that for inorganics, in that tissue concentration is treated as a linear
function of soil concentration rather than application or deposition rate,
as explained in Section D.3.4.2. Thus, the maximal soil concentration, LC
[from Equation (D-3)] of the contaminant is used in the above equation
rather than CD for the determination of the DI for organics or chemicals
subject to loss. The LC' should be compared with the phytotoxicity threshold
as described in Section D.3.4.1.
D.4.1.3 Input Parameter Requirements. The following individual
parameters are required to calculate the human daily intake for the
deposition-soil-piant-human toxicity pathway.
D. 4.1.3.1 Fraction of.food group assumed to originate from soil
contaminated bv municipal waste combustor emissions. All of the homegrown
food is assumed to originate from contaminated soil, but homegrown food
comprises <60 percent of the diet of the rural farm dweller (see Table D-4).
Therefore, values of FC (after rounding) are 0.60 for all vegetables (except
dried legumes), 0.45 for potatoes and 0.17 for dried legumes, based on the
average percent of annual consumption of homegrown foods. Some food groups
(such as mushrooms, peanuts and grains and cereals) are assumed to be
unaffected and the FC is set at zero.
D.4.1.3.2 Uptake response slope (UC). Uptake response data are
required for as many crops as possible in the food groups for which FC / 0.
If DC for the contaminants varies with soil pH, slopes appropriate to local
soil pH should be chosen as discussed in Step A.
D-16
-------
D.4.1.3.3 Daily dietary consumption of food group (DC). Values for DC
(in g DW/day) are needed for each food group for which FC jt 0. The values
chosen should be appropriate for the MEI. Consumption data presented in
Pennington and reanalyzed in Table D-l are mean values for each of eight
age/sex groups.13 One could define the MEI in terms of the age/sex group
having the highest consumption for all of the food groups combined or for
each individual food group. Alternatively, one could estimate a 95th-
percentile consumption level, based on variability of 1-day consumption;
however, this procedure risks overestimation of long-term consumption.
D.4.1.4 Example Calculations. In this section, examples will be
provided for this pathway using the metal cadmium, and the organic compound
benzo(a)pyrene [B(a)P]. Examples will be calculated for the model municipal
waste combustion facility assumed to be located in western Florida.
One of the most important steps in these calculations is the selection
of values for each of the parameters involved. 'Value selection must be
based on careful literature searches and evaluations of available data. For
many parts of this methodology, value selection is a very data-intensive
exercise.
Since the calculations presented in this section are only for the
purpose of examples, the values employed will not be based on literature
searches, but will rely on a few sources of readily available information.
Therefore, the results do not constitute actual recommendations for risk
assessment.
0.4.1.4.1 Cadmium.
Steps A and B. Sort Available Uptake Response Data for All Food Crops and
Determine Uptake Response Slopes for Each Food Group
Studies are sorted according to soil pH. Natural soils in the vicinity
of the southeastern United States would tend to have pH >6.0. The data of
Dowdy and Larson (pH >6.0) will be used for this example in lieu of all of
the available studies. Response data on carrot and radish are used to
derive a weighted mean.response slope for the "root vegetable" food group.
Consumption data are those for the 25- to 30-year-old male.
D-17
-------
Crop type carrot radish
Linear UC [ug/g (kg/ha)"1)] 0.20 0.056
Dry weight consumption (g DW/day) 0.55 0.021
Example Calculation of Weighted Mean
(0.20 x 0.55) + (0.056 x 0.021)
UCroot veg. ' 0.55 + 0.021
- 0.19 ug Cd/g [kg Cd/ha]"1
Step C. Determine Human' Daily Intake
CD for cadmium for 30 and 100 years of total deposition is calculated
-2 2
from Equation (D-l) using an AD = 1.088 x 10 g/m -year. The DI is
determined using Equation (D-4). UC values for each food group have been
derived from the data of Dowdy and Larson as illustrated above for root
vegetables. Values of DC for vegetables are obtained from Table D-l. DC
values chosen are those for the age/sex group with highest consumption of
that food group. Values of FC are those for the home garden scenario from
Table D-4.
Food Group UC DC FC UC x DC x FC
Potatoes 0.038 31.85 0.45 0.545 .
Leafy vegetables 0.605 2.78 0.60 1.009
Legume vegetables, nondried 0.0053 3.38 '0.60 0.011
Legume vegetables, dried 0.0053 8.51 0.17 0.008
Root vegetables 0.19 2.28 0.60 0.260
Garden fruits 0.073 5.94 0.60 0.260
Total 2.093
For T - 30 years, DI => CD (kg/ha) x 2.083 [ug x ha (kg x day)"1]
- 3.26 (kg/ha) x 2.093 [ug x ha (kg x day)"1]
=6.83 ug/day
For T = 100 years, Di = 10.88 kg/ha x 2.093 [ug x ha (kg x day)"1]
» 22.77 ug/day
D-18
-------
This represents the increase in daily human intake due to cadmium from
municipal waste combustor emissions with total deposition of 30 or
100 years.
D.4.1.4.2 Benzo(a)pvrene
Steps A and B. Sort Available Uptake Response Data for All Food Crops and
Determine Uptake Response Slopes for Each Food Group
Studies for organic compounds are not sorted according to soil pH.
Using the data in Connor and consumption data for the 25- to 30-year-old
male, the weighted mean response slope for the root vegetables are
17 18
calculated previously, in the example for cadmium. ' Leafy vegetable DC
19
were taken from Connor. Data for other food groups were not available and
were assigned the highest available (i.e., root vegetables UC). Calculated
values are shown in Step C below.
Step C. Determine Human Daily Intake
The DI is determined using Equation (0-4) with LC in place of CD. The
DC values are obtained from Table D-l. DC values chosen are those for the
age/sex group with highest consumption of that food group. Values of FC are
those for the home garden scenario from Table D-4.
Food Group UC DC FC UC x DC x FC
Potatoes 1.74 31.85 0.45 24.939
Leafy vegetables 0.42 2.78 0.60 0.701
Legume vegetables, nondried 1.74 3.38 0.60 3.529
Legume vegetables, dried 1.74 8.51 0.17 2.517
Root vegetables 1.74 2.28 0.60 2.380
Garden fruits 1.74 5.94 0.60 6.201
Total 40.267
The maximum soil concentration of B(a)P is determined from the annual
deposition using Equation (D-3). For B(a)P, the loss rate constant, k,
estimated from data for biotic loss only is 0.16 (year)"1, and may therefore
underestimate total soil loss.20 The AD is 5.66 x 10"4 g/(m2 x year). The
bulk density, B, of 1.5 g/cm for sandy clay loam is used.21 For 30 years,
D-19
-------
LC = 5.66 x 10"4 x [i-e~(°-16^30h x 102 x (1.50 x 20 x 0.16)"1
- 1.18 x 10"2 ug/g
For 100 years,
LC - 1.18 X 10"2 ug/g
The LC is compared with the phytotoxicity threshold, however, but data are
not available for benzo(a)pyrene. Therefore, for Equation (0-4) (as modified
for organics or chemicals subject to loss):
For 30 and 100 years,
01 - 1.18 x 10"2 ug/g x 40.267 (g/day) = 4.75 x 10"1 ug/day
This represents the increase in daily human intake due to B(a)P from
municipal waste combustor emissions.
D.4.2 Deposition-Human Toxicitv ("Pica") Exposure Pathway
0.4.2.1 Assumptions. In addition to many of the assumptions listed in
Table 0-2, some additional assumptions are made for this pathway, relating
to the degree of contaminated soil ingestion that could occur and the method
of assessing potential effects. These assumptions and their potential
ramifications are summarized in Table 0-5 and are further discussed in the
following sections.
0.4.2.2 Calculation Method. A human daily intake, 01 (in ug/day DW),
is determined as follows:
01 - LCT x Is Equation (0-5)
D-20
-------
TABLE D-5. ASSUMPTIONS FOR DEPOSITION-HUMAN TOXICITY ("PICA") EXPOSURE PATHWAYS
Functional Area
Assumptions
Ramifications/Limitations
Exposure assessment
o
i
IN)
Effects assessment
Deposited emissions are not necessarily
soil incorporated and may be concentrated
in the uppermost soil layer. Deposited
contaminant is assumed to be distributed
within the uppermost 1 cm of soil, and
ingested soil is assumed to originate
from the same 1 cm layer.
Deposition-contaminated soil may be
ingested by children at the rates
observed in studies of pica for soil.
It is assumed that pica may occur from
1-6 years of age. Cancer potency is
adjusted to reflect a 5-year rather
than 70-year exposure.
Overestimates exposure in situations
where soil incorporation occurs to any
depth >1 cm. If incorporation is to a
lesser depth, exposure is under-
estimated. For example, if exposure
is to fallout dust directly, ingested
soil could approach 100% deposited
particulate.
Overestimates exposure to the extent
the pica child frequents some areas
where deposition has not occurred.
If the child is more susceptible to
chemical carcinogenesis than the
adult, a 5/70 adjustment could result
in underestimation of hazard.
-------
where:
I - soil ingestion rate (g DW/day)
LCj » maximal soil concentration increment of pollutant after
time, T (ug/g)
Deposited contaminant is assumed to be within the uppermost 1 cm of
soil and ingested soil is assumed to originate from the same 1 cm layer.
Soil concentration, LC, is therefore calculated according to Equation (D-2)
for most inorganic chemicals or Equation (D-3) for chemicals that are
subject to loss processes using 1.50 g/cm as the bulk density and a depth,
D, of 1 cm.
D.4.2.3 Input Parameter Requirements.
D.4.2.3.1 Soil inqestion rate (I ). Soil ingestion has been
recognized as an important source of exposure to several pollutants. For
22
adults, a value of 0.02 g/day has been used to estimate ingestion.
Children may ingest soil by either inadvertent hand-to-mouth transfer or
by intentional direct eating. Pica is the term for frequent, intentional
eating of non-food objects. Lepow et al. estimated that children frequently
23
mouthing their hands may inadvertently ingest >100 mg of soil/day.
24
Children who eat soil directly may ingest as much as 5 g/day, thus
establishing a plausible typical-to-worse range of 0.1 to 5 g/day.
Studies aimed at more accurately determining the range of ingestion
rates have yielded some data, but are as yet inconclusive. Binder et al.
conducted a pilot study to establish methods for determining soil ingestion
25
rates in children living near a lead smelter. Based on these preliminary
data and the value used previously by EPA, a value of 0.5 g/day is suggested
as a reasonably protective value for I . This value represents an estimate
of the 95th percentile of soil ingestion in this study population.
0.4.2.4 Example Calculations.
D.4.2.4.1 Cadmium. The DI for cadmium is determined using
Equation (D-5), with I = 0.5 g/day. The LC for cadmium calculated
D-22
-------
according to Equation (D-2) using a soil depth of 1 cm is 21.76 and
72.53 ug/g for 30 and 100 years, respectively. For 30 years,
DI = 21.76 ug/g x 0.5 g/day
* 10.88 ug/day
For 100 years,
DI - 72.53 ug/g x 0.5 g/day x 1
= 36.27 ug/day
D.4.2.4.2 Benzofa)ovrene. The organic compound B(a)P is a carcinogen.
ihis example, I is 0.5 g/day and LC is 2.36 x 10
100 years of total deposition. Therefore, the DI is:
For this example, I is 0.5 g/day and LC is 2.36 x 10 ug/g for both 30 and
DI - 2.36 x 10"l ug/g x 0.5 g/day
= 1.18 x 10"1
D.4.3 Exposure Pathways for Herbivorous Animals for Human Consumption
D.4.3.1 Assumptions. Animal forage may be contaminated by uptake
through the plant roots of deposited pollutants (deposition-soil-plant-
animal -human toxicity) or by adherence to plant surfaces or roots
[deposition-soil-animal (direct ingestion)-human toxicity] of deposited
particulate or contaminated soil. In both pathways, humans are exposed by
consuming animal tissue, which has taken up the contaminants. In the second
pathway, soil .incorporation is not assumed and'direct ingestion of the
contaminant by farm animals may occur. Direct ingestion is also possible by
animals such as deer that will be taken by hunters. The amount of game
consumed by hunters, however, is assumed not to exceed the consumption of
home-produced meat by farm dwellers. Therefore, the farm dweller is taken
as the MEI for both pathways. Protection of the MEI is assumed to be
protective of hunters as well. Additional assumptions are listed in
Table D-6.
D-23
-------
TABLE D-6. ASSUMPTIONS FOR PATHWAYS DEALING WITH HERBIVOROUS ANIMALS
Functional Area
Assumptions
Ramifications/Limitations
Fraction of human diet
affected by MWC emissions
o
I
IM
Ingestion by animals of
parti culate-contaminated
soil or forage crops
The MEI is assumed to be an individual raising
much of his own meats, poultry, eggs and dairy
products. The percentage of home-produced
foods in the diet of the MEI is estimated from
a USOA (1966) survey of rural farm households,
which constituted 6% of all households.
Individuals consuming wild game that forage in
emissions-contaminated areas are assumed to
have no greater exposure than the MEI identi-
fied above.
Contaminant uptake by crops may affect all
animal feeds, but only grazing animals are
affected by adherence. Adulteration by soil
of harvested crops such as grains fed to non-
grazing animals is assumed to be minimal.
Deposited contaminant is assumed to be distri-
buted within the upper most 1 cm of soil, and
ingested soil is assumed to originate from the
same 1 cm layer. Direct ingestion of soil may
occur, and only animals consuming pasture crops
are affected.
The linear response slope of the most responsive
forage crop is used to represent all forage
crops in the animal diet.
More recent information,
if available, might show
significant changes in
both demographics and food
production habits of these
households.
Information to substan-
tiate this assumption is
not immediately available.
Will tend to overpredict
average crop response.
-------
D.4.3.2 Calculation Method.
D.4.2.2.1 Uptake pathway; deposition-soil-piant-animal-human
toxicitv. To determine the human daily intake of a contaminant .by this
pathway, an animal feed concentration increment of the contaminant (AFC, in
.ug/g DW) first must be determined. The equations differ for organics and
inorganics since the units of crop uptake differ. The AFC is calculated
from the deposition rate [for inorganics, Equation (D-l) or the soil
concentration [for organics or chemicals that are subject to loss,
Equation (D-2)] and the crop uptake slope as follows:
(for most organics) • AFC = CD x UC Equation (D-6)
(for chemicals subject to loss) AFC = LCj x UC Equation (D-7)
where:
CD - cumulative deposition of pollutant (kg/ha)
UC - linear response slope of forage crop [ug/g crop DW••(kg/ha) ]
or
[ug/g DW (ug/g)'1]
LCj = maximal soil concentration of pollutant, after time, T (ug DW)
Following the calculation of the AFC, the human daily intake
(DI ug/day) is calculated as follows;
n
DI - AFC x £ (UA. x FA. x DA.) Equation (D-8)
1 n n
where:
AFC - animal feed concentration of pollutant (ug/g)
UA. - uptake response slope of pollutant in the i animal tissue.
1 1
[ug/g tissue DW (ug/g feed DW) ]
DA.. = daily dietary consumption of the i animal tissue food group
(g DW/day)'
D-25
-------
FA1 - fraction of i food group assumed to be derived from animals
feeding on contaminated soil or feedstuffs
D.4.3.2.2 Adherence pathway; deposition-soil-animal (direct
inqestion)-human toxicitv. The animal feed concentration (AFC, in ug/g DW)
is calculated as for the uptake pathway [see Equation (D-6)], but the values
will differ because fewer human food categories will be affected. Soil
incorporation is not assumed; therefore, the upper parts of pasture crops
may be contaminated and may be harvested as feed for cattle or sheep or
grazed directly by these animals. The deposited contaminant or particulate
can also accumulate in the thatch layer of pasture, and be directly consumed
by grazing animals. Domestic animals (such as pigs, poultry) that consume
grains or other non-pasture crops are assumed not to be affected. When the
deposited contaminant is consumed from plant or soil surfaces, contaminant
intake by the grazing animal is related to the fraction of the animal diet
which that soil comprises. Deposited contaminant is assumed to be within
the uppermost 1 cm soil layer. Soil concentration (LC) is calculated
according to Equation (D-2) for most inorganic chemicals or Equation (D-3)
for chemicals which are subject to loss processes, using 1.50 g/cm bulk
density and a depth, D, of 1 cm. The animal feed concentration is derived
in terms of the soil concentration and fraction of the animal diet that is
adhering soil. AFC is derived as follows:
AFC - LCT x FS Equation (D-9)
where:
FS = fraction of animal diet that is adhering soil (unitless)
LCj = maximal soil concentration increment of pollutant after
time, T (ug/g)
D.4.3.3 Input Parameter Requirements.
D.4.3.3.1 Fraction of food group assumed to be derived from animals
feeding on soil or feedstuffs contaminated bv municipal waste combustor
emissions (FA). As was the case with FC in the crops-for-human-consumption
pathway (see Section D.4.1.3.1), this parameter determines which food groups
are included in the analysis.
D-26
-------
For the first of these two pathways (deposit!on-soil-plant-animal-
human toxicity), all meat groups except fish will be assumed to be affected,
including beef, lamb, pork, poultry, dairy products and eggs (see
Table D-4). The second pathway [deposition-soil-animal (direct ingestion)-
human toxicity] is assumed to affect only grazing animals (beef, lamb and
dairy food groups are included).
As for the pathways dealing with crops for human consumption, the MEI
for these two pathways is chosen as a farm family residing within 50 km of a
municipal waste combustor (in the area of maximal contaminant deposition)
raising a substantial percentage of their own meat or other animal products.
The choice of FA values is based on the percentage of homegrown foods
consumed by rural farm households (see Table D-4). As stated previously, it
is assumed that these FA values are sufficiently high that an individual
consuming wild game from contaminated areas will also be protected.
0.4.3.3.2 Daily dietary consumption of food group (DA). Values for
daily dietary consumption (DA, in g DW/day) are needed for each food group
for which FA / 0. As was described for DC (see Section D..4.1.3.3),
consumption data are taken from Table 0-1. The food item "beef liver"
includes various other organ meats consumed by humans in smaller amounts,
such as kidney, hearts, etc. Individuals with a preference for those organs
are expected to consume them at the rates given for beef liver. It is also
assumed that consumption of wild game does not exceed the values of DA for
other meats (beef, lamb) and will also protect hunters.
0.4.3.3.3 Fraction of animal diet that is adhering soil (FS). Studies
of grazing animals indicate that soil ingestion ordinarily ranges from 1 to
10 percent of dry weight of diet (but may range as high as 20%) for cattle
and may be >30 percent for sheep during winter months when forage is
27
reduced. It will be assumed that 100 percent of deposited contaminant is
retained in the uppermost soil layer (-1 cm depth) that animals may ingest.
Since lamb contributes relatively little to the United States diet, a value
of FL * 10 percent or 0.10, based largely on cattle, will be used to
represent a reasonably high exposure situation.
0-27
-------
D.4.3.3.4 Uou.
-------
The DI for cadmium is calculated using Equation (D-8). Values of UA for
29
various animal tissues were converted from wet to dry weight basis. The
value listed for "beef liver" is actually from sheep kidney; it is the
highest UA value for an organ meat, since the beef liver consumption data
are assumed here to represent any organ meat. No data were available for
pork, so an average of the values for beef and lamb was used. Data were
also unavailable for eggs and dairy products. In this example they are
assumed to be similar to poultry and beef muscle, respectively. The DA
values are derived from Table D-l. They include fat since the UA values are
on a dry weight basis (including fat). The FA values are from Table D-4;
they differ for the uptake and adherence pathways.
UA x DA x FA
Animal Tissue Group UA. DA FA (when FA £ 0)
Beef 0.003 53.0 0.44
Beef liver 9.9 1.54 0.44
Lamb 0.005 0.44 0.44
Pork 0.004 33.9 0.44
Poultry 0.08 U.7 0.34
Dairy 0.003 79.5 0.40
Eggs 0.08 8.1 0.48
Total 7.566
Therefore, the DI for cadmium for 30 years is,
DI - 0.45 ug/g x 7.566 g/day =3.48 ug/day
For 100 years,
Di =* 1.52 ug/g x 7.566 g/day = 11.50 ug/day
Benzo(a)ovrene--The animal feed concentration of B(a)P is determined
using Equation (D-7). The UC value of 0.42 ug/g (ug/g)"1 for spinach leaves
was chosen since it had the greatest UC value of available foliage data.30
For both 30 and 100 years,
D-29
-------
AFC = LCT (ug/g) x 0.42 ug/g (ug/g)"1
= 4.96 x 10"3 ug/g
_o
where LCj (1.18 x 10 ug/g) 1s calculated based on a soil depth of 20 cm.
B(a)P is not extensively bioaccumulated in animals, making UA
approximately zero. Therefore, DI of B(a)P by the uptake pathway from
consumption of animal tissue is not increased due to municipal waste
combustor emissions.
0.4.3.4.2 Pi for adherence pathway.
Cadmium--The AFC for this pathway is derived using Equation (D-9):
AFC - LCT x FS
where:
LCj = maximum soil concentration increment of pollutant after
time, T (ug/g)
• FS =» fraction of. animal diet adhering soil (unitless)
The value of FS will be 0.10 and the soil concentration of Cd (in 1 cm
of uppermost soil layer) is 21.76 and 72.53 ug/g for 30 and 100 years,
respectively.
For 30 years,
AFC - 21.76 ug/g x 0.10 - 2.18 ug/g
For 100 years,
AFC = 72.53 ug/g x 0.10 = 7.25 ug/g
Equation (0-8) and the UA and OA values from the uptake pathway are
also used. The FA values from Table 0-4 for meat and dairy (grazing animals
only) are used.
D-30
-------
Animal Tissue Group UA DA _FA_ UA x DA x FA
Beef 0.003 53.0 0.44 0.070
Beef liver 9.9 1.54 0.44 6.71
Lamb 0.005 0.44 0.44 0.001
Pork 0.004 33.9 0 0
Poultry 0.08 11.7 0 0
Dairy 0.003 79.5 0.40 0.095
Eggs 0.08 8.1 0 0
Total 6.876
The DI for cadmium is 14.99 and 49.85 ug/day, for 30 and 100 years,
respectively.
Benzo(a)ovrene--The AFC for this pathway is also calculated using
Equation (D-9). The value of FS is 0.10 and the soil concentration of B(a)P
in the upp<
100 years,
in the uppermost 1 cm soil layer is 2.36 x 10 ug/g. For both 30 and
AFC - 2.36 x 10"1 ug/g x 0.10 - 2.36 x 10"2 ug/g
Since UA for B(a)P is approximately zero, DI calculated according to
Equation (D-8) is zero, showing municipal waste combustor emissions do not
increase DI of B(a)P by this pathway.
D.4.4 Exposure Pathways for Toxicitv to Herbivorous Animals
These two pathways, deposition-soil-plant-animal toxicity and
deposition-soil-animal toxicity (direct ingestion), are similar to the
pathways for contaminant uptake by animal tissues consumed, by humans, as
discussed in Section D.4.3; however, since toxicity to the animal itself is
now the endpoint of concern, the list of animals to be considered is
broadened to include all herbivores. Herbivorous rodents and birds should
be considered, as well as large herbivores and other domestic animals. The
pathway for adherence of soils to plant roots is limited to grazing animals
that ingest significant amounts of soil (cattle, sheep).
D-31
-------
D.4.4.1 Assumptions. The assumptions pertaining to these pathways are
stated in Tables D-l through D-6.
D.4.4.2 Calculation Method. The first set of calculations for the
deposition-soil-plant-animal and deposition-soil-animal pathways are the
same as given in Section D.4.3.2 for the determination of an animal feed
concentration (AFC, in ug/g DW) from the cumulative deposition (CD, for
inorganics) or soil concentration (LC, for organics and/or chemicals subject
to loss) and linear uptake response slope for crops (UC) [Equations (D-6),
(D-7)]. The AFC is calculated for the deposition-soil-animal pathway from
the soil concentration (LC) and fraction of soil constituting the animal
diet (FS) as in Section D.4.3.2.2 [Equation (D-9)].
D.4.4.3 Input Parameter Requirements. The input parameters, LC and
CD, were defined in Section D.3.2, UC was defined in Section D.4.1.3 and FS
was defined in Section D.4.3.3.
D.4.4.4 Example Calculations.
D.4.4.4.1 Cadmium. The calculated AFC for cadmium for the uptake
pathway, as described in Section D.4.3.4.1 (Cadmium) is 0.46 and 1.52 ug/g
for 30 and 100 years, respectively. The AFC for the adherence pathway for
cadmium is 2.18 and 7.25 ug/g for 30 and 100 years, respectively, calculated
as described in Section D.4.3.4.2 (Cadmium).
For the uptake pathway (deposition-soil-plant-animal toxicity), a wider
variety of herbivores may be affected than for the adherence pathway
(deposition-soil-animal), which affects only grazing animals.
D.4.4.4.2 Benzofa)ovrene. The AFC for B(a)P for the uptake pathway is
lated as 4.96 x 10 ug/g for I
Section D.4.3.4.1 (Benzo(a)pyrene).
calculated as 4.96 x 10 ug/g for both 30 and 100 years, as described in
D.4.5 Deposition-Soil-Plant Toxicitv Exposure Pathway
D.4*5.1 Assumptions. The assumptions pertinent to this pathway have
been stated in Table D-2.
D-32
-------
D.4.5.2 Calculation Method. For this pathway, exposure is equivalent
to the deposition rates or soil concentrations of contaminants deposited
from municipal waste combustor emissions.
D.4.5.3 Example Calculations. No calculations are required for this
pathway, except in some cases to convert CD to LC.
D.4.6 Exposure Pathways for Toxicitv to Soil Biota and Their Predators
This section deals with two pathways: toxicity to soil biota
(deposftion-soil-soil biota toxicity) and toxicity to predators of soil
biota (deposition-soil-soil biota-predator toxicity). The term "soil biota"
refers to a broad range of organisms including microorganisms and various
invertebrates living in or on the soil. Their predators similarly include a
variety of organisms. The availability of data determines what species are
considered.
D.4.6.1 Deposition-Soil-Soil Biota Toxicitv Exposure Pathway. For
this pathway exposure is equivalent to the deposition rates or soil
concentrations of contaminants deposited from municipal waste combustors.
D.4.6.1.1 Example calculations. This pathway does not require
calculations, except conversion of CD to LC.
D.4.6.2 Deposition-Soil-Soil Biota-Predator Toxicity Exposure Pathway.
D.4.6.2.1 Assumptions. In addition to many of the assumptions listed
in Table"0-7, some additional assumptions relevant to this pathway are
stated Table D-8.
D.4.6.2.2 Calculation method. Calculations of criteria for this
pathway may take either of two forms, depending on the type of data
available concerning contaminant uptake by soil biota. If uptake response
(increase in concentration) in soil biota, U8, can be expressed in terms of
a contaminant deposition rate, the predator intake is calculated as follows:
D-33
-------
TABLE 0-7. ASSUMPTIONS FOR ECOLOGICAL EFFECTS FOR TERRESTRIAL FOOD CHAIN
Functional Area
Assumptions
Rami fications/Limi tations
Toxicity thresholds
for nonhuman organisms
All inhibitory effects should be
considered adverse.
The geometric mean of exposure
levels bracketing the appearance
of an adverse effect should be
used as the threshold.
The form of a contaminant used
in a study should not be.con-
sidered equally bioavailable
and toxic as form in soil, unless
suitable data using soil are not
available.
Some "inhibitory" changes might not
significantly affect the individual
survival or population dynamics.
Conversely, some ecologically important
effects might not be observed in toxicity
tests.
The true threshold may lie at any point
between these two levels, and may be over-
or underpredicted by this method.
Availability of chemicals in soil or
deposited on soil may differ; particularly,
it may be lower, so that toxicity is over-
predicted.
-------
TABLE D-8. ASSUMPTIONS FOR DEPOSITION-SOIL-SOIL BIOTA-PREDATOR TOXICITY EXPOSURE PATHWAY
Functional Area
Assumptions
Ramifications/Limitations
OJ
en
Contaminant uptake by
soil biota
Use of available data
to protect a variety
of species
Response in tissues of soil biota
can be represented by a linear
function.
It is assumed that use of the
highest available response slope
in soil biota and the lowest
available dietary threshold in
predators will result in protection
of untested species.
Probably oversimplifies a more
complex relationship.
This conservative assumption could
be overprotective of some species;
the extent to which it underprotects
others is unknown. A "match" of
data for a consumed organism and its
predator usually is not possible.
-------
RFC - CD x UB Equation (D-10)
where:
RFC = predator feed concentration (ug/g DW)
CD - cumulative soil deposition of pollutant (kg/ha)
UB - uptake response slope in soil biota (ug/g [kg/hg]
If the chemical is not subject to degradation or loss in soil, RFC is a
cumulative concentration. If soil biota response is measured in terms of a
soil concentration, the following equation is used:
RFC = LCT x UB Equation (D-ll)
where:
LCT - maximal soil concentration of pollutant at time, T (ug/g DW)
1 _i
UB = uptake response slope in soil biota (ug/g [ug/g] )
LC represents cumulative soil concentration (above background) due to
emissions from municipal waste combustors.
D.4.6.2.3 Input parameter requirements.
Uptake response slope in so.il biota (UB)--UB may take any of several
forms, depending on the characteristics of the chemical and the data
available. UB is derived by linear regression of tissue concentration by
either contaminant deposition rate or soil concentration. Two or more
points are needed to derive the slope for inorganics, while for organic
compounds a single data pair can be used to derive a bioconcentration fa-ctor
for plants and animal tissues.
The highest available uptake response slope will be used to estimate
the level of dietary contamination to which predators of soil biota would be
subject. This fact is important when evaluating data for earthworms. Some
studies distinguish between contaminant physiologically absorbed and that
which is due to gut contents or soil contamination, of the sample. There is
no need to distinguish between absorbed and unabsorbed contaminants, since
D-36
-------
a predator would ingest the gut contents as well as the rest of the
organism. Whenever possible, analyses used should be based on the whole
organism.
D.4.6.2.4 Example calculations.
Cadmium--The predator feed concentration (RFC) for cadmium is
calculated using Equation (0-11):
RFC = CD x UB
where:
CD » cumulative soil deposition of pollutant (kg/ha)
UB = uptake response slope in soil biota (ug/g[kg/ha] )
CD is 3.26 and 10.88 kg/ha for 30 and 100 years, respectively, for
ium, and
30 years is,
cadmium, and the UB is 13.7 ug Cd/g (kg/ha) . The RFC for cadmium for
For 100 years,
RFC - 3.26 kg/ha x 13.7 ug/g (kg/ha)"1 = 44.66 ug/g
RFC = 10.88 kg/ha x 13.7 ug/g (kg/ha)"1 = 149.06 ug/g
Benzofa)pvrene--Lack of data availability for UB of B(a)P prohibits
calculation of a RFC.
D-37
-------
D.5 REFERENCES FOR APPENDIX D
1. O'Flaherty, E. J. Toxicants and Drugs: Kinetics and Dynamics.
John Wiley and Sons. New York, New York. 1981.
2. U. S. Environmental Protection Agency. Effects of Sewage Sludge on the
Cadmium and Zinc Content of Crops. Council for Agricultural Science
and Technology. EPA Report No. 600/8-81-003. 1980.
3. Ryan, J. A., H. R. Pahren, and J. B. Lucas. Controlling Cadmium in the
Human Food Chain - A Review and Rationale Based on Health Effects.
Environ. Res. 28:251-302. 1982.
4. Logan, T. J., and R. L. Chaney. Utilization of Municipal Wastewater
and Sludge on Land - Metals. Proceedings of the Workshop on
Utilization of Municipal Wastewater and Sludge on Land. University of
California at Riverside, Riverside, CA. pp. 253-323. 1983.
5. Reference 3.
6. Page, A. L., T. J. Logan, and L. E. Summers. Draft Report of the
Workshop on Effects of Sewage Sludge Quality and Soil Properties on
Plant Uptake of Sludge - Applied Trace Contaminants. Las Vegas,
Nevada. 1986.
7. Reference 3.
8. Pennington, J. Revision of the Total Diet Study Food List and Diets.
J. Amer. Diet Assn. 82:166-173. 1983.
9. U. S. Environmental Protection Agency. Development of Risk Assessment
Methodology for Land Application and Distribution and Marketing of
Municipal Sludge. Draft Final Report. 1986.
10. Food and Drug Administration. Documentation of the Revised Total Diet
Study Food Lists and Diets. NTIS Publication No. 82-192154. 1981.
11. U. S. Department of Agriculture. Composition of Foods. Agricultural
Handbook No.. 8. 1975.
12. U. S. Department of Agriculture. Household Food Consumption Survey,
1965-1966. Report 12. Food Consumption of Households in the United
States - Seasons and Year, 1965-1966. 1966.
13. Reference 8.
14. Dowdy, R. H., and W. E. Larson. The Availability of Sludge-Borne
Metals to Various Vegetable Crops. J. Environ. Qua!. 4:278-282. 1975.
15. Reference 9.
D-38
-------
16. Reference 13.
17. Connor, M. S. Monitoring Sludge - Amended Agricultural Soils.
Biocycle 25:47-51. 1984.
18. Reference 9.
19. Reference 17.
20. Bosert, D. D., and R. Bartha. Structure - Biodegradability
Relationships of Polycyclic Aromatic Hydrocarbons in Soil. Bulletin
of Environmental Contamination and Toxicology 37:490-495. 1986.
21. National Resource Council. Mitigative Techniques and Analysis of
Generic Site Conditions for Groundwater Contamination Associated with
Severe Accidents. Battelle Memorial Institute. 1984.
22. U. S. Environmental Protection Agency. Air Quality Criteria for Lead.
EPA Report No. 600/8-83-0288.
23. Lepow, M. L., M. Gillette, S. Markowitz, R. Robino, and J. Kapish.
Investigations into Sources of Lead in the Environmental of Urban
Children. Environ. Res. 10:415-426. 1975.
24. Reference 22.
25. Binder, S., D. Sokal, and D. Ma.ughan. Estimating the Amount of Soil
Ingested by Young Children through Trace Elements. Draft Report.
Centers for Disease Control, Center for Environmental Health, Special
Studies Branch. 1985.
26. Reference 9.
27. Thornton, D., and P. Abrams. Soil Ingestion: A Major Pathway of Heavy
Metals into Livestock Grazing Contaminated Land. Sci. Total Environ.
28:287-294. 1983.
28. Telford, J. N., M. L. Thonney, and D. E. Hogue. Toxicological Studies
in Growing Sheep Fed Silage Corn Cultured on Municipal Sludge - Amended
Acid Subsoil. J. of Toxicol. Environ. Health 10:73-85. 1982.
29. U. S. Environmental Protection Agency. Environmental Profiles and
Hazard Indices for Constituents of Municipal Sludge: Cadmium. 1985.
30. U. S. Environmental Protection Agency. Environmental Profiles and
Hazard Indices for Constituents of Municipal Sludge: Benzo(a)pyrene.
1985.
31. Reference 30.
D-39
-------
APPENDIX E
SURFACE RUNOFF MODEL
-------
APPENDIX E
SURFACE RUNOFF MODEL
E.I INTRODUCTION
Contaminants associated with participates emitted by municipal waste
combustors are subject to deposition on surfaces downwind from the municipal
waste combustor at rates determined by meteorology, terrain and particle
physics. This fallout is subsequently subject to dissolution and/or
suspension in runoff after precipitation events. Runoff moves over the
surface of the earth to a surface water body, where it mixes with other
waters. As a consequence, humans utilizing water or eating fish from the
surface water body or aquatic life living therein may be exposed to runoff
transported contaminants.
The methodology derived to calculate exposure from the surface runoff
pathway was originally developed to evaluate impacts from, the application
of municipal wastewater sludge to land. A detailed discussion is available
in the document entitled "Development of Risk Assessment Methodology for
Land Application and Distribution and Marketing of Municipal Sludge. The
methodology is formulated in three successive tiers, which begin with simple
but very conservative estimates, and proceed to more detailed analyses if
the first tiers predict unacceptable exposures. Both acute events and
chronic exposure are evaluated using standard approaches to calculate runoff
volume and associated erosion potential.
The remainder of this appendix describes the assumptions, calculation
steps, and input requirements of the Surface Runoff Model. Example
calculations are also presented.
E.2 ASSUMPTIONS
A number of assumptions were required to formulate the risk-based
methodology both with- respect to runoff generation and subsequent mixing in
the receiving water. The key assumptions are provided in Table E-l with a
E-l
-------
TABLE E-l.
.RFACE RUNOFF METHODOLOGY ASSUMPTIONS
Functional Area
Assumption
Ramifications
Long-Term Concentrations:
Tier 1
Tier 2/3
General
Source area
All contaminant deposited
on an annual basis 1s trans-
ported to the receiving
water In a dissolved form.
Loadings to the receiving
water can be described as
a function of solids
loading.
Facility operates over a
sufficient period for
surface soil levels to
reach equilibrium where
annual losses equal annual
Inputs.
No settling of particles
In the deposition zone,
gross erosion reaches the
edge of field.
Event concentrations:
Tier 1
Tier 2/3
Source area
Stream
All contaminant emitted 1n
a year Is lost In a single
runoff event.
No klnetlcally limited
release of contaminant
from residual/soil mix-
ture; I.e., total con-
taminant concentration
Is fully equilibrated
Into adsorbed and dis-
solved phases.
Stream flow Is unchanged
by the storm unless arterial
velocity data are available
from the hydrograph.
Provides an extremely con-
servative estimate since
no losses are considered.
Mechanistically Inappro-
priate for contaminants
with low partition coef-
ficients.
Overpredicts contaminant
loading by Ignoring loss
mechanisms other than
those used in formulation.
namely, runoff and Infil-
tration.
Maximizes contaminant loss
Provides extremely conser-
vative estimate with no
provision for losses or In-
complete mobilization.
Maximizes contaminant con-
centration available for
runoff In dissolved phase.
thereby maximizing
loadings.
Overestimates stream
concentration, since the
storm will Increase stream
flow.
E-2
-------
discussion of their impact on the methodology. Because the science is not
exact, the assumptions are mostly conservative (i.e., overpredict
contaminant concentrations). In some instances, however, the nature of the
effect of a given assumption may vary with site-specific considerations.
E.3 CALCULATION STEPS
The methodology addresses both long- and short-term exposures as
illustrated in Figure E-l. In Tier 1, it is assumed that all contaminant
emitted in a given year is transported to the receiving water in that year.
The total mass flux of contaminant is distributed among the watersheds in
which the fallout is deposited. The downsteam boundary flow is then used .to
calculate the resulting concentration in the receiving water:
where:
Cix - [(Fa) (WAx)]/Vfx Equation (E-l)
Cix =» estimated receiving water concentration of contaminant for
watershed x (M/L3)
2
Fa * annual mass of contaminant in fallout/unit area (M/L -T)
2
WA. =- area of watershed receiving fallout (L )
Vf = total volume of flow of watershed outlet during the period
3
of observation; i.e., 1 year for chronic exposure (L /T)
For chronic exposure, total annual flow, Vf , is applied, while for .acute
^
exposure the total flow over the duration of the storm event, VS , is
A
substituted (e.g., VSX - Vfx/365), if the 24-hour rainfall event is
evaluated. (Ideally, VS is the actual hydrograph velocity for the event if
it is. known.) This very conservative calculation accounts for no losses
during transport and, therefore, overpredicts contaminant levels. If these
overpredictions do not result in excessive concentrations, no further
analysis is required. If the concentration predictions are high, a more
detailed Tier 2 or 3 analysis is required to determine probable receiving
water concentrations.
E-3
-------
Calcvlat* T»U1 Matt tf
CaMtMtMNt
Mttrlfcitt Ite taeiif Tfct
dt Mhtrt Fallwt
flccvr
Calculate Total Awni«1 Velim «f n«v
At th« Oomttrttn lountftry »f Uch
Mf»et»dJ«t*r»h»< (»*.)
bit
Uwl
F«r Cwitavfnanu In Cat*
fC-1
L
•»» Out t»
1* tart >Ut«r»hH
SUr« Event Fli
'•r CacO iaurrted
C«1cn1at* CwitMlMfit
T» Each
tt font tetflotnt IM»
for C>et»
Htw At OoMn«tft«B
Of Cacft *»t»m»*< (Tf.)
OT l*tM*n
Calenlttt Mass «f Hniculatt and
Tctal twnt n«v In
ttattr (VM)
Figure E-l. Surface Runoff Pathway Methodology
E-4
-------
Tier 2 and 3 calculations are identical. They differ only in the
origin of input values. Tier 3 is based on site-specific data from
empirical observations for such parameters as degradation rate and partition
coefficient. For the long-term (chronic) exposure analysis, receiving water
contaminant inputs are calculated from estimates of the soil contaminant
concentration and the bulk soil transport to the receiving water.
In order to estimate soil contaminant concentration, it is assumed that
the municipal waste combustors are in operation long enough for a
steady-state concentration to be reached. By definition, these conditions
will prevail when soil levels are high enough for the sum of zeroth and
first-order losses to equal the annual addition of contaminants in fallout.
Maximum soil contaminant mass per area during the period is calculated as
fol1ows:
k-d-S^l*)
Mm - — Equation (E-2)
ki
where:
2
Mm =- maximum contaminant mass per area of soil (M/L )
k~ = annual fallout rate for contaminant less any zeroth order
2
losses (typically none) (M/L -T)
k, = first-order loss rate includes infiltration losses, which
are controlled by partitioning, degradation and erosion
losses (T"2)
t = time span for analysis, typically the life of the combustor
and its replacements
The first-order loss mechanisms can be calculated as follows:
1) Infiltration (if equilibrium between soil and water is assumed)
Kn - Rc/(B d Kd) Equation (E-3)
where:
kn =• first-order loss rate coefficient for infiltration (T"1)
Re = recharge- (L/T)
E-5
-------
B - the bulk density (M/L3)
d = depth of incorporation (L)
Kd - the distribution coefficient (L3/M)
Once again, if infiltrate concentrations are thought to be solubility
limited, a zeroth order rate is more appropriate.
2) Surface runoff
klR = V(B d) Equation (E-4)
where :
^1R a first-order loss rate coefficient for surface runoff losses (T~ )
Xe = sediment loss rate (M/L2-T)
B = bulk density (M/L3)
d .= depth of incorporation (L).
3) Degradation
k1D
where:
In2/tj,2 Equation '(E-5)
kjg - first-order loss rate for degradation (T )
tl/2 = half-life due to degradation (T).
In general, volatilization would also represent a first-order loss
mechanism. For part icul ate fallout, however, it is assumed that volatile
species will not be present in the solids settling from the atmosphere.
Even if vapor pressures are measurable, the adsorption phenomena will
reduce their significance. Also, when field measurements are available for
deriving degradation rates, they may reflect all. first-order losses (k^) and
not just degradation (LCyj). Therefore, care must be taken in selecting
input values.
If tilling is prevalent in the watershed, concentrations are reduced by
dividing by d (depth of tilling in cm) to reflect that runoff only affects
the top centimeter of soil. When no tilling is practiced, d is set at
1.0 cm.
E-6
-------
Sediment losses to the receiving water are computed using the Universal
Soil Loss Equation (USLE):2
Xe - R K (LS) C P Equation (E-6)
where:
Xe - sediment loss rate (M/L2-T)
R - "erosivity" factor (T"1)
K - "erodability" factor (M/L2-T-unit 'R')
LS = "topographic or slope length" factor (dimensionless)
C = "cover management" factor (dimensionless)
P = "supporting practice" factor (dimensionless)
Guidelines for selection of input factor values for the USLE in each
watershed are provided by Wischmeier and Smith as well as the detailed
3 4
runoff methodology discussion. '
The total annual flux of sediment to a receiving water is the product
of the unit area sediment loss, X , and the area of the watershed, VIA . The
G A
concentration of the contaminant in the receiving water (Ci ) is determined
as the ratio of the .total contaminant flux and the annual volume of flow at
the downstream watershed boundary, Vf , or:
C1v " xo WAv Mm/(d vfv B) Equation (E-7)
A C J\ &
where:
Xa = sediment loss rate (M/L2-T)
2
WAV = area of watershed receiving fallout (L )
2
Mm = maximum contaminant mass per area of soil (M/L )
d = depth of tilling (L)
Vf = total volume of flow at watershed outlet during the
3
period of observation (L /T)
B » bulk density of soil (M/L3)
The point for calculating the annual flow should be the furthest downstream
point where the contaminated watershed feeds the receiving water. For lakes
or estuaries, the outlet flow is applied as Vf. Once again, these
» **
calculations are made for each affected watershed or subwatershed unit.
E-7
-------
For acute exposures, the methodology focuses on a specific storm event.
To estimate average contaminant mass per area of soil, it is once again
assumed that soil levels will increase until balanced by losses. The
maximum level (Mm) at any time (t) after the combustion is initiated can be
estimated as:
Ml-e'*!*)
Mm - — Equation (E-8)
ki
where the terms are the same as defined in Equation (E-2)- Once again, if
tilling is practiced to a depth of d cm, the contaminant mass per unit
volume of soil (Mm', in M/L3) is:
Ml-e'V)
Mm' = Mm/d = ——
dkl
Sediment loss to the receiving water is calculated from the Modified
Universal Loss Equation (MUSLE). ' In this approach, it is necessary first
to select a watershed retention factor (S) from the Soil Conservation
Service .(SCS) runoff curve number (CN) according to:
S = 2.54[(1000/CN)-10] Equation (E-9)
where:
S = watershed retention factor (cm)
Note that the units for S are centimeters and that S should be converted to
the proper length units for the remainder of the calculations. Using length
units of centimeters for D^, R^, Mt and S is preferred. Next, a runoff
depth value is estimated as:
(Rt + Mt - 0.2S)2
D0 - -r-± 5 Equation (E-10)
(Rt + Mt + 0.8S)
where:
D^ » depth of runoff in the watershed (L)
Rt - depth of total rainfall for the storm event (L)
Mt - depth of snowmelt during the storm event (L)
S - the watershed retention parameter (L)
E-8
-------
The storm runoff volume is calculated from the runoff depth according to:
Q - (WAX)DR Equation (E-ll)
where:
2
WAV » area of the watershed receiving fallout (L )
3
Q » volume of runoff (L )
DH « depth of runoff from the storm event (L)
A trapezoidal hydrograph is assumed so that the peak runoff rate can be
calculated as:
(WA ) DR Rt
q , * K_c_ Equation (E-12)
p Tr(Rt - 0.2S)
where:
q_ = peak runoff rate (L /T)
P 2
WA » area of the watershed (L )
A
Rt » depth of rainfall in the storm event (L)
DR - depth of runoff from the storm event (i)
Tr -duration of the storm event (T)
S = water retention parameter (L)
Sediment losses from the storm event are estimated with the MUSLE according
to:
Xs = 11.8(Q q )°'56K (LS) C P Equation (E-13)
where:
X = sediment loss from a single storm (metric tons)
3
Q =» volume of runoff (m )
q - peak runoff (m /sec)
K - "erodability" factor (tons/acre-year-unit "R")
LS - "topographic or slope/length" factor (dimensionless)
C = "cover management" factor (dimensionless)
P = "supporting practice" factor (dimensionless)
7 8
Again, selection of K, LS, C and P is discussed elsewhere. ' This equation
is an empirical relationship and the units must be consistent with those
shown above.
E-9
-------
If the risk evaluation is to be based on total contaminant exposure,
receiving water concentrations are calculated as in Equation (E-7):
Ci - Xc WAY Mm'/VSY B Equation (E-14)
A w A A
using the maximum soil contaminant level Mm' in place of the average soil
contaminant level MC. The value for VS (used instead of Vf ) would be the
A A
total volume of flow of the receiving water during the storm event rather
than a 1-year period. If criteria for risk differentiate between dissolved
and particulate contaminant, then the total mass of contaminant Mm' must be
partitioned between the adsorbed portion, Aa, and the dissolved portion, Da.
These are derived as:
Aa = [l/(l+(e/Kd B))]Mm' Equation (E-15)
and
Da - [l/(l+(Kd B/e))]Mm' Equation (E-16)
where:
2
Aa =» adsorbed contaminant mass in top cm of soil (M/L -L)
2
Da - dissolved contaminant mass in top cm of soil (M/L -L)
e = available volumetric water capacity of the top cm of soil
difference between wilting point and field capacity)
(dimensionless)
Kd = distribution coefficient for contaminant in soil water system
(L3/M)
B = bulk density of soil (M/L3)
Mm' = maximum level of contaminant in top centimeter of watershed
soil (M/L2-L)
The contaminant losses by each route are defined asr
Pxt - [XS/(WAX B)]Aa Equation (E-17)
and
P t - [DRARt + Mt)lDa Equation (E-18)
where:
2
P . » loss of contaminant in adsorbed form (M/L )
P j. - loss of contaminant in dissolved form (M/L )
Other terms are as defined in Equations (E-10), (E-14), (E-15) and (E-16).
For these cases, the receiving water concentrations would be derived from:
E-10
-------
Cix = Pxt WAX/VSX Equation (E-19)
and
Cix = Pqt WAX/VSX Equation (E-20)
where:
Ci = concentration in receiving water (M/L )
P^ and Pv+ - loss rates from Equations (E-17) and (E-18) (M/L2)
qt xi 2
WA =» area of watershed (L )
^
VS,, = total volumetric flow of receiving water during the
3
storm event (L )
E.4 REQUIRED INPUTS
For a Tier 1 analysis, the only input parameters required are the
annual mass of contaminant fallout on each"watershed or subwatershed unit
and the annual flow of the receiving water.
The equations for Tier 2 and Tier 3 analysis are the same; therefore,
they have the same input parameters. The difference is that many of the
parameters used in a Tier 2 analysis would be obtained from the literature, •
whereas Tier 3 requires'all site-specific input. All the input parameters
required for the runoff pathway analysis are shown in Table E-2. The only
input required for the receiving water analysis is the stream flow or the
flow into a lake or estuary.
E..5- EXAMPLE CALCULATION
The methodology presented above can best be illustrated with example
calculations. Two site-specific examples are provided below, one for a
long-term average case (chronic) and one for an event-loading (acute case).
For both cases, Tier 1 and Tier 2/3 analyses are made for two contaminants,
benzo(a)pyrene and cadmium. Contaminant deposition patterns were modeled
for emissions representative of a municipal waste combustor with an
electrostatic precipitator.
The annual mass of contaminant fallout (Fa) was calculated by averaging
deposition estimates for a given area over a 5-year period. Deposition
rates were estimated at points on rays spaced every 22.5° at ranges varying
E-ll
-------
TABLE E-2. INPUT PARAMETERS FOR THE RUNOFF PATHWAY METHODOLOGY
Symbol Function
foc Soil organic content (dlmenslonless)
UAX Area of land 1n a given watershed on which fallout will be
deposited (km*)
Fa Annual fallout rate/watershed unit (dry kg/ha-year)
t Time period over which facility and Its replacement will
operate (years)
Rc Recharge rate (in/year)
B Bulk density of the soil (g/cm3)
d Depth of soil Incorporation*
HI Total event snow melt (cm)
Rt Total storm rainfall depth (cm)
Tr Duration of storm event (hours)
Kd Distribution coefficient (craVg)
Vfx Total volume streamflow for the receiving water (i/year)
VSX Event streamflow volume (i/event)
R. K. LS. Input variables for the Universal Soil Loss Equation (USLE)
C and P
e Soil porosity (dlmenslonless)
t]/2 Contaminant half-life (I/years)
*d 1s used even If not tilled
E-12
-------
from 200 to 50,000 m from the municipal waste combustor. Hence, the sample
points form concentric rings with the combustor at their center. The
average annual mass of contaminant fallout was obtained by integrating the
2
measured data over the watershed area (1 km) being studied. The double
integral:
1AR2 llf(r,e) r dr de
where:
f(500,e) - f(200,e)
f(r,e) r + 5/3f(200,e)2/3f(500,e)
300
was evaluated for all data points on the 200 and 500 m rings (the area
2 •
within the 500 m ring is roughly the size of 1 km watershed). Integrating
r's from 0 to 500 m and e's from 0 to 2n-, the resulting average benzo(a)-
pyrene and cadmium fallout was 0.0020779 kg/ha-year and 0.039436 kg/ha-year,
respectively.
It is assumed that the particulate emitted by the municipal waste
combustor is spread over various watersheds; however, for the purpose of
illustration, the calculations are made for only a single watershed of an
2
area roughly 1 km that contains the MWC facility at its center.
All input parameters for the example calculations are shown in
Table E-3. Data values are for illustrative purposes only, but are, for
the most part, representative of a real site.
E.5.1 Tier 1
E.5.1.1 Lonq-Term Average Loading. For a Tier 1 analysis, calculate
the maximum possible receiving water concentration using Equation (E-l).
For benzo(a)pyrene:
Cix - [108(Fax x WAx)]Vfx
-108(0.0020779 kg/ha x 1.0 km2)/3.15 x 1010 I/year
- 6.5965 x 10"6 mg/1
For cadmium:
Cix - [108(Fax x WAx)]Vfx
'= 108(0.0039436 kg/ha x 1.0 km2)/3.15 x 1010 I/year
= 0.0001252 mg/1
E-13
-------
TABLE E-3. INPUT PARAMETERS FOR THE EXAMPLE CALCULATIONS
Location - western Florida
Soil Type - sandy clay loam
Organic matter content (foc) - 4%
Land use - agricultural, orchards
UA - watershed area. 1 km2 (100 ha)
Fa - annual mass fallout/unit area:
benzo(a)pyrene - 0.0020779 kg/ha-year
cadmium - 0.039436 kg/ha-year
t - elapsed time. 30 years
Re - recharge rate. 0.2S ra/year
Vfx - annual streamflow volume. 3.15x10*° t/year
8 - soil bulk density. 1.5 g/cm*
d - Incorporation depth. 1 cm
k0 . zeroth order loss rate. 0.0
k-j . first-order loss rate
benzo(a)pyrene - degradation * Infiltration and surface runoff,
0.278 year'*
cadmium - Infiltration * surface runoff, 0.168 year"1
k2 - Fa-k0 « Fa
LISLE Input Parameters:
R - 400 year"1
K - 0.21 tons/acre/year/unit R»0.21 tons/acre
LS - 0.179
C - 0.5
P - 1.0
CM - 78
RI - total event storm rainfall, 5 cm
HI - total event snow melt. 0 cm
Tr - storm duration. 6 hours
VSX - event/storm streamflow volume, 8.63xl07 I/event
Kd - partition coefficient:
benzo(a)pyrene - 3000 cmVg
cadmium - 300 cm»/g
6 - porosity. 0.2
E-14
-------
E.5.1.2 Event Loading. For event loading, V$x is defined as the total
volume of flow during the selected storm event rather than the 1-year flow
volume (Vf ) used for the long-term analysis. For this illustration, a
A
24-hour event is assumed. Therefore, the flow volume would be:
8.63 x 107 1 (3.15 x 1010 I/year -=- 365 day/year)
The concentration of benzo(a)pyrene in the receiving water would be:
Cix = [108 (Fax)(WAx)]/VSx
= 108(0.0020779 kg/ha)(1 km2)/8.63 x 107 I/event)
=0.002408 rag/1
The concentration of cadmium in the receiving water would be:
Cix = [108(Fax)(WAx)]/VSx
= 108(0.0039436 kg/ha)(1 km2)/8.63 x 107 I/event)
= 0.04570 mg/1
E.5.2 Tier 2/3
E.5.2.1 Lonq-Term Loading. To estimate soil contaminant concentration
it is first necessary to estimate rate constants for contaminant loss from
soils. For a degradable contaminant, k, is the combined first-order loss
rate constant for degradation, infiltration and surface runoff. For a
nondegradable contaminant, k, is based on the loss due to infiltration and
surface runoff only. For cadmium (nondegradable), k, was calculated as
fol1ows:
E-15
-------
1) Infiltration
k,T - Rc/(B d Kd)
- 0.25 m/year/(1.5 g/cm3)(l cm) (300 cm3/g)
0.0556/year
2) Surface Runoff
klR • V(B d)
- 10"4 [1688 mt/km2 -year/ (1.5 g/cm3)(l cm)]
- 0.1125/year
\f-\e j. U
.1 ~ K1I K1R
= 0.0556/year + 0.1125/year
= 0.168/year
For benzo(a)pyrene (degradable) a biodegradation rate in soil of 0.16/year
9
was estimated from a recent study, and k, was calculated as follows:
1) Infiltration
kn = Rc/(B d Kd)
=0.25 m/year/(1.5 g/cm3)(l cm) (3000 cm3/g)
= 0.0056/year
2) Surface Runoff
Same as 2) above where k1R » 0.1125/year
3) Degradation
kjp - 0.16/year
kl * kll + klR + klD
- 0.0056/year + 0.1125/year + 0.16/year
- 0.278/year
E-16
-------
Following estimation of degradation rates, the soil contaminant
concentration may be estimated.
For a degradable contaminant, such as benzo(a)pyrene, the soil
contaminant concentration can be estimated with Equation (E-2) as:
tyT-e'V)
Mm * —
kl
= 0.0020779 kg/ha-year [(l-e"<°-278 year"^ (3° years)]
0.278 year"
- 0.007473 kg/ha
For a nondegradable contaminant, cadmium, the maximum soil contaminant
concentration.can be estimated with Equation (£-2) as:
k2(l-e~klt)
Mm = —
kl
= 0.039436 kg/ha-year
0.278 year"1
= 0.23322 kg/ha
Bulk soil losses to the receiving water are calculated using
Equation (E-6).as follows:
Xe = 224.64(R)(K)(LS)(C)(P)
= (224.64)(400/year)(0.21 tons/acre/year)(0.179)(0.5)(1.0)
2
- 1688 mt/km -year
The total annual flux of sediment to a receiving water is the product
of the unit area sediment loss (Xe) and the area of the watershed (1 km2).
The contaminant concentration of benzo(a)pyrene in the receiving water is
calculated with Equation (E-2) as:
E-17
-------
(104)(X )(WA )(Mm)
Ci - — : - - - * -
X (d)(Vfx)(B)
(104)(1688 mt/km2-year)(l km2) (0.007473 kg/ha)
(1 cm)(3.15 x 1010 l/year)(1.5 g/cm3)
- 2.669 x 10"6 mg/1
For cadmium:
(104)(X )(WA )(Mm)
Ci - - -= - - -
X (d)(Vfx)(B)
(104)(1688 mt/km2-year)(l km2) (0.23322 kg/ha)
(1 cm) (3. 15 x 1010 I/year) (1.5 g/cm3)
- 8.332 x 10"6 mg/V
E.5.2.2 Event Loading. For the degradable contaminant, benzo(a)
*
pyrerie, the maximum contaminant level in the soil at any time can be
calculated as above using Equation E-8 as:
Mm » —
0.0020779 kg/ha-year [(l-e^0'278 ^r' ^30 years)]
0.278 year
7.473 x 10"3 kg/ha
If a soil incorporation depth of 1 cm is assumed, then:
Mm' - Mm/(l cm)
Mm' - 7.473 x 10"3 kg/ha.cm
E-18
-------
For the nondegradable contaminant cadmium:
Ml-a'V)
Mm - ~
kl
0.039436 kg/ha-year [(l-e^0'168 ^r'^ <30 years)]
0.168 year"
- 0.2332 kg/ha
Mm' - 0.2332 kg/ha.cm
where CN corresponds to hydrologic soil- Group B, moderately low runoff
potential, good hydrologic conditions and straight row crops.
The runoff depth can be calculated using Equation (E-10) as:
(R, + Mt - 0.2S)2
Q „ V. «•
[5 cm -t-. 0 - (0.2)(7.16 cm)]2
36 .1 i i
5 cm + 0 4- (0.8)(7.16 cm)
- 1.19 cm
Calculate the storm runoff volume from Equation (E-ll) as:
Q - 104(WAX)0R
- (104)(1 km2)(1.19 cm)
- 1.19 x 104 m3
E-19
-------
Calculate the peak runoff rate from Equation (E-12) as:
(2.78)(WAx)(Rt)(DR)
qp " rT (R - 0.2S)
(2.78)(1 km2)(5 cm)(1.19 cm)
3 — ^— —— — — — — — ^ ^— ^— ^— ^—
(6 hours) [5 cm - 0.2 (7.16 cm)]
=» 0.77 m /sec
The sediment losses from the storm event are calculated from the MUSLE
Equation (E-13) as:
Xs = 1.8 (Q qp)°'56 (K) (LS) (C) (P)
- 11. 8[(1. 19x10* m3)(0.77 m3/sec)]°'56(0.21 tons/acre-year)(0.179)(0.5)(1.0)
- 36.7 mt
The adsorbed and dissolved fractions can be calculated as follows.
First, the total mass of contaminant, Mm', is divided into the adsorbed and
dissolved portions using Equations (E-15) and (E-16).
Benzo(a)pvrene
Aa = [l/(l+(e/KdB))] Mm*
- [1/1+(0.2/(3000 cm3/g)(1.5 g/cm3))] 7.473 x 10"3 kg/ha.cm
=» 7.473 x 10"3 kg/ha.cm
Da = [l/(l+(KdB/e))] Mm'
= [1/(1+((3000 i
=0.00 kg/ha«cm
[1/(1+((3000 cm3/g)(1.5 g/cm3)/(0.2))] 7.473 x 10"3 kg/ha.cm
E-20
-------
[!/(!+((300 cm3/g)(1.5.g/cm3))] 0.2332 kg/ha.cm
Cadmium
Aa = [l/(l+(e/KdB))] Mm'
• [!/(!+((300 cm3,
- 0.2331 kg/ha»cm
Da - [l/(l+(KdB/e))] Mm'
- [I/(1+((300 cm3
=» 0.0001 kg/ha»cm
cm3/g)(1.5 g/cm3)/(0.2))] 0.2332 kg/ha.cm
Based on these calculations, virtually all the contaminant loss for
both benzo(a)pyrene and cadmium is in the adsorbed form. The losses for
each route, adsorbed (Pxt) and dissolved PQt), are defined by
Equations (E-17) and (E-18) as follows: "
Benzo(a)pvrene
Pxt = [10"4V(WAx)(B)1 Aa
- [(10~4)(36.7 mt)/(.l km2)(1.5 g/cm3)] 7.473 x 10"3 kg/ha.cm
-- 1.828 x 10"5 kg/ha
Pqt= [V(Rt^Mt^ Da
= [1.19 cm/(5 cm + 0 cm)] 0.00 kg/ha«cm
=0.00 kg/ha
Cadmium
Pxt = nO"4Xe/(WAx)(B)] Aa
- [(10"4)(36.7 mt)/(l km2) (1.5 g/cm3)] 0.2331 kg/ha.cm
- 5.7032 x 10"4 kg/ha
Pqt = tDR/(Rt + Mt)] Da
- [1.19 cm/(5 cm + 0 cm)] 0.0001 kg/ha»cm
- 2.3800 x 10"5 kg/ha
E-21
-------
Once P t and P t are known, the receiving water concentrations due to
adsorbed and dissolved contaminant, respectively, can be calculated from
Equations (E-19) and (E-20) as follows:
Benzo(a)pvrene
(Pxt)(WAx)
C1x
VSx
(108)(1.828 x 10"5 kg/ha)(l km2)
(8.63 x 107 1)
2.1182 x 10~5 mg/1
(Pat)(WAx)
=
..
sx
(108)(0.00 kg/ha)(1 km2)
(8.63 x 107 1)
0.00 mg/1
Cadmium
(Pxt)(WAx)
Cix
(108)(5.7032 x 1Q"4 kg/ha)(1 km2)
(8.63 x 107 1)
6.6086 x 10"4 mg/1
E-22
-------
vsx
(108)(2.3800 x 10"5 kg/ha)(1 km2)
(8.63 x 107 1)
2.7578 x 10"5 mg/1
The Ci for cadmium represents the maximum receiving water
concentration to which humans would be exposed.
E-23
-------
E.6 REFERENCES FOR APPENDIX E
1. U. S. Environmental Protection Agency. Development of Risk Assessment
Methodology for Land Application and Distribution and Marketing of
Municipal Sludge. Draft Final Report. 1986.
2. Wischmeier, W. H., and D. D. Smith. Predicting Rainfall Erosion
Losses - A Guide to Conservation Planning. USDA Handbook No. 537.
1978.
3. Reference 2.
4. Reference 1.
5. Williams. F. R. Sediment and Yield Prediction with Universal Equating
Using Runoff Energy Factor. Present and Prospective Technology for
Predicting Sediment Yields and Sources. USDA ARS-S-40. 1975.
6. Haith, D. A. A Mathematical Model for Estimating Pesticide Losses in
Runoff. J. of Environ. Qual. 9:428-433. 1980.
7. Reference 2.
8. Reference 1.
9. Bossert, D. D., and R. .Bartha. Structure.- Biodegradability
Relationships of Polycyclic Aromatic Hydrocarbons in Soil. Bulletin
of Environ. Contam. Toxicol. 37:490-495. 1986.
E-24
-------
APPENDIX F
GROUNDWATER INFILTRATION MODEL
-------
APPENDIX F
GROUNDWATER INFILTRATION MODEL
F.I INTRODUCTION
Contaminants associated with participates emitted from municipal waste
combustors are subject to deposition on surfaces downwind from the facility.
This fallout is subsequently subject to dissolution in rain or meltwater
from precipitation events. The dissolved portion can follow one of two
pathways: either move over the surface as runoff to a surface water body or
infiltrate into the ground and recharge the groundwater. As a consequence,
persons using the groundwater may be exposed to groundwater transported
contaminants. Aquatic life inhabiting surface water bodies fed by the
contaminated aquifer could be exposed as well.
The methodology derived to calculate exposure from the groundwater
pathway was.originally developed to evaluate impacts from the .landfill ing of
municipal sludge and to evaluate the groundwater pathway associated with the
application of sludges to land. A detailed discussion is available in the
document entitled "Development of Risk Assessment Methodology for Municipal
Sludge Landfilling". It is formulated in three successive tiers, which
begin with simple but very conservative estimates and proceed to more
detailed analyses if the first tiers predict unacceptable risks. Only
chronic exposure is evaluated using standard approaches to calculate
leachate generation and associated groundwater transport in the unsaturated
zone.
F.2 ASSUMPTIONS
A number of assumptions were required to formulate the risk-based
methodology both with respect to leachate generation and subsequent
transport in the unsaturated zone. The key assumptions along with a
discussion of their impacts on the methodology are provided in Table F-l.
F-l
-------
TABLE F-l. ASSUMPTIONS FOR THE GROUNDWATER PATHWAY METHODOLOGY
Functional
Area
Assumption
Ramifications
Source term
Unsaturated
zone transport
Equilibrium will be reached
wherein annual Inputs from
fallout are lost In re-
charge as leachate
Leachate pattern 1s level
over the entire year.
One-dimensional flow In
the vertical direction.
Hater flow Is steady-state.
Upper boundary has a con-
stant flux of recharge.
Soil characteristics are
constant with depth for
any layer.
Vertical hydraulic grad-
ient of unity.(not assumed
In one of two alternative
approaches).
Overpredlcts to the extent
that some contaminants are
Irreversibly bound to
solids and neglects other
loss mechanisms such as
runoff.
Underpredlcts peak concen-
trations but provides
estimates of average
annual loading.
Overpredlcts concentration
since 1t Ignores horizontal
dispersion.
Overpredlcts concentration
by accelerating flow In a
compressed time period.
Overpredlcts or under-
predlcts, depending on
the soil type.
It 1s Impossible to deter-
mine the effect of this
assumption., since 1t will
vary from site to site.
Overpredlcts concentration
to the extent that grad-
ients may be <1.0 and,
therefore, time of travel
Is slower, allowing for
more degradation.
F-2
-------
TABLE F-l. (Continued)
Functional
Area
Assumption
Ramifications
Unsaturated
zone transport
(cont.)
Saturated zone
(MINTEQ
Calculation)
Retardation of organic* Is
related to soil organic
fraction only.
All adsorption 1s revers-
ible.
First-order degradation
mechanism.
Groundwater conditions
dictate geochemistry.
The six groundwater pH-Eh
couplets modeled provide
an adequate set of alter-
natives.
Contaminant Interactions do
not affect geochemistry.
Overpredlcts contaminant
velocity for soils with
low organic content where
mineral Interaction may
predominate.
Overpredlcts concentration
arriving at aquifer.
Nlspredlcts degradation
where higher order rates
are functional: Overpre-
dlcts zero order mecha-
nisms.
Effects would be site-
specific depending on the
quantity and quality of
both the leachate and the
aquifer.
Some sites could have
extreme conditions beyond
those modeled. If pH
values are very low, the
model will underpredlct,
and If Eh conditions are
very low, the model will
overpredlct contaminant
concentrations.
Effects would be source-
specific.
F-3
-------
Due to the Inexact nature of the science, any assumptions made are
conservative (i.e., overpredict contaminant concentrations). In some
instances, however, the nature of the effect of a given assumption may vary
with site-specific considerations.
F.3 CALCULATIONS
The methodology for evaluating the groundwater pathway is illustrated
in Figure F-l. Tier 1 involves projecting leachate concentrations.
Leachate concentrations are predicted on the basis of annual fallout and
recharge rates. It is assumed that soil contaminant levels will increase
over time until leachate strength is sufficiently high to deplete the input
from fallout each year. Hence, leachate concentration is defined as:
X< - Fa/R Equation (F-l)
I V,*
where:
X. = leachate contaminant concentrations (M/L )
. 2
Fa = annual fallout flux of contaminant/unit area (M/L -T)
R = annual depth of recharge (L/T).
Any contaminant found to have an insignificant leachate concentration can be
eliminated from further consideration. Those with high leachate
concentrations are carried forward to a Tier 2 or 3 analysis. Currently,
the methodology can simulate geochemistry for seven metal contaminants
(Table F-2). For other inorganics, it is necessary to assume that no
precipitation reactions will occur.
The annual mass of contaminant fallout (Fa) should be calculated as the
average value that falls within a 200 m radius of the combustor. The 200 m
distance was chosen because it is the closest to the combustor that model
estimates can make accurately. In addition, the values at 200 m showed the
greatest deposition (most conservative). Also, a watershed of this size is
large enough to provide a potable water supply to several families (based on
a recharge rate of 0.25 m/year and the assumption that each person uses
100 gallons/day). Therefore, it is not unreasonable to assume that a single
well can be supplied by an area this size.
F-4
-------
For Each Contaminant 1
Calculate Contaminant Concentration
as Fallout Flux/Recharge
Tier 1
Health
Criteria
(xi)e > Health
Criteria
No
End
Distribution
Coefficient
Loss Rate
Yes
Determine Time of Travel and
Losses in Unsaturated Zone
Tier 3
Experimentally Determine
Retardation and Degradation
Values
Inorganic
Contaminant
Geochemical
Considerations
Determine Concentration in the
Aquifer Based on MINTEQ Curves
Tier 2
Yes
e Is
(xi)e > Health
Criteria
No
End
(xi)'
Concentration of Contaminant 1
FIGURE F-l. LOGIC FLOW FOR GROUNDWATER PATHWAY EVALUATION
F-5
-------
TABLE F-2. METAL CONTAMINANTS SIMULATED IN THE GEOCHEMICAL
PORTION OF THE GROUNDWATER PATHWAY
1. Arsenic
2. Cadmiurn
3. Chromiurn
4, Copper
5. Lead
6. Mercury
7. Nickel
F-6
-------
Presumably, the compliance point falls within the zone of deposition.
Hence, no saturated zone calculations are performed. The compliance point
is assumed to be that point at the base of the unsaturated zone where
leachate enters the saturated zone. For metals, the contaminant
concentration in the groundwater is adjusted based on geochemical reactions.
For organics, the contaminant concentration is adjusted due to degradation
as it travels through the unsaturated zone.
The Tier 2/3 analysis begins at the same point as Tier 1, calculation
of the source-term strength of the leachate from the fallout. Tier 2/3,
however,'allows for site-specific inputs to predict dispersion, degradation,
and retardation effects, which reduce resultant exposure levels. The
difference between Tiers 2 and 3 lies in the number of input parameters,
which are determined experimentally. In Tier 3, degradation rates and
retardation coefficients are measured directly.
The initial step in Tier 2/3 is to define the strength of contaminant
in recharge water. As with Tier 1, this is done by assuming that
equilibrium will be reached wherein the inputs from fallout each year will
be transported away from the soil in leachate. Equation (Frl) is used to
calculate the average leachate concentration over the fife of the facility
and its replacements.
For degradable contaminants, Equation (F-l) is modified to :
X1 - Fa(l-e"kt)/t k RC Equation (F-2)
where:
X. = average leachate contaminant concentration (M/L )
- 2
Fa = fallout contaminant flux (M/L -T)
k » first-order degradation rate for contaminant (T" )
t = 1 year (T)
RC = annual depth of recharge (L/T)
The next step is to calculate the time required for leachate thus
formed to move downward through the unsaturated zone to the aquifer. This
is accomplished by making a time of travel calculation with one of two
2
analytic approaches. The selected approach assumes constant moisture
F-7
-------
content In the soil with steady-state flow and a unit hydraulic gradient.
It uses the soil moisture, pressure and conductivity relationships described
by Campbell to solve for the moisture content of the soil:
f - fs(q/KSAT)1/(2b+3) Equation (F-3)
where:
f - field moisture content (L /L )
q = moisture flux, recharge rate in this case (L/T)
KJ.J =» saturated hydraulic conductivity of the soil in the
unsaturated zone (L/T)
fs » saturated.moisture content of the soil (L /L )
b - negative one times the slope of the log-log plot of matric
potential and saturated moisture content (dimensionless)
The velocity and travel times for flow in the unsaturated zone are related
according to:
V - q/f Equation (F-4.)
T =• hy/V = (hy)(f)/q Equation (F-5)
where:
V = velocity of flow (L/T)
T =» time of travel (T)
hy =• depth of the unsaturated zone (L)
For multiple layer systems, a travel time is calculated for each layer
and the total summed across the unsaturated zone. The total time is then
divided into the total depth to derive the equivalent velocity.
The velocity calculated from Equation (F-4) is for water traveling
through the unsaturated zone. If the contaminant interacts with the soil in
transit, it will travel at a retarded velocity, V^, defined as:
VR - V/[l + ((B/e)(Kd))] Equation (F-6)
where:
Vn » velocity of the contaminant (L/T)
V = velocity of water calculated from Equation (F-4)
F-8
-------
B - average bulk density of soil in the unsaturated column (M/L )
e - average porosity of soil in the unsaturated column
(dimensionless)
Kd » soil-water partition coefficient for contaminant (L /M)
If the contaminant is degradable, its concentration will change
according to:
x . x e[-k(hy)/VR] Equat.on {F.7)
a 1
where:
X, » contaminant concentration upon entry in the aquifer (M/L )
a
X. = concentration of leachate initially from Equation (F-2)
-2
k = first-order degradation rate for contaminant (T )
hy = depth of unsaturated zone (L) .
VR = retarded velocity form Equation F-6 (L/T)
e = base of natural logarithms, 2718 (unitless)
If the analyst wishes to account for dispersion as well as attenuation
and degradation in the uns.aturated zone; the water velocity V from.
Equation (F-4) can be input to the one-dimensional CHAIN code. The CHAIN
code is an analytical solution of the convective-dispersive transport
equation for a one-dimensional case that accounts for retardation and
degradation. The data inputs to the model include the average pore water
velocity, the dispersion coefficient, the water content, the pulse or
release time, the retardation factor, the decay rate, and several
coefficients describing the source term. The output from the code is the
concentration of the contaminant plume at the base of the unsaturated zone
for a period of time equal to several contaminant travel times through the
unsaturated zone.
For organic contaminants, the concentration calculated from
Equation (F-7) is the predicted concentration at the compliance point. For
metals, geochemistry can be considered using a series of input-output curves
generated by the MINTEQ code, which establishes the solubility limits for a
contaminant metal on the basis of groundwater (ionic) composition, pH, and
Eh conditions. The MINTEQ geochemical code can be applied to generate
predicted contaminant concentrations under selected groundwater conditions.
F-9
-------
MINTEQ Is a hybrid code that combines an efficient mathematical structure
with a large, well documented thermodynanric data base. Functionally, the
code models the mass distribution of a dissolved element between various
uncomplexed and complexed aqueous species; it also calculates the degree to
which the water is saturated with respect to the solids in the thermodynamic
data base. Adsorption, precipitation, and dissolution reactions can be
included in calculations. Detailed documentation of the MINTEQ code and
5 6 7 8
data can be found in Felmy et al. and Oeutsch and Krupka. ' ' Curves for
selected pH-Eh couplets and an explanation of how the relations were derived
Q
can be found in U.S. EPA. An example MINTEQ curve is shown in Figure F-2.
Output from the unsaturated zone [Equations (F-6) and (F-7) or the
CHAIN code and the MINTEQ input-output curves] are the predicted contaminant
concentration and timing at the compliance point. The pulse time is assumed
continuous for the period of operation of the combustor. If the output from
the Tier 2 analysis exceeds exposure-based criteria at the compliance point,
the analyst may choose to initiate a Tier 3 evaluation of inputs.
F.4 REQUIRED. INPUTS
For a Tier 1 analysis, the only data required are the contaminant
deposition rate and the recharge rate.
Both Tier 2 and Tier 3 have the same input parameter requirements. The
basic difference is that Tier 3 uses more site-specific data than Tier 2.
In Tier 2, data on hydraulic conductivity, recharge, depth to groundwater
and soil type should be determined empirically from samples. Other soil-
related properties can be selected on the basis of soil type, while
degradation rate constants and partition coefficients can be taken from the
literature. For Tier 3, the latter properties should also be determined
empirically with samples from the site.
The input parameters required for a Tier 2/3 analysis are shown in
Table F-3.
F-10
-------
f.t -4
5 1.0 —
? r
0.9
Unsaturated Zone Input Concentration (mg/1)
FIGURE F-2. EXAMPLE MINTEQ SPECIATION RESULTS FOR ENTRY OF A CONTAMINANT
INTO THE SATURATED ZONE FOR CONDITIONS OF pH - 7.0 AND Eh = 1.50 mv
F-ll
-------
TABLE F-3. INPUT PARAMETERS FOR GROUNDWATER PATHWAY
PATHWAY DATA
Symbol Source Data
Fa flux rate for contaminant fallout (mg/m'-year)
Re net recharge (m/year)
Unsaturated Zone
hy depth to groundwater (m)
X^ leachate contaminant concentration (mg/m3)
Material Type
IB Material layer thickness (m)
KSAJ saturated hydraulic conductivity (in/year)
b slope of matrix potential and moisture content plot (dimension-
less)
fs saturated soil moisture content (raVm*)
B bulk density (kg/m»)
Saturated Zone
pH groundwater pH
Eh groundwater Eh (mvolts)
CHEMICAL-SPECIF 1C DATA
Unsaturated Zone
Kd partition coefficient (I/kg)
k degradation rate constant (year'1)
F-12
-------
F.5 EXAMPLE CALCULATIONS
The methodology presented above can best be illustrated with example
calculations. A site-specific example calculation is provided below for two
contaminants, cadmium and benzo(a)pyrene. For both cases, a Tier 1 and a
Tier 2/3 analysis are illustrated.
All input parameters for the example calculation are shown in
Table F-4. Data values are for illustrative purposes only but are, for the
most part, representative of a real site. The annual mass of contaminant
fallout (Fa) was calculated by averaging electrostatic precipitator
measurements over a 5-year period. Measurements were conducted along
concentric rings at ranges varying from 200-50,000 m from the combustor on
rays spaced every 22.5°. The average Fa at the 200 m radium was chosen for
the reasons discussed in Section F-3.
F.5.1 Tier 1
For a Tier 1 analysis, the leachate contaminant concentration is
calculated using Equation (F-l) as follows:
Benzo(a)pvrene
X, = Fa/Rc
i
Cadmium
X
i
2
=» (0.33494 mg/m -year)/(0.25- m/year)
= 1.340 mg/m
Fa/Rc
(6.42288 mg/m2-year)/(0.25 m/year)
15.692 mg/m3
F-13
-------
TABLE F-4. INPUT PARAMETERS FOR THE EXAMPLE CALCULATIONS
Rx Annual depth of recharge. 0.25 in/year
Fa Fallout contaminant flux:
benzo(a)pyrene - 0.56642 mg/«a-year
cadmium - 10.880992 mg/m*-year
fs Saturated moisture content of soil. 0.4 m*/m*
q Moisture flux recharge rate. 0.25 m/year
KSAT Saturated soil hydraulic conductivity. 10* m/year
b Slope of matrix potential and moisture content plot. 4.0
hy Depth of unsaturated zone. 2 m
B Soil bulk density. 1.5 g/cm*
9 Soil porosity. 0.2
Kd Soil-water partition coefficient:
benzo(a)pyrene - 3000 cmVg
cadmium - 300 cm»/g
k First-order degradation rate:
benzo(a)pyrene - 0.16 year"1
cadmium - 0.00 year"1
pH Groundwater pH. .8.0
Eh Groundwater Eh. -200 mv
F-14
-------
F.5.2 Tier 2/3
A Tier 2/3 analysis begins at the same point as Tier 1, calculation of
the source-term strength of the leachate from the fallout. Equation (F-l)
is used to calculate the average leachate concentration over the life of the
facility.
For cadmium, which is nondegradable, Equation (F-l) is used to
calculate the average leachate contaminant concentration just as in Tier 1
above. For benzo(a)pyrene, which is degradable, Equation (F-2) is used to
calculate the average leachate contaminant concentration as follows:
X. -
(t)(k)(Rc)
=» (0.33494 mg/m2-year)[i-e-(0-16/year)(1 year)3
(1 year) (0.16/year) (0.25 m/year)
. = 1.238 mg/m3
The next step is to calculate the time required for leachate to move
downward through the unsaturated zone. First, the constant moisture content
is calculated using Equation (F-3):
f = fs _g_ 1/<2b+3>
KSAT
= 0.4 m3/m3 0.25 m/year V[2(4)+3]
T
10 m/year
- 0.15
The water velocity travel time through the unsaturated zone can be
calculated as follows:
T - (hy)(f)/q
• , - (2 m) (0.15)/(0.25
-1.2 year
F-15
-------
Since both berizo(a)pyrene and cadmium travel with a retarded velocity
compared to the groundwater, their velocities through the unsaturated zone
can be calculated with Equation (F-6) as follows:
Benzofa^pvrene
VR - V/[l+((B/e)(Kd))]
- 1.67 m/year/[!+((!.5 g/cm3/0.2)(3000 qn3/g))]
= 7.42 x 10"5 m/year
Cadmium
• VR = V/[H((B/e)(Kd))]
=1.67 m/year/[!+((!.5 g/cm3/0.2)(3000 cm3/g))]
-4
= 7.42 x 10 m/year
Since benzo(a)pyrene is degradable, its concentration will change as it
travels-through the unsaturated zone. The benzo(a)pyrene cpncentration at
the base of the unsaturated zone before it enters the aquifer can be
calculated with Equation (F-7) as follows:
x =, x e[-k(hy)/vR]
3 > 1^238 ng/n,3 .HO.M/iPt«r)(Z .)/(7.« x 10-5 ra/year)]
=0.00 mg/m
Because the benzo(a)pyrene travel time through the unsaturated zone is so
long, it degrades before it reaches the aquifer. The cadmium will not
degrade, but travels much slower (1800 times slower) than water. Assuming
no dispersion, the cadmium concentration may still be reduced due to
precipitation reactions in the saturated system. The MINTEQ results for
cadmium for pH and Eh values close to the values specified in the example
problem are shown in Figure F-3. An unsaturated zone concentration of 0.025
mg/1 for cadmium (Y axis in Figure F-3) corresponds to a concentration of
0.019 mg/1 (19 ug/1) in the saturated zone after accounting for speciation.
F-16
-------
0.2C98
pN • 8.1. Eh • ZW •»
0.01
0.125
0.250
0.499
Output Cadmium Concentration in the Saturated Zone after Speciation (mg/1)
FIGURE F-3. GROUNDWATER CADMIUM SPECIATION
F-17
-------
F.6 REFERENCES FOR APPENDIX F
1. U.S. Environmental Protection Agency. Development of Risk Assessment
Methodology for Municipal Sludge Landfill ing. Draft Final Report.
1986.
2. U.S. Environmental Protection Agency. Technical Support for
Development of Guidance on Hydrogeologic Criterion for Hazardous Waste
Management Facility Location. Draft Report. 1985.
3. Campbell, G.S. A Simple Method for Determining Unsaturated
Conductivity from Moisture Retention Data. Soil Science 117: 311-314.
1974.
4. Van Genuchton, M. Concentive-dispersive transport of Solutes Involved
in Sequential First-order Decay Reactions. Journal of Computers
Geosci. 11: 129-147. 1985.
5. Felmy, A.R., S.M. Brown, Y. Omishi, S.B. Yabusaki, and R.S. Argo.
MEXAMS - The Metals Exposure Analysis Modeling System. 1983.
6. U.S. Environmental Protection Agency. MINTEQ - A Computer Program for
Calculating Aqueous Geochemical Equilibria. EPA Report No.
600/3-84-032. 1984
7. Morrey, J.R. PRODEF - A Code to facilitate the Use .of the Geochemica.1
Code MINTEQ. Draft Report. Battelle Memorial institute. 1985.
8. Deutsch, W.J. and K.M. Krupka. MINTEQ Geochemical Code: Compilation of
Thermodynamic Database for the Aqueous Species * Gases, and Solids.
Containing Chromium, Mercury, Selenium, and Thallium. Draft Report.
Battelle Memorial Institute. 1985.
9. Reference 1.
F-18
-------
APPENDIX G
DERMAL EXPOSURE MODEL
-------
APPENDIX G
DERMAL EXPOSURE MODEL
G.I INTRODUCTION
The dermal exposure model refers to human skin contact with
contaminants from emissions of municipal waste combustors deposited on the
soil. The issue of dermal absorption of deposited contaminants is very
complex. There is a fundamental lack of data for percutaneous absorption of
chemicals in human skin from soil. Other factors important for estimation
of human exposure to contaminants by the dermal route also have many
uncertainties. The model described here is offered as a possible approach
for the estimation of human exposure and risk associated with dermal
exposure, but it recognized that iIT most, if not all cases, the available
data will not provide a satisfactory basis for risk calculations.
G.2 MOST-EXPOSED INDIVIDUALS (MEIs)
The MEI for this pathway of dermal exposure is an individual residing
within 50 km of a municipal waste combustor (in the area of maximal
deposition of emissions) who spends a majority of this daily activity
outdoors. Preschool children (between 1-6 years of age) who play outdoors
or rural farmers would most likely be the MEIs, since these groups have the
greatest opportunity for dermal exposure to particulates deposited on the
soil. These children are likely to be exposed in residential areas
(gardens, lawns, parks, etc.). Farmers or individuals who garden would also
have the potential for substantial skin contact with soil. Occupational
exposures of workers involved in the operation of municipal waste combustors
are not considered here since these workers can be required to use special
measures or equipment to minimize their exposure to possibly hazardous
materi als.
G-l
-------
G.3 DEPOSITION-HUMAN ("DERMAL") TOXICITY EXPOSURE PATHWAY
6.3.1 Assumptions
For this methodology, it will be assumed that the daily dermal intake
of the contaminant increases continually with contaminant concentration in
soil particles contacting the skin. It is assumed that 100 percent of the
.deposited contaminant particulate is retained in the uppermost 1 cm layer of
soil. The maximum concentration of deposited contaminant within 50 km of a
municipal waste combustor is used. It is further assumed that this soil
layer contacts the exposed skin of individuals involved in outdoor
activities. Consideration of dermal exposure during outdoor activity to
contaminated particulate deposited on soil most likely accounts for a more
substantial exposure to the individual than contaminated particulate
deposited in indoor (residential) dust. Indoor dust is therefore excluded
from this model at the present time.
Dermal intake of contaminants is assumed to be a function of the
fraction of the compound absorbed, contact time (or duration of daily
exposure), exposed skin surface area, contact amount (amount of soil
accumulated on skin) and soil contaminant concentration.
Calculation of the daily intake by. dermal exposure is the same for
organics and inorganics, since there is immediate exposure potential and
therefore no soil incorporation. Background concentration of contaminants
is not included because this methodology assesses only the risk associated
with the increase in exposure due to the contaminant from municipal waste
combustor emissions. The assumptions and uncertainties relevant to this
model and their ramifications/limitations are shown in Table G-l.
G.3.2 Calculation Method
A daily dermal intake (DDI, ug/day) is determined as follows:
DDI - Ct x SA x CA x AF x LCT x (10"3/24) x EDA Equation (G-l)
G-2
-------
TABLE 6-1. ASSUMPTIONS AND UNCERTAINTIES FOR DERMAL EXPOSURE MODEL
Functional
Area
Assumpti on/Uncertai nty
Ramifications/Limitations
Exposure
assessment
Dermal
Intake
Fraction of
contaminant
absorbed
Contact time
Deposited emissions are not
necessarily soil incor-
porated and may concentrate
in the uppermost soil
layer. Deposited contam-
inant is assumed to be
distributed within the
uppermost 1 cm of soil, and
the soil which contacts the
skin is assumed to
originate from the same
1 cm layer.
Linear with respect to soil
concentration.
Chemical and matrix
specific.
Data needed for dermal
absorption in humans may
not be available.
Dermal absorption varies
with age.
May be concentration
dependent.
The MEIs are exposed to
contaminated particulate
12 hours/day during
outdoor activity.
Overestimates exposure in
situations where.soil
incorporation occurs to
any depth >1 cm.
May overestimate exposure.
Uncertainty for many
contaminants and matrices.
Uncertainty in between
species extrapolation for
dermal exposure. Uncer-
tainty in route-to-route
extrapolation of
absorption.
Absorption data for a
contaminant at one age may
over- or underestimate
actual absorption at
another concentration.
Study data at one concen-
tration may over- or
underestimate actual
absorption at another
concentration.
May underestimate exposure
by excluding indoor expo-
sure to dust.
G-3
-------
TABLE G-l. (Continued)
Functional
Area
Assumpti on/Uncertainty
Ramifications/Limitations
Exposed skin
surface area
Soil contact
amount
Effects
assessment
Total intake is propor-
tional to exposed surface
area; adults = 2940 cm;
children = 980 cm.
2
Children =1.5 mg/cm ; -
Adults - 1.5 - 3.5 mg/cm.
Substantial dermal exposure
may occur only from 1 to 6
years of age. Cancer potency
is adjusted to reflect a
5-year rather than 70-year
exposure.
Lifetime exposure may need
to be adjusted for days/
year in contact with
contaminant during 70-year
lifetime.
May over- or underestimate
intake.
May over- or underestimate
exposure.
Hazard could be under-
estimated if child is more
susceptible to chemical
carcinogenesis than an
adult and 5/70 is used as
an adjustment factor.
Uncertainty as to appro-
priate adjustment.
6-4
-------
where:
DDI =» human daily dermal intake (ug/day)
CT - contact time (hours/day)
2
SA - exposed skin surface area (cm )
2
CA - contact amount (mg/cm )
AF - absorption fraction (%/day)
LCj =» maximal soil concentration at time, T (ug/g)
EDA » exposure duration adjustment (unitless)
3 63
(10/24) » conversion factor (10" x 1 day x 10 ug)(ug x 24 hours x mg)
The DDI represents the increase (above background) in daily human
dermal intake (ug/day) due to the contaminant from municipal waste combustor
emissions.
G.3.3 Input Parameter Requirements
G.3.3.1 Absorption fraction (AF). Dermal absorption of a contaminant
2 3
is both chemical -specific and matrix-dependent > Physicochemical
properties of the contaminant (e.g., lipid solubility) will affect dermal
absorption. Factors such as pH, molecular size, temperature and humidity
will also influence absorption. Absorption of a contaminant in a matrix or
vehicle is affected by the physicochemical properties of the matrix. There
is uncertainty as to how various soil or parti cul ate types or matrices would
influence absorption of deposited contaminants.. Ppiger and Schlatter
demonstrated that a soil matrix reduced the absorption of TCDD when TCDD was
applied to the skin of rats in a soil -water paste.
Dermal absorption in this case refers to the fraction of the applied
dose of the compound absorbed by human skin within a day (i.e., 6%/
24 hours). Such data for a contaminant in a particulate or soil matrix
should be used but are rarely available. Absorption studies of the
contaminant by the dermal route of exposure may not be available, while
exposure by other routes may be present in the literature. Guidelines,
however, for route-to- route extrapolation of absorption of toxicants have
G-5
-------
not been clearly delineated and pose uncertainties, especially where dermal
absorption is concerned. Extrapolation of percutaneous absorption data
among animal species would introduce additional uncertainty.
Absorption of contaminants may be altered if the skin is damaged
(diseased, lacerated, or abraded). Percutaneous absorption may also vary
with age. These issues pose uncertainties and need further research.
The fraction of the contaminant absorbed may vary with concentration.
The assumption of complete absorption of a contaminant irrespective of dose
is the most conservative approach (AF - 1), but may be unrealistic. For
example, Kimbrough et al. suggested human dermal absorption of TCDD from
soil was -1 percent. -
The estimation of the fraction absorbed of a contaminant is very
complex and dependent on many factors as discussed above. A paucity of data
regarding the dermal absorption of chemicals in humans, particularly in
particulate or soil matrices, will make estimation of the fraction of
contaminant absorbed from soil difficult for this methodology.
6.3.3.2 Contact Time (CTV. Contact time is defined as the amount of
time/day the MEIs would spend in association with contaminated soil.
Children playing outdoors and .adult farmers would contact soil very
frequently, and soil contact would continue until the cleansing of exposed
skin. The maximum contact time for these individuals will be assumed to be
o
12 hours/day. Contact time \
exposure was also considered.
g
12 hours/day. Contact time would probably be greater if indoor dust
G.3.3.3 Surface Area (SA). Total intake of a'contaminant would be
approximately proportional to the exposed surface area for absorption. The
anatomical region and surface area of skin that is directly exposed to the
contaminant will affect dermal intake of the contaminant since there is
q
anatomical variability in percutaneous absorption. Exposed surface area of
adults wearing short-sleeved, open-necked shirts, pants and shoes, with no
n
gloves or hat, is -2940' cm , whereas that of children wearing the same
clothing is 980 cm2.10
G-6
-------
G.3.3.4 Contact Amount (CA). The contact amount or amount of soil
2
accumulating on skin is considered to have an upper limit of 1.5 mg/cm for
children ' based on the reports of Lepow et al. and Roels et al. '
The contact amount for children was assumed to also apply to adults.
Hawley, however, suggested the soil coating on adults could be as great as
3.5 mg/cm2.15
G.3.4 Example Calculations
G.3.4.1 Cadmium. The DDI for cadmium for adults can be estimated
using Equation 3-38 (see Section X.2.2). The values for SA, CA and CT are
2 2
2940 cm , 1.5 mg/cm and 12 hours/day, respectively. For the purpose of
this example only, the AF will be arbitrarily set at 1%/day. The LC is
21.76 and 72.53 ug/g for 30 and 100 years, respectively.
For 30 years, DDI - 12 hours/day x 2940 cm2 x 1.5 mg/cm2 x 0.01/day x
21.76 ug/g x (10~3/24) xT- 0.48 ug/day.
For 100 years, DDI =» 12 hours/day x 2940 cm x 1.5 mg/cm2 x 0.01/day x
72.53. ug/g x (10"3/24) x 1 = 1.60 ug/day.
*
G.3.4.2 Benzo(a)Pvrene. Values of DDI for B(a)P for adults and
children are calculated in the same manner as above for cadmium; however,
the LC for B(a)P, 1.17 x 10"2 ug/g for 30 and 100 years, is substituted.
For this example, AF will be arbitrarily set at 1%/day. The RE is derived
as the reciprocal of the absorption fraction (50%) for ingested B(a)P.
The EDA remains at 1 for lifetime exposure.
DDI = 12 hours/day x 2940 cm2 x 1.5 mg/cm2 x 0.01/day x 2.36 x
10"1 ug/g x (10"3/24) x 1 > 5.20 x 10"3 ug/day
G-7
-------
G.4 REFERENCES FOR APPENDIX G
1. Hawley, J. K. Assessment of Health Risk from Exposure to Contaminated
Soil. Risk Analysis 5:289-302. 1985.
2. U. S. Environmental Protection Agency. Dermatotoxicity. EPA Report
No. 560/11-82-002. 1982.
3. Reference 1.
4. Poiger, H., and C. Schlatter. Influence of Solvents and Adsorbents on
Dermal and Intestinal Adsorption of TCDD. Food Cosmet. Toxicol.
18:477-481. 1980.
5. Reference 2.
6. Feldman, R. J., and H. I. Maibach. Percutaneous Penetration of Some
Pesticides and Herbicides in Man. Toxicol. Appl. Pharmacol.
28:126-132. 1974.
7. Kimbrough, R. D., H. Falk, P. Stehr, and G. Fries-. Health Implications
of 2,3,7,8-tetrachlorodibenzodioxin (TCDD) Contamination of Residential
Soil. J. of Toxicol. Environ. Health 14:47-93. 1984.
8. Reference 1.
9. Maibach, H. I., R. J. Feldman, T. H. Milby, and W. F. Serat. Regional
Variation in Percutaneous Penetration in Man. Arch. Environ. Health
23:208-211. 1971.
«
10. U. S. Environmental Protection Agency. Risk Analysis of TCDD
Contaminated Soil. EPA Report No. 600/8-84-031. 1984.
11. Reference 10.
12. Reference 1.
13. Lepow, M. L., M. Gillette, S. Markowitz, R. Robino, and J. Kashin.
Investigations into Sources of Lead in the Environment of Urban
Children. Environ. Res. 10:415-426. 1975.
14. Roels, H. A., J. P. Buchet, and R. R. Lauwreys. Exposure to Lead by
the Oral and Pulmonary Routes of Children Living in the Vicinity of a
Primary Lead Smelter. Envir. Res. 22:81-94. 1980.
15. Reference 1.
16. U. S. Environmental Protection Agency. Environmental Profiles and
Hazard Indices for Constituents of Municipal Sludge: Cadmium. 1985.
G-8
------- |